I S -

P A U L S C H E R R E R I N S T I T U T

—

J—

CH9900002

ISSN 1423-7296March 1999

PSI • Scientific Report 1998/Volume I

Particles and Matter

A new compact radiocarbon dating system has been developed at the Ion BeamPhysics Laboratory of PSI/ETH in collaboration with an industrial partner (NEC,Middleton, USA). The device exploits a novel method to destroy molecularinterference. Beam energies of less than 1 MeV are now sufficient to identifysingle 14C ions. The cover photo shows the achromatic mass spectrometer thatcombines magnetic and electrostatic filter elements following the 500 kVPelletron accelerator. Major components of the spectrometer, the electrostaticdeflector, the beam analyzing system and the beam line components have beendesigned and partly manufactured at PSI.

P A U L S C H E R R E R I N S T I T U T

—

LJ

——

ISSN 1423-7296March 1999

Scientific Report 1998Volume I

Particles and Matter

ed. by: U. Baltensperger, D. Herlach, P.-R. Kettle, R. Lorenzen

CH-5232 Villigen PSISwitzerland

Telephone: +41 56 310 21 11Telefax:+ 41 56 310 21 99

http://www.psi.ch

TABLE OF CONTENTS

Editorial : 1

Particle Physics 3Particle Physics News

Theory 4Facility for neutron decay studies with polarized cold neutrons 5

Ring Accelerator Experiments

Particle Properties and DecaysCommissioning of the PIBETA detector 6Muonium - antimuonium conversion: Final results 7Measurement of the transverse polarization of positrons from the decay of polarized muons 8Search for a neutral particle of mass 33.9 MeV in pion decay 9A precision measurement of the Michel parameter £" in polarized muon decay 10

Hadron-Hadron and Hadron-Nucleus ReactionsMultinucleon pion absorption cross sections on N, Ar and Xe in the A-resonance region 11Total cross sections of the charge exchange reaction (jt+, n°) on 2H, 3He and 4He across the DeltaResonance 12

Exotic AtomsDirect observation of Coulomb explosion in pionic and muonic atoms 13The induced pseudoscalar coupling in nuclei: Muon capture on UB and the hyperfine effect 14Kinetic energy distribution of theup(1S) atoms as a function of density 15Study of muon catalyzed dd-fusion in D2 and H/D gas mixtures 16Kinetic energy of jrp-atoms in liquid and gaseous hydrogen 17Doppler-broadening of gamma rays following muon capture: Search for scalar coupling 18Investigation of the metastability of the 2s-state in muonic hydrogen 19Precision measurement of the singlet \ip capture rate via p.- lifetime in hydrogen gas 20Measuring u.c/3He fusion 21

Injector I ExperimentsCPAA as an analytical tool in the biological field 22Precise measurement of the neutron-proton spin correlation parameter at forward angles 23

External Accelerator / Facility ExperimentsThe CPLEAR experiment 24Brookhaven experiment E865: Studies of rare and searches for forbidden K+-decays 25The pixel detector and electromagnetic calorimeter of CMS 26H1 -events with an isolated lepton, large hadronic jets and missing energy 27

Astrophysics 29HESSI successfully qualified 30The HESSI Solar Aspect System (SAS) 31The HESSI Roll Angle System (RAS) 32Irradiation of titanium and tungsten samples for HESSI project 33XMM-RGS calibrations completed 34Improved energy resolution with Giaever junction X-ray detectors 35Superconducting X-ray detector read-out using a dc SQUID array 36Operating of the Proton Irradiation Facility - concise summary. 37Heating of stellar coronae 38Radio counterparts to extreme X-ray young stellar objects 39On the structure of stellar coronae 40Microinstabilities and scattering of accelerated particles in plasmas 41Dark matter in the Universe 42X-ray properties of galaxy clusters 43REM measurements versus standard models 44

II

Who cares about weak geomagnetic storms? 45

Muon Spin Spectroscopy 47Magnetism: Heavy Fermion Systems, Magnetic Ordering and Dynamics, Spin Glasses

|j.SR studies of correlated electron ground states (I) 48uSR studies of correlated electron ground states (II) 49Heavy-fermion behaviour and stripe order in electron doped cuprates 50Study of the magnetic properties of Ce3Pd20Si6 and Ce3Pd20Ge6 compounds 51Moment collapse in UJ-a^S 52Study of U- and RE-intermetallic compounds exhibiting quadrupolar and magnetic ordering phenomena:UPd3 53Microscopic studies of the non-magnetic-magnetic transition in highly correlated electron systema 54|iSR study of the non-fermi-liquid behaviour in U2Pt2ln 55Study of the spin dynamics in axial magnets. 56Spin behaviour in novel transition metal magnets 57Internal fields in magnetically ordered Dy and Ho 58Study of the dynamic and structure features of magnetism in holmium 59Spin-fluctuations in YCo2 60u+SR studies of the triangular lattice LiNiQ, 61Magnetic behaviour of a two-leg organic spin-ladder compound 62(iSR study on the S= V2 quantum spin system KCuCI3 63Magnetic properties of the spinel ZnFe2O4 in crystalline and nanostructured modifications 64Stripe order and spin dynamics in nickelates 65Study of magnetism in novel hole-doped quasi-1 D cuprates 66^SR on (RC)2(NiB)2 67Possible muonium formation in the erbium halides ErClj and Erl3 68Condon domains in non-magnetic metals 69

SuperconductivityStudy of the magnetic properties of HoBajCUgO^ : 70(x+SR on La2.x.yRExSryCu04 71(J.SR investigation of the magnetic phase diagram of Bl, Sr, 85CaCu20e+5 72Two stage melting transition of the vortex matter in BL.Srj.Ca.puO^j 73Hole dynamics and localization in low-doped high-temperature superconductors 74Local magnetic order in YBa^Cu 1.xCox)3O7 75u.SR on "artificially aged" YBa,Cu3Oa6 76u.SR investigation of the magnetic phase diagram of organic superconductora 77jxSR studies of superconductivity and magnetism in organic molecular crystal systems 78Systematic study of the magnetic penetration depth in Chevrel phase compounds 79

Semiconductors, Hydrogen in MetalsEvidence for a novel muon species in crystalline silicon 80Relaxation rate and paramagnetic frequency shift of negative-muon spin precession in silicon 81Muons in sulphur 82jxSR in II-VI solar cell materials 83The JJ+ in metal-hydrogen (deuterium) systems 84

Chemistry: Radicals, Reaction Kinetics, Charge TransferReorientation dynamics and chemical reaction of cyclohexadienyl radicals in ZSM-5 zeolites 85Heterogeneous processes of environmental free radicals 86Radical chemistry without water - insights into cell membranes 87Weak organic charge transfer systems 88JJ.SR dynamics of liquid crystals studies by ALC 89Temperature dependence of the hyperfine coupling constant of muonium in liquid H2O and D2Oand in solid Si8O12-cages 90

Experiments with Low Energy MuonsRange and range distribution of low energetic muons in matter 91The use of slow muons for precision measurements of magnetic fields on a 10nm scale at a YBa2Cu307.ssurface 92

Low energy p.SR study of the field distribution of a flux line lattice crossing a YBapu3O7.8 surface 93Magnetism of thin chromium layers studied with low energy p+SR ......: 94Slow muons to study CMR and spin-glass films 95

Experimental TechniquesLow energy [i+ beam developments 96Development of a novel time-differential jiSR data-acquisition system and its application to radio-frequency n-SR 97

Micro and Nano Technology 99Replication of nanostructures with polymer injection molding 100Nanoreplication in thin polymer films 101Manufacture of moderator gratings for epithermal muons 102Diffractive optical elements generated by e-beam lithography. 103Nanostructured electrodes for electrochemical investigations 104Nanostruotured surfaces for biomolecular immobilization 105Thermogravimetric analysis of micro crystals: Experiment and finite element simulations 106Highly efficient solar cells on low cost sheet ribbon grown silicon 107Effects of spontaneous ordering on the elastic properties of GalnR, 108High frequency optical rectification in bulkGaAs 109Atomic structure of the carbon-induced c(4x4) reconstruction on Si(001) 110TEM study of an anticorrelation in corrugated layers of Si,.8.yGexC/Si superlattices 111Photoluminescence of C induced Ge quantum dots embedded in Si 112Intersubband absorption in Si/SiGe MOW 113Metal-insulator transition in a SiGe 2-dimensional hole gas 114Amperometric immunosensing of penicillins 115Functionalised glass chips for neuron cells 116SPM for functional identification of individual biomolecules 117One-dimensional confinement of molecules via selective adsorption 118

Radio- and Environmental Chemistry 119Heavy Elements

A 252Cf fission fragment source called Miss Piggy. 120Miss Piggy, the new califomium-252 fission source as a generator of short-lived radionuclides 121Thermochromatographic investigation of ruthenium with oxygen as carrier gas 122Studies of gaschromatographic behaviour of 169'174183ie4Re j n the gas system He/O/HCI 123Physicochemical characterization of seaborgium as oxide hydroxide 124On-line chromatography of short-lived W isotopes with the multi-column technique 125Production of 262Db in the reaction 248Crn(19F, 5n) 126Fluoride complexation of rutherfordium (Rf, element 104) 127Preliminary studies for the separation of263Rf in aqueous solution part I: Tracer experiments 128Preliminary studies for the separation of263Rf in aqueous solution part II: On-line experiments 129Thermochromatographic investigations of254Es bromides 130HIVAPSI: A program to estimate cross-sections in heavy ion reactions 131Calculation of cross sections in a-xn-reactions 132CORA - a new control program for the ROMA detection system 133

Surface ChemistryA linear scanning device for on-line radioactivity measurements 134Efficient transport of bromine and selenium nuclides from thermal neutron induced fission of235U usingthe SINQ gas-jet facility 135Synthesis of an Arctic halogen reservoir species: Hypobromous acid (HOBr) 136HONO formation on soot aerosol particles - influence of humidity and ozone 137Formation of HONO from the reaction of NO2 with diesel soot 138Reaction of NO2 to HONO on sulphuric acid coated soot aerosol particles 139Modeling the NO2+soot reaction in the lower troposphere 140The influence of heterogeneous processes on water soluble anions of carbonaceous aerosol particles 141

IV

Aerosol ChemistryThe influence of additives on the size distribution and composition of particles produced by diesel engines 142Aerosol measurements during summer smog events in the Milan area 143Submicrometer aerosol size distributions at a high-alpine site (Jungfraujoch, 3580 ma.s.l.) 144Condensation nuclei (CN) and ultrafine CN in the free troposphere to 12 km: A case study overthe Jungfraujoch 145Seasonal variation of 212Pb-activity on aerosols at the Jungfraujoch 146Measurement of reactive nitrogen (NOy) at the Jungfraujoch 147Chemical composition study of the fine atmospheric aerosol at a high-alpine site(Jungfraujoch, 3580 m a.s.l.) 148Historical record of carbonaceous particle concentrations obtained in a European high-alpine glacier(Colle Gnifetti) agree well with estimated emissions 149

Analytical ChemistryEnvironmental record preserved in an ice core from Cerro Tapado, Chile 150Ice thickness and firn temperature measurements on the Cerro Tapado glacier (Chile) 151Quantification of sulphate' sources from an alpine ice core 152Calibration of a historical aerosol concentration record from an alpine ice core by long-term aerosolmonitoring 153Comparison of dating methods for an ice core from an alpine site 154Sampling and analysis of ice crystals as a function of size 155Imaging and analysis of individual ice crystals 156Investigating scavenging processes of ice crystals 157Continuous melting and analysis of ice cores 158The influence of stratospheric intrusions on alpine ozone concentrations 159Characterization and use of the new irradiation facility at SINQ for neutron activation analysis 160

Project RadwasteGadolinium complexes for targeted molecular radiobiology and radiotherapy 161Measurement of the radionuclide inventory in accelerator waste 162

Cement ChemistryCement chemistry: Quality control and developments 1998 163Application of the parallel-plate squeezing-flow technique for quality control of bituminous samples 164

Ion Beam Physics 165The PSI/ETH tandem accelerator facility 166A new small accelerator for radiocarbon dating 167Exposure age dating of glacial deposits in theLitang area, Eastern Tibet 168Preliminary results on the use of pyroxene for 10Be surface exposure dating 169Catchment-wide denudation rates from cosmogenic nuclides in river sediment 170The determination of surface exposure ages in Queen-Maud-Land (Antarctica) via "in-situ" producedcosmogenic radionuclides 171Magnetic, sedimentological and geochemical properties of 2 loess profiles from the Loess Plateau ofChina 172Radiocarbon time scale of the Mediterranean core CT85-5: The last 40'000 years 173UC ages of terrestrial macrofossils from Lago Grande di Monticchio (Italy) 174Geomagnetic modulation of the 36CI flux in the summit GRIP ice core 175The role of gallium and indium at the aluminum-oxide interface 176Imaging of fluorine on geological samples 177MeV ion tracks in PMMA: Formation of nanowires 178Investigations on the mechanism of a novel focused ion beam based lithography technique 179Application of a novel lithography process to silicon and CoSi/Si heterostructures 180A CoY liquid alloy ion source for focused ion beam implantation 181

List of publications 183Contributions to conferences and workshops 201Lectures and courses 215Higher degrees awarded 216

EDITORIAL

The new department Particles and Matter, crea-ted 1 October 1998 aims to strengthen the twopillars of PSI, research at large facilities and re-search with interdisciplinary teams. For instance,Particle Physics and Astrophysics at PSI have anestablished tradition in the field of particle and X-ray detectors and both co-operate with the Labo-ratory for Micro and Nano Technology (LMN).Environmental topics are pursued in the Labo-ratories of Ion Beam Physics (LIP) and Radio-chemistry (LCH). As a further example of theinterdisciplinary nature of the department, thetechniques used for the new low energy muonbeam are spin-offs from particle physics. How-ever, further collaborations, outside the newdepartment are encouraged, for example X-rayoptics produced in LMN, can be used at the fu-ture Swiss Light Source (SLS) and the develop-ment of biosensors in LMN benefit from theknowledge of the Department of Life Scienceand from electrochemistry.

The Laboratory for Astrophysics, together withthe Institute for Astronomy at ETHZ were jointlyevaluated in summer 1998 and receivedexcellent ratings, the particular strength of PSIbeing instrument-building. The audit teamemphasized the fact that the cryogenic detectorshave revolutionary perspectives in almost allwavelength domains of astrophysics. An equallyhigh ranking was given by NASA to thelaboratory for the professional handling of theHESSI satellite project (solar mission satellite).

Particle physics activities at PSI are approxi-mately equally divided between collaborationwith teams at the high intensity pion and muonbeams at PSI, and external projects at CERN,DESY and BNL (Brookhaven). In 1998, the twoprojects CPLEAR at CERN and E865 at Brook-haven were completed, with final data beingevaluated. At PSI, we see a trend towards asmaller number of larger experiments whichsearch for rare processes involving severalmonths of data-taking. A proto-collaboration for anew search for the decay \i -» e y, with a sensi-tivity of 1014 has been formed. An increasinginterest is seen in fundamental experiments withneutrons at SINQ and the planning of an ultracold neutron source has begun. One beamline atPSI is devoted to the testing of detectors for theLarge Hadron Collider (LHC) experiments atCERN, namely CMS, ATLAS and LHCb.

The creation of a Laboratory for Muon Spin Rota-tion (MuSR) should provide the deserved visibil-ity. One highlight of MuSR has been thesuccessful operation of the low energy muonfacility, which allows the probing of surface ef-fects at the 10nm depth-scale with the MuSRtechnique. A decision will be taken in 1999 as towhether or not this will be run as a user facility,like other MuSR beam lines.

The Laboratory for Ion Beam Physics, togetherwith an industrial partner, have developed acompact 500kV Pelletron accelerator for radio-carbon dating. The device is fully functional andwill increase the radiocarbon dating capacity ofthe laboratory. The accomplished novel techno-logy will enable the industrial partner to commer-cialize the system and the company hopes to sellseveral to universities in the near future.

The Laboratory for Micro and Nano Technologyundertakes a large range of projects with part-ners in industry and universities, to whom it pro-vides its modern infrastructure. The laboratoryfocuses its own research on two projects: re-search on Si/SiGe and Si/SiC nanostructures,with the aim of both understanding their elec-tronic properties and eventually producing lightfrom silicon, and the interdisciplinary project,molecular nanotechnology, with emphasis onimmunosensors.

The Laboratory for Radio and EnvironmentalChemistry concentrates on the chemical analysisof super heavy elements and their homologuesproduced with ion beams at Injector I, and atSINQ, and the investigation of agglomeratesformed from nanoparticles (aerosols) in the at-mosphere for environmental and climate re-search.

Finally, I would like to express our thanks to Prof.H.K. Walter, former head of the F1 Division, forkeeping particle physics a strong part of PSIduring the restructuring and to Prof. K. Ensslin,former head of F3B, for building up the compe-tent research laboratory in micro and nanotech-nology.

PAQE{S)| i&H BLANK

P a r t i c l e P h y s i c sParticle Physics News

Ring Accelerator Experiments:Particle Properties and Decays

Hadron-Hadronand

Hadron-Nucleus ReactionsExotic Atoms

Injector I Experiments

External Accelerator/ Facility Experiments

THEORY

A. Borici, F. Cuypers*, A. Denner, D. Graudenz, Th. Jensen, K. Junker, M. P. Locher, V. E. Markushin,F. Niedermayei4, St. Pozzorini, R. Rosenfelder, M. Roth, St. v. Rotz§, D. Wackeroth

* now: Zurich Insurance Company, Reinsurance, General Guisan-Quai 26, CH-8022 ZurichI now: Institute Theoretische Physik, Universitat Bern, Sidlerstr. 5, CH-3012 Berni now: Solartechnik, Technikum Rapperswil, Oberseestr. 10, CH-8640 Rapperswil

At present the Theory group is mainly working in thefollowing fields: muon physics (exotic atoms), strong in-teraction physics (pions and antiprotons), lattice gaugetheory, the Standard Model of particle physics (and itslimits and extensions). Below a few examples are pre-sented in more detail; for further topics see the detailedlist of publications which also includes the work done incollaboration with visitors.

• The investigation of the electroweak interactionrequires accurate experiments as well as precise(one-loop) calculations. The calculation of thecorrections to 4-fermion production in electron-positron collisions, which are important for thepresent experiments at LEP2, has been contin-ued (PSI-PR-98-07). The electromagnetic cor-rections to polarized Compton scattering at lowenergies (widely used to measure the polarisa-tion of electrons and positrons at high-energy col-liders) have been investigated (PSI-PR-98-10) aswell as the corrections to Moller scattering (PSI-PR-98-16) and to the production of fermion pairsin 77- collisions (PSI-PR-98-28). The two latterreactions are well suited for the luminosity mea-surements at electron-electron colliders and phot-on-photon colliders, respectively. In view of a pre-cise measurement of the W-boson mass at hadroncolliders a study of the impact of electroweak cor-rections on W-boson production at proton-anti-proton (Tevatron) and proton-proton (LHC) col-liders has been finished (PSI-PR-98-15).

• Lattice simulations are the only direct way to in-vestigate chiral symmetry in strong interactions.Yet, due to well-known no-go theorems, there isa genuine difficulty to construct chirally symmet-ric lattice actions for QCD. The usual way of ap-proaching the chiral limit is therefore based onan extrapolation to zero quark mass but it suffersfrom large systematic errors. Recently, a new for-mulation of QCD with chirally propagating statesbased on the so-called Ginsparg-Wilson relationhas been developed (hep-lat/9810026). The priceto be paid for exact chiral symmetry, however, isa large increase of computation time for the chi-ral lattice Dirac operator. A new efficient numeri-cal method based on the Lanczos algorithm (PSI-PR-98-21) is therefore of great practical interest.

• The Bose-Einstein (BE) correlations in pion pro-duction in low-energy pp annihilation were stud-ied using the CPLEAR data for the exclusive re-action pp -4 27r+27T"7r°. Differential distributionsshow a strong enhancement for small invariantmasses of like-pion pairs. The traditional inter-

pretation of this effect in terms of the stochasticHanbury-Brown-Twiss mechanism was found tobe in a conflict with the data. It was shown thatthe pion (BE) correlations can be explained bydynamical models, i.e. by the annihilation mech-anisms with pZn and pair final states (PSI-PR-98-17).Detailed studies of the hadronic final states arepresently the main topic at the electron-protoncollider HERA. The Figure shows the measuredone-particle-inclusive cross section for the pro-duction of charged particles (circles, triangles andsquares) integrated over bins of the scaling vari-able xp for various values of Q2. The dashedcurve is the theoretical prediction (PSI-PR-98-29)in next-to-leading order of QCD perturbation the-ory which matches the data for large values ofQ2 and xp but fails at low values. The discrep-ancies can be accounted for by phenomenolog-ical power correction terms of the form 1/(1 +4^2/(xpQ)2) (Y. Dokshitser and B. Webber, Dur-ham workshop) which are included in the full curvefor a mass scale /J. of the order of a few hundredMeV. The data from e+e~-annihilation (stars) arealso shown but should not be compared directlywith the theoretical prediction.

HI 96 (PRELIMINARY) * * e+e-

CYCLOPS(NLO) CYCLOPS(NLO) + pwr correction

grange

0.02- 0.05 (xlO)

0.05-0.10 (X5)

•^+~* * ' 0.10-0.20 (x3)

* * 0.20-0.30

* A 0J0-0.40

0.40-0.50

10 10'

Q,E (GeV)

FACILITY FOR NEUTRON DECAY STUDIES WITH POLARIZED COLD NEUTRONS

ETH ZURICH - LNS - PSI - PNPI - CRACOW - MADISON

K. Bodek§, P. Boni*, N. Danneberg*, W. Fetscher*, W. Haeberli**, C. Hilbes*, K. Kirch*, St. Kistryn^, J. Lang*,M. Luthy*, M. Markiewicz*, O. Naviliat*, A. Pusenkov§ A. Schebetov§, A. Serebrov§, J. Sromicki*5

* Institute of Particle Physics, ETH Honggerberg, 8093 Zurich, Switzerlandt Laboratory for Neutron Scattering, ETH and PSI, 5232 Villigen, Switzerland| Paul Scherrer Institute, 5232 Villigen, Switzerland§ St. Petersburg Nuclear Physics Institute, 188350 Gatchina, Russiaf Jagellonian University, 30059 Cracow, Poland** University of Wisconsin, 53706 Madison, Wl, USA

Semileptonic weak decays are still mysterious en-ough to hide important secrets of particle physics. Am-ong them the neutron decay plays a particular role,serving as a model for all semileptonic weak processes.However, the more complicated observables involvingthe spin of the neutron are not yet known to the desiredprecision; some have not even been measured at all.For instance, the question of time reversal violation, ad-dressed to static properties of the neutron as well as tothe decay process, attracts considerable attention [1].

With the full scale operation of the SINQ source,excellent conditions have been created for the devel-opment of a new activity in the field of particle physicsusing very slow neutrons. These experiments are lim-ited by counting statistics, therefore a high intensitybeam of very slow neutrons is of the utmost impor-tance. At present, the SINQ source complex is beingequipped with a large aperture and large momentumacceptance beam guide, which is covered right fromthe exit of the D2O moderator by supermirrors, consist-ing of 450 layers of Ni/Ti, with a high index of reflec-tion (m « 3). The modification of the non-instrumentedearlier neutron channel (so-called Sector 50, containedin the shielding structure of SINQ), was completed insummer 1998. The "real" neutron flux density mea-sured at the exit of the SINQ shielding is around 109

n/(cm2s-mA); the corresponding "capture" flux is — 3 •109 n/(cm2smA). With a full current operation of SINQduring the second half of 1998, no signs of fatigue ofthe supermirror layers have been detected even thoughthe most upstream part of the guide is located in anunusually high radiation field. The installation of a fastneutron beam stop, which will be integrated into theSINQ block, is planned during the shutdown in early1999. This includes taking into operation an area ac-cess control system, which must conform to the safetystandards maintained at the SINQ facility, and prepara-tory work for building the experimental area.

The next stage of the project is the installation of theexternal beam tract [2] to guide and extract the neu-trons to the experimental station. The whole systemwas designed and approved in 1998, in a close collab-oration between PSI-Villigen, ETH-Zurich and PNPI-Gatchina. The assembly is far advanced, with commis-sioning planned for the calendar year 1999. The com-plex equipment consists of a multislit, supermirror po-larizer and bender, cold neutron beam stop, a focusingbeam guide (tapered from 8 • 15 cm2 down to 4-15 cm2),

radiofrequency spin flippers, a chopper device for time-of-flight measurements, and two additional cold neu-tron beam stops placed downstream. The whole beamtract is immersed in a guiding field for the neutron spin,provided by permanent magnets with appropriate yokestructure. The front of this beam channel, with pre-cisely adjusted polarizer and bender, will be integratedinto the bunker of SINQ to contain the radiation asso-ciated with the neutron capture reactions. This sys-tem has its own, independent vacuum system. A sand-wich radiation shielding (brass/polyethylene/iron /lead)slows down and absorbs fast neutrons penetrating thewalls of the guide as well as secondary gamma rays.The residual radiation will be contained in a massive,biological shield mounted around the neutron guide.With the proposed solution we expect to eliminate aseparate, safety controlled area. The main functionsof this relatively compact (total length ~ 7 m) neutronoptics system are: increase of the neutron flux density,reduction of transport losses, and provision of enoughroom for a series of experiments. The main objec-tives will be achieved at less expense than the fore-seen installation of equivalent equipment in the mainneutron hall, and with approximately an order of mag-nitude higher number of polarized cold neutrons avail-able for the experiments.

Obviously, the construction of such a complex facil-ity is a formidable task and invovles substantial invest-ment. We are grateful to the PSI management for pro-viding financial support for this project. We very muchappreciate the engagement of PSI technical staff, inparticular D. Graf and J. Lederman (construction of-fice), G. Brunner (workshop), and help in coordinationof this project by G. Holtkotten, J. Duppich, K. Geiss-mann and W. Wagner.

REFERENCES

[1] P. Harris: Electric Dipole Moment, J. Gordon:"D" Correlation, J. Sromicki: "R" Correlation, Inter-national Workshop on Particle Physics with SlowNeutrons, Oct. 1998, ILL, Grenoble.

[2] [4]. A. Schebetov, A. Serebrov, V. Pusenkov,J. Sromicki, P. Boni et al. "Polarizers with Con-densers for Sector 50 at SINQ", PNPI Reports,Sept. 1997 and June 1998.

COMMISSIONING OF THE PIBETA DETECTOR

R-89-01.1, PIBETA Collaboration: VIRGINIA - PSI - SWIERK - ARIZONA - DUBNA - TBILISI - ZAGREB

E. Frlez*, J. E. Koglin*, W. Li*, A. V. Lebedev*, R. C. Minehart*, D. Pocanic*, S. Rift**, P. L Slocum*, L C. Smith*W. A. Stephens*, K. O. H. Ziock*, W. Bertl*, C. Bronnimannt, M. Daum*, T. Flugelt, R. Horisbergei-t, D. Renkert,R. Schnydert, H.-P. Wirtzt, T. Kozlowski*, D. W. Lawrence^, B. G. Ritchie§, V. V. KarpukhinU, N. V. KhomutovU,A. S. Korenchenko", S. M. Korenchenko^, N. P. Kravchukfl, N. A. Kuchinsky1', D. Mzhavia**, Z. Tsamalaidze**,I. Supek"

* Department of Physics, University of Virginia, Charlottesville, VA 22901, U.S.A.f Paul Scherrer Institut, CH 5232 Villigen PSI, Switzerlandt Institute for Nuclear Studies, PL-05-400 Swierk, PolandI Department of Physics and Astronomy, Arizona State University, Tempe, AZ 85287-1504, U.S.A.f Joint Institute for Nuclear Research, Dubna, Russia** Institute for High Energy Physics, Tbilisi State University, Tbilisi, Georgiaf t Rudjer Boskovic Institute, Zagreb, Croatia

The PIBETA experiment at PSI proposes to mea-sure the pion beta decay (ir+^n°e+iy) branching ratiowith an accuracy of about 0.5% in the current phaseof the project. Owing to low theoretical uncertaintiesstemming mainly from axial vector loop corrections, thisis one of the most precisely calculated observables inthe Standard Model. The best experimental value atpresent has error limits of 4%, far exceeding the theo-retical uncertainty of < 0.15%.

PIBETA Detector: tiev 2nd Pass, With Target Correction» 800 i

L=1.0X10*7vVsecTT —>)i—>e Events

10 20 30 40 50 60 70 80 90 100

Fit with POP Lifetime

T,=26.03 ns

"0 10 20 30 40 50 60 70 80 90 100TDCoK-TDCc (ns)

Figure 1: Timing spectra for ir -> p, ->• e events (toppanel) and n ->• ev events (bottom panel). Fitting func-tions using the Particle Data Group values for the p, andIT lifetimes are superimposed on the data histograms.

The PIBETA detector system was assembled dur-ing the first half of 1998. The heart of the detector isa 3TT spherical calorimeter consisting of 240 pure Cslcrystals. The central part of the apparatus contains ac-tive beam degrader and pion stopping target counters,surrounded by a charged particle (CP) tracking systemconsisting of a 20-bar scintillator hodoscope and a pairof concentric cylindrical wire chambers. All detector

components, including trigger electronics and front-endDAQ computers, are mounted on a compact platform.Details of the detector design and construction can beseen at the experiment URL site:http://psw340.psi.ch/"pibeta/.

PIBETA Detector: n/J Events (preliminary)

"0 25 50 75 100 125 150 175 200 225 250DPGrtf Csl ADC (MeV)

0 10 20 30 40 50 60 70 80 90 100

4 6 8 10DPGtt/?TGTADC(MeV)

Figure 2: 7r/?-spectra from the 1998 data: Csl calorime-ter energy spectrum (top), timing histogram for eventsthat pass 7r/?-cuts (middle), and active target energyspectrum (bottom panel).

First calibration and commissioning data with thewhole detector were taken during a six-week long sum-mer 1998 run. After a lengthy initial setup and calibra-tion, the response of all detector components to five dif-ferent 7r+ stopping rates between 0.3x105 and 8x105

7T+/s was studied. The total number of pions stopped inthe active target was 5.6x 1010. The CP detection inef-ficiency was less than 3x10~5. The DAQ live time was85% and higher, enabling us to collect ~80k n ->• evevents (Fig. 1) and ~120 ^-events (Fig. 2). Furtheranalysis of the commissioning run data, as well as in-strumental improvements and fixes are under way.

MUONIUM - ANTIMUONIUM CONVERSION: FINAL RESULTS

R-89-06, HEIDELBERG - ZÜRICH - PSI - DUBNA - TBILISI - YALE

L Willmann*, RV. Schmidt*, H.P. Wirtzf, R. Abela*, V. Baranov§, J. Bagaturia11, W. Berti*, R. Engtet,A. Großmann*, V.W. Hughes**, K. Jungmann*s, V. Karpuchin«, I. Kisel§, A. Korenchenko§, S. Korenchenko§,N. Kravchuk5, N. Kuchinsky§, A. Leuschnert, V. Meyer*, J. Merkel*, A. Moiseenko§, D. Mzavia^, G. zu Putlitz*W. Reichartt, I. Reinhard*, D. Renker*, T. SakhelashvilP, K. Träger*, H.K. Walter*

* Physikalisches Institut der Universität Heidelberg, D-69120 Heidelberg, Germanyt Physik-Institut der Universität Zürich, CH-8057 Zürich, SwitzerlandI Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland§ Joint Institute of Nuclear Research, RU-141980 Dubna, Russia,f Tbilisi State University, 380086 Tbilisi, Georgia** Yale University, New Haven Ct. 06520, USA

Launched in 1990, the experiment has witnessedseveral stages of improvements [1] until it reached itsultimate sensitivity in 1996. Final results and implica-tions are available now [2].

The conversion process muonium (M = v+e~) toantimuonium (M = n~e+) provides a sensitive testof lepton family number conservation laws and probesphysics beyond the standard model. The search wasmotivated by the wide spectrum of speculative modelspredicting detectable rates and the opportunity to in-crease the experimental sensitivity by 3 to 4 orders ofmagnitude. The latter became feasible by improvingthe LAMPF detection principle [3] with a large solid an-gle electron spectrometer and making use of PSI's highquality muon beams.

In total NM = 5.6(l]_x 1010 muonium atoms in vacuowere investigated for M~ decays. They emerged froma SiO2 powder target where they were formed by elec-tron capture after stopping a //+ beam. Candidate eventsfor the conversion process were filtered out using threestandard deviation cuts applied to the relevant distrib-utions which were determined from ordinary muoniumdecay prior to processing the antimuonium search data.Only a single event survived all cuts, in agreement withthe expected 1.7(2) events from accidental background.An upper limit on the conversion probability in an exter-nal magnetic field of 0.1 T was derived:

PMM(0.1T) < 8.3 x 10-11 (90 % C.L),representing an improvement of the LAMPF result bya factor of 2740 (both results normalised to zero field).The new bound has been turned into an upper limit forthe coupling constant GMJJ of a hypothetical M -M~interaction assuming Lorentz transformation propertiesof the (V - A) type to account for the strongest possi-ble rate suppression induced by the external magneticfield [4]:

GMM < 3 0 X 1 0 " 3 GF ( 9 0 % C - L )(Gp being the Fermi coupling constant).

Although no evidence for an exotic interaction hasbeen found, the new upper limit increases the domainsof validity of the standard model by postponing the scaleof new physics to higher energies:

A Z8 symmetric model for leptons with radiativelygenerated lepton masses is ruled out at least in itssimpliest version [5]. In the class of models with neu-

tral scalars mediating the MM interaction the one ex-changing a supersymmetric r sneutrino is of particularinterest here. In comparison to earlier experiments theupper bound on the coupling parameters |A3i2A521| <3 x 1CT4 has improved by a factor of 15 [6],

In grand unified theory models (GUT) postulatingthe existence of bileptons (gauge bosons carrying alepton number of two units) the new MM result causesa lower limit on the mass of flavour diagonal gaugebosons mx±± > (2.6 TeV/c2)#3( well beyond the val-ues extracted from direct searches [7]. A chiral bileptonmodel with the standard model embedded in a largergauge group (331 model) is disfavoured in its minimalversion as the upper bound mx±± > 850 GeV/c2 im-plied by our result exceeds the limit predicted by themodel (800 GeV/c2) which has been derived from pre-cision electroweak data [8].

In minimal left-right symmetric models the MM limitinfluences the upper bound on the "forbidden" muondecay fi+ -> e+^Fj. Thus, this decay cannot be re-sponsible for the observed excess of neutrino countsin the LSND experiment [9].

In the future, significant improvements in sensitivityare conceivable, if enhanced muonium production (invacuo) becomes feasible and/or high intensity pulsedmuon beams are available.

REFERENCES

[1] See PSI Newsletter 1993, 1994, 1995 and 1996.[2] LWillmann et al., Phys.Rev.Lett. 82 (1999) 49.[3] B.E.Matthias et al., Phys.Rev.Lett. 66 (1991)

2716.[4] G.G.Wong etal., Phys.Lett. B 357 (1995) 145.[5] G.G.Wong et al., Phys.Rev. D 50 (1994) R2962;

W.S.Hou et al., Phys.Rev. D 53 (1995) 1537.[6] A.Halprin et al., Phys.Rev. D 48, (1993) 2987.[7] H.Fujii et al., Phys. Rev. D 49 (1994) 559;

F.Cuypers et al., Eur.Phys.J C 2 (1998) 503.[8] RH.Frampton, Phys.Rev.Lett. 69 (1992) 2889;

P.H.Framptonetal., Phys.Rev.D58(1998)95013.[9] P.Herczeg, Los Alamos Science 25 (1997) 128.

8

MEASUREMENT OF THE TRANSVERSE POLARIZATION OF POSITRONS FROM THEDECAY OF POLARIZED MUONS

R 94 -10.1, ETH ZURICH - PSI - CRACOW - KATOWICE

K. Bodekt, A. Budzanowski§, N. Danneberg*, W. Fetscher*5', Ch. Hilbes*, M. Janousch*, L. Jarczyk*, K. Kirch*St. Kistryn*, A. Kozela§, J. Lang*, M. Markiewicz*, X. Morelle*, Th. Schweizer*, J. Smyrskit, J. Sromicki*,E. Stephan 1f, A. Strzalkowski^, J. Zejmat

* Institut fur Teilchenphysik (IPP), ETH Zurich, CH-5232 Villigen and CH-8093 Zurich, SwitzerlandI Institute of Physics, Jagellonian University, Cracow, Polandi Paul Scherrer Institut, CH-5232 Villigen-PSI

H. Niewodniczanski Institute of Nuclear Physics, Cracow, PolandInstitute of Physics, University of Silesia, Katowice, Polandf

This experiment aims at improving the present sen-sitivity [1, 2] on the transverse positron polarization byone order of magnitude in orderto achieve a better sen-sitivity on non-standard couplings and time reversal in-variance in weak interactions. In 1998, data were takenwith an almost complete experimental setup (a few ofthe BGO crystals were still missing) and a speciallydesigned cluster recognition unit (CRU). This CRU se-lects events consisting of at least two clusters on theBGO wall corresponding to the expected signature ofannihilation events.

The BGOs were calibrated online using cosmic rays.Fig. 1 shows the resulting Michel spectrum of the de-cay e+ from JX+ decay. For comparison the e~ spec-

Michel Spactrum |

|220

•°200

180

180

140

120

100

80

60

40

20

0

I ¥••

~ ft - . .

r «""-" 1EU.i J • • • ( « • • i » |

10 ?0

Cro*se*:)isSolkf:|i-

- - t YJi

Figure 1: Normalized energy spectra of electrons fromstopped ju~ and //+. The energy spectrum of the e+

from fi+ decay is shifted to the right due to the annihi-lation of the stopped e+ in the BGOs.

trum from fi~ decay is also shown. Note that the e+spectrum is shifted to the right due to the additional en-ergy from the subsequent annihilation at rest.

The measurements with stopped pr are a valuabletool in testing the annihilation trigger, since they canonly produce background events. By subtracting thespectrum obtained from n~ from the spectrum obtainedfrom n+ one gets the annihilation energy spectrum (cros-ses) of Fig. 2. The other distribution (shaded area) isobtained by demanding the opening angle between thetwo 7S, as calculated by the impact points on the BGOwalls, agree to within 7°, with the angle as calculatedfrom the two cluster energies (see Fig. 3). This resultsin a strong suppression of the background and will beimplemented as a hardware trigger in future measure-

Annihilation Energy Spoctrum |

350

300

250

wn

150

100

Im.l

E_ ijjft

I ^+jjf

\ii

„> ;

I• .

fMevf

Figure 2: "Golden" annihilation events. The solid distri-D U t i o n is obtained by requiring the opening angle be-tween the two photons tf34, as calculated from theirenergies, to be consistent to within 7° with the open-ing angle as calculated by their impact points (for thecut see Fig. 3). Also shown is a spectrum obtained bysubtracting the normalized annihilation spectrum mea-sured with fi~ from the normalized annihilation spec-trum measured with n+ (crosses).

Figure 3: Distribution of the spatial polar opening angletf34 as determined from the impact points of the twophotons on the BGO wall versus ^34 as determinedfrom the two photon energies. A cut requiring thesetwo angles to be equal to within 7° serves to preparethe "golden" (2-cluster) annihilation events.

ments.This project is supported in part by the Polish Commit-tee for Scientific Research under Grant No. 2P03B05111.

REFERENCES

[1] H. Burkard et al., Phys. Lett. 160 B (1985) 343.

[2] I. Barnett et al., in PSI Newsletter 1997.

SEARCH FOR A NEUTRAL PARTICLE OF MASS 33.9 MeV IN PION DECAY

R-96-05.1, ETH ZÜRICH - PSI - VIRGINIA - ZÜRICH - BEIJING

M. Daum*s, M JanouscM, P.-R. Kettle*, J. Koglin**, D. Pocanic*, J. Schottmüller, C. Wigger*,D. Wyler§,Z. G. Zhao^

* PSI, Paul Scherrer Institut, CH-5232 Villigen-PSI, Switzerlandf IPP, Institut für Teilchenphysik, Eidgenössische Technische Hochschule Zürich, CH-5232 Villigen-PSI,

Switzerlandt Physics Department, University of Virginia, Charlottesville, Virginia 22901, U.S.A.§ Physik-Institut der Universität Zürich, Winterthurerstr. 190, CH-8057 Zürich, Switzerlandf Institute of High Energy Physics, Chinese Academy of Science, Beijing 100039, The People's Republic of

China

In 1995, the KARMEN collaboration [1] reported ananomaly in the time distribution of neutrinos from apulsed beam-stop source, with a speculative explana-tion that these events could originate from a rare piondecay process,

w+-^fi++X, (1)

where X is a heavy neutral particle with a rest mass ofmx = (33.9049 ± 0.0009) MeV. Also in 1995, we un-dertook a measurement of the muon momentum spec-trum from 7T+-decay in flight using the beamline itselfas a spectrometer and placed an upper limit on thebranching fraction r] for this decay at 2.6-10~8 with aconfidence level of 95% [2].

In 1997, we conducted beam studies on a new ex-perimental setup, similar in principle to the 1995 setup,but with an extended beam line, an improved vacuumsystem and the addition of active and passive collima-tors [3]. The beamline consisted of the pion transportchannel, the pion decay region and the muon spec-trometer. The coincidence of three beam counters inthe muon spectrometer, in addition to appropriate tim-ing and veto counter cuts defined our events. After adetailed analysis of the 1997 data and further simula-tion studies, we concluded that a significant fraction ofour background could be attributed to scattered muonsoriginating from pions decaying through the normal de-cay mode within the pion decay region and also in thefirst dipole magnet of the spectrometer (ASL52) whichserves to separate out the beam pions from the decaymuons.

In 1998, a fullscale experimental run was conducted.We used the same basic 1997 setup with a widenedASL52 pole-gap and the addition of several strategi-cally positioned active veto counters to reduce the pos-sibility of accepting scattered muons. After optimizingour setup, we were able to reduce the background bymore then a factor of seven. Our data can be dividedinto three sets, each of which were fitted by a back-ground distribution plus the expected distribution formuons from the decay (1). The fitted momentum spec-trum of muons from the second data set is displayed inFig. 1 with a peak indicating the signal we would expectfor the branching fraction of the previously publishedupper limit [2]. Combining the preliminary fit results foreach data set (see Fig. 2), we find a new upper limit of

r-n-r i "n

I105 110 115 120 125

MUON MOMENTUM [MeV/c]

Figure 1: The solid line indicates the background fit todata set B. The dashed curve shows the peak expectedfrom the decay (1) for the previously published upperlimit on the branching fraction of 2.6-10~s[2].

B

, I- 1 0 1 2 3 4

BRANCHING FRACTION [109]

Figure 2: Preliminary fit results for each data set. Theshaded area indicates the region excluded by the newupper limit on the branching fraction of 2.5-10~9.

REFERENCES

[1] B. Armbruster et al., Phys. Lett. B 348 (1995) 19.

[2] M. Daum etal., Phys. Lett. B 361 (1995) 179.

[3] M. Daum et al., PSI Newsletter (1997)

< 2.5 • 10~9 (C.I. = 95%). (2)

10

A PRECISION MEASUREMENT OF THE MICHEL PARAMETER £" IN POLARIZED MUONDECAY

R-97-06, LOUVAIN - PSI - ZÜRICH - FRIBOURG

P. Van Hove", N. Danneberg*, J. Deutsch*, J. Egger4, W. Fetscher*, F. Foroughit, J. Govaerts*, M. Hadri*,C. Hübest K. Kirch*, P. Knowles^, J. Lang*, G. Llosa*, M. Markiewicz*, R. Medve*, X. Morellet, O. Naviliat*,A. Ninane*, R. Prieels*, T. Schweizer*, J. Sromicki*,

* Université catholique de Louvain, B-1348 Louvain-la-Neuve, Belgiumf Paul Scherrer Institute, CH-5232 Villigen, Switzerland% Institut für Teilchenphysik, Eidgenössische Technische Hochschule Zürich , CH-8093 Zürich, Switzerland§ Institut de Physique, Université de Fribourg , CH-1700, Fribourg, Switzerland

PSI experiment R-97-06, "A precision measurementof the Michel parameter £" in polarized muon decay" [1,2] is progressing. A Silicon Telescope Array (SITAR)to determine the energy of positrons in a uniform mag-netic field has been developed and tested and the analy-sis is in progress. Problems concerning forces on thecoils of the PSC magnet and stray fields in the regionof the BGO photomultipliers have been solved and suc-cessfully tested.

TEST OF SILICON TELESCOPE ARRAY

After an initial crude energy selection, positrons frommuon decay are guided inside a uniform magnetic field,where they spiral. A Monte Carlo study showed that themeasurement of the positron position in three planeswith a resolution of 1 mm enables the determination ofits energy with a FWHM resolution of 1 MeV. SITAR,the device built for the three position measurements,consisted of three planes of four double-sided multi-strip silicon detectors read by VA-Rich chips (IDE-AS,Norway). These chips were encapsulated in a PGAbox, in France, and can be easily plugged into or outof a specially designed board. Each board makes theconnection between the strips of one face of a silicondetector and the VA chip via de-coupling capacitors,drives the biases of the silicon detector and finally, setsthe working parameters of the VA via computer control-lable digital potentiometers. The VA chip can be tunedvia six parameters, the stability of which proved to becrucial for low noise and so an improved version of theoriginal board has been designed. It provides for bettergrounding and noise suppression.

In a first application at PSI, parameters giving anoptimal signal to noise ratio were found for one sili-con element (2 VA chips) and those parameters wereused for the remaining 22 chips. The resulting signalto noise ratios for the p-side data ranged from 8:1 up to17:1. The n-side readout fared worse, with the best ra-tio at 13:1, and four units operating below 6:1, the limitfor efficient usage in the energy reconstruction. A firstanalysis tends to suggest that this behaviour is due tothe chip, and not to the attached Si detector. Individualoptimization or replacement of those four chips was notpossible due to the loss of test beam but work is under-way.

MAGNETIC FIELD AND SHIELDING

Computations with POISSON made in parallel at

PSI and at LLN, as well as investigations with GEANTat LLN, selected a setup where the forces on the coilsof the PSC magnet were reduced to a level well un-der the maximum allowed. This setup consists of theCyclotron TRAP, extended by 16 cm and with an im-proved shielding, the PSC Magnet and finally STAR.The distances between these magnets are such thatthe forces on the PSC coils due to TRAP and STARjust compensate each other. The additional shieldingand the extension of TRAP, as well as a new supportfor the PSC magnet, needed because of the small dis-tances (about 30 cm) between the magnets, has beenrealized at PSI.

The other concern was the 20 G stray-field that re-mains in the region of the polarimeter. This stray-fieldwould destroy the gain of the 127 BGO photomultipli-ers, which cannot be individually shielded. The solutionfound consists of a (3 x 3) m2 cross section and 1.8 mlength Big Iron Box (BIB) which was built at LLN. Thedimensions of the box were optimized using a simpli-fied cylindrical geometry of the box in the POISSONcode and led to a remaining field of about 3 G.

This whole setup was mounted during our last runat PSI. The results were rather a success: the magnetswere powered without any break and the magnetic fieldin the region of the BGO photomultipliers was such thatfor a given light input, the output ranged from 3.6 V to3.3 V. For comparison, for the same light input but inthe earth's magnetic field where the BGO are usuallyused, the output variation ranged between 4.4 V and3.9 V depending on the orientation.

We are most grateful to Leopold Simons and BrunoLeoni for their pleasant and efficient help and discus-sions, and to Manfred Werner for computations of mag-netic fields and forces. We also wish to thank all thepeople in the workshop who enabled the realization ofall the mechanical parts.

REFERENCES

[1 ] P. Van Hove et al., A precison measurement of theMichel Parameter £" in polarized muon decay" PSIproposal R-97-06, unpublished.

[2] P. Knowles et al., PSI Ann. Rep. NPP (1997) 16and included references.

11

MULTINUCLEON PION ABSORPTION CROSS SECTIONS ON N, Ar AND Xe IN THEA-RESONANCE REGION

R-87-13(A), BASEL - KARLSRUHE - LAMPF - MARYLAND - MIT - NEW MEXICO STATE - OLD DOMINIONPSI - ZAGREB

D. Androic, G. Backenstoss, D. Bosnar, H. Döbbeling, T. Dooling, M. Furic, P.A.M. Gram, N.K. Gregory, A. Hoffart,C.H.Q. Ingram, A. Klein, K. Koch, J. Köhler, B. Kotliriski, M. Kroedel, G. Kyle, A. Lehmann, A.O. Mateos,K. Michaelian, T. Petkovic, M. Planinic, R.P. Redwine, D. Rowntree, U. Sennhauser, N. Simicevic, R. Trezeciak,H. Ullrich, H.J. Weyer, M. Wildi, K.E. Wilson

The 4TT solid angle Large Acceptance Detector Sys-tem (LADS) was built at PSI to study multi-nucleon pionabsorption at energies around the A-resonance [1]. Inaddition to the light targets 3He and 4He (see e.g.[2,3]),measurements with heavier targets, N, Ar and Xe, werealso performed. Earlier [4, 5], we reported first heavytarget results which showed the existence of the initialstate interaction in pion absorption. More recently, theheavy target data have been fully analyzed in order todetermine the breakup of the absorption cross sectioninto individual channels. Pion absorption events lead-

100 150 200 250 300 350

T, (MeV)

Figure 1: 2N, 3N, 4N and 5N partial absorption crosssections for the 7r++Ar pion absorption as a function ofpion kinetic energy.

ing to final states with energetic protons, neutrons anddeuterons were selected from the N, Ar and Xe targetdata. Absorption events were identified by requiring nopion in the final state and were sorted according to thenumber of protons, neutrons and deuterons in the finalstate. Particles with kinetic energy below 30 MeV wereignored. About 26 channels were identified, includinghigh particle multiplicity states (e.g. 6p).

All channels were simultaneously analyzed with thehelp of Monte Carlo simulations. This procedure re-sulted in acceptance corrected cross sections. In orderto present results of a large number of channels wegroup them according to the total number of particlesin the final state. For example the label 3N means finalstates with three particles (3p, 2p1 n or 1 p1 d1 n).

In Figure 1, the ratio of the measured cross sectionsdivided by the total absorption cross section is plotted

versus pion beam energy for Argon. One sees a grad-ual increase of the particle multiplicty with pion energy.The 2N fraction of the absorption cross sections, whichdominates at the lowest energy, decreases in impor-tance and becomes negligible at the highest energywhile the 3N fraction peaks around 160 MeV. Othermultiplicties gain strength with increasing pion energy.In Figure 2, these ratios are plotted as a function of tar-get mass at one pion energy. An important feature isthat fractions of the pion absorption cross section go-ing to final states with 2N, 3N, 4N and 5N reach "satu-ration" values already for Nitrogen. The fractions stayrelatively constant for heavier nuclei.

10Target A

Figure 2: 2N, 3N, 4N and 5N partial absorption crosssections for the TT+ pion absorption at Tw = 239 MeV asa function of the target mass.

REFERENCES

[1] T. Alteholz et al., Nucl. Instrum. Methods A373,374(1996).

[2] A. Lehmann et al., Phys. Rev. 56 1872 (1997).

[3] A.O. Mateos et al., Phys. Rev. 58 942 (1998).

[4] D. Androic et al., Phys. Rev. 53 (1996) R2591.

[5] B. Kotlinski et al., Euro. Phys. Journal A1 435(1998).

12

TOTAL CROSS SECTIONS OF THE CHARGE EXCHANGE REACTION (7r+,7r°) ON 2H, 3HeAND He ACROSS THE DELTA RESONANCE

R-87-13(B), BASEL - KARLSRUHE - LAMPF - MIT - NEW MEXICO STATE - OLD DOMINION - PSI - ZAGREB

D. Androic, G. Backenstoss, D. Bosnar, T. Dooling, M. Furic, P.A.M. Gram, N.K. Gregory, A. Hoffart, C.H.Q. Ingram,A. Klein, K. Koch, J. Kohler, B. Kotliriski, M. Kroedel, G. Kyle, A. Lehmann, A.O. Mateos, K. Michaelian, T. Petkovic,M. Planinic, R.P. Redwine, D. Rowntree, N. Simicevic, R. Trezeciak, H. Ullrich, H.J. Weyer, M. Wildi, K.E. Wilson

In the range of the A(1232) resonance the mostimportant reactions of a pion with a nucleus are elas-tic and inelastic scattering, absorption and charge ex-change. While charged pion scattering and absorptionon various nuclei were studied in detail during the lastdecades and a large body of data exists, the data baseof the charge exchange reactions (7r+,7r°) and (7r~,7r°)(SCX) is considerably smaller. Even for the intensivelyinvestigated xNN system only few experimental dataof the SCX reaction 2H(7r+,7r<>)pp are available, mostof them being measurements of differential cross sec-tions. For the basic observable of the total cross sec-tion the situation is even worse: to date only four datapoints were measured for SCX on 2H, none for 3Heand 4He. This is mainly due to the required detectionof the two photons of the decaying TT° which is difficultto achieve with good efficiency.

£50

|

I 40

bc 30o<u

10208

s10

XoCO

n(7t+,n°)

50 100 150 200 250 300 350 400Pion Kinetic Energy T, (MeV)

tion of neutral particles (e.g. ~ 30% for photons) thisdetector also allows an investigation of SCX reactions.The good neutron-gamma discrimination, the low de-tection threshold for charged particles, and the largesolid angle coverage makes it an excellent detector fora total cross section measurement of (71-+, TT0) reac-tions.

The SCX total cross sections on the nuclei 2H, 3Heand 4He were determined for incident pion energies of70, 118, 162, 239, and 330 MeV (see Fig. 1). As an-ticipated the charge exchange cross section per neu-tron on a nucleus is usually lower than that on the freeneutron. Further, it was found that on 2H the SCXcross sections exhaust about one quarter of the fullpion inelastic cross section, which is significantly largerthan one would expect for an unbound p - n pair (~17%). On the helium isotopes a strong damping ofthe SCX cross sections in the A(1232) energy regionis observed. These observations and also the relativestrengths of SCX on 2H, 3He and 4He suggest that inpion scattering nuclear medium effects are already sig-nificant on the helium isotopes. However, double scat-tering contributions to the SCX total cross sections on3He and 4He are substantial (~ 30%) only above theA-peak energy. This indicates that at and below theA(1232) resonance multiple pion scattering is not theonly reason for the damping of the SCX cross section,but that other nuclear medium effects like, e.g. absorp-tion, are also important [2].

REFERENCES

[1] T. Alteholz, et al., Nucl. Instrum. Methods A373(1996)374.

[2] A. Lehmann, et al., PSI-Preprint PSI-PR-98-30;submitted to Phys. Rev. C.

Figure 1: Energy dependence of the SCX total crosssections of the reactions 7r++2H —> K°PP, 7r++3He•—> n°ppp, and 7r++4He —»• •w°pp{pn/d), as deter-mined by this work. The data of 3He for 70, 239, and330 MeV are shifted slightly to make them distinguish-able from the data of 2H and 4He. Note that the crosssection is displayed per neutron of the target nucleus.

The 4TT solid angle Large Acceptance Detector Sys-tem (LADS) [1], at PSI, was built to study the pion ab-sorption reaction in detail. However, due to its goodperformance parameters including an efficient detec-

13

DIRECT OBSERVATION OF COULOMB EXPLOSION IN PIONIC AND MUONIC ATOMS

R-94-01, IONANNINA - JÜLICH - PARIS - PSi - NEUCHÄTEL

D.F. Anagnostopoulos*, M. Augsburgert, G. Borchert*, D. Chatellard*, M. Daum§, J.-R Egger*, P. El-Khoury^,D. Gotta*, P. Hauser§, P. indelicate1', K. Kirchs, o.W.B. Schuld, Th. Siems*, L.M. Simons^

* Department of Physics, University of loannina, 45332 loannina, Greecet Institut de Physique de PUniversite, Breguet 1, CH-2000 Neuchäteli Institut für Kernphysik, Forschungszentrum Jülich, D-52425 Jülich§ Paul Scherrer Institut, CH-5232 Villigen PSI% Institut du Radium, Universite Pierre et Marie Curie, F-75252 Paris

Measurements of pionic and muonic X-ray transi-tions in nitrogen, oxygen, and neon at atmospheric pres-sure have been performed in the TTE5 area at the PaulScherrer Institut (PSI) using the Jülich crystal spec-trometer [1] together with the new cyclotron trap [2, 3].

The pionic neon transitions reflect the experimentalresponse function of the crystal spectrometer, since nomolecular effects are present (Fig. 1). In case of pio-nic nitrogen and muonic oxygen the line widths of theX-ray transitions are significantely broadened (Fig. 2).This broadening is interpreted as a Doppler broaden-ing, originating from a Coulomb explosion in diatomicmolecules.

The observed additional width of the irN (5g-4f)transition is (540±60) meV, where the main uncertaintyoriginates from details based on the assumption of theline shape. This broadening corresponds to a kineticenergy of the TTN atom of about (90-150) eV, originat-ing from the Coulomb potential energy at the time ofseparation. Assuming that the (7rNN)* molecule disso-ciates at distances comparable to the molecular bondlength of 1.1 • 10~8 cm, one obtains for the product ofcharges q, of the decomposing atoms q r q 2 « 15-22,i.e. the range of charge states at the time of explosionis 3.8 to 4.7. For N2, this corresponds to the removal ofabout 7 electrons before dissociation.

CO

I 200o

175

150

125

100

75

50

25

0

1 Si

-

-

1t-

-

220

eV |!

1

TTN (5-4)

TIN (5g-4f)

TIN (5f-4d)

An

CO

1 2°0o

175

150

125

100

75

50

25

0

1 Si 220• |

- I

: 1 eV1 1 1

r

• . . J .

TTNe (6-5)

7T2°Ne (6h-5g)

Tt^Ne (6g-5f)

\ / Tt^Ne (6h-5g)

40 50 60 70 80 90 100 110 120 130 140

Energy (Channels)

Figure 2: Position spectrum of the 7rN(5g-4f) transitionat Ex = 4.055 keV, pressure = 105 Pa

REFERENCES

[1] S. Lenz et al., Phys. Lett. B 416 (1998) 50.

[2] D.F. Anagnostopoulos et al., Annual report Jül-3505, IKP 1997, 85.

[3] LM. Simons, PSI annual report 1996, p. 50.

40 50 60 70 80 90 100 110 120 130 140

Energy (Channels)

Figure 1: Position spectrum of the 7rNe(6h-5f) transi-tion at Ex = 4.510 keV, pressure = 105 Pa

14

THE INDUCED PSEUDOSCALAR COUPLING IN NUCLEI: MUON CAPTURE ON B ANDTHE HYPERFINE EFFECT

R-94-02, DUBNA - LOUVAIN - PRAGUE - PSI - ZURICH

V. Wiauxt, V. Brudanin*, J. Deutscht, V. Egorov*, J. Govaertsf, J. Lehmann*, T. Otto*, C. Petitjean*, R. Prieelst,J. Rak', P.Truol§.

* Nuclear Problems Laboratory, Joint Institute for Nuclear Research, 141980 Dubna, RussiaI Universite catholique de Louvain, B-1348 Louvain-la-Neuve, Belgiumt Paul Scherrer Institute, CH-5232 Villigen, Switzerland§ Physik-lnstitut des Universitat Zurich , CH-8001 Zurich, Switzerlandf Institute of Nuclear Physics, Rez, Prague, Czech Republic

We report on the results of a muon capture exper-iment on n B which aims to determine the ratio gp/gAof the pseudoscalar and axial couplings, relevant to thequestion of the possible renormalization of the couplingof the nucleon in the hadronic medium. This experi-ment uses the hyperfine effect and the hyperfine con-version in the muonic atom fiuB. One measures theratio A+/A~ of partial capture rates, from each of thetwo hyperfine states to the first excited state of n B eat 320 keV, which then immediately emits a 320 keVgamma ray. A+/A~ is related to gp/gA through nu-clear models. The principle of this experiment and themethod of analysis are explained in detail in the previ-ous report [1]. Let us recall that we have to extract twodifferent signals from our data: the time evolution ofthe emission rate of the 320 keV gamma rays, and thetime evolution of the emission rate of the decay elec-trons from the / / n B atom. The experimental informa-tion is contained in two different histograms. Their the-oretical shapes are fitted simultaneously. The gammaray signal is extracted from an energy-time histogram.The time and energy evolution of the background isconstrained below and above the 320 keV peak, andinterpolated under the gamma ray signal. The elec-tron signal is extracted from a time histogram. It in-cludes in addition the decay electrons from muons cap-tured by impurities in the target. We have since im-proved the evaluation of the different contributions tothe electron spectrum. The theory is fitted by the leastsquare method simultaneously to the gamma ray andthe electron spectra. We use the CERN minimisationsoftware, MINUIT. Throughout the analysis, some pa-rameters are set to fixed values. They have both sta-tistical and systematical errors. The error propagationis made by setting the value of each parameter to thelimit of its error interval, keeping the other parametersat their central value. The statistical errors obtained inthis procedure are added quadratically to the dominantone given by MINUIT, to provide the statistical error onthe final results. In a similar way, the systematical errorof the final result is the sum of the individual contribu-tions.

The value of the conversion rate resulting from thisexperiment gives a much better precision than the ex-isting data:

Up to now, only approximate values of the disappear-ance rate from the lower hyperfine level were available.Our result is

\p = 478.70 ± 0.72(stat) ± 0.25(syst) • 10V (2)

This experiment is also sensitive to the amplitude ofthe hyperfine effect on the total capture rates in n B ,previously unmeasured.

AA = -13.2 ± 1.7(stat) ± 0.7{syst) • 10V (3)

This result is in reasonable agreement with existingtheoretical predictions [2].

The main result is the value of the ratio of partialcapture rates,

A+- = 0.028 ± 0.021{stat) ± 0.003{syst) . (4)

A

According to the dependence computed by T. Suzu-ki [3] and by V. Kuz'min et al., [4], this value of A+/A~corresponds to the following ratio of the pseudoscalarand axial weak couplings :

± 0A{syst) . (5)

R = 181 ± I6{stat) ± l(syst) 10V1(1)

The latter value is in agreement with the prediction fromPCAC and the pion pole dominance hypothesis,{gp/9A)pcAC = 6.8. However, as with other measure-ments in complex nuclei, this result is lower, by approx-imately two standard deviations, than the value froma recent RMC measurement in hydrogen, {gp/gA) =9.8 ±0.7±0.3 [5]. The disagreement between the latterresult and the PCAC prediction should be understoodbefore any conclusion concerning an eventual renor-malization is made.

REFERENCES

[1] V. Wiaux et al., PSI Ann. Rep. NPP (1997) 26.

[2] K. Koshigiri et al., Prog. Th. Phys. 71 (1984) 1293.

[3] T. Suzuki, Proceedings of the International Sym-posium on non-nucleonic degrees of freedom de-tected in nucleus (Sep 2-5,1996) Osaka.

[4] V. Kuzmin et al., Phys. At. Nucl. 57 (1994) 1063.

[5] D. H. Wright et al., Phys. Rev. C57 (1998) 373.

15

KINETIC ENERGY DISTRIBUTION OF THE ^p(1S)-ATOMS AS A FUNCTION OF DENSITY

T w < 0.5 eVT w < 1.0 eVT w < 2.0 eV/(expected)/(lower limit)

1/16 hPa0.48 ±0.080.70 ±0.060.79 ±0.040.70 ±0.10

0.30

16hPa0.17 ±0.020.30 ±0.030.47 ±0.030.25 ±0.05

0.09

Table 1: Fraction of ^p-atoms with kinetic energiesT»p < T and the metastable 2S-fraction / .

formed when the /ip(2S) are slowed down to below0.31 eV. Table 1 gives the expected metastable 2S-fraction / , based on a theoretical calculation of the A«P(2S)slowing down and quenching [4]. The last line gives alower limit corresponding to Tw(2s) < 0.31 eV.

R-94-04.1, MUNICH - ETH ZURICH - PSI

F.J. Hartmann*, P. Hauser*, F. Kottmann*, Ch. Maierl*, V. Markushin*, M. Muhlbauer*, C. Petitjeant, R. Pohlt,W. Schotr.D.Taqqut

* Technical University of Munich, Physics Department, D-85748 Garchingf Paul Scherrer Institut, CH-5232 Villigen PSI| Institut fur Teilchenphysik, ETH Zurich, CH-8093 Zurich

An experiment making use of the low energy p-beam, produced at PSI, provided for the first time, de-tailed information on the kinetic energy distribution of//p-atoms. This information is relevant for two funda-mental experiments in preparation at PSI, the 1S-widthof the 7rp-atom [1] and the Lamb shift of the /^p-atom[2].

In an ~20 cm long cylindrical target of low densityhydrogen, axial injection of the low energy fi~ beamresults in a pencil-like stopping volume. The formedpp-atoms travel to the wall of the gold-coated innercylinder. Their time-of-flight distribution is measured bydetecting the X-rays emitted after the transfer reactionof /ip-atoms with Au. In the cylindrical geometry a largefraction of the yup-atoms reach the wall by a path whichis only slightly longer than the path perpendicular to theaxis. This results in a good velocity resolution allowingthe energy distribution of the /ip-atoms to be deduced.Using variable inner wall diameters consistency checkscan be performed. Information on the elastic cross sec-tion of the /up(1S)-state in H2 can be deduced from thehigher density targets.

Measurements previously done with inner wall di-ameters of 20 and 58 mm have been recently extendedto smaller diameters (7 and 12 mm) providing a highersensitivity to energies TMP < 0.5 eV. Detailed data analy-sis has been performed to determine the contributionof the background from muon decay electrons in a selfconsistent and reliable way. Fig. 1 shows the kinetic en-ergy distribution obtained at the lowest pressure (1/16hPa at room temperature) and at 16 hPa where targetswith small inner diameter have been used.

Information on the atomic formation of exotic hydro-gen atoms, which takes place at high excitation lev-els n, has been obtained here for the first time. Pres-sures below 0.1 hPa are required to ensure that atleast part of the formed /ip-atoms de-excite without col-lisions. The observed energy distribution and its pres-sure dependence allow the range of n levels initiallypopulated to be deduced. At 16 hPa the kinetic en-ergy distribution extends to quite high values (15% withr w > 10 eV). This indicates a strong acceleration ofthe /ip-atom that can only be reached by Coulomb de-excitation involving relatively low n levels. The deducedstrength of the interaction is much larger than that pre-dicted by calculations [3] using H-atoms as collisionpartners.

The results (see Table 1) are sufficiently precise toprovide essential conclusions on the fraction of meta-stable ^p(2S)-states as a function of density. The initialMP(2S) kinetic energy distribution is practically identi-cal to that of the j«p(1S) one. Metastable ^p(2S) are

0.175 r

0.15 •

0.125 :

0.1 r

0.075 ;

0.05 -

0.025 :

n -

":

\ ^ ^ - 1/16 hPa

n : s 16hPa

\ ^ _ _ _ _ _ _ _ _ _ _ _

0 1 8 9 10Tw(eV)

Figure 1: Density distribution p(TMP) of kinetic energiesof the /ip-atom at 1/16 hPa and 16 hPa H2 gas pres-sure.

REFERENCES

[1] PSI Proposal R-98-01.1 (1998).

[2] PSI Proposal R-98-03.1 (1998).

[3] L.I. Ponomarev and E.A. Soloviev, JETP Lett. 64(1996) 135.

[4] G. Carboni and G. Fiorentini, Nuovo Cimento B39(1977)281.

16

STUDY OF MUON CATALYZED dd-FUSION IN D. AND H/D GAS MIXTURES

R-94-05, BERKELEY - GATCHINA - MUNICH - PSI - VIENNA

D.V. Balin*, W.H. Breunlich*, T. Case*, K.M. Crowe*, V.A. Ganzha*, B. Gartner*, FJ. Hartmann§, P. Kammel*,S. Kozlov*, B. Lauss*, E.M. Maev*. V.E. Markushin" Yu.A. Misko*, M. Muhlbauer§, C. Petitjean^5, G.E. Petrov*,W. Prymas*, S.M. Sadetsky*, G.N. Schapkin*, W. Schotte G.G. Semenchuk*, Yu.V. Smirenirf, M.A. Soroka*,V.A. Trofimov*, A.A. Vasiliev*, A.A Vorobyov*, N.I. Voropaev*, J. Zmeskal*

* Petersburg Nuclear Physics Institute, Gatchina 188350f Austrian Academy of Sciences, Institute for Medium Energy Physics, A-1090 Vienna| University of California and Lawrence Berkeley Laboratory, Berkeley CA, 94720§ Technical University of Munich, Physics Department, D-85748 Garchingf Paul Scherrer Institut, CH-5232 Villigen PSI

Between 1994-96, we performed, at the //E4 chan-nel, experimental runs to study muon catalyzed dd-fusion in pure deuterium, HD gas and in H2(+HD)+D2

mixtures [1]- Most data are now analyzed, and the re-sults were presented at the workshop on Exotic Atoms,Molecules and Muon Catalyzed Fusion in Ascona, July19-24,1998 [2-5]. Here we summarize some importantresults:

(1) The most precise data set (to ±1%) of absoluterates of dp.d muonic molecule formation in D2 is nowavailable, see Fig. 1, allowing the parameters of reso-nant and non resonant dd-fusion to be exactly determ-ined [2]. The final analysis is still on-going.

(2) A temperature dependence of the spin flip pro-cess /JC/(F = 3 /2) + d -> fj,dSF~xl2^ + d' in D2 was clearlyobserved for the first time and the contributions from in-elastic scattering {\scat) and from resonant dpd-forma-tion followed by back decay (\back) quantitatively de-termined, see Fig. 2, resolving longstanding discrep-ancies between experiments and theory [3].

(3) In pure HD gas, ^- format ion at thermal ener-gies proceeds only via the (non resonant) Auger mech-anism. However, we observed a large formation peakduring the first 60 ns after the //-stop, when ^cf-atomsare still in epithermal states. The peak is well repro-duced by Monte Carlo assuming epithermal resonancespredicted by theory [4].

(4) The high statistics of accumulated dfid-fus\onevents were also used to determine the energy depen-dence of /^.-stripping after sticking in dpd -¥ ^3He+n.The result (10% stripping probability) allows the re-eva-luation of our former /za-sticking data in cft-fusion andyields a slightly modified, improved value for^a-stickingus = (0.58 ±0.04)% [5].

REFERENCES

[1] See R-94-05 PSI Ann.Reports/Annex 11994-97.

[2] C. Petitjean et al., Hyperfine Int. (1999) in print.

[3] N. Voropaev et al., Hyperfine Int. (1999) in print.

[4] G.G. Semenchuk et al., Hyp. Int. (1999) in print.

[5] T. Case et al., Hyperfine Int. (1999) in print.

m1

5:

4-i

3:

2:

1 :

0 :

low t e m p . (

e q u i l i b . * _ S

Wtgas

Yniquid

g?sol!d

^ ^

•A

F=3/2

F=1/2 •

o PSI 19873 Dubna19964 TRIUMF1996• this w o *

0 100 200 300 400TEMPERATURE [K]

Figure 1: Measured rates of cfytd-formation in D2

(overview). The curves are a preliminary fit by theory.

50

' 40

30

20

10

A.A0AMC2AK*tol....--'

....--"#'=0.61(1)

HI. FAIFMAN THEORY

50 100 150 200 250 300TEMPERATURE (K)

Figure 2: New results on ßd spin flip rates in pure deu-terium. The fit (solid line) is a linear combination of thetwo theory contributions.

17

KINETIC ENERGY OF 7T~p-ATOMS IN LIQUID AND GASEOUS HYDROGEN

R-95-03, PSI - ETH ZURICH - ZURICH - VIRGINIA

A. Badertschert, M. Daum*s, P.F.A. Goudsmitt, M. Janouscht, R-R. Kettle*, J. Koglin**, V. Markushin*,J. Schottmuller*§, Z. G. Zhao*

* PSI, Paul Scherrer Institut, CH-5232 Villigen-PSI, Switzerlandf IPP, Institut fur Teilchenphysik der ETHZ, CH-5232 Villigen-PSI, Switzerland% Institute of Nuclear and Particle Physics, University of Virginia, Charlottesville§ Physik-lnstitut der Universitat Zurich, Winterthurerstr. 190, CH-8057 Zurich, Switzerland

Our recent measurement of the Doppler broaden-ing in the neutron time-of-flight (TOF) for the reactionn~p ->• 7T° + n confirmed the existence of 'high energy'components (Tnp > 1 eV) in both liquid and gaseoushydrogen, as well as, producing evidence for Coulombde-exciation transitions [1] with An = 2 in liquid.

The measured TOF-spectra were fitted by distrib-utions generated by a detailed GEANT Monte Carlosimulation which accounted for the stopping distribu-tion in the target, geometric effects and neutron scat-tering. The results of the fits are presented in Table 1.Here Tn>6 is the upper energy bound for all Coulomb

Transitionn —> n'

n > 66-^55-+44->33-J-2

6-*45->3

X=7DOFC.I. [%]

TiTn>6

LH2

27 ±29± 17± 17± 13±18± 13±10.96111

1.0±0.17.6±0.3

H2 Gas19±59±37±45±35±19±40t2

00.9672.2

1.6±0.26.7±2.2

'[%]LH2

26 ± 27± 112± 114± 14±1

//

1.0613.1

1.0±0.17.9±0.4

H2 Gas21 ± 56 ± 314±310±24± 1

//

0.9672.3

1.5±0.36.8±2.3

Table 1: Fitted yields Ann> of Coulomb de-excitationpeaks in the kinetic energy distribution f{Twp) for thetransitions n ->• n' in liquid and gaseous hydrogen. Fitresults including An = 2 (left) and An = 1 only (right).

de-excitation transitions with n > 6 in liquid hydrogen.In the case of gaseous hydrogen Tn>6 is the mean en-ergy of these transitions.

The discrepancy between the predicted and the ob-served transition energies Tnn> reported in Ref. [2] canbe accounted for and made to vanish in our presentanalysis if we include the An = 2 Coulomb de-exci-tation transitions. Coulomb de-excitation is at presentthe only cascade process known that would lead to asubstantial kinetic energy increase of the pionic hydro-gen atom in a collision with a proton of the surroundinghydrogen. In addition, a method which makes no as-sumptions about the transition energies was used toobtain the kinetic energy distribution f(%p). Here, thedata measured in liquid and gaseous hydrogen werefitted using a kinetic energy distribution consisting of 16energy bins from T^x to T, with T{ = i2 (i = 1 , . . . , 16).The resulting kinetic energy distributions are shown inFig. 1.

64 81 100121144169196225256

Figure 1: Model independent kinetic energy distri-bution f(Twp) for liquid and gaseous hydrogen ob-tained by fitting 16 kinetic energy bins to the measuredTOF-spectra. The number n -» n' indicate the posi-tions of the theoretical values of the kinetic energy forthe Tr-p -atoms derived for the Coulomb de-excitationprocess. The errors of the yields are indicated by verti-cal bars.

REFERENCES

[1] L. Bracci et al., Nuovo Cimento 43A (1979) 9.

[2] A. Badertscher et al., Phys. Lett. B 392 (1997)278.

18

DOPPLER BROADENING OF GAMMA RAYS FOLLOWING MUON CAPTURE: SEARCHFOR SCALAR COUPLING

R-97-03, DUBNA - LOUVAIN - PSI - ORSAY

V. Egorov*5, Ch. Briancont, V. Brudanin*, J. Deutsch*, T. Filipova*, J. Govaerts*, C. Petitjean§, R. Prieels*,Yu. Shitov*, Ts. Vylov*, V. Wiaux*, I. Yutlandov*, Sh. Zaparov*

* Joint Institute for Nuclear Research, 141980 Dubna, Russiaf Centre de Spectrometrie Nucleaire et de Spectrometrie de Masse, 91405 Orsay, Francei Universite Catholique de Louvain, B-1348 Louvain-la-Neuve, Belgium§ Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland

Some scenarios of leptoquark exchange beyond theStandard Model imply the presence of a genuine scalarcoupling Cs which would contribute to observables inmuon capture summed with the induced scalar cou-pling gs (expected to be small): Cs + gs- Experimentallimits on the scalar coupling deduced from muon cap-ture rates on hydrogen and 3He depend on the valueassumed for the induced pseudoscalar form factor gP.Therefore, experiments insensitive to this assumptionare called for.

Instead of absolute capture rates, our experimentobserves the angular correlation in the first forbiddentransition of ordinary muon capture (OMC) on oxygennuclei and is independent of gp\

ß-+16O{0+) °^% 16N*{l-,mkeV) +isß (1)| 7— transition

16N (0-, 120keV) +-,'{277keV)

The successful test [1] allowed us to carry out themain experiment with an improved setup and methodof measurement:

• the measurement was done in O2 gas at nor-mal pressure with a new low mass gas container andan entrance window of 3mg/cm2 of aluminized mylar.This reduced both eventual background in the 277 keVregion and bremsstrahlung background produced bymuon decay electrons in the surrounding materials;

• a new construction of scintillator counters mounteddirectly in the target resulted in the decrease of mate-rials crossed by the muons before their stop in O2. Itallowed us to stop 96% of the incoming muons in thegas. Besides, at this pressure, the recoil slowing downeffect is negligible;

• 3 co-axial 100 ccm HP Ge detectors similar to theone which provided the best result in the previous run,were used in the experiment;

• the energy signal from each detector was regis-tered by two independent amplifiers and ADCs with dif-ferent energy ranges and shaping times. This allowedboth, to monitor a wider part of the energy spectrum forthe identification of background lines, and to improvethe reliability of the electronics. The response func-tions of all 3x2=6 channels were tuned to be as closeas possible to a Gaussian. The energy resolution wasconsidered as the second priority;

• the data were stored in a simplified format (sev-eral spectra instead of "event-by-event" structure). It

CalibralionSourea 1 6 9 Y b

Delayed 7 -spectrum

following the OMC:

Gamma-Energy, keV290 300 310

Uncorrelated-1 (beam ON)

Uncorrelated-2 (beam OFF)

Promptl_

Figure 1: A portion of the measured 7-spectra includ-ing the line-of-interest, as well as, some backgroundspectra.

allowed us to decrease the dead-time and to increasethe beam intensity;

• the "uncorrelated" background spectra were mea-sured simultaneously with the "correlated" spectra andare used for the background monitoring and subtrac-tion;

• in order to optimize the off-line data analysis, the"delayed" events were grouped into three parts ("fast","middle" and "slow" spectra). During 3 weeks of mea-surement we collected 4.5-105 events in the 277 keV7-line.

The run performed under these improved conditions,allows us to suppress possible sources of systematicerrors and to control the remaining ones through thedata analysis which is in progress now.

REFERENCES

[1] V.Egorovet al., PSI Ann. Rep. NPP (1997) 30.

19

INVESTIGATION OF THE METASTAB ILITY OF THE 2S-STATE IN MUONIC HYDROGEN

R-97-04, MUNICH - ETH ZÜRICH -PSI

F.J. Hartmann*, P. Häuser*, F. Kottmannt, G. Llosa*. C. Petitjean1, R. Pohl1, W. Schott*, D. Taqqu*

* Technische Universität München, D-85747 Garching, Germanyf Institut für Teilchenphysik, ETHZ, CH-8093 Zürich, Switzerland| Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland

In view of a new experiment proposed at PSI tomeasure the r.m.s. proton charge radius, by determin-ing, the Lamb Shift in muonic hydrogen ^p, it is of vitalimportance to have a good knowledge of both the pop-ulation and lifetime of the metastable ^p(2s)-state. Inorder to obtain this information we performed an exper-iment where we investigated the muon transfer to neonin a mbar gas mixture of hydrogen with several percentof neon.

Our setup is shown in Fig. 1. The JTE5 muon beamis compressed in the inhomogenous stray field of the5 Tesla superconducting PSC solenoid. The muonsare slowed down to kinetic energies below 100 keV, de-tected in the parallel plate avalanche chamber (PPAC)and pass the stack of 6 ultra-thin carbon foils. In thestack the muons are, on the one hand, cooled by 'frac-tional cooling1 [1] and on the other hand eject electronswhich are detected in the micros-sphere plate (MSP).Thus, the stack also serves as a second entrance de-tector. Finally, the muons are stopped in a 15 cm longcylindrical target with diameters of 7, 12 and 16 mm.The target is filled with 4, 8, or 16 mbar of hydrogenand a 3, 6,12, 25, or 50% admixture of neon.

PMT

Figure 1: Setup of the experiment. PPAC: parallel plateavalanche chamber, MSP: micros-sphere plate, PMT:photomultiplier tube.

We measured the muonic neon K-series X-rays fol-lowing the transfer reaction

/Ltp(ls or 2s) + Ne —> p + juNe* (1)

exploiting the fact that the transfer signal of /j.p(2s) toneon is enhanced by two effects: the transfer rate of^p(2s) to neon is about an order of magnitude largerthan that for /ip(1s) [2, 3], and the /up(1s) vanish intothe walls of the target chamber quickly due to the smallelastic cross section for scattering of jup(1s) on H2.

We also recorded the signal of the /ip(1s) arrivingat the target chamber walls which were coated with athin layer of gold:

This way we obtained valuable information on the timeevolution of the //p(1s).

The time spectra of both the neon and the goldtransfer reaction were compared with detailed MonteCarlo simulations including knowledge of the initial ki-netic energy distributions of muonic hydrogen [4], scat-tering of /*p(1s) on H2 and the measured transfer rateof /up(1s) to neon at near-thermal energies. We finda number of excess events at late times which canneither be attributed to 1s-transfer nor to background(which we measured carefully in pure H2, He, and Negas). The use of the various pressures, concentra-tions and target diameters allows multiple cross-checksof the hypothesis, that these excess events have tobe attributed to the presence of long-lived metastabley«p(2s), to be made. A lower limit (95% C.L.) of 0.4%for the fraction of long-lived metastable //p(2s) (per ^pformed) results. A more detailed analysis of the data ispresently being performed.

REFERENCES

[1] M. Mühlbauer et al., Hyper. Int. 101/102, 607(1996).

[2] R. Jacot-Guillarmod, Phys. Rev. A51, 2179(1995).

[3] S. Romanov, private communication.

[4] F.J. Hartmann et al., R-94-04, this Report.

+ Au —> p + (2)

20

PRECISION MEASUREMENT OF THE SINGLET fip CAPTURE RATE VIAHYDROGEN GAS

LIFETIME IN

R-97-05.1, GATCHINA - PSI - BERKELEY - MUNICH - LOUVAIN - SACLAY - MOSCOW - KRAKOW

A. Adamczak**, V.A. Andreev*, D.V. Balin*, T. Case*, K.M. Crowe*, J. Deutsch11, P.U. Dickt, A. Dijksman*,J. Egger*, T. von Egidy§, M.P. Faifman**, A.A. Fetisov*, V.A. Ganzha*, J. Govaerts^, F.J. Hartmann§, W.D. Herold*,V.l. Jatsioura*, P. Kammel*5, A.G. Krivshich*, E.M. Maev*, V.E. Markushin*, J. Martino**, C. Petitjean*s,G.E. Petrov*, L.I. Ponomarev**, R. Prieels11, S.M. Sadetsky*, G.N. Schapkin*, R. Schmidt*, W. Schöps*,W. Schott§, G.G. Semenchuk*, A.A. Vorobyov*, N.I. Voropaev*

* Petersburg Nuclear Physics Institute, Gatchina 188350, Russiat Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandt University of California and Lawrence Berkeley Laboratory, Berkeley CA, 94720, USA§ Technische Universität München, Physik Department, D-85748 Garching, Germanyf Universite catholique de Louvain, B-1348 Louvain-La-Neuve, Belgium

CEN Saclay, F-91191 Gif-sur-Yvettes CEDEX, FranceMUCATEX, Moscow 123479, RussiaInstitute of Nuclear Physics, 31-342 Krakow, Poland

**ft

Nuclear muon capture, JJ. + p —> n + vß, is a fun-damental electroweak charged-current reaction whichcan be studied at PSI's low energy muon beams underideal conditions. The capture rate is sensitive to theweak form factors of the nucleon, in particular to theinduced pseudoscalar coupling constant gP. A precisemeasurement of gP would rigorously test modern ef-fective field theories of low energy QCD, which predictQP with ~ 2 % precision [1]. This goal can be reachedby a pr lifetime measurement in H2 gas to a precisionof 10"5 or better.

' i

' umSfi *

Figure 1: First setup of the prototype TPC in area

Based on the development of three small test cham-bers in 1997 [2], a full size prototype time projectionchamber (TPC) was constructed at Gatchina and shippedto PSI for first tests in the muon beam (Fig. 1). The TPChas a sensitive volume x-y-z = (15 • 8 • 30) cm3 with avertical electrical drift field of 2.2 kV/cm. At the beamentrance two x-y MWPCs (nr. 1,2) are placed to localizearriving muons and below the TPC, two (15x15) cm2 x-z MWPCs (nr.3,4) act as external electron detectors.All chambers sit inside a pressure vessel filled with 10bar of pure hydrogen gas.

In the TPC, the ionization electrons from tracks ofmuons and muon decay electrons drift towards a MWPC

plane of 75 anode wires, where electronic signals aregenerated with a gas gain of ~ 104. The idea of using aTPC is to unambiguously identify the tracks of stoppingmuons and their decay electrons without using globalpileup rejection, thus allowing eventually to reach beamrates up to 50 kHz required for high statistical accuracy.

•x

1

JtSC ' "JÜLJ/L _

—- ft:— -1 ' 1

-JA

., - A-

i i i

Figure 2: Sample of a tracked ^-e event in the TPC

Figure 2 shows an event from our first test run inDecember 98. The muon can be seen passing the en-trance wire chambers (bottom cross-hair) and stoppingwith high energy loss on anode 6 (time scale in /J,S).The signals on anodes A1-A6 are saturated and scaleddown eightfold with respect to A7-A12. The decay elec-tron is seen on anodes 7-12 going to the upper left,where it then exits the TPC and hits MWPCs 3+4 asshown by the top cross-hairs. The negative (clipped)pulses on the wires at the muon signal are cross talkfrom the muon signal.

REFERENCES

[1] V. Bernard et al., Phys. Rev. D50 (1994) 6899;H.W. Fearing et al., Phys. Rev. D56 (1997) 1783.

[2] PSI Annual Report 1997 / Annex I, p. 32.

21

MEASURING ^<23He FUSION.

R-98-02, FRIBOURG - JINR - PSI - MUNICH - CRACOW

A. Del Rosso*, M. Augsburger*, V.F. Boreiko^, V.M. Bystritskyt5, W. Czapliriski', C. Donche-Gay*, M. Filipowicz1',O. Huof, P. Knowles*, F. Mulhauser*s, V.N. Pavlov*, F.M. Pen'kov*, C. Petitjean*, N.P. Popov§, V.G. Sandukovskyt,L A. Schaller*, H. Schneuwly*, V.A. Stolupint, J. Wozniak^

* Institut de Physique, Universite de Fribourg, Perolles, CH-1700 Fribourgt Joint Institute for Nuclear Research, Dubna, Russia| Paul Scherrer Institut, CH-5232 Villigen PSI§ Sektion Physik der Universität München, München, Germanyf Institute of Physics and Nuclear Techniques, Cracow, Poland

Nuclear fusion from /id3He may occur via two dif-ferent channels: emission of a proton with an effectiverate \f or emission of a hard 7

/Kd + 3He - ^ fi + a(3.66MeV)+p(14.64MeV) (1)

//Li + 7 + 16.39 MeV. (2)

The predictions for the reaction rate of Eq.(1) differ byan order of magnitude and the existing experimentaldata are insufficient to resolve the dispute [1, 2, 3, 4].The reaction of Eq.(2) is expected to proceed with abranching ratio on the order of ~ ICH4*5) [5], i.e., fourorders of magnitude smaller than A/.

We performed a first test run in October 1997 inorder to evaluate the background and the sensitivity ofour experiment. The results obtained were good (cf.[6J)and subsequent improvements were made. We testedthe new equipment at the/iE4 channel during Fall 1998using both pure D2 and the gas mixture D2+3He (c3He ~5%, target temperature ~ 32 K at <j> = 7.8% of LHD).

The target and the detectors are schematically sh-own in Fig. 1. Of note are the gas volume T (~66 mminner diameter) and the three pairs of silicon detectors(0.36 and 3.65 mm thick respectively) installed aroundit, just behind Kapton windows (55-135 fxrr\ thick to al-low the radiation of interest to reach the detectors; notshown in the Figure). Details of the target and the gassupply system are found in Ref. [7].

Muons of interest passed a lead collimator, fired anentrance scintillator and stopped in the gas volume.Four plastic scintillators, EUP, EDO, ERI, and ELE de-tected the decay electrons and were also used as acharged particle veto in conjunction with the neutronand 7 detectors.

The thin silicon detectors give the energy loss, dEjdx, for the particles reaching the thick detectors. Theenergy loss and total energy allow us to determine theparticle type and hence select the protons of interest.The y«d3He fusion proton at 14.7 MeV will be finallystopped and detected by the thick silicon detectors.The data will be cleaned by requiring that the decayof the muon causing the event of interest be seen in agiven time interval after the event, which proves that themuon was not destroyed by the event of interest. Thisdele condition is a very powerful tool in the suppressionof background coming from muon nuclear capture inthe target walls.

The energy and time distribution of the 6.85 keVX-rays emitted by the de-excitation of the ßd3He mole-cule has been measured by the Ges germanium detec-

Figure 1: The arrangement of the equipment used inthe fxEA area shown from the perspective of an arrivingmuon.tor. This permits us to measure the fxd3 molecular for-mation rate in the Do+3He mixture and to monitor theother muonic X-rays providing essential information onthe purity of the target. The GeB (122 cm3) was usedto detect hard 7 (cf. Eq.( 2)) and the Ge^ (75 cm3) tomonitor impurities. The neutron detector NE213, hasbeen used to monitor the drf-fusion cycle.

In two weeks of beam time we collected a total of~ 1O10 good muons, i.e., muons with a good entrancedefinition and no pileup. The analysis will focus on thesilicon detector performance and results as well as the6.85 keV time distribution. We will be studying alsohard 7-rays, neutrons and possible impurities in thegases. Preliminary results of the analysis are foreseenin late Spring 1999.

REFERENCES

[1] D. V. Baiin et al., Muon Catal. Fusion 7,1 (1992).

[2] P. Ackerbauer et al., PSI Annual Report Annexe I,32(1996).

[3] D. V. Baiin et al., Gatchina Preprint 2221 NP-7(1998).

[4] E. M. Maev et al., accepted for publication in Hy-perfine Interact. (1999).

[5] F. E. Cecil et al., Phys. Rev. C 32, 690 (1985).

[6] A. Del Rosso et al., accepted for publication in Hy-perfine Interact. (1999).

[7] V. F. Boreiko et al., Nucl. Instrum. Methods A 416,221 (1998).

22

CPAA AS AN ANALYTICAL TOOL IN THE BIOLOGICAL FIELD

Z-94-07.1, MILAN - NEUHERBERG

M.C. Cantone*5, D. de Bartolo*, A. Giussani*, P. Roth*, E. Werner*

* Dipartimento di Fisica, Universitä degli Studi di Milano, and INFN, Via Celoria 16, 1-20133 Milano, Italyf GSF-Forschungszentrum für Umwelt und Gesundheit, Institut für Strahlenschutz, D-85764 Neuherberg

The year 1998, has seen the continuation of thestudies of biokinetics and metabolism of Mo, Zr andRu, elements of interest in the field of nutrition andradiation protection. As explained in previous annualreports, these studies are performed using stable iso-topes as tracers and activation analysis with chargedparticles (protons) as the analytical technique for thedetection of the tracers in biological samples.

A new series of tests on the biokinetics of molyb-denum have been started, which aim to further investi-gate some peculiarities found in the first preliminary in-vestigations. In these studies a larger group of healthyvolunteers have been recruited. The results have givencrucial information about molybdenum biokinetics in hu-mans. It was shown that the current model for ingestionof molybdenum radionuclides recommended by ICRPcontains several inaccuracies and requires some mod-ifications [1] An improved description of the excretionprocesses and a better evaluation of the transfer coef-ficients enable a successful reproduction of the exper-imental measurements, while maintaining the generalfeatures of the general ICRP biokinetic model. The useof separate sets of parameters for solids and liquidsis recommended. The improved realism of the modelmeans a more reliable application in the field of radia-tion protection for the evaluation of the consequencesof accidental incorporations of molybdenum radionu-clides [2] and a more reliable application in the field ofnutrition. The biological samples are also measuredat GSF, using mass spectrometric techniques (ICPMSand TIMS), but up to now only activation analysis canachieve the required accuracy and precision for tracermeasurements in blood plasma samples.

After completion of the animal experiments[3], theanalysis of the samples from preliminary studies on thebiokinetics of zirconium and ruthenium in humans hasbeen started.

The activation analysis technique has also been tes-ted for the detection of two stable Gd isotopes in plasmasamples. The interest in Gd relies on the fact that it canbe used as an analogue of selected actinides in bioki-netic studies. Both (p,n) and (p,2n) reactions on theisotopes Gd-154, Gd-155, Gd-156 and Gd-157 fulfillour analytical requirements, i.e. no reciprocal interfer-ence of gamma lines from the radionuclides producedand also no interference with the matrix background,reasonable irradiation and measuring times requiredfor a satisfactory sensitivity etc.

During the year we have also carried out some mi-nor repairs and made improvements to the activationchamber which is coupled to the cyclotron beam chan-nel. Firstly, the portion of the rotating disc where thecooling liquid flows, which is necessary to spread theheat released in the samples by the incoming beam,

has been replaced. Secondly, new, more reliable tem-perature sensors are now employed to monitor the in-coming and outcoming cooling fluid, and lastly, NTCsensors are embedded in standard blood plasma sam-ples to have an estimation of the temperature variationin the samples during the irradiation. Some improve-ments are still needed for the signal transmission ofthe sample temperature to the control room during theexperiments.

Finally, in 1998 the official results of the IAEA in-tercomparison study on the determination of toxic andother main and trace elements in algae have been pre-sented [4]. Three sets of algae, with different enrich-ment in toxic elements (IAEA-391 low level, IAEA-392environmental level and IAEA-393 elevated level) weredistributed for analysis. The (p,n) reaction was utilizedfor the determination of As, Ca, Cr, Fe, Hg, Zn; (p,2n)reaction for Cd, Cu and (p,3n) for Pb. Our results arewell in line with those of the other laboratories involved.We would like to kindly acknowledge the support givento us by Dr. J.Gobrecht and his co-workers of the NanoLab for the use of their clean rooms during samplepreparation.

REFERENCES

[1] A.Giussani, M.C.Cantone, D.de Bartolo, P.Rothand E.Werner, Health Phys. 75 (1998), 479.

[2] A.Giussani, P.Roth,E.Werner, H.Greim, M.C.Cantone and D.de Bar-tolo, Radiat. Prot. Dosim. 79 (1998), 367.

[3] M.C.Cantone, D.deBartolo, A.Giussani, L.Garlaschelli, Ch.Hansen,RRoth, E.Werner, J.Radioanal. Nucl. Chem. 217(1997)279.

[4] Preliminary statistical evaluation of the intercom-parison run IAEA-0390, IAEA/AL/101P(1998).

23

PRECISE MEASUREMENT OF THE NEUTRON-PROTON SPIN CORRELATIONPARAMETER AT FORWARD ANGLES

-0.1

-0.8100

[degrees]ISO

Z-96-02.1, BASEL - PSI

S. Buttazzoni*, C. Carasco*, M. Hauger*, P. Hautlet, A. Honegger*, J. Jourdan*, J.A. Konter*, G. Kubon*,S. Mango*, M. Mühlbauer*, H. Mühry*, Th. Petitjean*, D. Rohe*, I. Sick*, B. van den Brandt*, G. Warren*,H. Wöhrle*5

* Department of Physics and Astronomy, University of Basel, Basel, Switzerlandt Paul Scherrer Institut, Villigen, Switzerland

The effect of the isoscalar tensor force, a funda-mental component of the nucleon-nucleon force andrelevant for nuclear structure studies, is the mixing ofthe 3Si and 3Di-state. In a phase shift analysis, it is de-scribed by the mixing parameter s\. Its precise knowl-edge is crucial for the understanding of the structure ofnuclear systems. The observable most sensitive to ex

is the spin correlation parameter Azz at forward angles[1]. At these angles the correlation of ex with the b i -phase - which is also poorly known - is minimal (seefig. 2).

Here we report on the first measurement of Azz us-ing a new technique. In contrast to our previous exper-iments [3] we used a newly developed scintillating po-larized target which enabled us to measure the energyof the recoil proton. The detection of the scattered neu-tron in coincidence with the recoil proton strongly re-duces the background contribution in the neutron time-of-flight spectra and allows the determination of Azz

with much higher accuracy (compared to the previousexperiments [3]), only limited by the decreasing recoilenergy with decreasing angle. Due to the progressmade over the last year the scintillating target material(TEMPO doped polystyrene based scintillator) is nowavailable in blocks of up to (5 x 18 x 18) mm3 with pro-ton polarizations up to 80 % [2].

The first measurement using such an active targetwas performed in August 1998 at the PSI injector cy-clotron. The scattered neutrons were detected withan array of plastic counters placed 2 m from the tar-get, allowing the measurement over a wide range ofscattering angles simultaneously. Figure 1 shows theenergy of the recoil protons for different scattering an-gles t?cm = 34 to 57°. Each of the spectra containsapprox. 3 -105 events, about one third of the events ofthe final data.

From the number of events N + and N~ for the two

Figure 2: Neutron-proton spin correlation parameterAzz as a function of the scattering angle tfcm.

spin states the asymmetry

A - l N+

2000

0800200 400

proton energy [chn]

Figure 1: Energy of the recoil protons for different neu-tron scattering angles ticm.

Pn pT N+ + N- - 2iVB

can be calculated, where yVB is the background con-tribution and pn and pT denote the neutron and targetpolarization, respectively. The neutron beam polariza-tion was flipped every 3 s in order to minimize system-atical errors. It is known to ±1.1 % [4], and the targetpolarization was measured to ±4 % using nuclear mag-netic resonance. Preliminary results are shown in fig.2 together with the previously measured values [3, 5].The dotted and dashed lines show the variation of Azz

caused by a 1° variation of lP\ and t\. The error barsshow the statistical errors estimated for the full data set.The systematic effects are being studied in detail in theongoing analysis. So far the new results lie well withinthe error bars of our previous values [3], but systemat-ically above them.

The preliminary evaluation of the data taken so fargives very promising results and already shows that ex-periments on spin observables like Azz gain from us-ing an active polarized target. We plan to continue ourmeasurements at smaller angles yielding new insightson Azz and ei at the most forward angles (see fig. 2).

REFERENCES[1] J. Binstock et al., Phys. Rev. D9 (1974) 2528.

[2] B. van den Brandt et al., Proc. 12th Internat. Symp.High Energy Spin Physics (SPIN96), Amsterdam,The Netherlands, Sept 10-14, 1996, p. 238.

[3] B. Zihlmann et al., PSI Annual Report (1995) 83.

[4] M. Zeier et al., PSI Annual Report (1997) 93;M. Zeier et al., submitted to Nucl. Phys. A.

[5] M. Hammans et al., Phys. Rev. Lett. 66 (1991)2291.

24

THE CPLEAR EXPERIMENT

PS-195, THE CPLEAR COLLABORATION

P. Bargassat, F. Blanc*T, P.-R. Kettle*, T. Nakada*, O. Wigger*

* Universite de Fribourg, CH-1700 Fribourgt Paul Scherrer Institut, CH 5232 Villigen PSI

The CPLEAR experiment was designed to studyCP- and CPT-symmetries in the neutral kaon system.The neutral kaons were produced in the pp annihilationat rest:

K - T T + K 0

each with a branching fraction of ~ 2 x 1O~3. The initialflavour and momentum of the neutral kaon are mea-sured by the accompanying charged particles.

The main data taking of the CPEAR experiment tookplace from 1991 to 1995. In 1996, a special run wasmade to investigate the effect of regeneration.

A detailed description of the detector can be foundelsewhere [1]. During the 1994 data taking, a small up-grade was made to the detector. The spherical targetfilled with 15 bar H2 gas (radius 7 cm) was replacedby a cylindrical one with a radius of 1.1 cm and filledwith 27 barH2 gas. A cylindrical multi-wire proportionalchamber with a radius of 1.5 cm was added aroundthe new target. This was used in the trigger to rejectbackground, resulting in a smaller dead time which in-creased the rate of recorded signal events by a factorof two.

As described above, the initial flavour of the neu-tral kaon is identified by the charge of the accompa-nied kaon at the production point. When the neutralkaon decays semileptonically, its flavour at the decaycan also be identified by the charge of the lepton fromthe decay point. Since a K° (K°) contains s (s) quark,the decay produces a positively (negatively) chargedlepton.

By identifying the both initial and final flavours, thefollowing probabilities can be measured:

• initial K° oscillates to K° at time t: P_ (i),

• initial K° remains as K° at time t: P+{t),

• initial K° oscillates to K° at time t: P+(<),

• initial K° remains as K° at time t: P_ (t).

In the CPLEAR experiment, semileptonic decays withe+ and e~ are used.

P_ (t) can be obtained by applying a time reversal(T) or CP-transformation to P+(t). Similarly, P+(t) canbe obtained by applying a CPT- or CP-transformation toP_ (t). Therefore, any deviation of the time dependentasymmetry AT(<)» defined as

P+(t)-P_(t)

Figure 1 shows the measured AT(t) by the CPLEARexperiment [2] which clearly indicates that Ar{t) ^ 0.This is the first time that violation of the time reversalsymmetry has been directly observed. T-violation hasso far been shown only indirectly in the neutral kaonsystem by applying the Bell-Steinberger relation, whichis derived from unitarity.

-o.oi E-

-0.028 10 12 14 16 18 20Neutrol-kaon decoy time ( TS ]

Figure 1: Measured asymmetry between the probabili-ties that an initial K° oscillates to a K° and that an initialK° oscillates to a K°, as a function of the decay time.

asIn a similar way, any deviation of of Acpr(t), defined

C P T U P ( t ) + P ( t )from 0, is a signal of CP- and CPT-violation. Figure 2shows the measured ACPT(<) by the CPLEAR experi-ment [3]. It shows no sign of CPT violation.

0.1

0.08

0.06

0.04

-0.02-0.04

-0.06

-0.08n 1

I ,

D.I 2 4 6 8 10 12 14Neutral—kaon decaytime

iJ

\

16

f

18

[ T j

from 0, indicates that CP and T are violating in the K°-K° oscillation.

Figure 2: Measured asymmetry between the probabili-ties that an initial K° remains as a K° and that an initialK° remains as a K°, as a function of the decay time.

REFERENCES

[1] R. Adleret al., (CPLEARCollaboration) Nucl. Inst.A379(1996)76.

[2] A. Angelopoulos et al., (CPLEAR Collaboration)Phys. Lett. B444 (1998) 43.

[3] A. Angelopoulos et al., (CPLEAR Collaboration)Phys. Lett. B444 (1998) 52.

25

BROOKHAVEN EXPERIMENT E865: STUDIES OF RARE AND SEARCHES FORFORBIDDEN ÜT+-DECAYS

E-91-02.1, BNL - NEW MEXICO - PITTSBURGH - YALE - INR MOSCOW - BASEL - PSI - ZÜRICH

G.S. Atoyan**, V. Atoyan*, R. Appell K. Baker*, B. Bassalleck™, E. Battiste*, D. Bergmann^, S. Dhawan§, H. Do§,S. Eilerts™, J. Egger*, C. Felder*, H. Fischer™, M. Gach*, W.D. Herold*, V.V. Isakov**, H. Kaspart§, D.E. Kraus*,D.M. Lazarus*, L. Leipuner*, J. Lowe™, J. Lozano§, H. Ma*, W. Menzel, S. Pislak**§, A.A. Poblaguev**,V. Postoev**, A. Proskurjakov**, P. Rehak*, P. Robmann **, A. Sher *, M. Shubin*, J.A. Thompson*, P. Truöl**§,A. Walid§, H. Weyer^t, D. Wolfe™, M.E. Zeller§,

* Brookhaven National Laboratory, Upton, N.Y. 11973Paul Scherrer Institute, CH-5232 VilligenUniversity of Pittsburgh, Pittsburgh, PA 15260Yale University, New Haven, Connecticut 06511University of Basel, CH- BaselUniversity of Zürich, CH- ZürichUniversity of New Mexico, Albuquerque, NM 87131Inst. for Nuclear Research, Moscow

t

**ttT +

The broad band charged particle spectrometer andparticle identification system of the E865 experiment iswell suited for the study of rare K+ decays into mul-tilepton final states [1]. The final long data taking runfor this experiment, at the Brookhaven AGS ended onnew years eve 1998. A sufficiently large number ofevents were recorded to reach a sensitivity at the 10~n

branching ratio level for the lepton number violating de-cay K+ ->• n+n+e~, which was the primary goal of thisexperiment. The 1998 run nearly tripled the statistics,which can be attributed to, an upgrading of the dataacquisition system, and an increase in primary protonbeam intensity and spill length. The analysis of thedata has only recently started.

During our shorter 1997 run a scintillator pixel de-tector was installed in the primary beam, which permit-ted decays with missing neutrinos or photons to also belooked at. With dedicated trigger settings we recordedlarge event samples (see [1]) for Ke3 decays (K+ -»n°e+pe, to improve the accuracy on Vus, comparingKe3 to Kw2 (K+ -> TT+TT0)), for the radiative decaysK+ -» /j,+i/ße+e~ and K+ -> e+i/ee

+e~ (probing theA'+ electromagnetic structure), for K+ ->• 7r+e+e~ andK+ -> ir+fi+fi~ decays (flavor changing neutral currentdecays dominated by long distance effects and calcula-ble in chiral perturbation theory) and lastly Ke4 decays(K+ -> n+ir~e+ve). As an example of the quality ofthe 1997 data, we show in Figure 1 a mass spectrumindicating a clear peak from about 400 K+ -»7r+^+fi~events.

The Ke4 (K+ ->• ir+TT-e+i;e) decay is the only avail-

able source of clean information on mr S-wave scat-tering near threshold and has for this and other rea-sons some special significance in the theory of chiralsymmetry. Figure 2 shows preliminary results from ananalysis of these data (300000 events).REFERENCES

[1] G.S. Atoyan et al., PSI Annual Report/Annex 1(1995) 89; ibid (1997) 99.

[2] L. Rosselet et al., Phys. Rev. D15 (1977) 574.

[3] J. Gasser, and H. Leutwyler, Phys. Lett. B125(1983)325.

E865, K: ••••> i ' i i V , 1S97 DATA

Figure 1: irfifi invariant mass spectrum after appropri-ate likelihood cuts. The exponentially decreasing back-ground arises from K+ ->• 7r+7r~7r+ decays with a sub-sequent TT* ->• Ju

±i//J(i7/() decay.i • • • i

. , i

0.32 0.34 0.36 0.38 0.4M„ [MeV]

Figure 2: TTTT s-wave scattering phase shift SQ as a func-tion of the 7T7T invariant mass. The triangles and theupper line refer to the data of Rosselet et al., [2], thecircles and the middle line represents the best fit to thepreliminary analysis of this experiment. The solid lowerline corresponds to the chiral perturbation theory pre-diction [3].

26

THE PIXEL DETECTOR AND ELECTROMAGNETIC CALORIMETER OF CMS

E-94-09, CMS COLLABORATION

R. Baur*, W. Berti*, K. Deiters*, R. Eichler*, K. Gabathuler*, J. Gobrecht*, G. Heidenreich*, B. Henrich*,R. Horisberger*, Q. Ingram*, R. Kaufmann*, D. Kotlinski*, R. Morf*, D. Renker*, R. Schnyder*, D. Zuercher*,

* Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland

The Compact Muon Solenoid (CMS) is one of thetwo large multi-purpose detectors to be built at CERN/LHC. PSI contributes to the tracking (pixel detector)and electromagnetic calorimeter (avalanche photo-dio-des and crystals) sub-detectors.

Pixel Detector:

PSI has overall responsibility for the innermost partof CMS, a silicon detector with 150 micron square pix-els, and is designing and will build the cylindrical lay-ers of this detector (in collaboration with ETHZ and theUniversities of Aachen, Basel, Vienna and Zürich). Inspring 1998, the CMS tracker community submitted itsTechnical Design Report[1], with chapter 2 describingthe Pixel Detector System. The complete CMS Trackerproject was approved by CERN in autumn 1998.

For the readout of the pixel detector a column drainarchitecture has been chosen after extensive computersimulations. Many crucial functional blocks of the read-out have been implemented on small radiation hardchips which proved to be mostly within the specifica-tions. A 20x30 pixel frontend chip with a simpler read-out architecture, but already equipped with many fea-tures needed in the final chip (e.g. addressing in oc-tal analogue code, calibration circuitry, individual pixelthreshold trimming with programming capabilities, timestamp counter and trigger mechanism), shows goodperformance characteristics.

A bump-bonding technique needed to connect pixelsensor and readout chip has been developed at PSI.All necessary steps in this complicated procedure wereperformed in various laboratories at PSI. Fig.1 showsa section of the 20x30 pixel chip with 20//m diameterIndium bumps. The chip has been bump-bonded to aSi-sensor using a device developed and built at PSI.Fig.2 shows the response of the sensor to 14 keV X-rays when covered with a brass mask.

Electromagnetic Calorimeter (ECAL):

For the ECAL, PSI has overall responsibility for theavalanche photo-diodes (APDs) and takes part in thecrystal and electronics R&D. In 1998, the choice wasmade of the type of APD to be used in the final detector.Two companies had carried out R&D based on differ-ent underlying structures and prepared samples whoseperformance and reliability were evaluated before thedecision was made. Nevertheless, since these APDsare new developments, intensive R&D work is still con-tinuing. A related difficulty is the optical coupling of theAPD to the high refractive index (n=2.2) lead tungstatecrystal: The ideal coupling should also have a high re-fractive index, must be radiation hard, should be trans-parent into the UV, must stick reliably onto the smoothcrystal surface and finally should be simple to use sinceit will be used for over 120,000 APDs. Various mate-rials were examined and a promising candidate with a

refractive index of over 1.6 identified. This is a commer-cial product used as a mounting material in microscopy.Further testing is still needed before a final choice ismade. For the ECAL subdetector as a whole, approvalof the Technical Design Report[2] was a major land-mark of 1998, while the crystal R&D made impressiveprogress in crystal quality and reproducibility, and inthe preparations for the mass production, in both Rus-sia and China.

REFERENCES

[1] CMS Tracker TechnicalCERN/LHCC 98-6.

Design Report,

[2] CMS ECAL Technical Design Report,CERN/LHCC 97-33.

Figure 1: Section of radiation-hard chip with Indiumbumps.

Figure 2: Response of a masked 20x30 pixel array.The area of the elements is proportional to the regis-tered hits.

27

H1 -EVENTS WITH AN ISOLATED LEPTON, LARGE HADRONIC JETS AND MISSINGENERGY

H1 COLLABORATION

R. Eichler*, K. Gabathuler, J. Gassner*, R. Horisberger*,

* Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland

In 1997, H1 reported on an excess of events atQ2 > lO'OOO GeV2 which could have been a statisticalfluctuation, however, this type of event was also seenin the ZEUS detector. In the combined data-taking sta-tistics of 1994-1997, we also observed an excess ofevents with an isolated muon, missing transverse en-ergy and a hard hadronic jet, which again could be at-tributed to a statistical fluctuation [1].

Figure 1 shows the distribution of events with a miss-ing transverse energy greater than 25 GeV and a highmomentum charged particle with transverse momen-tum larger than 10 GeV. Plotted is Djet:

(</>track ~

the distance of the high momentum track to the clos-est jet and Dtrack, the distance to the nearest track.Six events appear to be isolated from the rest of theevents and it turned out that all are leptons, one is anelectron and five are muons (isolated positrons weredeliberately excluded from the analysis).

Figure 2 shows the distribution of the 6 exotic lep-ton events in the plane of transverse hadronic energyversus transverse lepton-neutrino mass, together withthe event distribution, as expected, from SM-processes(T^-production, 77 events). The total number of fivemuon events is larger than the 0.8 ± 0.2 expected fromSM-processes, the latter dominated by W-production.It will be very interesting to see whether with larger sta-tistics this is a statistical fluctuation or a sign of newphysics, such as lepton flavour violation e.g. ep ->• rX,suggested in models of R-parity violating processes.

REFERENCES

[1] H1 collaboration, Eur.Phys.JC5 (1998) 575.

•3 4

a3

Figure 1: Correlation between the distances Djet andto the closest hadronic jet and track, for alltracks in the inclusive event sample.

^ 8 0

5

40

20

a)

10 10MTev(GeV)

60 -

40

20 -

Figure 2: Distribution of events in Pjf (transverse mo-mentum of the hadron system) and My" (transversemass of the high pT lepton-neutrino where the mea-sured missing transverse momentum is attributed to ahypothetical neutrino): a) electron channel; b) muonchannel. The crosses indicate the 1-sigma uncer-tainty on the measured kinematic parameters of thesix observed events (for event MUON-5 the 2a lowerlimit for Mf" is shown). The dominant SM contribu-tions (open-circles for W production, triangles for pho-ton-photon processes in the muon channel) are shownfor an accumulated luminosity which is a factor of 500higher than in the data. In the muon channel no signifi-cant contribution is expected at p^ < 25 GeV becauseof the data selection in the trigger.

29

A s t r o p h y s i c s

30

HESSI SUCCESSFULLY QUALIFIED

M. Guedel, R. Henneck, A. Mchedlishvili, P. Ming, K. Thomsen, A. Zehnder (PSI), in collaboration with theSpace Science Laboratory Berkeley, the GSFC, Greenbelt, and the University of Delft

A Qualification Model (QM) of the High Energy Solar Spectroscopic Imager (HESSI) has been built at PSIand tested with success.

HESSI is a NASA mission with a single instrument ona small spin-stabilized spacecraft in low earth orbit.The primary scientific aim of the mission is to explorethe basic physics of particle acceleration and explosiveenergy release in Solar Flares. HESSI will producehard X-ray images with an angular resolution as fineas 2 arcseconds and a temporal resolution in the orderof 10 ms; a detailed image can be obtained in 2 sec-onds. At the same time HESSI will provide energyresolution below 1 keV over the range from 3keV to400 keV. For the first time HESSI will perform hard X-ray and Gam ma-Ray imaging above 100 keV with anenergy resolution of a few keV up to energies as highas 20 MeV.

The imaging capability of HESSI is based on aFourier-transform technique using a set of 9 pairs ofgrids each spaced 1.55 m. Transmission through agrid pair is modulated as the spacecraft rotates aroundits axis. The different grid pairs have different slitwidths, corresponding to 9 partly redundant Fouriercoefficients. For the finest grids an alignment and itsstability to better than 20 arcsec is mandatory to yieldsufficient modulation depth. Given the exact timing ofsingle photons together with the precise knowledge ofthe attitude of the imager it is possible to reconstructan image of the source. To do this no spatial resolutionis needed in the detectors proper, only fluxes have tobe measured.

The energy resolution of the instrument is achieved byusing Ge detectors, which are cooled to liquid nitrogentemperature by means of a mechanical cooler. HESSIwill have the finest angular and spectral resolution ofany hard X-ray or Gamma-ray instrument flown so far.

For PSI the work on HESSI started at the beginning of1998 with the design of a qualification model (QM)imager including the telescope tube, grid trays for bothends of the telescope with all necessary mechanicalparts like fixtures and an adjustable tray mount.

With the encouragement of a series of successful re-views the mechanical design and the production ofalmost all needed components was performed at PSI.In parallel a solar and a roll aspect system (SAS andRAS) have been designed, see accompanying contri-butions, as well as a dedicated optical system to verifythe alignment; the TMS, twist measurement system.

In autumn 1998 a thermal test was performed meas-uring the stability of the alignment between the trayswhile changing the overall temperature between 0 and40 degrees centigrade. Excellent stability was con-firmed. In December 98, the QM assembled with flightrepresentative trays and grids was subjected to vibra-tional qualification tests. All of the hardware under PSIresponsibility performed well. No undue change incharacteristics, especially alignment, has been de-tected.

Fig. 1: HESSI imager QM at Contraves in occasion ofthe vibration test with PSI team.

1620-

220-

210-

]

:

i

I

t

• Tray® Grid 3

11

t

\

l

it

1

t •

Before Vib. After Z-Vibration After X-Vibration After Y-Vibration

Fig. 2: Rotational alignment "twist" before, during andafter the vibration test. Note: 1 arcsec corre-sponds to ~1|im at the circumference of thetray. The allowed twist is 20 arcsec.

31

THE HESSI SOLAR ASPECT SYSTEM (SAS)

J. Bialkowski, M. Fivian, ft Henneck, A. Mchedlishvili, P. Ming, K. Thomsen, A. Zehnder, (PSI), G. Hurford,D. Curtis (UC Berkeley), B. Dennis (GSFC, Greenbelt)

We describe the design and construction of the HESSI Solar Aspect System. It is based on the optical ob-servation of the sun image with 3 line CCDs and it will provide a position determination with accuracy <0.4arcsec.

The Solar Aspect System (SAS) is part of the HESSIinstrument [1] to be launched by mid-2000 at solarmaximum. It will provide continuous information on thepointing of the rotating spacecraft (15 rpm). Preciseknowledge (0.4 arcsec on a 1a level) of the pointing isa necessary ingredient of the image reconstruction inorder to obtain the HESSI imaging resolution of 2arcsec. The SAS consists of 3 identical lens/sensorsubsystems (spaced at 120°) which interface with thespacecraft computer via the Aspect Data Processor(ADP). Each subsystem consists of a 4 cm diameterlens located on the front grid tray that focusses a solaroptical image onto a 2048-element x (13\if linear CCDlocated on the rear grid tray (focal length 1.55 m).Simultaneous exposures of the solar images are madeabout every 10 ms (programmable). For each image,there are two locations (limb crossings) where thesolar image intersects the CCD. A digital threshholdalgorithm is used to select N pixels that span eachsolar limb for inclusion in the telemetry. Determinationof the 6 limb crossing locations provided by the 3 sub-systems (over-) defines the position offset of the sun inthe rotating frame. The SAS can also used as a real-time sun acquisition sensor to back-up the fine sun-sensor.The lenses are made from fused silica and aremounted in such a way as to allow for a determinationof the optical center to within 5 \i with a coordinatemeasuring machine. Since the positions of the CCDsand of the grid fiducial marks will also be measuredwith micron accuracy, we shall know the SAS viewingdirections with respect to the individual grid pair view-ing directions on a level of 1 arcsec. A blocking filterand a bandpass filter (center wavelength at 700 nm,10 nm FWHM wide) are coated on opposite sides ofthe lens in order to reflect most of the light and toreduce the chromatic aberrations. The CCD (Loral 145EDC) is glued directly to an Al stand-off to guaranteestable positioning on a micron level. This has beenverified during vibration testing [1]. The integrationtime is around 0.1 ms to eliminate time-smearing forlarge pointing offsets. The CCD front-end electronics isessentially identical to the RAS front-end electronics[2] and consumes about 0.5 W per SAS. Allsubsystems, SAS as well as RAS are controlled by theADP which contains a DSP for on-line data processingand an FPGA (Actel 14100) for communication withthe Instrument Data Processing Unit (IDPU).We have simulated the response of the CCD aroundthe limb crossing in order to test different limb-findingalgorithms and to estimate the obtainable relative ac-

curacy (see Fig.1). Due to diffraction and chromaticaberration the limb amplitude increase is spread overseveral pixels. Random noise with 1 ch rms wasadded to the pixel amplitudes which were then fittedwith a pre-determined 'limb model' function to deter-mine the position accuracy. The resulting uncertaintyusing 3 pixels is about 0.08 arcsec and is dominatedby the digitalization effect.

An unambiguous aspect solution at a given time canbe derived from any 2 SAS pairs, yielding 3 independ-ent solutions. Differences between these solutions canbe traced back to variations of the positions of thelenses and of the CCDs. The internal consistency ofthe final solution can be checked by looking at severalcontrol parameters and their time derivative (since thespacecraft movements are slower than the SAS datataking rate). For instance, the inferred solar radiishould be independent of pointing direction. Similarly,we will be able to eliminate false solutions due to localflares or sunspots near the limbs which will producemomentary variations at 1 limb only. The SAS will becalibrated on ground using real sun images via a he-liostat. Since the final spin-axis cannot be predictedwith good accuracy and since with launch vibrationssome tiny position variations may occur, we shall haveto determine some calibration parameters in-flight. Oursimulations as well as the experience with the HEIDIexperiment [31 show that this should be possible onthe sub-arcsec level.

Fig.1: Simulated CCD response around a limb crossing.

REFERENCES[1] R. P. Lin et al., SPIE 3442, 2-12, 1998 and J.

Bialkowski et al., Ann. Report PSI 1998, thisvolume.

[2] J. Bialkowski et al., this volume.[3] C.J. Crannell et al., Proc. Int. Balloon Technology

Conf., 1991, AIAA-91 -3653, 1-11 (1991).

32

THE HESSI ROLL ANGLE SYSTEM (RAS)

J. Bialkowski, M. Fivian, ft Henneck, A. Mchedlishvili, P. Ming, K. Thomsen, A. Zehnder, (PSI), G. Hurford,D. Curtis, D. Pankow (UC Berkeley), B. Dennis (GSF Greenbelt)

We describe the design and construction of the HESSI Roll Angle System. Based on the observation of atleast one star per satellite rotation, it will provide an absolute roll angle determination with accuracy < 1arcmin.

The Roll Angle System (RAS) is part of the HESSIinstrument [1] to be launched by NASA by mid-2000 atsolar maximum. Its purpose is to provide informationon the roll angle of the rotating spacecraft (15 rpm).Precise knowledge (1 arcmin on a 1a level) of the rollangle is a necessary ingredient for imagereconstruction. The RAS is a star scanner whichpoints out radially and observes stars at 75° from thesun direction. Stars within a field-of-view of 30° x 1.2°(parallel x orthogonal to satellite spin axis) arefocussed by a lens onto a 2048 pixel line CCD. Thepassage of a star image over the CCD (orthogonally)will raise the charge in one or several pixels abovethreshold and the timing of this signal defines the rollangle, once the star has been identified by comparingits pixel position and amplitude with a star map. Duringeach passage of 6.4 ms duration we will make 2snapshots of ~ 5 ms each. Determination of thecentroid of this time sequence will allow us toreconstruct the corresponding momentary roll anglewith arcmin accuracy for a given object. Roll angles atintermediate times are inferred by assuming uniformrotation.

We selected the fast Dalsa IL C6 with rectangularpixels (500 \i wide x 13 (j. long) to provide for the 6.4ms long image passage and thus a high sensitivity. Incombination with a powerful lens (Leica Summilux,f/1.4, 50 mm with a quartz plate for radiationprotection) we will be able to observe stars down tomagnitude m=3. For m=3 the signal expected from apoint image is ~ 5 mV (slightly depending on starclass) corresponding to a S/N at room temperature of15 (verified experimentally). In practice we observelens vignetting as well as an off-axis broadening of theimage point-spread-function. At the very edge thisleads to a loss of sensitivity per pixel of almost 10. Wedid simulations based on the Hipparcos input starcataloque, taking this into account as well as the orbitparameters (see Fig.1). With a limiting magnitude ofm=3.1 we observe at least 1 star per revolution (duringdirect sun view) over 1 year; on the average we willdetect about 10 stars/ revolution.

The system is protected from earth-reflected light forviewing angles > 30° to the RAS axis. The opticalbaffle was designed for the specific case when HESSIis located exactly between sun and earth, requiring astraylight reduction of about 5*10"7. The design isbased on detailed calculations with the ASAP/ APARTstraylight codes. The baffle consists of CFRP,

containing medium- and high-module fibres. The latterprovide thermal conductivity to radiatively cool theCCD. We have done total reflectivity measurementson several CFRP samples and it was found that CFRPstructures lifted off a microbead-blasted replica surfacehave 'solar' reflectivities around 0.06, only about 50%higher than special absorptive coatings. The baffle isfixed directly to the central Ti frame which supports thelens as well as the CCD board and is in turn clampedto the satellite interface via 3 Ti legs which provide athermally stable (and thermally insulated) angularposition of the RAS with respect to the HESSI imager.A mechanical shutter directly in front of the CCDprotects the CCD from an accidental sun illuminationin the initial phase. The CCD board contains theFPGA, which provides all necessary CCD controlsignals, reads the ADC and sends the data to theSAS/RAS main processor (ADP). The CCD has beenpassively irradiated with 60 MeV protons; we observeda 20% variation of the dark level after 8 krad. TheCCD has also survived the HESSI vibration level. Thepower consumption is 1.2 W, the mass 1 kg.

f

ss

{dev

io

E

10

0

-10

:+ :

i- ;

: + :r [

-10

stars

<85

seen by RAS on Q-th

= °^ +

180rotation angle phi

day

-t-

-.

275 370

Fig. 1: Stars seen (crosses) and stars obscured bythe earth (squares) on 1.1.2001 during onesatellite revolution. The earth center is givenby a diamond.

REFERENCE

[1] R. P. Lin et al., SPIE 3442, 2-12, 1998 and J.Bialkowski et al., Ann. Report PSI 1998, thisvolume.

33

IRRADIATION OF TITANIUM AND TUNGSTEN SAMPLES FOR HESSI PROJECT

W. Hajdas, N. Frey, K. Thomsen, A. Zehnder (PSI)

Modules, made of titanium and tungsten, comprise a large part of HESSI detectors' surroundings. Theiractivation in the Earth radiation belts will increase the detectors' background. Samples of Ti and W wereirradiated at PIF and analyzed with Ge(Li) gamma spectrometer. The activity increase due to satellite'spass through the proton rich South Atlantic Anomaly is equal to 14 Bq/kg for Ti and 138 Bq/kg for W.

SATELLITE ORBIT AND RADIATION BELTS

The HESSI mission will be launched into a circularorbit at the distance of 600 km above the earth. Fororbit's inclination of 38° some trajectories will bepassing through the South Atlantic Anomaly SAA - aregion with increased proton fluxes- see Fig. 1.

so75

60

45

30

15

0

-15

-30

- 4 5

- 6 0

- 7 3

- 9 0

hasa. radation data_ l 1 1 1 1

- wSL " " ^

- p a~ - \ -j- •" x D

r ?/

~~ '*>- " ''- -*gr~'-"~*"~~**1- - - " " " " "~ ""

1 L-

*- ••:

- ^ , ^ ^ ^

i . . -

I

from the ""Sc decay after passing through the SAA isequal to 14 (6) Bq/kg.

The tungsten activation was higher than the titaniumone though the spectra - see Fig. 2 - were recorded60 min. after the irradiation.

•3 3 -

Tungsten gamma spectrum

Sample mass • 23 g,D&tance from Ge{Li) - 5 cm,Accumulation -1070 sec.Time after proton exposure - 60 min."

-1S0 -120 -60 120 180

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Energy / keV

Fig. 1: HESSI single orbit path through the SAAregion of increased proton radiation.

The SAA proton fluxes (above 30 MeV) are up to fewthousands per p/cm2 /sec. Nuclear activation ofmaterials can lead to emission of gamma rays, givingrise to a time dependent background in the HESSI Ge-detectors. Material that surrounds the detectorsconsists mainly of tungsten and titanium apart of Al.

IRRADIATION PROCEDURE AND ENERGY BINS

The SAA spectrum was generated in the PSI ProtonIrradiation Facility PIF, where small samples oftitanium and tungsten were exposed to protons. Thespectrum was divided into 7 energy bins covering therange between 35 and 300 MeV. For activitymeasurement purposes the total exposure dose wasequivalent to 100 days in orbit (ca. 70 rad(Si)).

ACTIVITY MEASUREMENT AND ANALYSIS

Activity measurements after proton exposures weredone using Ge(Li) detector of the SU spectroscopystation in the Experimental Hall.

The titanium sample was measured 20 min afterproton run. Two spectral lines were identified: the1157 keV of **Sc and the strong annihilation oneattributed to ^Sc and "So isotopes. Resulting activity

Fig. 2: Gamma spectrum from the tungsten sample.

200

175

150

125

100

75

50

25

0

• r

IIIIffillIlllill"'\I

10 20 30 40 50

Time in orbit /hours

60 70

Fig 3: Tungsten sample activity as a function of thetime in orbit.

In total, 28 gamma ray lines were localized between65 to 2000 keV. They were attributed to 12 differentisotopes. Their life times vary between 13 min and 20hours. Resulting activity of Tungsten after passingthrough the SAA is equal to 138 (28) Bq/kg.

Taking into account successive decay and activationprocesses during subsequent SAA passages, theactivity as a function of orbital time was calculated.Results for tungsten are shown in Fig. 3.

34

XMM-RGS CALIBRATIONS COMPLETED

K. Thomsen (PSI) and XMM collaboration

Resolving power and effective area of the XMM-RGS instruments have been measured.

Two of the three X-ray telescopes onboard the ESAcornerstone X-ray Multimirror Mission (XMM) are fittedwith a Reflection Grating Spectrometer (RGS)instrument. This instrument is capable of providinghigh resolution X-ray spectra in the range of 5 - 35Angstrom. Parts of the X-rays, after being focussed bythe mirrors, are deflected by a set of 182 reflectiongratings. The spectral orders (-1,-2,-3) are projectedonto a strip comprised of 9 cooled CCDs in eachdetector. Using the intrinsic energy resolution of theCCDs the various spectral orders, which overlapspatially on the camera, can be very well separated.PSI joined the RGS collaboration in 1989 anddelivered the flight hardware of detector housings,cooling systems and front-end electronics 1996. Sincethen PSI took part in the calibration of the instrumentand in the preparation of the operations in orbit.

The instrument resolving power and effective area arenow based on measurements, rather than onsimulations. Work is still ongoing to understand themeasured excess above model predictions of theeffective area at short wavelengths. The effective areaof the gratings in the range 5-10 A is about 10% abovemodel predictions.

Overall agreement between theory and measurementsperformed at the long X-ray facility Panter in Munich isof order 10%, which is reasonable given the knownuncertainties. The maximal effective area for a singleRGS is in first order about 70 cm2, Fig. 1).

One of the original requirements on the performanceof RGS was that it would be able to resolve the He-like'triplets' of the abundant elements for which these linesare in the RGS band (O through Si). From the

800

measured line ratios densities in the emitting regionscan be derived. The oxygen VII He-like triplet at lowand high density plasma can be resolved; figure 2). Mgline ratios cannot be extracted from the first orderspectrum, however it is possible to extract and resolvethe ratio using the 2nd order spectrum.

Fig. 1: Expected theoretical effective area andmeasured data for one RGS. The range of thesolid curves reflects the true range inwavelengths for the different orders of theRGS for on-axis sources.

800

600 -

100 -

200

T. = 2 MK n.. = 1.2x10" cm"

20.0 20.5 21.0 21.5 22.0 22.5 23.0wavelength (A) (uncolibroted)

23.5 20.0 20.5 21.0 21.5 22.0 22.5 23.0wavelength (A) (uncolibroted)

23.5

Fig. 2: Simulated spectral images for the Helium-like Oxygen 'triplet' in order m=-1 for RGS1 at lowdensity (left) and high density (right) at a constant temperature of 2MK.

35

IMPROVED ENERGY RESOLUTION WITH GIAEVER JUNCTION X-RAYDETECTORS

Ph. Lerch, E. Kirk, J. Bialkowski, A. Mschedlichvili, F. Burri, A Zehnder (PSI), /-/.ft Ott (ETHZ& PSI)

We fabricated superconducting devices as energy dispersive photon detectors. We could isolate oneenergy resolution degradation mechanism from several others and obtain a value ofAE*= 58 eV at 6 keV.

We are developing superconducting tunneling junction(STJ) devices, called Giaever junctions, as energydispersive detectors for optical, ultraviolet and X-rayphotons. Energy absorbed in a superconductor breaksCooper pairs, and releases excess charge carriers(ECC). A STJ measures n x« Ex / e the number of ECCproduced upon the absorption of energy Ex, where £ =2 - 3 meV is the minimum energy required to release 2ECC. The energy resolution is set by the statisticalfluctuations in nxi and by tunneling noise in the STJ [1],

AE f w h m °cE

It has been shown recently [2] that AE ~ 6.5 eV at Ex

= 700 eV. This result is very close to the expectedlimit, and makes this class of detectors promising forapplications. However, for E > 1 keV several factorsdegrade the energy resolution. Spatial variations in theresponse to energy deposition increase the variance ofnx. Moreover, the fabrication of these thin film devicesdetermines the dynamics of the ECC (recombination,tunneling and trapping rates, phonon losses, etc.)which, in turn, affect the resolution.

In order to study the physics causing the resolutiondegradation distinctly, we instrumented 2 STJs at theend of 200 x 40 u.m2 large absorber of epitaxial Ta [3].The devices were kept at 550 mK, illuminated with 5.9and 6.4 keV photons of a 55Fe source. Each STJ wasbiased at ~ 300 p.V. In our configuration, X-rays hit theSTJs as well as the absorber. The ECC released inthe absorber diffuse in either direction. The chargeratio collected by each STJ depends on the position ofimpact on the absorber. The total charge collected byboth STJs depends on the initial energy released bythe photon, on the diffusion losses along the absorber,and on the response function of each STJ. The ECCproduced by photons hitting one or the other STJ di-rectly are weakly affected by diffusion losses in theabsorber and carry more information on processesand losses occurring within the STJ.

We have measured the response of two spectrome-ters, whose epitaxial Ta absorbers had different elec-tronic mean free path. The devices have aTa/AI/AIOx/AI/Nb structure. The Ta, Al, and Nb layersare 40 nm, 50 nm, and 30 nm thick, respectively. Aland Nb are polycrystalline. The two spectra below

were measured from absorber events and illustrate theimpact of the quality of the epitaxial absorber on theenergy resolution. The electronic mean free path ofthe best sample was of the order of 40 nm.A careful study of the ECC dynamics in a supercon-ducting absorber, distinct from processes occurringwithin the STJ is now possible and should guide thefuture development strategy for this class of detectors.

1.0

0.8

0.6

0.4

0.2

0.0

i

\ \

i

I I

— - RRR = 60,

• RRR = 20,

f

I

AE tahm =

i +

-

58 eV

= 104eV

-

5600 5800 6000 6200 6400 6600 6800 7000

Energy (eV)

Fig. 1: Spectra recorded with 2 detectors fabricatedwith Ta absorbers of different quality. Thecontribution of electronic noise was not re-moved. The energy scale was correctedquadratically.

REFERENCES

[1] D.J. Goldie, P.L Brink, C. Patel, N.E. Booth, andG.L Salmon, Appl. Phys. Lett. 64, 3169 (1994).

[2] J.B. le Grand, C.A. Mears, LJ. Hiller, M. Frank,S.E. Labov, H. Netel, C. Chow, S. Friedrich, M.A.Lindeman, Appl. Phys. Lett. 73, 1295 (1998).

[3] H. Kraus, F. von Feilitzch, J. Jochum, R.L. Mb'ss-bauer, Th. Peterreins and F. Probst Phys. Lett.B231, 195(1989).

36

SUPERDONDUCTING X-RAY DETECTOR READ-OUT USING A DC SQUID ARRAY

J. Olsen, Ph. Lerch, E. Kirk, K. Thomsen, F. Burri, A. Zehnder, B. van den Brandt, S. Mango (PSI),H.R. Ott (ETHZ)

Arrays of dc SQUIDs were used to read-out current excursions produced in a superconducting tunnelingjunction X-ray detector. The potential of this amplification scheme was investigated.

Future energy dispersive soft X-ray spectrometerswith very high energy (and/or position) resolution willmost likely be cryogenic devices [1]. In a supercon-ductor, the minimal energy required to produce ex-cess electronic excitations is = meV. Thus, a largefraction of the absorbed energy is carried by excita-tions that result from broken Cooper pairs. Theexcess charge that is produced can be measured witha superconducting tunneling junction (STJ) andenables an energy resolution that cannot be reachedwith semi-conducting devices.

Signals produced by STJ are usually detected bycharge sensitive amplifiers (CSA) using low noisejunction field effect transistors (JFET) cooled at =100 Kas the input stage. The equivalent noise charge (ENC)of such an arrangement ranges between 0.5 - 5 -104

e" FWHM with an STJ in operation and is still smallenough when compared to the signal charge Q = 5 •106 e' that is collected after the absorption of a 6 keVphoton. This approach has disadvantages:

• the ENC is dominated by the input capacitance CD

» 200 pF of a 100 x 100 jim2 STJ device. Thus, anincrease of the photon collection area correlateswith an increase in noise.

• the power, « 3 mW, dissipated by each JFET inthe very vicinity of an STJ will impose severe con-straints on multiple pixel-devices.

• the measurement of the total charge Q = J i(t) dtmakes the precise study of the qp dynamics,expressed in i(t) and containing relevant physicalinformation, rather difficult.

The use of an array of dc SQUIDs may relax some ofthe limits imposed by the CSA. The purpose of thisexperiment was to investigate for the potential of theSQUID approach. The arrays of dc SQUIDs [2] arebased on a design developed at NIST [3]. The array ismade with 200 dc SQUIDs connected in series anddelivers a large output signal (mV).

Our STJ device is voltage biased and inductively cou-pled to the SQUID through an input coil. The deviceand the SQUIDs are kept at 500 mK and 2.2 Krespectively. A ^Fe source is placed near the STJ, andemits 6 keV photons. Absorption events in the STJproduce current excursions recorded by the SQUIDs.

2 3

Temps [us]

Fig. 1: Output generated by one 6 keV photon hittinga Nb/AI/AIOx/AI/Nb tunneling junction meas-ured with an array of dc SQUIDs.

The main results are: the ENC does not depend onCD, the power dissipation of the SQUIDs is negligible.The current can be measured accurately provided avery stiff voltage bias is applied to the STJ. Taking intoaccount the gain of the system, 104 V/A, 1 mV in Fig.1 corresponds to i(t) = 10"7 A. The shielding of thelines coupling the STJ to the input coil of the SQUIDsrequires particular care. The bandwidth is of the orderof 2 MHz and can be increased. A pulse height spec-trum (PHS) of the charge cannot be simply obtainedby the PHS of the current pulse because of variationsin the shape of i(t). Note the fast decay time of theevent in Fig.1.

This read-out technique is promising, however, severeSQUIDs reliability problems were identified and arecurrently investigated in collaboration with the manu-facturer and NIST.

REFERENCES

[1] In: Low Temperature Detectors, edited by H. R.Ott and A. Zehnder, NIM A 370 (1996).

[2] Devices fabricated by HYRPES, Inc, 10523Elmsford, NY (USA), http://www.hypres.com.

[3] R.P. Welty, J.M. Marinis, IEEE Trans. Appl. Su-percond. 3, 2605 (1993).

37

OPERATING OF THE PROTON IRRADIATION FACILITY - CONCISE SUMMARY

W. Hajdas, F. Burri, J. Biaikowski, A. Mchedlishvili, K. Thomsen, A. Zehnder (PSI),R. Sorenson, B. Nickson, L. Adams (ESA-ESTEC)

In 1998, PIF experiments were conducted in NA2 and OPTIS areas and used more than 70 beam shifts. Allruns were grouped in 22 beam blocks resulting in 60 irradiation days. Research groups from spaceagencies, scientific institutes, industrial firms and universities carried out about 30 different experiments.

PIF particle beam experiments in 1998 covered aperiod between 14 March and 11 December. Theywere arranged into 22 irradiation blocks of variableduration and resulted in 60 days with the beam ontarget. Despite of the prolonged shutdown of the RingCyclotron the experimental load didn't decrease. It waspossible due to large interest of using low energyprotons in the OPTIS area. The beam time was almostequally shared between NA2 and OPTIS irradiationfacilities. More than 30 types of irradiation tests by 17different groups of researchers were performed.Required proton fluxes ranged from 104 to 1010

/cmVsec and radial beam profiles varied from 16 to100 mm FWHM. Beam energies covered values from5 to 590 MeV.

Beam time and test area utilization by PIF

Beam blocks

Shifts

NA2

12

42

OPTIS

10

29

Total

22

71

IRRADIATION EXPERIMENTS

Four major blocks of experiments characterized PIFactivities. The largest amount of beam time for ESArelated activities was used for calibration and tests ofthe proto-flight model of the SREM detector. It will belaunched in 1999 on board of the STRV-1c satellite.The next one is related with the HESSI satellitemission. Together with NASA-GSFC and BerkeleyUniversity the PSI, represented by Laboratory forAstrophysics, is preparing the HESSI launch in theyear 2000. Two other experiments were carried out byindustrial research teams in frame of new spacetelecommunication systems like e.g. Teledesic orIridium. Carls Zeiss GmbH studied proton effects inoptical glasses while Bosch-Telecom GmbHcharacterized total dose and single event effects inOptical Terminal electronic components. In theeducational field the data for one dissertation and onediploma work were gathered during PIF tests.

List of PIF experiments:

• SREM STRV-1c calibration

• CREDO particle monitor calibration

• SRAM-based proton dosimeter calibration

• SRAM/DRAMs proton SEU characterization

• Proton damage effects in GaAs solar cells

• CCD, DSP, ADC total dose and SEE testing

• Total dose effects in optical glasses

• Irradiation of Optical Terminal electronics

• Total dose damage and SEE in Opto-Encoder

• Optical-coupler SEE determination

• Sensitivity test of the ISS Air-fan electronics

• Activation measurement of Ti and W plates

• Flash-disks PCMCIA proton effects tests

• Radiation effects in high frequency devices

• Feasibility tests of proton induced productionof spallation neutrons

• (40Ar+p) -•• (26AI + ...) cross sectionmeasurements

SYNOPSIS OF USERS AND COLLABORATIONS

The following list contains research groups involved inPIF experiments.

No Research Institution

1 ESA/ESTEC, Nordwijk

2 PSI/GSFC/Berkeley University

3 Technical University Munchen

4 Bosch-Telecom GmbH, Backnang

5 Carl Zeiss GmbH, Oberkochen

6 Contraves Space, Zurich

7 CNES, Toulouse

8 HIREX Engineering, Toulouse

9 Matra Marconi Space, Paris

10 ONERA-CERT, Toulouse

11 Bradford Engineering, Heerle

12 NLR, Amsterdam

13 Thomson&Nielsen Co., Nepean

14 SCK/CEN, Mol

15 SOREQ NRC, Yavne

16 DERA Space, Farnborough

17 EEV Ltd., Chelmsford

FACILITY DEVELOPMENT IN 1998

• Further elaboration of the PIF OPTIS irradiationsite - new dosimeters and automatic beam-stop.

• Preliminary version of universal testing unit forradiation effect studies in semiconductor memories.

38

HEATING OF STELLAR CORONAE

M. Audard, M. Gudel (PSI), E.F. Guinan (Villanova Univ., USA)

A few thousand kilometers above the surface of a star, the average temperature of the corona reaches upto a few millions of degrees. The mechanism of coronal heating is unknown. We have investigated thequestion of "flare heating" on two young active stars whose coronae are thought to be scaled-up versionsof the solar corona.

Stellar coronae are interesting objects when we try tofind answers about questions related to our Sun or tocosmic plasmas in general. Despite their distance,active stars provide a much larger range of magneticphenomena than the Sun.

We have investigated the question of stellar coronalheating by flares using data from the ExtremeUltraviolet Explorer (EUVE) satellite. We observed twoyoung active stars, 47 Cas and EK Dra, that arebelieved to be identical to the Sun in its early age.Explosive flare energy releases heat the plasmatemporarily and emit an important portion of theirradiation in the soft X-rays and the Extreme Ultraviolet(EUV) range. Flares on the Sun and stars areobserved over a wide range of radiated energy. Largeflares on the Sun emit on the order of 1030-1032 ergs.Smaller-scale events are called ' microflares'. Thewhole family of flares is thought to originate fromenergy released by reconnection of the magnetic field.The 'microflare' hypothesis proposes that the smallevents are responsible for a large fraction of thecoronal heating because they would occur very oftencompared to the more energetic events (flares).

We have developed a statistical method that identifiespossible flares in our EUVE data [1]. A total of 28events were identified during our two week-longobservations. Their EUV+X-radiated energy wereestimated numerically, allowing us to constructcumulative flare rate distributions in energy (i.e. therate of flares exceeding a given energy). Thesedistributions have been found to be power laws of the

type N(> E) = k2E~a+l (where N(>E) =j~^E). WeCltL

found for both stars and their combined distributions a

value of a that exceeds 2. Since | " ——EdE can beAo dE

arbitrarily large for small Eo and a > 2, the "microflare"hypothesis may be a valid explanation of the coronalheating of these two young active solar analogs. Itmay indicate that much, if not all, of the observablecoronal energy in stars is derived from magneticenergy stored in the outer stellar atmospheres.REFERENCE

[1] M. Audard, M. Gudel, E.F. Guinan (1998), inProc. of the First XMM Workshop on "Sciencewith XMM", Noordwijk, NL, M. Dahlem (ed.), URLhttp://astro.estec.esa.nl/XMM/news/ws1 /ws1 _papers.html.

coun

tsC

ount

rat

0.25

0.20

0.15

0.10

0.05

0.00

0.25

0.20

0.15

0.10

0.05

0.00

472

: 0V*t

474 476

47 Cas

EKDra

478

-

-

-

-

58 6460 62Days after JD 2450000.0

Fig. 1: EUVE light curves of 47 Cas and EK Dra.

47 Cas

Energy (E) [erg]

Fig. 2: Cumulative distributions in energy of the flareoccurrence rate (dashed: before somecorrection due to overlapping flares).

39

RADIO COUNTERPARTS TO EXTREME X-RAY YOUNG STELLAR OBJECTS

K.W. Smith, M. Gudel (PSI & ETHZ), A.O. Benz (ETH)

Extremely bright X-ray sources have recently been found in star-forming regions. We search for radiocounterparts for two of these radio and X-ray luminosities, considered together, have implications forcoronal models.

The recent discovery [1,2,3] of surprisingly large X-rayflux from Young Stellar Objects (YSO's) raisesimportant questions on the role of coronae at earlystages in pre-main sequence stellar evolution. Youngstars are usually rapid rotators, and the associatedkilogauss magnetic fields seem to be a typical featureof these systems. More evolved young stellar objects(Classical T Tauri and Weak-lined T Tauri stars, CTTSand WTTS) are seen to be X-ray sources, and aresupposed to emit their X-ray flux from a corona. Thenew YSO's are, however, considerably more X-raybright than either WTTS or CTTS, and in fact are moreX-ray bright than any previously known coronal X-raysource. This raises the question of whether the coronalmodel needs to be modified to account for such brightsources, or whether an entirely different model isnecessary to explain them.

One feature of stellar coronae is an empiricalobservation that the X-ray luminosity is closelycorrelated with continuum radio luminosity [4].Although purely empirical, the physical inference ofthis observation is that the high energy electronsresponsible for the synchrotron radio emission are alsoresponsible for heating the coronal plasma, and arehence indirectly responsible for the thermal X-rayemission. We set out to find radio counterparts for thetwo most extreme X-ray YSO's by analysing VLAarchive data. If the radio flux from these objects fits theLX-LR relation for stellar coronae, this would be strongcircumstantial evidence that coronal models areapplicable.

less radio flux as the fast electrons are thermalised bycollisions, whilst the coronal plasma is still efficientlyheated by magnetic processes.

In Figure 1 we show the LX-LR relation for stellarcoronae, with the two new sources added. Onesource, IRS-95, has a radio counterpart which is closeto the relation, but radio weak (X-ray bright) by aboutan order of magnitude. The other source, SVS-16, hasonly an upper limit in the radio.

In the case of SVS-16, we can conclude that the X-rayemission is definitely not coronal in origin. In the caseof IRS-95, the object lies close to the coronal relation,but sufficiently removed from it that we cannotconclude we are seeing a normal stellar corona. Weare therefore developing models for a modifiedcorona, in which the circumstellar material leads to

28 "

14 16Log L, (erg S-' Hz"1)

Fig. 1: The LX-LH relation for stars, amended toinclude the X-ray luminous YSO's. The SVS-16 point is shown with an upper limit in theradio. The other stars have been divided intotwo categories. dMe stars, dKe tars and BYDra stars, which typically have lowluminosities, are plotted as crosses. WTTS,RS CVn's, Algols, FK Com stars and AB Dorare shown as pentagons. The best-fit line withslope 1 for the low-luminosity sources isplotted, and extrapolated into the high-luminosity regime for ease of comparison withthe YSO points.

REFERENCES

[1] Preibisch, Astron. & Astrophys. 324, 690 (1997).

[2] Preibisch, Astron. & Astrophys 338, L25. (1998).

[3] Preibisch et al, Astron. & Astrophys 338, 923(1998).

[4] Gudel & Benz, Astrophys. J. 405, L63 (1993).

40

ON THE STRUCTURE OF STELLAR CORONAE

M. Gudel (PSI), A. Benz (ETHZ), J.L. Linsky & A. Brown (Univ. Colorado), J. Conway (Onsala Observatory,Sweden), F. Nagase (ISAS, Japan)

The structuring of the hot outer atmospheres of cool stars is of crucial importance for plasma heatingprocesses, particle acceleration, and the physics of plasma or MHD instabilities. We have addressed thisquestion using two widely different methods, namely Very Long Baseline Interferometry at radiowavelengths and theoretical modelling of an X-ray flare.

The outer atmosphere of cool stars like the Sunconsists of very hot (>106 K) plasma confined inmagnetic loops that are anchored on the stellarsurface. Apart from the Sun, there is little directinformation about the detailed structuring of stellaratmospheres that are much more active than the Sun.

Our collaboration obtained the first ever spatiallyresolved, direct image of a magnetized stellar coronaof a solar-like star. We used the Very Large BaselineArray (VLBA, consisting of 10 radio dishes spread overthe USA) together with the Very Large Array (27antennas in New Mexico, USA) to obtain direct imagesof the radio corona of one of the nearest stars, thelow-mass dwarf UV Cet. Surprisingly, the image showstwo very large blobs that seem to be aligned with thepolar axis. From an interpretation of this synchrotronradiation, we have proposed a model based on aglobal dipole into which relativistic electrons areinjected and are trapped. Such structure is completelyunknown on the Sun, where the magnetic loops aresmall and concentrated toward the equator. Thetransformation between the two arrangements mayoccur during stellar evolution, or may be related to theinternal structure of the star, in particular its internalconvection.

In another observation, using the ASCA X-rayastrophysics satellite ASCA and the ExtremeUltraviolet Explorer, we investigated the structuring ofthe hot corona of a very active subgiant star byapplying an MHD model. The flare event wasobserved over several hours and reached a peakenergy output of 2x1032 ergs s'1; extremely hot plasma,up to 108 K, was identified. Our model took intoaccount estimates of the surface magnetic field, theobserved time scales, and magnetic and thermalpressure effects in the loops. We found that the onlysolutions fulfilling all requirements were loops with asize scale on the order of one stellar radius. Thisagain contrasts with the structures found on the Sun.Flare electron densities reached about 1011 cm'3 whiledensities inferred from spectral line ratios duringquiescence were higher. This may suggest that flaresoccur in giant magnetic arches while much of thesteady emission is produced in low-lying compactmagnetic loops.

REFERENCES[1] Benz, A. O., Conway, J., Gudel, M., 1998, Astron.

& Astrophys. 331, 596.[2] Gudel, M., Linsky, J. L, Brown, A., Nagase, F.

1999, Astrophys. J. 511, 404 (1999).

as

0.0

Fig. 1:

x/R.

Top: VLBA image of the atmosphere (corona)of the low-mass star UV Cet; the dashed circlegives the size of the optical star.Bottom: Calculation of the magnetic X-ray loopstructure at different times, modeled from X-ray data of a flare on a subgiant star.

41

MICROINSTABILITIES AND SCATTERING OF ACCELERATED PARTICLESIN PLASMAS

G. Paesoid (ETHZ& PSI)

We are studying plasma microinstabilities in fully ionized plasmas to investigate of accelerated particles tohigher pitch angles in velocity space. These studies will be important for application to solar coronal plas-mas that will be observed with HESSI. This project is part of a PhD Thesis which is carried out under theguidance of Arnold O. Benz (ETHZ).

Acceleration of particles (in particular of electrons) isan ubiquitous phenomenon in many astrophysicalplasmas. It may be of fundamental importance sincethe primary release of energy stored in magnetic fieldsmay be via accelerated particles. Further, acceleratedparticles could be efficient in heating plasmas to hightemperatures. We study plasma theoretical conceptsof particle acceleration with applications to solarcoronal plasmas.

According to different acceleration models [1,2,3] anecessary property of the accelerator to be efficient isthe presence of a mechanism scattering the acceler-ated particle species to higher pitch angles in velocityspace. Microinstabilities of anisotropic plasmas withrespect to various wave modes can lead to pitch anglescattering to higher pitch angles.

It is assumed that these microinstabilities, arising froman anisotropy intrinsic to the accelerator mechanism,produce the required scattering. In order to investigatethe properties of anisotropic plasmas with respect toinstabilities to various wave modes, a fast and efficientmethod to solve the plasma dispersion function provedto be a necessity. We decided to use an existing com-puter code called WHAMP developed by KjellRoennmark in 1982 [4]. This code solves the disper-sion equation (1)

k2c2 = u; —

^k f r "na J Vw-

" ~ kv\\

0)

where co is the complex frequency of the plasma modeand k the appropriate wavenumber. Up to six plasmaparticle species a which are described by velocity

distribution functions f<f of the form (2) can be con-

sidered.

fo(v±,v\\) =

= exp - - ^ -

Aj 1-Aj

exp -

(2)

This enables us to consider plasma species which are

drifting (vda - drif t ) , anisotropic in temperature

V uthenn ) a

(Aa - ~ ) a r |d having a loss cone (size: P2o, depth:

M).Due to the difficult and untemporary usage of thiscode, an update of the user interface was necessary inorder to maintain an adequate usage. This was real-ized on an IDL platform and was called IDLWhamp.This provides us with a powerful tool for the investiga-tion of microinstabilities in magnetized plasmas.First contacts with international researchers workingon the same topic were established during the visit oftwo meetings:

• The CESRA meeting 1998, 'Workshop on CoronalExplosive Events' in Helsinki (Finland), 8.6.98-13.6.98 . A poster with the title 'Particle Accelera-tion by Fermi Acceleration' was presented there.

• The 208. Heraeus-Seminar: 'Energetic Processeson the Sun and in the Heliosphere' in Bad Honnef(Germany), 19.10.98-21.10.98.

REFERENCES

[1] La Rosa et al. 1993, Astrophys. J. 418, 912.[2] La Rosa et al. 1996, Astrophys. J. 467, 454.[3] Miller et al. 1996, Astrophys. J. 461, 445.[4] Roennmark 1982, KGI Report, 179.

42

DARK MATTER IN THE UNIVERSE

Ph. Jetzer, D. Puy, L. Grenacher (PSI), M.Strassle, N. Straumann (Univ. Zurich), F. De Paolis (Bartol Res.Inst), G. Ingrosso (Univ. Lecce), M. Roncadelli (Univ. Pavia)

Recently a diffuse y-ray emission from the galactic halo has been detected with the instrument EGRET.The observed y-ray flux is in remarkably good agreement with the flux predicted in the framework of ourmodel for the halo dark matter. In order to explain the observed pulse period variations in X-ray pulsars westudied an extension of the disk-fed model and we compared our results with the measurements for Cen X-3. Moreover, using microlensing data we found evidence that globular clusters contain an importantamount of dark matter in form of brown dwarfs or white dwarfs.

Dark Matter in the Halo of our GalaxyIn collaboration with F. De Paolis, G. Ingrosso and M.Roncadelli we further elaborated on our scenario fordark matter in the halo of our Galaxy, according towhich MACHOs (Massive Astrophysical Compact HaloObjects) are clumped into dark clusters along with coldH? clouds [1]. An important prediction of our scenario isthat there is a y-ray flux arising from the halo cloudsdue to the interaction of high-energy cosmic-ray pro-tons coming from the galactic disk with the cold Hz

clouds [1]. Recently, Dixon et al. [2] have reanalyzedthe data taken with the EGRET instrument, which is aY-ray detector on board of the satellite ComptonGamma Ray Observatory (CGRO). Dixon et al. founda statistically significant diffuse y-ray emission from thegalactic halo. Motivated by this discovery wecomputed in more detail the expected y-ray flux, takingalso into account the contribution due to inverseCompton scattering [3]. The agreement between theobservation and our more detailed calculation for theflux is very good, making thus our scenario for the halodark matter worth further consideration. In particular,the next generation of y-ray satellites like AGILE orGLAST can test our predictions, thanks to the highersensitivity and the better angular resolution.

On the Variation of Pulsar Periods in Close BinarySystemsX-ray pulsars, which form together with a main se-quence star or a giant a close binary system, showstrong variations in their pulse period. Epochs of spin-up and spin-down interchange on timescales of sometens up to some thousand days. We studied, togetherwith M. Strassle and N. Straumann this phenomenonin the framework of the disk-fed model by Ghosh andLamb, which we generalized by allowing for aninclination angles between the magnetic dipole

moment fi and the rotation axis [4]. Moreover, we

took into account the relatively small magnetic fieldinduced by currents in the magnetosphere flowingalong the dipole field lines. This component, which wecomputed numerically, leads to a torque perpendicularto the rotation axis and induces a precession of it. Dueto this additional degree of freedom, the calculatedpulse period history is much smoother than in theclassical Ghosh and Lamb model. Within this extendedmodel we computed, as an example, the pulse periodhistory of Cen X-3 using the observed X-ray flux as ameasure for the mass accretion rate (see Figure 1).

We found good qualitative accordance with themeasured pulse period data as obtained with theBATSE instrument on CGRO.

S5OO

JD-2440000.5

Fig. 1: Calculated pulse period history for Cen X-3using the BATSE flux as input.

Gravitational Microlensing by Globular Clusters

Stars in globular clusters can act either as sources forMACHOs located along the line of sight or as lensesfor more distant background stars. Globular clustersoffer the advantage that either the distance to the lensor the one to the source is precisely known, hencethey allow for a much more accurate mass de-termination of the lens. We analysed the known mi-crolensing events towards the galactic bulge, which lieclose to the three globular clusters NGC 6522, NGC6528 and NGC 6540 [5]. We found evidence thatsome microlensing events are indeed due to MACHOslocated in the globular clusters, suggesting, therefore,that these clusters contain a significant amount of darkmatter in form of brown dwarfs or white dwarfs. This isa new result, which indicates that globular clusterscontain a substantial amount of dark matter, compara-ble to the luminous mass.

REFERENCES[1] F. De Paolis, G. Ingrosso, Ph. Jetzer, M. Ronca-

delli, Phys. Rev. Lett. 74,14 (1995).[2] D. Dixon et al., New Astronomy 3, 539 (1998).[3] F. De Paolis, G. Ingrosso, Ph. Jetzer and M. Ron-

cadelli, Astrophys. J. Lett. 510, L103 (1999).[4] Ph. Jetzer, M. Strassle and N. Straumann, New

Astronomy 3, 619 (1998).[5] Ph. Jetzer, M. Strassle and U. Wandeler, Astron.

and Astrophys. 336, 411 (1998).

43

X-RAY PROPERTIES OF GALAXY CLUSTERS

D. Puy, L. Grenacher, Ph. Jetzer (PSI), M. Signore (Obs. Paris-France), F. Melchiorri (Univ. of Rome),Y. Rephaeii (Tel Aviv University)

We studied the equilibrium temperature for cold clouds in cooling flows of galaxy clusters. Moreover, westarted an international collaboration with the aim to measure the Hubble constant via the Sunyaev-Zeldo-vich effect. To that purpose we selected some galaxy clusters, which we plan to observe with the XMMsatellite. These observations will also be very useful to study the total mass of the clusters and other prop-erties of the hot intracluster gas.

Cooling Flows: The observation of the X-ray fluxallows to evaluate the total mass of galaxy clustersand to give lower limits on the baryonic fraction incluster, moreover an interesting application is thestudy of their cooling flows. In several clusters of gal-axies one observes a soft X-ray excess towards thecentral regions. This excess is interpreted as beingdue to hot intracluster gas with a cooling time less thanHubble time. This gas cools and falls quasi-hydrostati-cally into the center of the cluster potential well. In-deed, radial profiles of temperature and density in theintracluster gas inferred from X-observations show thatcooler denser gas is present in the central region ofthe cluster. Cooling flows in clusters of galaxies de-posit large quantities of cool gas around the centralgalaxy, which is still continuing to grow today andcould just accumulate as cool dense clouds.We have investigated the minimum temperature whichcan be reached by clouds in cooling flows by comput-ing the cooling function due to H2, HD and CO mole-cules in the cooling flows of the Centaurus, Hydra andPKS 0745-191 clusters [1]. Our results suggest thatclouds can indeed reach very low temperatures of lessthan ~ 10 K (see Figure 1). We started to investigatewhether thermal instability can take place and thuslead to a fragmentation of the cooling gas. This lastprocess could provided an explanation to the exis-tence of very cold clouds in the center of the clusters.These investigations are of relevance in view of theinterpretation of the observation which we will do in theframework of the XMM core program of clusters ofgalaxies and with SAX on the A478 cluster, for whichwe submitted a proposal that has been accepted.

Sunyaev-Zeldovich Effect: The Sunyaev-Zeldovich(SZ) effect offers the promise of measuring the Hubbleconstant Ho within an accuracy of about 10 %. Thetechnique relies on combined X-ray observations of ahot intracluster gas and microwave measurements ofthe Compton scattering between Cosmic Backgroundphotons and the energetic electrons in the intraclustergas. This last effects appears as a distortion in theCosmic Background Radiation (CBR) with a brightnessdecrement in the Rayleigh-Jeans region and an incre-ment in the Wien region.We started an international collaboration with F.Melchiorri (Univ. of Rome), M. Signore (Paris Obser-vatory) and Y. Rephaeii (Tel Aviv University) with theaim to perform measurements of the SZ effect. Wehave already selected a list of possible clusters ofgalaxies for which we plan to perform X-ray observa-

tions with XMM in order to get a precise map of theelectron distribution and the other relevant parame-ters. The measurements of the background thermalSZ effect in several bands of the millimetric region willbe carried out by the group of Prof. F. Melchiorri atMITO (Millimetric and Infra Red Testa Grigia Obser-vatory) at Testa Grigia-Cervinia (Italy).

300

H\ \

200 fO

0

\\ HYDRA A

ft \\\

A \ \\ \ \

\V\ \ \

V T"\ \ ;

A \\ \ \ T

. ,, ,1.1.1 , .\ . ,*. .

\

I , , , ! , , ,

0 20 40 60 80 100 120Teq[K]

Fig. 1: The equilibrium radius R^ for molecular cloudsis shown as a function of the cloud tempera-ture T^ for the Hydra cluster (see Ref. [1]).

Primordial Clouds: It has been suggested that forma-tion of molecules such as H2, HD and LiH can takeplace immediately after the recombination of cosmo-logical hydrogen [2]. We have shown [3] that duringthe early stages of gravitational collapse, for somecollapsing masses, HD molecules were the maincooling agent. We have analytically estimated themolecular cooling during the collapse of protocloudsand we discussed the dynamic of the early phase ofgravitational collapse. We show [4] that a thermal in-stability is possible at the beginning of the collapse of aprotocloud. This process could explain the formation ofprimordial clouds.

REFERENCES[1] D. Puy, L Grenacher, Ph. Jetzer, preprint (1998).[2] D. Puy et al., Astron.&Astrophys. 267, 337(1993).[3] D. Puy, M. Signore, New Astron. 3, 27 (1998).[4] D. Puy, M. Signore, New Astron. 3, 247 (1998).

44

REM MEASUREMENTS VERSUS STANDARD MODELS

P. Buhler, L. Desorgher, A. Zehnder (PSI), A. Glover (MSSL), E. Daly (ESA/ESTEC)

Radiation analysis in spaceflight are based on modeis from the late 60's and early 70's. Comparison ofrecent measurements of REM from PSI with these models reveals deficiencies.

RADIATION ENVIRONMENT MODELS: WHATFOR?

The charged particle environment in space is a seriousconcern for space missions. The hazardous effects onsatellites reach from charging/discharging with associ-ated system anomalies to penetration of high energeticparticles into sensitive parts, which can cause per-formance degradation, transient or even permanentfailures of electronic devices. The ionizing particles arealso harmful to human beings, which in the era ofregular manned space flights must be of special con-cern. In order to take appropriate measures to protectthe crew members and sensitive equipment aboard aspace vehicle a so-called radiation analysis is per-formed in the planing phase of a space mission. Itprovides an estimate of the radiation environment tobe encountered during the missions lifetime. Suchestimates are based on radiation environment models.The high energetic particle environment in the innermagnetosphere is dominated by electrons and protonstrapped in the earth magnetic field, forming the earth'sradiation belts. The most commonly used models topredict these particles are AE8 for electrons [1] andAP8 for protons [2]. These models have been com-piled in the late 60's and early 70's from a restrictedset of observational data. There are models for solarminimum and solar maximum conditions but otherwisethe models are static and are supposed to representthe average trapped particle environment. In order totest AE8/AP8 for their validity we compared them withmeasurements from the two Radiation EnvironmentMonitors, REM from PSI [3] aboard the UK satelliteSTRV-1 b in a nearly equatorial Geostationary TransferOrbit and the Russian space station Mir in a Low EarthOrbit. STRV-REM was operational between July 1994and September 1998 and Mir-REM between Novem-ber 1994 and November 1996.

MODELS VERSUS OBSERVATIONS

Whereas the experimental data set provides a largenumber of energy deposit histograms at differentpoints along the orbits, the radiation models provideelectron and proton energy spectra at a given point inspace. In order to compare models and observationswe used the model spectra to compute expected en-ergy deposit histograms and using the calibrated aver-age energy deposit per detector channel the total en-ergy deposit in each histogram was calculated. Theneither the observed and modeled histograms or thecorresponding total doses could be directly compared.

In order to reduce the amount of information, the ob-servational data was binned into L-bins (L: L-shellparameter) and for each bin the observations wereaveraged over the entire mission lifetime. The figureshows summary plots of the total doses as function ofL for the STRV-1 b orbit. The asterisks are model val-ues and the diamonds are the experimental values.The left panel shows results for the p-detector with ashielding of > 2.1 g/cm2 and the right panel, for the e-detector with a shielding of > 0.8 g/cm2, respectively.Whereas there is a rather good agreement betweenmodel and observations in the inner radiation belt (±30%) the models underrate the observations in theouter belt.. The discrepancy is due to the variability ofthe electron population trapped in the outer belt. Theratio between maximum and minimum dose rate ob-served between August 1994 and August 1998 in agiven L-bin was up to a factor 1000, showing thatstatic models are not adequate to describe the radia-tion belts and more sophisticated modeis are needed.

STRV, electrons & protons STRV, electrons & protons

a

10

10

2.3 3.S 4.7 5.9 7.1L-values

1.1 2.3 3.S 4.7 5.9 7.1L-values

Fig. 1: Comparison of STRV-REM average dose rate,experimental (diamonds) and model(asterisks) values. Left panel shows data for p-detector (>2.1 g/cm8 shielding) and the rightpanel for the e-detector(> 0.8 g/cm2 shielding).Differences are most important in the outer ra-diation belt which is highly dynamic.

REFERENCES

[1] J.I. Vette, NSSDC/WDC-A-R&S 91-24,1991.

[2] D.M. Sawywer, and J.I. Vette, NSSDC/WDC-A-R&S 76-06, 1976.

[3] P. Buhler et al., Nuclear Instruments & Methods,A 386, 825, 1996.

45

WHO CARES ABOUT WEAK GEOMAGNETIC STORMS?

P. Buhler, L Desorgher, A. Zehnder (PSI), E. Fluckiger (UNI Bern), E. Daly (ESA/ESTEC)

Although reiativistic electron enhancements do occur together with geomagnetic storms, the size of theenhancements does not correlate with the size of the storms. Better correlations are found with parametersmeasuring substorm activity.

The earth magnetic field is subjected to continuousvariations which are driven by the solar wind and theinterplanetary magnetic field, IMF. These variationsare known as magnetic activity and show differenttypical features. The most common type of activity aregeomagnetic substorms whose most prominentsignature are the brightening of the auroral arcs.Especially during periods of strong substorm activity ithappens that a geomagnetic storm develops, which ischaracterized by the build up of a ring current. Thering current consists of ions in the 100 keV rangewhich are trapped in the earth magnetic field andencircle the earth at a distance of typically 3 to 6 earthradii. These particles induce a southward orientedmagnetic field, such that the geomagnetic field in theinner magnetosphere is reduced. In magnetogramsrecorded at the earth equator this shows up as typicaldip which decreases to its minimum value withintypically a day and then recovers over days back to itsinitial level. The size of a storm is measured by thedisturbance storm-time index, Dst, which is thedeviation of the magnetic field H-component at theearth equator compared to the quiescent state and isgiven in nT. According to the minimum Dst reached,magnetic storms are classified into the three classesweak (-30<Dstmln), moderate (-30<Dstmin<-50), andgreat (Dstmin<-100).

Geomagnetic activity can influence our life on earth invarious manners. Telecommunication for example isaltered by magnetic activity induced changes in theionosphere and electric power systems can beknocked out by geomagnetically induced currents. Thenumber of such effects normally increases withgeomagnetic activity and are most severe during greatmagnetic storms. So one is most concerned about thelarge magnetic storms which threathen to break downcritical systems.

Besides life on earth also space born systems can beeffected by geomagnetic activity. Magnetic storms arenormally associated with significant variations of thereiativistic electrons trapped in the earth's outerradiation belt [1]. The initial magnetic field drop in thefirst phase of a magnetic storm is accompanied by adecrease of the electron fluxes, mainly due to betatrondeceleration, but during the recovery of the magneticfield particle fluxes increase to levels which can bemuch higher than the prestorm value. These electronenhancements can cause failures in satellites.

Concluding by analogy it could be expected, thatreiativistic electron enhancements do also scale withthe size of a magnetic storm which however is not thecase. This is shown in the figure (left panel) where

minimum Dst measured during a magnetic storm isplotted versus the > 1 MeV electron count rateincrease measured with REM from PSI aboard theSTRV-1b satellite in the outer radiation belt for 42magnetic storms from 1995 to 1997. The correlation isobviously poor. There are moderate to great magneticstorms with only small enhancements but worth, alsoweak magnetic storms with large electronenhancements. Thus Dst is no good predictor for thesize of the electron enhancements.

Better correlations are found with geomagneticparameters measuring the substorm activity. For theinvestigated set of 42 magnetic storms the bestcorrelation was found with Apint, which is thegeomagnetic index Ap integrated between storm onsetand the time the maximum electron rate was reached.Ap measures the averaged 3-hourly range of magneticfield variations at different geomagnetic observatoriesaround the world. The results suggest that theacceleration process which produces the electronenhancements is linked with substorm activity anddepends much less on the size of a magnetic storm.

So the answer to the question posed in the title of thisreport is: people who are concerned about the effectsof charged particles in space should be aware thatlarge reiativistic electron fluxes can occur withoutassociated large magnetic storm.

weak moderate great10

JlO J

<L A

» ' .

^ A

& A

t

A

A

A

A

. . . .

10'

50 100[nT]

10

10'200

Aplnt [days]150 0

a) " D s t - [ n T ] b)

Fig. 1: Minimum Dst a) and integrated Ap b) versuselectron count rate enhancement measured byREM in the earth's outer radiation belt for 42magnetic storms from 1995 to 1997. There isobviously a much better correlation of the sizeof the electron enhancements with Apint thanwith Dstmin.

REFERENCE

[1] P. Buhler et al., PSI Annual Report 1997, AnnexIHA, p. 145 and 146.

47

Muon Spin SpectroscopyMagnetism:

Heavy Fermion SystemsMagnetic Ordering and Dynamics

Spin Glasses

Superconductivity

Semiconductors / Hydrogen in Metals

Chemistry:Radicals

Reaction KineticsCharge Transfer

Experiments with Low Energy Muons

Experimental Techniques

48

AtSR STUDIES OF CORRELATED ELECTRON GROUND STATES (I)

RA-90-05, ETH ZURICH - LEIDEN - LOS ALAMOS- PSI - RIVERSIDE- SAN DIEGO

A.Amato*, C. Baines*, R.H.Heffnerts, D.E. MacLaughlints, M.B.Maple§, G.J.Nieuwenhuys11, H.R. Otf* s ,A.Schencktt, J.E.Sonier^, J.L Smith*

* Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerlandf Los Alamos National Laboratory, Los Alamos, New Mexico 87545, U.S.A.t University of California, Riverside, California 92521, U.S.A.§ University of California, San Diego, La Jolla, California 92093, U.S.A.f Kamerlingh Onnes Laboratory, Leiden University, 2300 RA Leiden, The Netherlands** Laboratory for Solid State Physics, ETH-H6nggerberg, CH-8093 Zurich, Switzerlandtf Institute for Particle Physics, ETH Zurich, CH-5232 Villigen PSI, Switzerland

UBe13 and alloys. The current understanding ofbroken time-reversal symmetry (TRS) pairing states inuncoventional superconductors is far from complete. Atheoretical consensus is emerging that aTRS-violatingstate can be induced by impurities or by application ofa magnetic field [1]. Whether these predictions are in-deed relevant remains to be verified. From our own//SR measurements on the heavy fermion supercon-ductor UBei3 we find novel behaviour below the tran-sition temperature Tc which is qualitatively consistentwith some of the most recent ideas concerning the for-mation of a broken TRS state. Experimental verifica-tion of TRS-violating states is an important step to un-derstanding unconventional superconductivity.

At TRIUMF in 1997 we performed transverse-field/uSR measurements on single-crystal UBei3 where weidentified the muon site as (0,0,1/4). In an appliedfield, both a positively and negatively frequency-shiftedprecession signal were observed due to the hyperfinefields produced by the U /-electron moments. BelowTc one would expect the magnitude of the shift to de-crease in a singlet-pairing superconductor, i.e. the posi-tively-shifted line should decrease in frequency and thenegatively-shifted line should increase. Instead, weobserve that both lines undergo a positive (paramag-netic, PM) shift, indicative of the formation of an inter-nal magnetic field. This additional field is characteristicof a superconducting state which breaks TRS. A com-plete temperature scan was obtained at an applied fieldH = 0.75 T.

Additional temperature scans in UBei3 were takenat PSI in 1998 at H = 1.0, 1.5, and 2.0T. As Fig. 1shows, the PM shift scales with magnetic field.

We also found that the muon depolarization rategenerally decreases with temperature below Tc. Thisis opposite to the expected behaviour of the supercur-rent response. We currently have no explanation forthis latter finding. Transverse-field/iSR measurementsat low fields and in the normal state above Te are stillneeded for a complete interpretation of the anomalousbehaviour in UBe13.

UBe,

-3.9

-4.1

-4.3

-4.5

11

io 'e

9 ^

8 I

0 0.2 0.4 0.6 0.8 1.0

TEMPERATURE (K)

Figure 1: The experimental Knight shift for the twomagnetically inequivalent muon sites denoted by theprecession frequencies vL and v\\. The frequency ofthe muon precession in the silver backing is denotedi/Ag- The data are for two different transverse fields,i.e., H = 1.0T (open circles) and 1.5T (stars).

REFERENCES

[1] M. Sigrist etal., Phys. Rev. Lett. 74 (1995) 3249;A.V. Balatsky, Phys. Rev. Lett. 80 (1998) 1972;G. Preosti and M. Palumbo, Phys. Rev. B 55 (1997)8430;J.X.Zhu etal., Phys. Rev. B 57 (1998) 13410.

49

txSR STUDIES OF CORRELATED ELECTRON GROUND STATES (II)

RA-90-05, ETH ZURICH - LEIDEN - LOS ALAMOS - PSI - RIVERSIDE - SAN DIEGO

A.Amato*, C.Baines*, R.H.Heffnerts, D.E.MacLaughlin*5, M.B.Maple§, G.J.Nieuwenhuys11, H.R.Ott**s,A.Schencktt, J.E.Soniert

* Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandf Los Alamos National Laboratory, Los Alamos, New Mexico 87545, U.S.A.i University of California, Riverside, California 92521, U.S.A.§ University of California, San Diego, La Jolla, California 92093, U.S.A.i Kamerlingh Onnes Laboratorium, Rijksuniversiteit Leiden, 2300 RA Leiden, The Netherlands** Laboratorium fur Festkorperphysik, ETH-H6nggerberg, CH-8093 Zurich, Switzerlandff Institut fiir Teilchenphysik, ETH-Zurich, CH-5232 Villigen PSI, Switzerland

Non-Fermi liquids. Evidence has emerged fromour /iSR and NMR experiments [1, 2] that in atleast some cases non-Fermi-liquid behaviour in heavy-fermion compounds [3] is due to random disorder inthe material. The disorder causes a wide distributionof local characteristic energies A^; this in turn impliesa broad distribution of the local spin polarization XJ °<1/Aj at low temperatures, direct signatures of whichare correspondingly broad ^SR and NMR lines. In thesimplest "Kondo disorder" picture [1] A,- is the Kondotemperature TK, disordered by variation of the local-moment/conduction-electron hybridization.

Our results have stimulated a considerable amountof theoretical work on disorder-driven mechanisms fornon-Fermi-liquid (NFL) behaviour. The Kondo disordermodel was put on a firm theoretical footing by Mirandaera/. [4], who found that the effect of structural disorderwas enhanced by electron correlation effects. A recentextension of the model by Castro Neto et al. [5] takesRKKY interactions between uncompensated spins intoaccount.

We have begun a study of zero- and longitudinal-field ^SR relaxation to probe spin dynamics and mag-netic ordering in NFL materials. ZF- and LF-//SR mea-surements were carried out in the LTF on the NFL com-pound UCu4Pd. The relaxation function could be fittedtothe"stretched-exponential"formG(i) = exp[-(Ai) /?],suggesting inhomogeneity in the relaxation rate. Fig-ure 1 gives the temperature dependence of A and /? forlongitudinal field HL=0 and 300 Oe. For HL - 300 Oethe relaxation rate is depressed but not eliminated, in-dicating that dynamic spin-lattice relaxation is involved.

These results are very significant for understand-ing disorder-driven NFL behaviour. Within the setof disorder-driven NFL theories the role of RKKY-likeinteractions between uncompensated local momentsand consequent cooperative behaviour has provedcontroversial. An estimate of the muon relaxation ratecan be made assuming the simple Kondo disordermodel [4] which neglects RKKY interactions. The re-sult is a / i + relaxation rate ~10~3 /us"1, which is morethan two orders of magnitude smaller than observed inour recent experiments. This implicates slowing downof local-moment fluctuations by RKKY interactions,which is an important ingredient of the "Griffiths-phase"model [5]. The observed inhomogeneous (stretched-

£ 1 . 0CD

I0.5 -

10) 0.02•g 0.5

I<r

0.0

• G(t) =

, (a)

* • •: (b)

UCu4Pdn D a n D o D D D

• •

(

* • . At . . ?

H L = O :

° a D :A

* • •

= 300 Oe ;

0.0 0.5 1.0

Temperature (K)

Figure 1: Temperature dependence of muon relaxationparameters A and /? in UCu4Pd below 1 K. (a) Longitu-dinal field EL = 0. (b) HL = 300 Oe.

exponential) relaxation suggests cluster formation, alsoin agreement with the Griffiths-phase picture.

REFERENCES

[1] O.O. Bernal et al., Phys. Rev. B 54 (1996) 13000.

[2] D.E. MacLaughlin, O.O. Bernal, and H.G. Lukefahr,J. Phys.: Condens. Matter 8 (1996) 9855.

[3] See, for example, articles in Proceedings of theConference on Non-Fermi Liquid Behaviour in Met-als, Santa Barbara, California, 1996, edited by P.Coleman, M.B. Maple, and A.J. Millis, J. Phys.:Condens. Matter 8 (1996).

[4] E. Miranda, V. Dobrosavljevic, and G. Kotliar, J.Phys.: Condens. Matter 8 (1996) 9871; Phys. Rev.Lett. 78 (1997) 290.

[5] A.H. Castro Neto, G. Castilla, and B.A. Jones,Phys. Rev. Lett. 81 (1998) 3531.

50

HEAVY-FERMION BEHAVIOUR AND STRIPE ORDER IN ELECTRON DOPED CUPRATES

RA-94-12, BRAUNSCHWEIG - KARLSRUHE - KOLN

H.-H.KIauf35*, D.Baabe*, W.Kopmann*, W.Wagener*, J.Littersts*, RAdelmannt, G.Czjzek*, H.Berg*

* Institut fur Metallphysik und Nukleare Festkorperphysik, Technische Universitat Braunschweig, Germany.f Institut fur Nukleare Festkorperphysik, Forschungszentrum Karlsruhe, Germany.i II. Physikalisches Institut, Universitat zu Koln, Germany.

Nd2-3/Cej/Cu04 belongs to the superconducting cu-prates in which the carriers of the superconducting cur-rent are electrons. These compounds RE2_yCej/Cu04

(RE=Pr, Nd, Eu) show T' structure with two-dimensio-nal sheets of Cu-0 squares and no apical oxygen.

In this system the Ce-doping changes at the sametime the Cu-3d magnetism by introducing charge carri-ers as well as the Nd-4f spin system by diluting it. Todifferenciate their influence on the low energy excita-tions we studied the system (Ndi_.rREx)2_yCej/Cu04

(with RE=La,Pr,Eu) where we can independently dilutethe Nd spin lattice and change the carrier doping inthe CuO2 planes. This systematic study with macro-scopic and microscopic methods enables us to differ-entiate between the cooperative Nd spin wave modelby P.Thalmeier [1] and single ion models (e.g. by Igara-shi et al. [2]). In 1998 we extended the 1997 exper-iments on (Nd0.6La0.4)2-yCej/CuO4 to further dilutedNd-f4 spin systems. Since the La ion radius is too largefor stronger substitution (a transition to the T structureof La2Cu04 would occur) further dilution of the Nd-lattice is possible by doping with non magnetic rareearth ions like Eu and Pr. This leads to systems (Pri-^Nd;c)i_j/Cej/CuO4 and (Eu1_;rNd:r)1_yCej/CuO4. Evenin these systems a Ce-doping of y = 0.2 has not beenachieved without precipitations of the T structure of La2

CuO4. Therfore we used a maximum Ce-doping ofy =0.15.

We performed zero field (ZF) and longitudinal field(LF) /iSR experiments on a Ndi.85Ce0.i5CuO4 singlecrystal with a superconducting transition temperatureof Tc « 12 K at the GPS and LTF spectrometer. Themeasurements show an increasing //+ relaxation ratebelow 3 K (Figure 1), which can be understood in termsof a development of a quasistatic field distribution withdecreasing temperature. Between T =0.1 and 2.3 Kwe found a fluctuation rate of 3 MHz. Compared to thefluctuation rate of 0.5GHzforNdi.8Ce0.2CuO4 [3] in thesame temperature range the observed fluctuation ratediffers by two orders of magnitude. This can be tracedback to the reduced charge carrier concentration andthe proximity to the magnetically ordered state (y < 12).

To separate the Nd-Nd and Nd-Cu exchange inter-actions we have performed ZF and LF fj,SR investiga-tions on polycrystalline samples (Pro.9Ndo.i)i.85Ceo.i5CuO4, Pri.85Ceo.i5Cu04, (Euo.gNdo.ih.ssCeo.isCi^and Eui.85Ce0.i5CuO4.The measurements on (Pro.9,Nd0.i)i.s5Ce0.i5CuO4 between 0.1 and 1 K show a qua-sistatic /*+ relaxation with vSiuc «3MHz. The magni-tude of this fluctuation indicates a magnetic origin sim-ilar to the undiluted Ndi.85Ce0.i5CuO4.

1 10temperature (K)

0.25 rr

Figure 1: Temperature dependence of the muon-spinrelaxation rate (top) and ZF/LF ^SR spectra at 0.1 K(bottom) for Ndi.85Ce0.i5CuO4.

The transition to this magnetic state occured at atemperature of Tmag «50K. In the reference samplePri.ss Ceo.i5 CuO4 we found Tmag «70K. Similar ex-periments have been performed on (Euo.gNdo. 1)1.35Ce0.i5CuO4. There we found a magnetic transition atTmag >20K. In summary, this behaviour indicates thecoexistence of superconductivity and (short range or-dered) magnetism in the Pr and Eu based compoundswith an electron doping of y =0.15. The doping de-pendence of the electronic groundstate is slightly differ-ent from the pure Ndi_!/Cej/CuO4 system. Similar ex-periments are planned on samples with increased Ce-contents to supress magnetic short range order downto the Kelvin range.

REFERENCES

[1] P. Thalmeier, Physica C 266 (1996) 89.

[2] J. Igarashi, K. Murayama and P. Fulde, Phys. Rev.B 52 (1995) 15966.

[3] M. Hillberg et al., Hyp. Int. 104 (1997) 221.

51

STUDY OF THE MAGNETIC PROPERTIES OF Ce3Pd20Si6 AND Ce3Pd20Ge6 COMPOUNDS

RA-97-04, PSI - DUBNA - MOSCOW

A.Amato*, Ch.Baines*, V.N.Duginovt5, A.V.Gribanov*, I.A.Krivosheev§, V.N.Nikiforov*, V.Yu.Pomjakushint,A.N.Ponomarev§5, Yu.D.Seropegin*

* Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland| Joint Institute for Nuclear Research, 141980 Dubna, Moscow region, Russiai Moscow State University, Moscow, 119899, Russia§ RSC "Kurchatov Institute", Kurchatov sq.1,123182 Moscow, Russia

The Ce3Pd2oSi6 compound is one of the heaviest-electron systems with 7 reaching 8J(molCe)~1K~2 at0.2 K. We have studied a 14g, polycrystalline sampleprepared by arc melting under argon atmosphere.

Zero-field measurements were undertaken to gaininformation on magnetic ordering at low temperatures.The measured muon-spin depolarization was fitted toa Gaussian G(t) = exp{-a2t2).

54.15

0.4

0.00.00 0.20 0.40 0.60 0.80 1.00 1.20

Temperature, K

Figure 1: Temperature dependence of the relaxationrate a in zero field.

The temperature dependence of the depolarizationparameter is shown in Fig. 1. Strong longitudinal fieldsapplied at 0.04 K did not reveal any field dependence ofa- above 0.5T: cr(0.5T) « <r(2.5T) = 0.18 (ns)~K Below0.4 K the increase of the depolarization rate representsthe development of quasi-static ordering of magneticmoments of electronic origin, supposedly random ori-ented.

Crystal electric field (CEF) calculations show thatin this compound the Ce moments may undergo frus-tration, which could explain such a "spin-glass" behav-iour. The depolarization caused by such a distributionshould be described by a Kubo-Toyabe function withthe typical 1/3 recovery of polarization. The absenceof the recovery is usually caused by fluctuations of themagnetic moments. We plan to study this effect bymeasurements in low longitudinal fields.

In the paramagnetic temperature region, transversefield measurements were performed using the GPD set-up. A clear frequency shift was seen (Fig. 2).

54.10

oa3cr£54.05

54.00

Ce3Pd20Si6

TF 0.4 T

50 100 150 200 250Temperature, K

300 350

Figure 2: Temperature dependence of muon spin pre-cession frequency.

This fact may be attributed to the increase of totalmoments of the superparamagnetic cube (SPC) con-taining 8 Ce-2 atoms and their ferromagnetic orderingwith decreasing temperature. The muon Knight shifton Pd is too small to explain our results. It should benoted also that after switching of the 0.4T field at 20 K,a remaining field of 2G was seen. This fact requiresfurther study.

SPC formation may be confirmed by measurementsof the dependence of the frequency shift on the exter-nal field at various temperatures. The frequency shiftshould deviate from being linear in the external field,due to the increase of the total moment of the cube. Inorder to exclude possible methodical errors, we hopeto perform additional TF measurements using the Ja-nis cryostat instead of the closed-cycle refrigerator.

52

MOMENT COLLAPSE IN

RA-97-10, MUNICH - ETH ZURICH

G.Grosse*, G.M.Kalvius*5, E.Schreier*, A.Kratzer*, F.J.Burghart*, K.Mattenbergert, O. Vogt*

* Physik-DepartmentE15, Technische Universitat Munchen, D-85747 Garching, Germanyf Laboratorium fur Festkb'rperphysik ETH Zurich, CH-8093 Zurich, Switzerland

The pseudo-binary system UKLai_xS, in which theNaCI structure is maintained throughout, shows quiteunusual magnetic behaviour. Bulk magnetic [1] andneutron diffraction [2] studies revealed:(1) The change of lattice constant is smooth and linear.(2) The effective U moment is practically independentof x at ~ 2 . 2 5 ^ B . while the paramagnetic Curie tem-perature decreases linearly from 184K {x = 1) to 52 K{x = 0.08). The change in sign occurs at x = 0.3.(3) The Curie temperature first decreases linearly from177K(ar = 1) to ~100K for x = 0.6. In this temperaturerange the ordered moment seen by neutrons changeslittle {fiord « 1.5/JB)- Below x = 0.6, magnetic Braggreflections could no longer be detected. The orderedmoment must have collapsed to fiord < 0.4 fiB, whichwas the sensitivity limit. Extrapolation of high field (9T)magnetization to zero field yields moments around 0.3-0.5 fiB below the region of collapse, but this procedureis considered unreliable due to domain effects.To clarify the situation we have performed fj.SR mea-surements on single crystals with x = 1, 0.8, 0.55, 0.4,0.15 and 0 between 300 K and 0.1 K at the TTM3 beam-line. From previous /iSR studies on related UX com-pounds with the NaCI structure, the muon is known tobe stationary (up to 300 K) at its stopping position lo-cated in the center of a cube formed by two U and twoX ions. At this position the local field 5M cancels for asimple AFM structure. In a FM like US only the contactfield will be present. Results of the present work are:LaS: Relaxation in ZF is given by a static GaussianKubo-Toyabe function arising from the nuclear dipoleson 139La. Full decoupling can be achieved in LF = 10 G.US: The precession frequency and the muon spin re-laxation rate in TF rise sharply on approaching Tc, asexpected when moving towards a second order mag-netic transition. The TF signal amplitude breaks downat Tc as a consequence of the appearance of sponta-neous magnetization. In ZF a spontaneous spin pre-cession pattern is seen below Tc. The frequency at100 K is Vvt = 63.4 MHz (SM = 4.7 kG).Uo.5Lao.2S: The TF results for this compound are simi-lar to those of US except for the shift in Tc, but sponta-neous spin precession could no longer be seen in ZF.Uo.55Lao.45S: According to the neutron data this com-pound already exhibits moment collapse, yet the TF-/iSR parameters clearly show the presence of a mag-netic transition at 110 K. The frequency shift near Tc isgreatly reduced, while for x = 0.8 practically no changehad been observed in this respect.Uo.4Lao.6S: Peaking of the relaxation rate A and of theprecession frequency v in applied field together with a

loss of signal strength Ao show that a magnetic transi-tion still takes place at 60 K. The peak frequency shiftis further reduced.Uo.15Lao.85S: TF measurements indicate a magnetictransition at around 0.45 K. The signal amplitude doesnot break down, however. The likely reason is the for-mation of an AFM state (no spontaneous bulk magneti-zation) as suggested by the negative Qp. The maximalfrequency shift is even further reduced.

"c(DEoE

| > 0.1-

"SEV,HI

0.01-

(a)

from ND (x) :

morriBntcollapse

I

\/I

iiUU'

180-

^ 160-

S 140-13•g 120-0)Q. 100-

o B0-H.g 60-

20-

n-

( b )

/ •

<• /

. /Y

, - • '

. . . • • • • '

0.0LaS

0.2 0.4 0.6 0.8 1.0us

0.0LaS

0.2 0.4 0.6 0.8 1.0US

U concentration (x)

Figure 1: Temperature dependencies of the U momentand the transition temperatures for

Due to the absence of spontaneous spin preces-sion in the pseudo-binaries U^Lai^S we estimatedthe magnitude of the U moment from the maximum ofthe frequency shift at Tc. We find that the collapse ofmoment between x = 0.6 and x = 0.55 does not leadto vanishing moments, but they do become quite small(see Fig. 1). Furthermore, we observe magnetic transi-tions in all compounds which vary roughly linearly withtemperature, meaning that the known high temperatureresult just continues. A theoretical study [4] predicts themoment collapse combined with the absence of mag-netic order. The latter result is not verified.

REFERENCES

[1] J. Schoenes etal., Phys. Rev. B53 (1996) 14987.

[2] F. Bourdarot etal., 27. J. d. Actinides, Dijon, 1997.

[3] L. Asch, Hyperfine Interactions 64 (1990) 351.

[4] B.R. Cooper and Y.-L Lin, J. Appl. Phys. 83 (1998)6432.

53

STUDY OF U- AND RE-INTERMETALLIC COMPOUNDS EXHIBITING QUADRUPOLAR ANDMAGNETIC ORDERING PHENOMENA: UPd3

RA-98-06, ETH ZURICH - OSAKA - LONDON - JERUSALEM

D. Andreica*, F.N. Gygax* M. Pinkpank*, A. Schenck*5, K.A. McEwent, A. Amato*

* Institut fur Teilchenphysik der ETH Zurich, CH-5232 Villigen PSI, Switzerlandf Department of Physics and Astronomy, University College London, London WC1 E6BT, U.K.i Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland

The hexagonal compound UPd3 with the dhpc-TiNi3structure has attracted considerable attention for nowabout 20 years [1], due to its rich low temperature prop-erties. Two low temperature phase transitions are ob-served at ~4.5K and ~6.5K. The lower one is be-lieved to signal the onset of some small moment AF-magnetic order (/j,ord ~ 0.01 ^B) and the upper one theevolution of quadrupolar order. UPd3 is also the onlyU-intermetallic compound in which a crystalline elec-tric field (CEF) splitting of the (5/2 )3#4 ground statemultiplet is seen in elastic neutron scattering, implyingrather well localized /-electrons. Actually, because ofthe ABAC-stacking of the hexagonal U-layers two in-equivalent U-sites are present with different CEF split-ting. In any case the lowest level is a non-magneticsinglet state. No direct evidence for AF order below4.5 K has yet been found. The quadrupolar order is stillunder investigation [2].

In view of this situation we have thought it worth-while to investigate UPd3 by a microscopic, local tech-nique, i.e. juSR. Here we report some first results onthe fx+ Knight shift, measured with the primary aim ofdetermining, as a first step, the //+ site or sites in thiscompound. We found a single component for S||c-axisand a split signal for Bl_c-axis. One of the two com-ponents disappeared for T > 30 K. On rotating B inthe basal plane no angular dependence of the Knightshifts was seen. This result implies that the fx+ site(s)must possess axial symmetry which restrict the possi-ble ^+-site(s) to the fe-sites and the /-sites (Wyckhoffnotation). These measurements also clearly revealedthe two low temperature phase transitions as can beseen from Fig. 1 which displays the frequency shifts onexpanded scales between 1.5 K and 10 K. It should benoted that the lower transition is only very weakly seenin the magnetic susceptibility. Overall the p+-Knightshifts do not scale with the bulk susceptibility which isnot so surprising in view of the two inequivalent U-sitesassociated with different atomic susceptibilities.

REFERENCES

[1] K. Andres et al., Solid State Commun. 28 (1978)405; H.R. Ott etai, Physica 102 B (1080) 148.

[2] K.A. McEwen et al., JMMM 140-144 (1995) 1411and references therein; M. Lenkewitz et al., J. Al-loys a. Compounds 232 (1996) 67.

NX

X

N

81.04

81.02

81.00

80.98 -

80.96 -

81.17

81.16-

81.15-

81.14-

81.13-

81.12

81.56-

81.54-

81.52-

81.50-

81.48-

81.46-

81.44-

,o -o--©- -©^

-4— $--•+-

TN

*•©-«

'N

Bic-axissignal 0

TQ

Bic-axis

TQ

2 3 4 5 6 7 8 9 10

Temperature (K)

Figure 1: Temperature dependence of the //+-spin pre-cession frequencies (B = 0.6T) for B-Lc-axis (a,b) and5||c-axis (c).

54

MICROSCOPIC STUDIES OF THE NON-MAGNETIC-MAGNETIC TRANSITION IN HIGHLYCORRELATED ELECTRON SYSTEMS

RA-98-07, ETH ZURICH - PSI - GENEVE - VIENNA - DRESDEN

D. Andreica*5, A. Schenck*, A. Amato ts, F.N. Gygax*, M. Pinkpank*, J. Sierro*, D. Jaccard*, G. Wiesinger§,Ch. Geibel11, C. Reichl§, E. Bauer§

* Institute for Particle Physics, ETH Zurich, CH-5232 Villigen PSI, Switzerlandf Paul Scherrer Institut (PSI), CH-5232 Villigen PSI, Switzerland% Universite de Geneve, DPMC, 24 q. E. Ansermet, CH-1211 Geneve 4, Switzerland§ Inst. of Exp. Phys., TU-Vienna, Wiedner Hauptstrasse 8-10, A-1040 Vienna, Austriaf Max-Planck-lnstitut fur chemische Physik fester Stoffe, Bayreuther Str. 40, D-01187 Dresden

We have studied the nonmagnetic-magnetic transi-tion in the Yb(Cu1_jrNi;c)2Si2 series. The compoundswere prepared and characterised by X-Ray diffraction,resistivity and thermoelectric power measurements atthe University of Geneva [1].

The transition is induced by the Ni substitution of Cubecause of the increase of the chemical pressure at theYb ion site, in agreement with the Doniach's magneticphase diagram [2]. Temperature scans in zero mag-netic field (ZF) were performed for different Ni concen-trations. The transition was clearly detectable in thejuSR spectra from the occurrence of a fast depolaris-ing component. This is also true for the £=1 compoundwhich possesses, according to neutron and Mossbauerdata, a well defined helical structure.

From the LTF and GPS runs we observed that themagnetic correlations develop in a fraction of the sam-ple even at the lowest Ni concentrations. Fig. 1 showsthe monotonic increase of the transition temperature ofthe magnetic fraction of the sample as a function of thenickel concentration x. The YbCu2Si2 sample is para-magnetic down to the lowest temperature in the LTFspectrometer, approximately 10 mK. In this compoundthe relaxation of the muon's spin is due to the static nu-clear field distribution arising in principle from the Cunuclei.

An interesting point is that the volume fraction whichexhibits a static magnetic response is clearly depen-

0.0 #

dent on the concentration of Ni ions. A preliminaryanalysis shows that the magnetic fraction is directlyproportional to the fraction of Yb ions having at leasttwo Ni-ion nearest neighbours (see Fig. 2).

1.0

0.00 0.25 0.50 0.75

Ni concentration x1.00

0 25 50 75

Cone, of the Ni atoms in Yb(Cu1 -*Nix)2Si2

Figure 2: The probability, as a function of the Ni con-centration x, that an Yb ion has 0...8 Ni ions as nearestneighbours. The right axis gives the magnetic fractionof the sample.

The curves in Fig. 2 were computed assuming astatistical distribution of the substituted Ni atoms in theavailable sites. This distribution reflects the disorder inthe Yb(Cui_3;NU)2Si2 system, observed:

• in transport properties - large residual restivities

• in /uSR - wide field distribution at the muon sitein the magnetic phase leading to huge relaxationrates.

REFERENCES

[1] D. Andreica etal., to appear in Physica B.

[2] S. Doniach, Physica B 91 (1977) 231.

[3] G. Andre et et al., J. of Alloys and Comp., 224(1995)253.

Figure 1: Dependence of the transition temperature onthe nickel concentration x.

55

fiSR STUDY OF THE NON-FERMI-LIQUID BEHAVIOUR IN U2Pt2ln

RA-98-14, AMSTERDAM - LEIDEN - PSI - ETH ZURICH

P. Estrela*5, A. de Visser*, G.J. Nieuwenhuys*, A. Schenck*, A. Amato§

* Van der Waals-Zeeman Institute, University of Amsterdam, The Netherlandsf Kamerlingh Onnes Laboratory, University of Leiden, The Netherlands| Institute for Particle Physics, ETH Zurich, Switzerland§ Paul Scherrer Institute, Switzerland

U2Pt2ln is a heavy-fermion compound showing pro-nounced non-Fermi-liquid (NFL) behaviour in its ther-mal and transport properties [1]. It deserves specialattention because it is a stoichiometric compound incontrast to most other NFL materials. The yuSR tech-nique is an excellent tool to investigate NFL behaviour.On the one hand, the muon may serve as a local probeto measure the internal field distribution related to thedisorder, while on the other hand it is an extremely sen-sitive probe for the detection of tiny ordered momentsand the proximity of a quantum critical point.

We have characterized extensively U2Pt2ln singlecrystals prepared by L.C.J. Pereira (ITU, Karlsruhe).The thermal and transport properties show pronouncedNFL behaviour. The specific heat varies as T\n(T)in the range 0.1-6 K [1] and the resistivity varies asTa, with a ~ 1(0.5) for /||a (c) [2]. The magnetic sus-ceptibility shows enhanced Pauli paramagnetism anda weak maximum at Tmax = 8K for B\\c, suggestingantiferromagnetic correlations.

The //SR experiments confirm the absence of staticmagnetic order. Transverse field (TF) /iSR experimentswere performed in the temperature range 5-300 K ina field of 0.6 T applied along the crystallographic a-and c-axes (tetragonal symmetry). An inspection ofthe Fourier transforms of the spectra shows that possi-bly 4 peaks are present in the frequency range 80.9-

2.5

2.0

0.5

u2pynsingle-crystal

81.5 MHz. As the temperature is increased above 10 Kthe peaks appear to "collapse" on each other. Due tothe small frequency shifts observed, a proper analysisof the data was so far not possible.

The zero-field (ZF) /iSR spectra were fitted to a two-component function containing an exponential and aGaussian term. The Gaussian component has a linewidth of <rG ~ 0.3 fis'1 at high temperatures, which in-creases slightly at the lowest temperatures. This valueof aG might be attributed to the nuclear moments of115ln (/ = 9/2, (i = 5.5/J.N)- The temperature depen-dence of the line width of the exponential component(Fig. 1) yields evidence for magnetic fluctuations belowJ-max-

Analysis of the longitudinal field (LF) measurementsat 60 mK suggests that the Gaussian signal has a sta-tic origin (decoupling of magnetic moments), while theexponential component corresponds to dynamic mo-ments. This could be an indication that in one stoppingsite the main contribution arises from a distribution of Innuclear moments, while at the other site the magneticfluctuations are dominant.

REFERENCES

[1] P. Estrela et al., to appear in Physica B (Proc.SCES, Paris 1998).

[2] P. Estrela ef al., J. Phys.: Condens. Matter 10(1998)9465.

0.1 1 10 100

T(K)

Figure 1: Linewidth of the exponential component ofthe zero-field ^SR signal of U2Pt2ln. The solid line isto guide the eye (note the logarithmic scale).

56

STUDY OF THE SPIN DYNAMICS IN AXIAL MAGNETS

RA-95-14, DELFT - GRENOBLE - AMSTERDAM

C.T.Kaiser*5, P.C.M.Gubbens*5, RDalmas de Reotier1, A. Yaouanc*, A.A.Menovsky*.

* Interfaculty Reactor Institute, Delft University of Technology, 2629 JB Delft, The Netherlandsf Commissariat a I'Energie Atomique, DRFMC/SPSMS, F-38054 Grenoble Cedex 9, France| van der Waals Zeeman Laboratory, University of Amsterdam, 1018 XE Amsterdam, The Netherlands

The purpose of our study is the investigation of thespin dynamics in axial magnets using //SR. The com-pounds under study have the tetragonal ThCr2Si2 struc-ture. The magnetic structures are less complex as com-pared to the rare earth metals, which will facilitate anunderstanding of the /iSR data.

A theoretical model has been developed, which hasproven to successfully describe the /iSR data of thesimple axial ferromagnet GdNi5 [1]. At low tempera-tures the muon spin is depolarized by a two magnonprocess. Near Tc the thermal behaviour is controlledby the dipolar interaction between the Gd+ 3 ions. Thiswork has shown the advantages of ^SR because infor-mation is obtained at wave vector q very close to zeroi.e. very near the wave vector describing the magneticstructure.

Our intention is to use a similar model for the studyon the antiferromagnet NdRh2Si2, the long range com-mensurate antiferromagnet PrCo2Si2 and the incom-mensurate antiferromagnet NdRu2Si2.

In 1996 we have investigated the temperature de-pendence of the frequency and the damping rate belowTN in PrCo2Si2 with the initial muon beam polarizationperpendicular to the c-axis. In contrast to the muonprecession frequency, the damping rate clearly showedsome anomaly at the expected transition temperatures[See Fig. 1]. The first transition at T\ =9.6K sepa-rates a short from a long range commensurate mag-netic structure, whereas the second one at T2 = 16.4 Kis in between two long range commensurate magneticstructures [2].

Although one expects a magnetic field distributionat the muon site, due to the incommensurate magneticstructure, these data were analysed with a single muonprecession frequency and an associated exponentialdamping rate A. Dalmas de Reotier and Yaouanc [3]derived a formula to describe the depolarization func-tion for an arbitrary field distribution at the muon site. Inorder to compute the magnetic distribution at the muonsite, the muon site itself is needed.

To obtain the muon site, Knight shift measurementshave been performed in November 1998. In general anangular scan gives information on the symmetry at themuon site, due to the angular dependence of the dipo-lar interaction. A scan in the aa-plane at 35 K and anexternal field of 6 kG, shows little, if any, angular depen-dence. This would indicate that the muon is situatedat an axial site, e.g. the expected ( | , | ,0 ) site. Thecoupling tensor at the muon site, can be deduced fromtemperature scans along the three axes and the sus-ceptibility. For this reason a temperature scan with the

field along the a-axis was done. It shows that the muonfrequency distribution develops from one single peakT < 35 K to two peaks for T > 55 K [See Fig. 2]. Thisunexpected behaviour could indicate that the (^,0,0)site becomes populated in this temperature range. Fur-ther analysis is needed to get a consistent picture.

Din

g ra

te (1

£CDQ

100

10

1

1

r

r

|

0

PrCo2Si2

Tj | J41 • /

j

1o

, , , , i , , , , i

10 20Temperature

r

(K)

' ' ' ' '

1 •

= ^

, i , , ,

30

Figure 1: Temperature dependence of the exponentialmuon damping rate of PrCo2Si2l with the initial muonbeam polarization perpendicular to the c-axis

4000

I 3000

g* 2000

1000

1 ' 1 ' 1

: PrCo2Si2

; o

0. o #.<: o o .**- oo -••

1 , 1 , 1

80.8 81.0 81.2

i •

&

tu° o0

\ .

1

81.4Frequency (MHz'

t ' 1

o T=140 K :

o T= 55 K "

• T=32K :

o

•••!§ :i , i

81.6 81.8

Figure 2: The muon frequency distribution, withBext || a, at three temperatures.

REFERENCES

[1] A. Yaouanc etal., Phys. Rev. B 53 (1996) 350.

[2] T. Shigeoka etal., Physica B 156 (1989) 741

[3] P. Dalmas de Reotier and A. Yaouanc, J. Phys.:Condens. Matter 9 (1997) 9113.

57

SPIN BEHAVIOUR IN NOVEL TRANSITION METAL MAGNETS

RA-98-04, LEIDEN - LOS ALAMOS - RIVERSIDE - COLUMBIA

R.H. Heffner*s, J. Gardner*, D.E. MacLaughlint, GJ. Nieuwenhuys*, G.M. Luke§ J.E. Sonier*, Y.J. Uemura11

* Los Alamos National Laboratory, Los Alamos, New Mexico 87545, U.S.A.f University of California, Riverside, California 92521, U.S.A.t Kamerlingh Onnes Laboratory, Leiden University, 2300 RA Leiden, The Netherlands§ McMaster University, Hamilton, Ontario, Canadaf Columbia University, New York, New York, 10027 U.S.A.

The goal of these experiments is to elucidate thestatic and dynamic spin behaviour in a variety of ma-terials exhibiting strong coupling between spin, chargeand lattice degrees of freedom. Previously, we havestudied [1] the 3D perovskite materials (La,A)MnO3, A= Ca, Sr or Ba, which exhibit colossal magnetoresis-tance (CMR) near their paramagnetic/insulating to fer-romagnetic/conducting transition temperature Tc. Cur-rently we are focussing on the 2D Ruddlesden-Popperstructures [2] (Lai-^Sr^MnOs^SrO, n = 2 [3], in or-der to compare the effects of reduced dimensionalityon the physics of the manganites. In 1998 zero andlongitudinal field data were collected on a single crystalof Lai.2Sri.8Mn207, (x = 0.4), which is a ferromagnetwith the spins aligned in the a-b plane [4]. Fig. 1 showsa preliminary analysis of the temperature dependenceof the muon precession frequency z^(T) (proportionalto the sublattice magnetization). Fig. 2 shows the spinlattice relaxation rate l /T i for the muon spin orientedalong the c-axis. For our sample Tc = 114.5 K. A cur-sory inspection of these data reveals signatures of their2D nature: (1) the temperature dependence of J^(T) isunlike the sublattice magnetization found in the 3D fer-romagnet La0.67Ca0.33lvlnO3, and (2) there is no signof critical slowing down in the relaxation rate, usuallyfound in 3D ferromagnets. Furthermore, the slow Mn-ion spin fluctuations found [1] by us below Tc in the fer-romagnet Lao.67Ca0.33Mn03, and attributed to spin po-larons, is completely absent in Lai.2Sr18Mn2O7. Thelatter is consistent with l /Xi behaviour discovered pre-viously in Lai.4Sri.6Mn207 (x = 0.3), which undergoescomplicated magnetic order along the c-axis, and at-tributed to the effects of the 2D nature of this system[5].

REFERENCES

[1] R.H. Heffner etal., Phys. Rev. Lett. 77 (1996) 1869.

[2] S.N. Ruddlesden and P. Popper, Acta Crystaliogr.11 (1958)54.

[3] T. Kimura eta!., Science 274 (1996) 1698.

[4] T.G. Perring etal., Phys. Rev. Lett. 78 (1997) 3197.

[5] R.H. Heffner etal., Phys. Rev. Lett. 81 (1998) 1706.

20 30 40 50 60 70 80 90 100110120T(K)

Figure 1: Temperature dependence of the muon pre-cession frequency v^ in La12Sr18Mn2O7.

0.2 •

0.1

0.0

0 50 100 150 200 250 300

T(K)

Figure 2: Temperature dependence of the spin latticerelaxation rate l /T i in Lai.2Sr1.8Mn207 in applied fieldsof zero and 3000 Oe.

58

INTERNAL FIELDS IN MAGNETICALLY ORDERED Dy AND Ho

RA-90-18, MUNCHEN - UPPSALA

E.Schreier*, M.Ekstrom1, O.Hartmann^ R.Wapplingt5, G.M.Kalvius*, A. Marelius*, S. Henneberger*,F.J.Burghart*, A.Kratzer*5

* Physik-DepartmentE15, Technische Universitat Munchen, D-85747 Garching, Germanyf Institute of Physics, University of Uppsala, S-75121 Uppsala, Sweden

For several years we have carried out systematicstudies of the magnetic properties of the heavy rareearth (RE) metals Gd, Dy, Ho and Er covering the para-magnetic as well as the magnetically ordered statesincluding especially in Gd the critical regime aroundTc- Furthermore, we have performed high pressurestudies, mainly in the magnetically ordered state. Sin-gle crystalline samples were used in all cases whichallowed in addition measurements of orientational de-pendencies of relaxation rates. Results on Gd and Erhave been reported previously. Here we give a briefsummary of the data obtained for Ho and Dy in theirmagnetically ordered states.

Dysprosium: Large axial anisotropy confines themoments on Dy to the basal plane. Between TN = 178 Kand Tc =85K, a helical AFM spin structure is present.The helix angle decreases with lower temperature. AtTc an orthorhombic distortion of the hep lattice occursand the transition into the FM state is of first order. Thespins point along the orthorhombic a axis. The ampli-tude of the spontaneous muon spin precession signalbelow Tjy is larger in the SM|| c geometry, meaning thatthe local field lies predominantly in the basal plane aswell. The temperature dependence of the precessionfrequency is shown in Fig. 1. The stopping site of themuon is not known, but calculations of the dipolar fieldsshow that a small shift of the local muon field shouldtake place at Tc (which is observed) if the muons oc-cupy tetrahedral interstitial sites, but not if the site is oc-

250

200-

120 -

100-

/ -

I 1

ti

I

i

1ti

Ki r

_TS4

i >

j

s- 170

- 165

1

\

I 1

I

\

\

]

\

0 20 40 60 80 100 120 140 160 180 200

Temperatur [K]

Figure 1: Temperature dependence of the sponta-neous precession frequency in magnetically orderedDy (single crystalline sample.)

20 40 60 80 100 120 140Temperatur [K]

Figure 2: Temperature dependence of the sponta-neous precession frequency in magnetically orderedHo (single crystalline sample.)

tahedral. The detailed behaviour in the ferromagneticregime, however, is more complex. An unusual reduc-tion in frequency on approaching lowest temperaturesis clearly visible. A first guess as to its origin is a slightre-orientation of spin directions. Dipolar field calcula-tions were not successful in reproducing the observedeffect under this assumption. At present we are forcedto consider the behaviour of spin precession (i.e. of BM)in the ferromagnetic state as not understood. Furtherwork is in progress.

Holmium: This RE metal possesses a rich varietyof spin structures which are often almost continuouslychanging with temperature. In the helical spin struc-ture formed below TN = 125 K the helix becomes dis-torted with decreasing temperature. Before reachingTc = 18 K, which leads into a shallow conical ferromag-netic structure pointing along the c axis, several spinslip transitions occur. The temperature dependence ofthe internal field (Fig. 2) shows no apparent irregulari-ties at the spin-slip transitions {e.g. 97, 40, 25 K). As inthe case of Dy an irregularity is seen around Tc- Thefield decreases again in the FM regime. The changein cone angle is the most obvious cause, but dipo-lar field calculations were unable to reproduce the ob-served effect. More complex interactions are probablypresent.

59

STUDY OF THE DYNAMIC AND STRUCTURE FEATURES OF MAGNETISM IN HOLMIUM

RA-98-01, DUBNA - PSI - MOSCOW

V.N.Duginov*5, K.I.Gritsaj*, D.Herlacht, I.A.Krivosheev*, A.A.Nezhivoy*, B.A.Nikolsky*, V.G.OIshevsky*,V.Yu.Pomjakushin*, A.N.Ponomarev*5, U.Zimmermann1

* Joint Institute for Nuclear Research, 141980 Dubna, Moscow region, Russiat Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland| RSC "Kurchatov Institute", Kurchatov sq.1,123182 Moscow, Russia

The magnetic helicoid of holmium is a spiral struc-ture whose wavelength along the c-axis and the dis-tance between the basal planes of the crystal latticeare incommensurate. This type of magnetic structureis described by the spin-slip model.

The ^SR method was used to measure the inter-nal magnetic fields at interstitial sites of the crystal lat-tice, where muons are localized. In a simple helicoidstructure all interstitial sites are magnetically equiva-lent, and in a //SR experiment only one muon preces-sion frequency is observed at a given temperature. Inthe spin-slip structure, the interstitial fields in differentsections of the helicoid are different and the frequencyspectrum of the muon signal is found to be more com-plicated.

We have performed earlier [1] ZF //SR measure-ments with a holmium sample consisting of a texturedpolycrystal with prefered orientation along the hexag-onal c-axis and with approximately 0.01% impurities.During the measurements the c-axis was oriented par-allel to the polarization of the muon beam. The experi-mental spectrum was well described by a single muonspin precession frequency. This fact characterized themagnetic structure of holmium as a simple spiral.

Low statistics experiments with a sample made of 3single crystals showed similar results [2]. At the sametime, the high values of the relaxation rate which de-scribed the width of spectral line (up to 60 s " 1 at 10K),could be due to several overlapping frequencies for thespin-slip structure.

ZF measurements were performed this October onthe GPD setup with a holmium single crystal which wascut from the textured sample used earlier. High statis-tics spectra were measured below the Neel tempera-ture. At every temperature two time-spectra were mea-sured - with the c-axes oriented perpendicular and par-allel to the polarization of the muon beam.

We have found that each frequency spectrum con-tains several broad peaks (Fig.1). This fact may be re-lated to the expected spin-slip structure but should beinvestigated and proven by additional measurements.Special attention should be paid to the temperaturepoints corresponding to the simplest expected spin-slipstructures.

0.3 -i

0.2 T70 K

0.0195 295 395 495

Frequency, MHz

Figure 1: Fourier spectra of the muon-spin precessionsignal measured on single crystal holmium at the indi-cated temperatures.

REFERENCES

[1] I .A. Krivosheev et al., JETP Lett. 65 (1997) No. 1,p. 81

[2] O. Hartmann, private communication

60

SPIN-FLUCTUATIONS IN YCo2

RA-98-03, WIEN-CRACOW

G. Wiesinger*5, E. Gratz*, Ch. Reichl*, P. Mietniowskyt, H. Figiel*, C. Kapusta*

* Institut fur Experimentalphysik, Technische Universitat Wien, A-1040 Wien, Austriaf Department of Solid State Physics, University of Mining and Metallurgy, PI-30-059 Cracow, Poland

Currently, the "enhanced paramagnetic" systemYCo2 has again found considerable attention. Fol-lowing preliminary inelastic neutron scattering experi-ments, spin fluctuations are supposed to occur in thecompound. At low temperatures single particle exci-tations may be dominant, whereas at more elevatedtemperatures, particularly at temperatures close to themaximum in the susceptibility (T « 300 K), collectiveexcitations are presumed to occur. Since YCo2 seemsto be close to static magnetic order, n+SR is antici-pated to exhibit a time scale suitable for studying thescattering due to thermally induced moments associ-ated with the cobalt sites. Due to the short beam timeallocated in late 1998 the data are of somewhat prelim-inary character.

0.25

0.15-

T i m e

Figure 1: Zero-field /z+S.R-spectrum at 4K fitted to astatic Kubo-Toyabe yielding A = 0.493(2) ^ s - 1

At low temperatures (T<4K) , zero-field /x+SR -measurements could be well described using a sta-tic Kubo-Toyabe relaxation function (Fig. 1) indicatingsmall static moments (in view of the muon time scale)randomly distributed. Following the theory of Kubo-Toyabe the starting dynamics in the zero-field mea-surements can be sufficiently well described usingthe dynamical Kubo-Toyabe function, which takes themuon diffusion into account. This process is best ex-plained by a "strong collision" model, as proposed byHayano et al. [1]. The value of A (see Fig. 2) wasfound to slowly decrease with increasing temperature(roughly proportional to T"1 /8) with a plateau at aboutA = 0.4/is"1 occuring in the range of above 50 K. Thehopping rate u was found to be below 1 (is*1 in theentire temperature range experimentally covered (seealso Fig. 3). A hint was obtained for the occurence of amaximum of v around 100 K. The real part of the initialsusceptibility (measured at / = 80 Hz, H = 220 A/m)

shows a maximum as well at a somewhat lower tem-perature around 70 K (although of a very low value «10-7m3/kg).

Reflectance spectroscopy [2] shows an unusualsharp excitation near A"1 = 120 cm-1 . There could besome relation, as this value leads to approximately T =115 K using ffcT = f.

0.6

0 5-jfc AW=°-5 8 T" a iV1

0 10 20 30 40

0 50 100 150 200 250 300temperature [K]

Figure 2: Temperature dependence of A for thezero-field scan. The inset shows a fit to the power law.

0 50 100 150 200 250 300temperature [K]

Figure 3: Temperature dependence of the "hoppingrate" v during zero field scans.

This work was supported by the Austrian ScienceFoundation (grant 13076).

REFERENCES

[1] R.S. Hayano etal., Phys. Rev. B20 (1979) 850-859.

[2] D. Crandles, Dept. of Physics and Astronomy,Franklin and Marshall College, Lancaster, PA, pri-vate communication.

61

/x+SR STUDIES OF THE TRIANGULAR LATTICE LiNIO2

RA-97-02, GRENOBLE -PSI

Tapan Chatterji*5, W. Henggeler*

* Institut Laue-Langevin, 156X, 38042 Grenoble CEDEX, France| Paul-Scherrer-lnstitut, CH-5232 Villigen PSI, Switzerland

The triangular lattice with antiferromagnetic inter-action is the famous example of a topologically frus-trated system. The simplest system is an antiferro-magnetic triangular lattice of Ising spins interacting withtheir nearest neighbours only. Such a system is knownto have a finite entropy at T = 0 and has no long rangemagnetic order at any finite temperature. Anderson [1]suggested that the ground state of the S = 1/2 Heisen-berg antiferromagnetic lattice may have resonating va-lence bond (RVB) state, which is a quantum liquid com-posed of randomly distributed movable singlet pairs ofspin. It has been further suggested [2] that the groundstate of anisotropic Heisenberg systems may be alsoa RVB state, even in the Ising limit. Lithium nicke-late Lii_ rNii+. rO2, which is always nonstoichiometric,forms a 5 = 1/2 triangular lattice for small value of xand has been suggested [3] to be a candidate for thespin liquid ground state.

We reported zero-field /uSR investigations on thenearly stoichiometric Lio.9sNio.02O2 [4]. We have ex-tended these measurements by applying a longitudi-nal magnetic field of 600 mT on the GPS spectrometerat PSI. The polycrystalline samples were pressed intocylindrical plates of radius of about 8 mm and thick-ness of about 2 mm by applying high pressure. Onesuch plate shaped sample was fixed on top of the sam-ple holder made of silver plate of very high purity. Thesample holder was mechanically fixed to the cold tip ofthe helium flow cryostat of the GPS beam line. Theincident muon beam was perpendicular to the sam-ple plate. The initial muon polarization was parallel tothe beam. The muon signals were recorded at sev-eral temperatures in the range from 2 to 140K. Thesignal showed no oscillations but showed relaxationswhich could be fitted by a stretched exponential func-tion Ae^~Dt^. Fig. 1 shows the temperature variationof the muon relaxation rate D and the exponent k. Therelaxation rate at T = 2K is 0.4 ±0.1 (^s)"1. The re-laxation rate increases rapidly with temperature andshows a maximum value of 1.5 ±0.1 (/is)"1 at aboutT = 10 K and then decreases with increasing tempera-ture. It becomes 0.053 ± 0.006 (/^s)"1 at T = 50 K. Theexponent k of the stretched exponential is 0.34 ± 0.01at T = 2 K. It increases continuously and becomes « 1at T = 30 K, which means that the relaxation becomesexponential. The temperature variation of both the re-laxation rate and the exponent of the stretched expo-nential suggests that some sort of spin freezing occursbelow about 10 K. We concluded this already from ourzero-field measurements. But the zero-field measure-ments showed huge relaxation rates at low tempera-

0 10 20 30 40 SOTemperature (K)

0 10 20 30 40

Temperature (K)

Figure 1: Temperature variation of the muon relaxationrate D and the exponent k of the stretched exponen-tial function used to fit the muon relaxation rate of thenearly stoichiometric Lii_a;Nii+xO2 with x = 0.02, un-der applied longitudinal magnetic field of 600 mT.

ture which we now ascribe to the nuclear spins of Ni.Similarly, zero-field measurements showed an expo-nent of about 2 at higher temperatures which suggestsa Gaussian relaxation rate, again most probably dueto the nuclear spin of Ni. We have eliminated the nu-clear spin dynamics by the longitudinal field of 600 mTat which we are left with the effect of the electronic spinof Ni.

REFERENCES

[1] P.W. Anderson, Mater. Res. Bull. 8 (1973) 153.

[2] P. Fazekas and P.W. Anderson, Phil. Mag. 30(1974)423.

[3] J.B. Goodenough et al., J. Phys. Chem. Solids 5(1958) 107.

[4] T. Chatterji etal., PSI Annual Report 1997/Annex I,p. 52.

62

MAGNETIC BEHAVIOUR OF A TWO-LEG ORGANIC SPIN-LADDER COMPOUND

RA-98-19, SUSSEX - BARCELONA

D. Arcon*, A. Lappas*, S. Margadonna*, K. Prassides*s, E. Ribera*, C. Rovira*

* School of Chemistry, Physics and Environmental Science, University of Sussex, Falmer, Brighton BN1 9QJ,United Kingdom

f Institut de Ciencia de Materials de Barcelona, Campus de la UAB, 08193-Bellaterra, Spain

Rovira et al. [1] have recently synthesised the or-ganic charge-transfer salt, [(DT-TTF)2]tAu(mnt)2] (DT-TTF = dithiophentetrathiafulvalene, mnt = maleonitriledithiolate) whose structure is suggestive of a two-legspin ladder system. Magnetic measurements [1] re-vealed that the electronic static susceptibility shows aCurie-Weiss behaviour at high temperatures and a ther-mally activated temperature dependence below 70 K.The susceptibility data were successfully fitted to theTroyer expression [2] for a two-leg spin ladder system.However, the calculated inter-ladder transfer integral,t' = 6 meV implies that the ladder-ladder interaction, J'is quite sizeable and its effect on the magnetic proper-ties cannot be ignored. If J1 is larger than the criticalvalue, Jc, the true ground state becomes a Neel or-dered rather than the spin-liquid state.

In this contribution, we employed the positive muonspin relaxation Cu+SR) technique in its ZF (zero field)and LF (longitudinal field) variants to authenticate thetrue magnetic ground state of [(DT-TTF)2][Au(mnt)2].

Down to 85 mK the ZF p+SR spectra fitted best todouble-relaxation function:

(1)

where A is the magnitude of the asymmetry, and Aand a are Lorentzian and Gaussian relaxation rates,respectively. No oscillating signal is seen at any tem-perature. The slowly relaxing Gaussian term with a re-laxation rate, a = 0.44(1 )(is~1 is characteristic of thepresence of weak static nuclear dipole moments. In ad-dition, the //+ spin is also relaxed by the rapidly fluctuat-ing electron spins which give rise to the exp(-Atf) com-ponent in eq. (1) with A = 0.04(2) ps'1. Above 20 K, Ais temperature independent (Fig. 1). This suggests thatthe relaxation at high temperatures is probably due tofast exchange fluctuations of the ladder spins.

Below 20 K, the ZF A<+SR spectra begin to changeconsiderably. We find that the double-relaxation func-tion of eq. (1) still describes the data well. However,while the Gaussian relaxation rate, a remains tempera-ture independent, the exp(-Atf) part damps faster withdecreasing temperature. A increases continuously asT decreases, approaching a value of 0.29(1 )ns~1 be-low 1 K (Fig. 1). The source of the observed relaxationcan be either fluctuating and/or static internal fields.

In order to distinguish between the static and dy-namic nature of the internal fields, we performed com-plementary longitudinal field (LF) /j+SR experimentsbetween 10 and 1000 G at 2 K. These showed the pres-ence of a slow relaxation of the n+ spin polarisation up

1 10

Temperature (K )

Figure 1: Temperature dependence of the zero field (•)and longitudinal field (25 G, • ) exponential relaxationrate, A.

to a longitudinal field of 100 G, providing support for thedynamic nature of the local fields even below 20 K.

In conclusion, our ZF and LF /u+SR experimentsshowed the dynamic nature of the local fields on muonsites ruling out any static long-range antiferromagneti-cally ordered or disordered ground state in [(DT-TTF)2][Au(mnt)2]. This means that the ladder-ladder interac-tions are not strong enough to destroy the spin-liquidground state. The general temperature and field de-pendence of A are very similar to that found in theSr2Cu406 two-leg ladder system where the increase ofA at low temperatures has been associated with crystalimperfections. Since magnetic susceptibility measure-ments revealed a Curie-tail below 15 K which is dueto chain-breaks, we also assign the observed slow fi+

spin relaxation to sample defects. The present resultssupport the spin-liquid ground state proposed originallyfor [(DT-TTF)2][Au(mnt)J.

REFERENCES

[1] C. Rovira et a/., Angew. Chem. Int. Ed. Engl. 36(1997)2324.

[2] M. Troyer etal., Phys. Rev. B SO (1994) 13515.

63

STUDY ON THE S = 1/2 QUANTUM SPIN SYSTEM KCuCI3

RA-96-10, BERN - ETH ZURICH

K. Kramer*5, H.U. Gudel*, F.N. Gygaxt, M. Pinkpank*, A. Schenctf, N. Cavadini*

* Institut fur anorganischen Chemie, Universitat Bern, CH-3000 Bern 9, Switzerlandf Institute for Particle Physics, ETH Zurich, CH-5232 Villigen PSI, Switzerlandi Laboratory for Neutron Scattering, ETH Zurich & PSI, CH-5232 Villigen PSI, Switzerland

The 5 = | copper compound KCuCI3 has recentlyattracted much experimental and theoretical interest be-cause of its quantum spin properties, leading to a non-magnetic singlet ground state and a spin excitation gap.Neutron scattering studies revealed well defined three-dimensional magnetic excitations above a spin gap ofthe order« 2.6 meV. A dimer model based on disper-sive singlet-triplet excitations was proposed to accountfor the measured data [1].

Inspired by earlier yuSR results on KCuCI3 [2], wedecided to perform exploratory studies in the GPS. Ourmeasurements in zero field showed a rapid decreaseof the asymmetry below T «5 K, which we ascribed tothe presence of a distribution of static internal fields.Additionally, fluctuating fields have been observed inform of a slow spin lattice relaxation. Upon raising thetemperature from 1.8 K the static part of the asymme-try is depressed in favour of the fluctuacting part, whichfinally dominates the signal above 5 K (Fig. 1). The ab-sence of any muon precession at low temperatures in-dicates that the static magnetic moments are randomlyordered. They might give rise to a distribution of fieldwidths at the muon place, because the characteristicdip of the static Kubo-Toyabe function has not been ob-served.

0 2 4 6temperature [K]

2 4 6temperature [K]

Figure 1: Temperature dependence of the asymmetriesrelated to static fields (•, gaussian function) and fluc-tuating fields (o, exponential function), and dephasingrate of the static component (right, gaussian approxi-mation).

The persistence of a depolarization at longitudinalfields which are much bigger than the internal field dis-tribution determined from Fig. 1 is interpreted as a di-rect proof for the dynamical nature of the slow relax-ation. Whereas the increase of the relaxation rate downto 5 K may indicate a slowing down of the fluctuationscommon to all external field strengths, the behaviour inthe low temperature range is strongly field dependent(Fig. 2).

0 2 4 6

temperature [KJ0 0.05 0.1 0.15 0.2 0.25

inverse temperature [K1]

Figure 2: Temperature dependence of the spin latticerelaxation rate at different longitudinal fields (left), andof the frequency shifts at 6 kG transverse field (right).

We found the transverse field signal to split into fourcomponents distinct by their paramagnetic shifts. At30 K we measured the angular dependence of the fre-quency shifts by rotating the sample around the (1,0,2)axis with the applied field perpendicular to this axis.The data showed the expected cos2 (0) dependence.Temperature scans between 2K and 30 K were per-formed for the two extremal orientations. The shiftsabove 5K followed very closely a Curie-like 1/T sus-ceptibility dependence (Fig. 2) quite in contrast to thebulk susceptibility [3].

The above results are difficult to put into the contextof a singlet quantum ground state, where for instanceno spin moment is expected. They may indicate thatthe presence of muon destroys the spin singlet stateof the nearest dimer spin pair or pairs. The onset ofsome kind of static magnetic order at around 5 K in thispicture would be restricted to just the immediate vicinityof the muon which is difficult to justify. It is important inthis respect to identify the muon-site(s) and also to findout whether the static fields are completely randomlyoriented or whether there is some preferred orientation.Next years measurements will serve this purpose.

REFERENCES

[1] N. Cavadini etal., Eur. Phys. J. B in press.

[2] W. Higemoto etal., Phys. Lett. A 243 (1998) 80.

[3] H. Tanaka etal., J. Phys. Soc. Jpn. 65 (1996)1945.

64

MAGNETIC PROPERTIES OF THE SPINEL ZnFe2O4 IN CRYSTALLINE ANDNANOSTRUCTURED MODIFICATIONS

RA-94-08, MUNCHEN - JULICH - CANBERRA - JENA - LEIPZIG

F.J.Burghart*. W.Potzel*5, G.M.Kalvius*, E.Schreier*, A.Kratzer*, G.GroBe*, W.Schafert, W.Kockelmannt,S.J.Campbell*, W.A.Kaczmarek§, A. Martin1', and M.K.Krause**

* Physik-DepartmentE15, Technische Universitat Munchen, D-85747 Garching, Germanyt Mineralogisches Institut der Universitat Bonn, D-53115 Bonn and Forschungszentrum (KFA) Julich, D-

52425 Julich, Germanyi Department of Physics, University College, The University of New South Wales, ACT 2600, Australia§ Research School of Physical Sciences and Engineering, The Australian National University, Canberra ACT

0200, Australia% Chemisch-Geowissenschaftliche Fakultat, Friedrich-Schiller-Universitat Jena, D-07743 Jena, Germany** Fakultat fur Physik und Geowissenschaften, Universitat Leipzig, D-04103 Leipzig, Germany

During the last four years we have investigated themagnetic properties of three differently prepared sam-ples of the spinel ZnFe2O4 with //SR, neutron diffrac-tion and Mossbauer effect measurements. In this con-tribution we provide a short summary of the results ob-tained.

The spinel structure is characterized by two dif-ferent cation sites: the A sites are tetrahedrally andthe B sites octahedrally coordinated by oxygens. Themagnetic coupling between the Fe atoms by superex-change via oxygen atoms is weak within the A andthe B sublattices whereas the A-B coupling is muchstronger.

The first sample investigated had been carefully an-nealed during its preparation. Neutron diffraction mea-surements showed that in this sample the Fe ions onlyoccupy the B sites in the spinel lattice (inversion < 3%).As the B-B interaction is weak, the Neel temperatureof this sample is low (TN = 10.5K). However, neutrondiffraction and ,uSR experiments proved that already atmuch higher temperatures (~80 K) parts of the sam-ple exhibit magnetic short range order. The correlationlength of this short range order is only about 2.7 nm.The volume fraction of the short range ordered regionsincreases with decreasing temperature and reaches75 % at TN. At TN the paramagnetic part of the sam-ple orders antiferromagnetically. The volume fraction ofthe short range ordered part decreases rapidly belowTN. But even at the lowest measured temperature of3 K about 20 % of the sample stays in a short range or-dered state. The fiSR experiments in longitudinally ap-plied fields show that short range order and long rangeorder are dynamic, producing field fluctuation rates atthe muon site of at least 1.5 GHz.

The second sample was rapidly quenched duringpreparation. The inversion of this sample was deter-mined to be 11 % by neutron diffraction. The neutrondiffraction measurements show that the sample doesnot exhibit any magnetic long range order down tothe lowest measured temperature of 2 K. Below ~25 K,however, one observes short range magnetic order-ing of a part of the sample. This transition into ashort range ordered state is also visible in /aSR and

Mossbauer effect data. The volume fraction of theshort range ordered parts of the sample increases be-tween 25 K and 10 K continuously with decreasing tem-perature. As the ^SR data show, below 10 K the sam-ple is magnetically short range ordered over its wholevolume. With longitudinally applied fields part of therelaxation rate of the jiSR signal could be decoupledat a temperature of 30 K. Therefore the fluctuation rateof the fields at the muon site must be much lower thanin the annealed sample and can be estimated to be~100MHz.

Finally a nanostructured ZnFe2O4 sample was in-vestigated. The average size of the particles in thissample is about 9nm. The inversion was determinedto be 44%. At present two different explanations forthe experimental observations are possible. In the neu-tron diffraction spectra an increase of intensity in someBragg reflections is visible. This can be explained byferromagnetic ordering of part of the sample, the Curietemperature being about 463 K. In the yuSR spectra theferromagnetic component is noticeable by a loss of ini-tial asymmetry. Neutron diffraction and ^SR data arefully consistent with an increase of the volume frac-tion of ferromagnetic ordered regions with decreas-ing temperature. This behaviour indicates a distribu-tion of Curie temperatures in the sample caused bya broad distribution of particle sizes. The constantvalue of the initial asymmetry in the fiSR spectra be-low 130K can be explained by a critical particle size,below which magnetic ordering is not possible. TheMossbauer spectra show that the magnetic momentsof the one-domain nanocrystals are not static but ex-hibits superparamagnetic fluctuations. The superpara-magnetic fluctuations freeze out below 50 K. All obser-vations can also be explained by the assumption thatthe missing magnetic moment in the neutron diffractiondata at low temperatures is caused by the special prop-erties of the nanostructured modification. In this casethe reduced but constant signal in the /^SR spectra atlow temperatures has to be interpreted as 1/3 signalof the ferromagnetic phase extending over the wholesample. The spin-glas like behaviour below 50 K isalso valid for this interpretation of the data.

65

STRIPE ORDER AND SPIN DYNAMICS IN NICKELATES

RA-98-11, BRAUNSCHWEIG - KOLN

H.-H. Klau(3*s, D. Baabe*, M. Birke*, W. Kopmann*, W. Wagener*, H. Waif*, F.J. Litterst*5, M. Huckert, O. Fried*,B. Buchnert

* Institut fur Metallphysik und Nukleare Festkorperphysik, Technische Universitat Braunschweig, Germany.f II. Physikalisches Institut, Universitat Koln, Germany.

The aim of this experiment is to examine the for-mation of spin stripes in Sr-doped lanthanum nicke-lates La2-tfSrxNi04+,5 with ^SR. From x-ray and neu-tron scattering the existence of static charge and spinstripes is known for a hole doping between 0.15 and0.3 [1, 2, 3]. This work is also very helpful to identifysimilar order phenomena in the cuprates, since a typ-ical signature of the stripe correlations is enhanced inthe nickelates due to a stronger hyperfine coupling be-tween the magnetic ions and the muon.

In 1998 we focused on two La2_.rSr:!;Ni04 singlecrystals with x = 0 and x = 0.33. In these crystalswe examined the internal field distribution as a functionof temperature and initial muon spin polarization. Theundoped crystal showed a very slowly damped (A m0.2 ^ s " 1 in an exponential fit) spin precession with asaturation value of 36 MHz at 10 K reflecting a very ho-mogeneous long-range antiferromagnetic order. TheNeel temperature of this crystal is above 300 K sinceeven at 290 K we found a spontaneous frequency of24 MHz (Fig. 1). From the amplitude analysis of the ro-tating component, which changes by a factor of 4 fordifferent initial muon spin polarizations, we concludethat the local field vector is oriented at an angle of «25degrees with respect to the a-b plane.

50

45

40

35

30

N§20I 1C

10

5

0

1 'h igh

T» m lowK S M

-

•

1 '

frequencyfrequency

X

ID X

in a

fflrX

0. 3 3. 3 3

Sr H<—i .Sr «—iSr i-*—i "

K

50 100 150 200 250temperature (K)

300

Figure 1: Spontaneous muon-spin precession fre-quency in zero field for La2-rSra;Ni04 single crystalswith x = 0.0 and x = 0.33

The x = 0.33 doped crystal showed a spontaneousmuon spin precession below 190K with a very strongrelaxation (A « 30 //s"1). The oscillation is damped tozero within 0.25 yus! A typical frequency spectrum withthe c-axis orientated parallel to the initial muon spinabove 100 K is shown in Fig. 2. The spectra exhibita characteristic change below 100K. Between 100Kand 185 K we identified a second precession frequency

of about 10 MHz which could be associated with thehole-rich stripes separating the antiferrromagnetic do-mains. This strongly broadened two-component fre-quency spectrum can be identified as the signature ofthe inhomogeneous ordered stripe phase known fromthe neutron scattering experiments. This crystal is thefirst stripe-phase material showing two distinct peaksin internal field distribution seen by the muon. The dis-appearance of the two-frequency pattern below 100 Kcan be associated with the formation of a spin-glasslike phase as proposed from neutron scattering below40 K. This could be caused by a freezing of the spin-|holes.

30 60 90 120

Frequency (MHz)

150

Figure 2: Typical zero-field frequency spectrum for

Due to time limitations the lower temperatures havenot been studied in much detail. Preliminary experi-ments in an orientation with pmuon(t = 0) parallel tothe a-axis show nearly identical results. We concludea broad distribution of internal field orientations due tothe strong inhomogeneous magnetic structure.

Work supported by the German Federal Minister ofEducation, Research and Technology (BMBF).

REFERENCES

[1] S.-W. Cheong etal., Phys. Rev. B 49 (1994) 7088.

[2] J.M. Tranquada et a/., Phys. Rev. B 54 (1996)12318.

[3] J.M. Tranquada etal., Phys. Rev. Lett. 79 (1997)2133.

66

STUDY OF MAGNETISM IN NOVEL HOLE-DOPED QUASI-1D CUPRATES

RA-97-11, UNI ZURICH - ETH ZURICH - BELGRADE

A. Shengelaya*5, C. M. Aegerter*, J. Karpinski*, H. Schwer*, E .M. Kopnint, I. M. Savic*, H. Keller*

* Physik-lnstitut der Universitat Zurich, CH-8057 Zurich, Switzerlandf Laboratorium fur Festkorperphysik, ETH Zurich, CH-8093 Zurich, Switzerland| Faculty of Physics, University of Belgrade, 11001 Belgrade, Yugoslavia

One-dimensional (1D) spin systems have recentlyreceived considerable attention because of the quan-tum mechanical nature of their ground states. In thepresent proposal we study the magnetism of the novel1D cuprates Sr0.73CuO2 and Ca0.82CuO2 recently syn-thesized at ETH Zurich [1]. These compounds con-tain copper-oxygen chains which are doped with a highhole concentration: 0.34 holes/Cu in Ca0.s2CuO2 and0.56 holes/Cu in Sr0.73CuO2. Our previous magneti-zation and Electron Paramagnetic Resonance (EPR)measurements on these compounds indicated unusualmagnetic transitions at low temperature [2]. Here, westudy these compounds with ^SR, in order to detect aninternal local field in zero applied field (ZF).

We performed ZF //SR experiments on Sr0.73CuO2

and Ca0.82CuO2 polycrystalline samples. The most im-portant results can be summarized as follows: bothSr0.73CuO2 and Ca0.82CuO2 undergo a magnetic tran-sition below Tc = 7 K and 11 K, respectively. Above Tc

the time spectra are essentially the same: the muonspins are depolarized slowly with a ~0.1 /JS"1 due tointeraction with a copper nuclear moments. Below Tc

the muon-spin depolarization rate increases sharply bymore than two orders of magnitude, accompanied bythe appearance of oscillations. These changes in the/JSR spectra clearly indicate the onset of magnetic or-der below Tc, in agreement with our previous EPR andmagnetization studies. In Fig.1 we show the tempera-ture dependence of the muon precession frequenciesobserved in Sr0.73CuO2 and Ca0.82CuO2.

In order to estimate the volume fraction of the or-dered phase, we performed measurements in a weaktransverse magnetic field. It was found that below Tc

the magnetic state develops and its volume fractionapproaches nearly 100%. This implies that the mag-netic order takes place in the bulk of the samples andis not due to some impurities. We would like to point outseveral unusual features of the observed magnetic or-der. In contrast to typical precession signals observedin well-ordered ferromagnets and/or antiferromagnets,these novel compounds exhibit strongly damped pre-cession signals. This indicates a large spatial inhomo-geneity of the local field at the muon site. Fourier trans-forms of the ^SR time spectra show a broad distributionof local fields in the samples centered around ~20 mTin Sr0.73CuO2 and 120 mT in Ca0.82CuO2. Of particularimportance is the fact that when fitting the oscillationsto cos(ut + <j>), the phase <f> assumes large unphysi-cal values. Instead, we used successfully an alterna-tive fitting function, i.e. Bessel function, Jo(ut), whichis expected for an incommensurate magnetic structure

150

Figure 1: The temperature dependence of the inter-nal magnetic field at the muon site for Sr073CuO2 andCa0.82CuO2.

described by one g-wave vector [3].Our zero field /zSR results provide conclusive ev-

idence that magnetic order is present in Sr0.73CuO2

and Ca0.82CuO2 below Tc = 7K and 11.5 K, respec-tively. The observation of magnetic order is very un-usual in view of the high hole doping in these com-pounds, where each hole renders a CuO2 unit to be-come nonmagnetic. This is, to our knowledge, the firstexample of the appearance of long-range magnetic or-der when a substantial number of holes are introducedinto CuO2 chains. Further experiments are necessaryin order to clarify the microscopic mechanism of themagnetism observed in hole-doped quasi-1 D cuprates.

REFERENCES

[1] J. Karpinski etal., PhysicaC 274 (1997) 99.

[2] A. Shengelaya etal., Phys. Rev. Lett. (1998), 80(1998)3626.

[3] A. Amato etal., Phys. Rev. 52 (1995) 54.

67

ON (RC)2(NiB)2

RA-98-09, BRAUNSCHWEIG - RIO - MARINGA

W. Kopmann*, H. Waif, M. Hillberg*, D. Baabe*, H. Luetkens*, H.-H. KlauB*, F.J. Litterst*5, M. Fontes*,E. Baggio-SaitovitcM, M. de Melo*

* Institut fur Metallphysik und Nukleare Festkorperphysik, Technische Universitat Braunschweig, Germany.f CBPF, Rio de Janeiro, Brazil| Universidade Estadual de Maringa, Maringa, Brazil

The RC(NiB)2 compounds (R=rare earth) havebeen studied intensively, since they possess in partrelatively high superconducting Tc both for magneticand non magnetic R [1]. The layered structure per-mits coexistence of magnetic order and superconduc-tivity. (RC)2(NiB)2 (often written as RNiBC) belong tothe same homologous series of compounds having anadditional R-C layer inserted between the Ni2B2 layers.The space group changes from 14/mmmto P4/nmm.

The differences in the magnetic properties ofRC(NiB)2 and (RC)2(NiB)2 are mainly related to differ-ences in the interplane interactions. While we have sin-gle RC-planes separated by the non-magnetic (NiB)2-planes in RC(NiB)2 there are double planes of RC sep-arated by the (NiB)2-planes in (RC)2(NiB)2. So there isan interplane interaction within the double planes andbetween neighbouring double planes, the intraplaneferromagnetic interactions (in the simple RC planes)staying more or less unchanged.

Decoupling experiments on YNiBC (which is dia-magnetic as revealed by susceptibility measurements[3]) could prove that an eventual Ni magnetism is neg-ligible in (RC)2(NiB)2. Comparing the purely nucleardamping rate in this compound with Monte-Carlo sim-ulations we were able to localize the muon site in thecenter of the tetrahedron built by two B-atoms and twoR-atoms.

Neutron scattering experiments on TbNiBC showedan antiferromagnetic transition at 12 K and a broaderferromagnetic transition which starts at ~17K. Below12 K the ferromagnetic reflections and the antiferro-magnetic ones appear simultaneously. DyNiBC showsa similar behaviour. In ErNiBC there is just one ferro-magnetic transition at 4.6 K [4].

In 1998 we used ZF and field-cooled ZF / SR onTbNiBC, DyNiBC and GdNiBC. The ZF spectra arewell described by a stretched exponential relaxation

function Gz{t) = Aoexp (- (Ai)^V The asymmetryAo is temperature independent in the high temperatureregime.

We observe a transition between 10 and 20 K ineach compound. Below this transition the muon asym-metry is reduced by more than 50%.

We cannot decide whether the reduction of themuon asymmetry below the ordering temperature iscaused by corresponding precession signals above theupper limit of resolution of the spectrometer or by staticdampings too large to be measured. As a first order es-timation of the field expected at the muon site, we get

I

I

' : * :*

0

• • •

100temperature (K)

-•-'200

1.2

1

l e a -'sg 0.6"

" 0.4"

0.2

°0

I

I*

I

100 200 30temperature (K)

Figure 1: Temperature dependence of the dampingrate A (left) and the exponent /? (right) of the ZF sig-nal for TbNiBC.

2 - 3 T for these compounds (considering only nearestneighbours).

For the Tb and Dy compounds, A increases withdecreasing temperature (Fig. 1, left). This behaviourcorresponds to a slowing down of the electronic spinfluctuations. The exponent /? decreases down to about0.25 - 0.30 (Fig. 1, right). This muon spin relaxationbehaviour is typical for frustrated spin systems [2],and is similar to that found in TbNi2B2C [5]. For theGd compound /? decreases only slightly from 1 to 0.8upon cooling. This is also similar to the behaviourof GdNi2B2C [6]. This similar behaviour of RC(NiB)2

and (RC)2(NiB)2 excludes an anisotropy competitionbetween different magnetic coupling mechanisms asthe origin of the distribution of relaxation rates.

Both the distribution of damping rates (/?) and themagnitude of A suggest a distribution of splittings be-tween the low lying crystal electric field (CEF) levels (inGd there is no CEF splitting). The origin could be acertain degree of disorder, for example a possible sub-stitutional disorder among C and B.

REFERENCES

[1] Cava etal., Nature 367 (1994) 252.

[2] I.A. Campbell ef a/., Phys. Rev. Lett. 72 (1994)1291.

[3] J.C.T. Mondragon, Thesis, CBPF (1998)

[4] J.W. Lynn, private communication

[5] D.R. Sanchez, etal.: Phys. Rev. B 57 (1998) 10268.

[6] R. Cywinsky, etal., ISIS Anual Report 1995-96 Vol.II, RB 7801

68

POSSIBLE MUONIUM FORMATION IN THE ERBIUM HALIDES ErCI and Erl,

RA-96-10, BERN - ETH ZURICH

K. Kramer*, H.U. Gudel*5, F.N. Gygax* M. Pinkpank*, A. Schenck*,

* Department fur Chemie und Biochemie der Universtat Bern, CH-3000 Bern 9, SwitzerlandI Institut fur Teilchenphysik der ETHZ, CH-5232 Villigen PSI, Switzerland

The original aim of the erbium halide project con-cerned the study of the magnetically ordered phasesof ErCI3 and Erl3 which occur below 350 and 300 mK,respectively and which were also investigated by neu-tron diffraction [1].

However, the juSR measurements do not reflect themagnetic ordering. Our 1998-results on ErCI3 confirmthe previous measurements on Erl3 [2]. In the ZF mea-surements in the LTF between 50 mK and 1 K we de-tected two exponentially decaying components at zeroprecession frequency. One component is very fast andhas a relaxation rate of 45(5) MHz in Erl3. For ErCI3 it istemperature dependent and decreases from 45(5) MHzat 50 mK to 25(5) MHz at 900 mK. The second, slowcomponent is temperature independent and shows re-laxation rates of about 0.2 MHz for Erl3 and 0.5(2) MHzfor ErCI3. For both compounds the total observed asym-metry is only about half of the full value. In ErCI3 theslow signal has an asymmetry of 1.5(5)% and that ofthe fast one decreases from 12(1)% at 50 mK to 8(1)%at 900 mK. This behaviour together with last year's TF-observation of huge paramagnetic shifts [2], displayinga 1/T dependence, suggest the formation of paramag-netic muonium like centers. In fact it seems that we arestudying the solid state chemistry of the implanted fi+

in these insulating compounds.In order to study the nature of these paramagnetic

states further we performed longitudinal field meaure-ments both as a function of field strength and temper-ature.

As in ZF a fast and a slowly damped componentare observed. Fig. 2 and 3 show the temperature de-pendence of the two relaxation rates and associatedasymmetries, respectively in an applied field of 0.4 T. Aquantitative interpretation of these results has not yetbeen attempted. In continuation of last year's TF mea-surements of the paramagnetic shifts we have mea-sured the angular dependence at 80 K rotating the ap-plied field in the a - c plane (see Fig. 1). The resultsconfirm the presence of three distinct components im-plying up to three different states or sites, respectively.

REFERENCES

[1] K. Kramer et al., LNS Annual Progress Report1997, pages 25, 26.

[2] K. Kramer et al., PSI Annual Report 1997, AnnexI, p. 45.

8 2 - - 7 2

- 7 0

81 -

100 150X

200 250 / 300

Figure 1: Angular dependence of the frequenciesin Erl3 at 80 K and 6kG TF. Triangles refer to theright-hand scale.

2 0 -

S 1 5 -

la£ 1 0 -

I 5H

ofCE

So

I-BH

I-BHI-BH

B

e m

I-BH

(I

I

CD

i

CD

J

I-BH

CD 35

0-I I I I I I0 50 100 150 200 250 300

temperature / K

Figure 2: Relaxation rates in Erl3 at 4kG LF.

N 2110

so

2cS 0gio°-CO

CDCD

m

tS tD_i«i

m Bmon era

i

3 4 5 6 7 1

10 temperature / K 10

Figure 3: Asymmetry of the relaxation rates as shownin Fig. 2. Below 80 K the total asymmetry is fixed to0.23.

69

CONDON DOMAINS IN NON-MAGNETIC METALS

RA-94-14, PSI - MOSCOW

G. Solt*5, C. Baines*, V.S. Egorov*, D. Herlach*, U. Zimmermann*

* Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland| RSC "Kurchatov Institute", 123182 Moscow, Russia

Dia- and paramagnetic (Condon) domains [1], oc-curring in non-magnetic metal single crystals at lowtemperatures and high magnetic fields, are due to or-bital, Landau-type magnetism, basically different fromany form of spin ferro- or antiferromagnetism.

The phase transitions reappearing in every periodof the de Haas - van Alphen (dHvA) oscillations and'breaking up' the uniform phase, the unknown structureof the domains and of the phase diagram have moti-vated several theoretical studies [2]. As to the exper-iment, after a first observation of Condon domains [3]in Ag, 28 years elapsed until a second case of domainformation could spectroscopically be detected, in beryl-lium, by yuSR [4]. Quite unexpected is the result for thephase diagram: dia- and paramagnetic domains havebeen observed in Be even at such points of the (B, T)plane which lie deep in the predictedly 'uniform phase'.F-ig. 1 shows the calculated phase diagrams, assumingdifferent Dingle temperatures (crystal perfection) forthesample. (In Be, due to two nearly equal beating dHvAfrequencies, the predictions differ for beat maxima andminima.)

1 2 3 4 5 6 7induction B (T)

Figure 1: Magnetic phase diagram [4] for Be. Domainsshould occur below the solid curves for the beat max-ima of x{B) and below the dashed lines for beat min-ima; the numbers are Dingle temperatures TD. Thepoints for beat maxima (•) and minima (o) show ob-served domain phases for the sample with TD «2.6 K.

The positions of the measured points show the failure,in the case of Be, of the commonly used 3D formulaforthe dHvA magnetization amplitude. In fact, a recentcalculation based on a 2D rather than a 3D electrongas model [5] results in more extended domain regions,in agreement with our results.

In view of its band structure, domains in white tinshould be formed at comfortably lower external mag-netic fields than in Be, at «1 T. The oscillations of theline width in the ,uSR spectrum of an extreme high pu-rity Sfi single crystal (Fig. 1) is indeed a signal of thefact that the line becomes a doublet in a section of eachdHvA period: the implanted muons 'observe' two in-ductions in the sample, corresponding to the dia- andparamagnetic domains. To find a (H, T) region v/here

10.00 10.02 10.04 10.06 10.08 10.10

field H (kOe)

Figure 2: Damping rate of the yuSR signal (linewidth) for white Sn with iJ||[100], at temperature T= 0.1 K. The oscillations indicate periodic domain for-mation, with a period coinciding with the length of thede Haas - van Alphen cycle AH « H2/FCx = 23 G (FCi= 440T, [6]).

the components of the doublet in Sn become sepa-rately observable, new experiments are under progress.

REFERENCES

[1] J.H. Condon, Phys. Rev. 145, 526 (1966)

[2] M.A. Itskovsky et al., Phys. Rev. B50, 6779, 1994and references therein

[3] J.H. Condon and R.E. Walstedt, Phys. Rev. Lett.21,612(1968)

[4] G. Solt et al., Phys. Rev. Lett. 76, (1996) 2575, andPhys. Rev. B59, March 1999

[5] A. Gordon, M.A. Itskovsky, I.D. Vagner andP. Wyder, Phys. Rev. Lett. 81, 2787 (1998)

[6] M.D. Stafleu and A.R. de Vroomen, phys. st. sol.23,675(1967)

70

STUDY OF THE MAGNETIC PROPERTIES OF HoBa2Cu3Oe+x

RA-95-03, ETH ZURICH - PSI

M.Pinkpank*5, A. Amatot, D.Andreica*, F.N.Gygax*, H.R.Ott*, A.Schenck*

* Institute for Particle Physics, ETH Zurich, CH-5232 Villigen PSI, Switzerlandt Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerlandt Laboratory for Solid State Physics, ETH-H6nggerberg, CH-8093 Zurich, Switzerland

The field and temperature dependence of the LFrelaxation rate in HoBa2Cu3O6+;r was measured in theGPS using the new muons on request (MORE) device(see Fig. 2).

There are two muon stopping sites in REBa2Cu3

Oe+x (RE=rare earth): in REBa2Cu307 all muons stopin 1 A distance from the chain oxygen. lnREBa2Cu3O6,because all chain oxygen is extracted, the muons stopin 1 A distance from the oxygen 0(1). For sampleswith intermediate oxygen content, the probability for themuon to stop at one of the two sites changes linearlywith the oxygen content. This change of the stoppingsites will lead to a linear decrease of the / SR relaxationas function of the oxygen content, because the secondmoment of the dipolar field near the oxygen 0(1) is big-ger than the one near the chain oxygen. This is theonly effect of oxygen reduction on the yuSR relaxationrate found in GdBa2Cu306+x [1].

The oxygen dependence of the LF relaxation ratein HoBa2Cu3O6+z is very different from that (Fig. 1).This is a result of the strong effect of the crystal electricfield (CEF) on the Ho3+ multiplet in HoBa2Cu3O6+:c,and the change of the CEF with the oxygen content.The ^SR spectra are fitted with a stretched exponen-tial (Fig. 2). The analysis of this data is still in progress.Different from GdBa2Cu306+x, a strong field depen-dence of the relaxation rate is found at low temper-atures, while at higher temperatures this field depen-dence disappears. There are drastic variations in thefield dependence of A, if the oxygen content of the sam-ples is reduced (Fig. 3). The field dependence of thesuperconducting HoBa2Cu306+2; samples can be de-

0.1 --

0.01

i

•~ _ _

i

t(ns)

Figure 2: Histogram of HoBa2Cu3O7 at 6K in a LF of750 G measured with MORE.

N0.1 •

0.01 •s *X

~—\ 1

•

B

A

o

$i

— i —

HoBa2Cu3O7

HoBa2Cu3O66

HoBa2Cu3O6.4HoBa2Cu3062

\ ,

500 1000 2000

B(G)

6.0 6.2 6.4 6.6 6.8 7.0

oxygen content

Figure 3: Field dependence of the yu+-depolarisationrate of HoBa2Cu306+:r at 6 K for samples with differentoxygen content.

scribed by a power law (A(B) = CB7), with 7 =1.95and 7 = 1.16 for the samples with x = 1 and x =0.6,respectively. This is not possible for the non supercon-ducting samples. In the case of HoBa2Cu3062, the re-laxation rate stays constant for fields up to 1000 G, de-creases for higher fields, and is constant again for fieldsabove 2000 G. In comparison with GdBa2Cu306+a;, theincrease of the relaxation rate between HoBa2Cu306.4

and HoBa2Cu3O6.2 is much to large to simply be ex-plained by the different population of the two muon sites.

Figure 1: Oxygen dependence of the yw+-spin depolar-isation rate of HoBa2Cu3O6+:c at 6 K in a LF of 1000 G.

REFERENCES

[1] M. Pinkpank etal., Physica C, in press.

71

ji+SR ON La2-x-yRE;cSryCu04

RA-93-05, BRAUNSCHWEIG - KOLN

W. Wagener*, H. Waif, H. Luetkens*, H.-H. KlauB*, W. Kopmann*, D. Baabe*, M. Birke*, RJ. Litterst*s,B. Biichnert

* Institut fiir Metallphysik und Nukleare Festko rperphysik, Technische Universitat Braunschweig, Germany.I II. Physikalisches Institut, Universitat Koln, Germany.

The physics of the LTT phase of doped La2Cu04 incomparison to the LTO phase can provide useful andimportant hints that lead to a better understanding ofsuperconductivity in the cuprates. Surprising effectslike static stripe order and suppressed superconductiv-ity have been found in the LTT phase. jz+SR as an out-standing method to observe local magnetic order hasplayed an important role in these investigations. Theexistence of magnetic order in La2-z-yREa;SryCu04with y « 0.15 has been proven by this project and themagnetic phase diagram of the LTT phase has beenestablished. The last beam period, we have concen-trated on three topics:

We studied single crystals of Lai.8_j/Euo.2SryCu04

with y = 0.08, y = 0.125 and y = 0.15. Using ju+SR,we could prove the onset of magnetic order in thesesamples. The saturations frequencies are in agree-ment with the values found for polycrystalline samples(see Table 1). For y = 0.08 and y = 0.15, the onsettemperature of magnetic order is, just so, of compara-ble size, for y = 0.125, however, To is more than 10 Klower in the single crystal than in the polycrystal. NMRexperiments of Hammel et al. confirm this surprisinglylow value of To. It seems that this single crystal whichwas the first that was grown is of low quality, althoughthe x-ray spectra look very nice.

Srconc.

yy = 0.08y = 0.12(5)« = 0.15

polycrystal, (MHz) To (K)

3.0 ±0.4 6 ± 13.4 ± 0.3 27 ± 23.2 ± 0.3 25 ± 2

single crystal,( (MHz) To (K)

3.0 ±0.3 7 ± 13.2 ±0.3 15 ± 23.1 ± 0.3 25 ± 3

Table 1: Comparison of saturation frequencies vsat andordering temperatures To in polycrystals and singlecrystals of Lai.8_j/Eu0.2SryCuO4

When rotating the sample relative to the initial di-rection of the muon spin, the relative asymmetry of theprecessing signal changes. The local magnetic field atthe muon site is mainly oriented perpendicular to thec axis. As expected for the magnetic structure of thestripe order, no difference between (100) || beam and(110) || beam is visible. At present, we have no cluesfor a better fit function than the one we used previously(two frequencies). Further angular scans on one ofthe crystals (La1.65Eu0.2Sr0.i5CuO4) are needed to getmore information about the direction of the local field atthe muon site and the possible angular dependence of

the frequency distribution.A Ba concentration of y =0.12 has been investi-

gated. The saturation frequency vsat is about 3.4 MHz,which is in very good agreement with the value found inthe Sr-doped system. The onset temperature of mag-netic order, however, is reduced to 17 ± 3 K (compareto To =25 ± 2 K in Lai.65Eu0.2Sr0.i5CuO4 polycrystal).This could again be a problem of sample quality.

The results on Lai.8Eu0.2CuO4 and Lai.6Ndo.4Cu04

are compared on Fig. 1. The low-temperature struc-tural phase transition at about 130 K in the Eu-dopedcompound and at about 75 K in the Nd-doped com-pound is visible as a kink in v{T), XT{T) and XL(T).The transition is more pronounced in Lai.6Ndo.4Cu04

supposedly due to the higher TN. At low tempera-tures, the Nd moments lead to an increase of the muonspin precession frequency and of the damping rates.In contrast to samples with Sro.i5, however, the low-temperature increase of the muon spin precession fre-quency is rather small. Therefore we conclude that noreorientation of the copper spins from the a-b plane tothe c axis takes places in Lai.6Ndo.4Cu04.

Muon spin precession frequency v

3-

• LawEu02CuO4

A La16Nd04CuO,

50 100 150 200 250

temperature (K)

300

Figure 1: T-dependence of the muon spinprecession frequency in Lai.8Eu0.2CuO4 andLai.6Ndo.4Cu04

72

/xSR INVESTIGATION OF THE MAGNETIC PHASE DIAGRAM OF Bi2.15Sr1.85CaCu2O8+<s

RA-94-04, ST. ANDREWS - ZURICH - BIRMINGHAM - GRENOBLE - PARIS

C. M. Aegertert C. Ager*, R. Cubitt§, E. M. Forgan*, S. T. Johnson^, H. Keller*, S. L. Lee'5, P. G. Kealey*,F. Y. Ogrin*, T. M. Riseman*, I. M. Savic*

* School of Physics and Astronomy, University of St. Andrews, North Haugh, St. Andrews, Fife, KY16 9SS,U.K.

f Physik-lnstitut der Universitat Zurich, CH-8057 Zurich, Switzerlandt School of Physics and Space Research, University of Birmingham, Birmingham B15 2TT, U.K.§ Institut Max von Laue - Paul Langevin, 38042 Grenoble, Cedex, France.% Laboratoire de Physique des Solides, Universite Paris-Sud, Orsay, France.

In a type II superconductor, an applied magneticfield may enter the bulk of a superconductor in the formof quantised flux lines or vortices. In the high Tc-super-conductor Bi2.15Sr1.85CaCu2Os.f5 (BSCCO), the behav-iour of these flux lines is very complicated showing arich variety of phases, including a liquid of vortex lines[1]. It is the aim of this study to elucidate the natureof these different phases by p.SR. In the past year wehave continued investigating the vortex arrangementand the melting transition in the vicinity and above thedimensional crossover field. At fields close to the cross-over, the subtle competition between pinning induceddisorder and thermal fluctuations can be inferred fromthe increase in correlation of the flux lines with tem-perature [2]. This takes place at a temperature, wherethe thermal energy is comparable to the pinning energyleading to a depinning and hence ordering of the pan-cakes [3]. A further increase in temperature howeverleads to the melting of the vortex lattice. In contrastto magnetisation measurements [4], we still observe amelting transition above the crossover field. The ob-served melting temperature is virtually independent ofthe applied field (see Fig. 1). Moreover the value of thistemperature is in good accord with the predictions ofa Kosterlitz-Thouless like melting of a two dimensionallattice of pancake vortices, which is also expected to befield independent [2]. A similar transition at the sametemperature has also been observed by microwave ab-sorption measurements [5]. The different time-scalesinvolved in these two measurements indicate that theobserved change in the vortex arrangement is indeeda phase transition.

We have also continued the investigation of BSCCOcontaining irradiation induced pinning centres. Hereour findings of the strong suppression of both dynamic(thermal) and static fluctuations have been acceptedfor publication [6]. The pinning centres take the formof cylindrical tracks and are produced with 17.2 GeVU ions. Due to the suppression of the superconduct-ing order parameter in the amorphous region of thesetracks, they act as very strong pinning sites. In pris-tine BSCCO, transverse fluctuations of pancake vor-tices lead to a field-dependent modification of the tem-perature dependence of the /iSR linewidth [7]. Due tothe pinning caused by the CD, these thermal fluctua-tions are strongly suppressed, leading to a temperature

T(K)

Figure 1: The experimental phase diagram of the vor-tex lattice in BSCCO, as determined from the disap-pearance of the scattered neutron intensity and froma sharp change in the ^SR lineshape. The shape ofthe melting line changes to a 2D Kosterlitz-Thoulesskind of melting around the crossover field Bcr, whichincreases slightly at high temperatures, due to thermaldepinning.

dependence, which is independent of the applied field.

Work supported by the Science and EngineeringResearch Council of the United Kingdom, Swiss Na-tional Science Foundation, and by a special grant fromthe British/Swiss Joint Research Programme.

REFERENCES

[1] S.L. Lee etai. Phys. Rev. Lett. 71 (1993) 3862.

[2] S.H. Lloyd etai., submitted to Phys. Rev. Lett.

[3] D. Ertas and D.R. Nelson, Physica C 272 (1996)79; A.E. Koshelev and V.M. Vinokur, Phys. Rev. B57(1998).

[4] E. Zeldov etai., Nature 375 (1995) 373.

[5] J.J. Wingfield, private communication.

[6] S.L. Lee etai., Phys. Rev. Lett. 81 (1998) 5206.

[7] S.L. Lee etai., Phys. Rev. Lett. 75 (1995) 922.

73

TWO STAGE MELTING TRANSITION OF THE VORTEX MATTER IN Bi2Sr2Ca2CuO8+<5

RA-96-07, KONSTANZ - LOWER HUTT - STUTTGART

T. Blasius*5, Ch. Niedermayer*, J.L Tallon*, D.M. Pooke*, A. Golnik* and C. Bernhard*

* Fakultat fur Physik, Universitat Konstanz, D-78457 Konstanz, Germanyf The New Zealnd Institute for Insdustrial Resarch, Lower Hutt, New Zealand

Max-Planck-1 nstitut fur Festkorperforschung, D-70569 Stuttgart, Germany

The vortex matter in cuprate high-Tc superconduc-tors exhibits a complex phase diagram as a function ofmagnetic field, temperature and doping. With the mag-netic field applied perpendicular to the CuO2 planesthe flux lines can be viewed as stacks of pancake vor-tices. Both the intra-planar vortex order and the inter-planar coupling between the pancake vortices can beovercome by thermal fluctuations. In principle, theseprocesses may occur independently and at differenttemperatures. Last year we reported on the flux line lat-tice melting in single crystals of overdoped and under-doped Bi2Sr2Ca2Cu08+,5 (Bi-2212) where we obtainedfor the first time from yuSR measurements evidence fora two-stage melting transition of the vortex matter un-der equilibrium conditions [1].

Here we summarize our ^SR data for nearly op-timized (Tc = 90 K) Bi-2212 single crystals which givefurther evidence for a two-stage melting scenario. Fig. 1illustrates the asymmetry a = <AB 3> 1 / ' 3 /<AS 2> 1 / 2 ,the second moment < A 5 2 > and the shift of the cuspfield with respect to the external field of the measuredinternal field distribution n(B) versus temperature. Inthe irreversible regime well below the irreversibility line(IL) we observe a transition in the vortex matter whichwe associate with the intra-planar melting transition.The resulting flux-line liquid persists over a sizeabletemperature interval before the inter-planar decouplingof the individual vortex lines takes place in a secondstep at the IL, resulting in a pancake liquid [2].

An interesting question is how our observation fitsin with the reports of a one-stage melting transition. Itseems likely that the key to the answer lies in the c-axis coupling strength, which varies strongly amongstthe different high-Tc compounds, while the condensa-tion energy or the super-fluid density [3] does not varystrongly. In fact, by rescaling the (H, T)-phase diagramwith the critical field H* of the dimensional crossover,we obtain a rather unique (i7,T)-phase diagram fromthe ^SR and the DC-magnetization (Fig. 2) data for awider series of Bi-2212 single crystals [1]. This wouldimply that a two-stage melting process occurs if the ap-plied magnetic field is comparable to or larger than H*,while a seemingly one-stage melting process occurswhen Tab and Tdc do collapse for H <c H*. In fact,such a trend is even visible from our data on the Bi-2212 crystals for which the Tab{H)- and Tdc(#)-linesseem to merge for H <c i?*[1]. For the less anisotropicY-123 compound H* should be in excess of 10T andaccordingly most of the experiments have been donefor H < H* where in fact a single-stage transition isobserved.

•0.5

1.0

I 0.5

H0.5

; • • • •

• • • • * •

•

• • *

r* •

•«••+

*

0.0 0.2 0.4 0.6 0.8t=T/Tc

Figure 1: T-dependence of a, the cusp shift and< AS 2 > of n(B) on nearly optimized Bi-2212 sin-gle crystals for n0Mext = 27.5 mT. Tab (Tdc) denotes thetemperature for the in-plane (decoupling) transition [2].

VSM-Data

00 0.2 0.4 0.6 0.8

t=T/Tc

Figure 2: Rescaled VSM data (scaling factor H*) fordifferent doping states of Bi-2212 single crystals [1].

REFERENCES

[1] T. Blasius etai, submitted to Phys. Rev. Lett.

[2] T. Blasius etal., to be published.

[3] J.L. Tallon etal., Phys. Rev. B (in press).

74

HOLE DYNAMICS AND LOCALIZATION IN LOW-DOPED HIGH-TEMPERATURESUPERCONDUCTORS

RA-98-10, BRAUNSCHWEIG - KÖLN

W. Wagener*, W. Kopmann*, D. Baabe*. H. Waif*, H. Luetkens*, M. Birke*, H.-H. Klauß*, H. Waif*, F.J. Littersfs,M. Hucket, B. Büchner*

* Institut für Metallphysik und Nukleare Festkörperphysik, Technische Universität Braunschweig, Germany.f II. Physikalisches Institut, Universität Köln, Germany.

Susceptibility and resistivity measurements haveshown that doping with zinc not only introduces local-ized spin holes on copper sites but also strongly local-izes the charge holes produced by Sr doping. This in-fluences both the Néel temperature TN and the sublat-tice magnetization curve. Concerning the investigationof magnetic order, the big advantage of /*+SR in com-parison to methods like susceptibility is the easy acces-sibility of the sublattice magnetization via the sponta-neous muon spin precession frequency. Any irregulari-ties in the v(T) curve can be identified easily. Becauseof the local character of the probe, phase separationand strong inhomogeneities are detectable.

We have started our .+ SR experiments on two se-ries of samples, one with a Sr concentration of y =0.017, the other one with y - 0.021.

Lai.983Sro.oi7Cui_zZn204: Three different Zn con-centrations, z = 0.00, z = 0.10, and z = 0.15, havebeen investigated. The Néel temperatures determinedby our p+SR experiments are in very good agreementwith the values from susceptibility measurements. Forlow Zn concentrations, the hole localization caused byZn doping leads to an increase of TN; for higher dop-ings, however, this effect is overcompensated by thedecrease of TN due to spin dilution.

Muon spin precession frequency

' n L a i .983S r0.017C U1.ZZ n

Z°4

"N" 4-

•AX

z=0.00z=0.10z=0.15

20 40 60 80 100

temperature (K)120 140

Figure 1: Temperature dependence of the muon-spinprecession frequency in L.ai.983Sro.oi7Cui_2Znz04

withO < z < 0.15

The sublattice magnetization curve cannot be de-duced from susceptibility, but +SR is an ideal tool toextract the sublattice magnetization from the sponta-neous muon spin precession frequency in zero applied

field. Fig. 1 shows the temperature dependence of themuon frequencies. For z = 0, the transition from lo-calized holes to mobile holes (at about 20 K) is clearlyvisible, whereas for z = 0.10 and z = 0.15, there are nosigns for an onset of hole delocalization.

Muon spin precession frequency

La i .979Sr0.021CU1-zZnzO4

•AXV

z=0.00z=0.005z=0.02z=0.04

20 40 60 1 80temperature (K)

100

Figure 2: Temperature dependence of the muon-spinprecession frequency in Lai.989Sro.o2iCui_^Zn204with 0 < z < 0.04

Lai.979Sr0.o2iCui_zZn^04: The Sr concentration y= 0.021 lies at the transition between the anti-ferromag-netic (AFM) and the cluster spin glass regions. Thesample without zinc shows a magnetic ordering tem-perature of about 15K. As expected from susceptibilitymeasurements, TN strongly increases with zinc dop-ing. For z = 0.04, a value of TN ~ 90 K is reached. Theevolution of the v(T) curve with zinc doping is shownin Fig. 2. The series La1.979Sr0.02iCui_2ZnzO4 with in-creasing z resembles the series La2_ySr3/Cu04 withdecreasing y and y < 0.021. The system goes backfrom the cluster spin glass region into the AFM region;even for z = 0.04 the charge holes are not completelylocalized. Higher zinc concentrations have to be inves-tigated in the next beam period. At present, no state-ments about the disappearance of the delocalizationpeak in the longitudinal damping rate are possible. Thescatter of the data is too large. Temperature scans inapplied longitudinal fields of about 200 mT could possi-bly provide a lower scatter.

75

LOCAL MAGNETIC ORDER IN YBa2(Cu1_;ECoa;)3O7

RA-98-15, SACLAY-DELFT-GRENOBLE

J.A. Hodges*5, RC.M. Gubbensf, C.T. Kaiser*, P. Dalmas de Réotiei4, A. Yaouanc*.

* CEA, DRECAM-SPEC Centre d'Etudes de Saclay, F-91191 Gif sur Yvette, Francef Interfaculty Reactor Institute, TU Delft, NL-2629 JB Delft, The Netherlands| CEA, DRFMC-SPSMS, CENG, F-38054 Grenoble, France

When fully oxidised, YBa2Cu307 is partially substi-tuted at the Cu(1) chain sites, the charge transfer to theCu(2) - O planes is modified, the system becomes un-derdoped and the superconducting transition temper-ature (Tsc) decreases below the optimum value. Forhigh substitution levels of Co, superconductivity disap-pears and it is replaced by a state where the Cu(2)show long range magnetic order [1]. This behaviouris consistent with the widely held belief that supercon-ductivity and Cu(2) based magnetic moments are mu-tually exclusive. However, in contradiction with this, wehave recently observed that for low Co substitution lev-els where TSc remains high, fluctuating short rangecorrelated magnetic moments (including at the Cu(2)sites) are present over the whole sample volume [2].

Figure 1: Muon depolarisation in superconductingYBa2(Cuo.98Co0.02)307. The solid line is a fit to aKubo-Toyabe (nuclear) function.

0.15

0.10

0.05

0.00

—1—'—1—•-

• vT1

I . I .

"1—'—1

1

W

T 1'—1—'—1

zero-field j

Mj ,

• = 6 K :

iijii . l t •uUeilfti '

1 . 1

0.0 0.2 0.4 0.6 0.8 1.0 1.2Time (us)

Figure 2: Muon depolarisation in superconductingYBa2(Cu0.98Coo.02)307. The solid line is a fit to a prod-uct of Kubo-Toyabe (nuclear) and stretched exponential(electronic) functions.

Zero field /JSR measurements were made in therange 100 to 2K on two samples : YBa2(Cui_.rCo^O?with x = 0.04 (Tsc = 68K) and x = 0.02 (Tsc = 84K).

The data for the sample x = 0.02 at 100K and 6K areshown in Figs 1 and 2. The muon depolarisation at hightemperatures (Fig. 1) is controlled uniquely by the nu-clear moments (Kubo-Toyabe dependence), whereasat low temperatures (Fig. 2) the depolarisation is fur-ther influenced by electron-based fluctuating magneticmoments. The solid lines are data fits to the functionP^t) = aoexp[-(\\)ß] x GürT(A,t) with the stretched ex-ponential term present only at low temperatures. Thevalues for the damping rate (A) and the exponent of thestretched exponential (ß) are shown in Fig. 3.

40-

30-

X20-

«< 10-

0-

1.0-

0.8

^ 0.6-

0.4-

0.2

• 5

o 8

fill

0 5 10 15 20

Temperature (K)

Figure 3: Temperature dependence of the damp-ing rate (A) and the stretched exponent (ß) inYBa2(Cui_a;Co;i;)307.

ß is generally below 1.0, and although this is linkedto distributions in the coupling strengths, it is probablyalso linked to the presence of distributions in the localfluctuation rates of the correlated moments that wereevidenced in [1]. We observe an unusual tendency forß to increase as the temperature is lowered.

REFERENCES

[1] P. Zolliker, D.E. Cox, J.M. Tranquada, G. Shirane,Phys. Rev. B 38 (1988) 6575.

[2] C. Vaast, J.A. Hodges, P. Bonville, A. Forget, Phys.Rev. B 56 (1997) 7886.

76

ON "ARTIFICIALLY AGED" YBa2Cu3O3O6.6

RA-98-16, ORSAY - TECHNION - SACLAY - KAZAN

P. Mendels *5 , A. MacFarlane *s, J. Bobroff *, H. Alloul *, N. Blanchard *, A. Kerens \ A. Dooglav *, G. CollinJ.F. Marucco * s

* Laboratoire de Physique des Solides, URA2 CNRS, Université Paris-Sud, Orsay, 91405 Orsay, Francef Department of Physics, Technion, Haifa, Israëli Laboratory of Magnetic Resonance, Kazan State University, 420008 Kazan, Russia§ L.L.B, C.E.N. Saclay, CEA-CNRS, 91191 G if-sur-Yvette, Francef Laboratoire des Composés non Stoechiométriques, UPS, 91405 Orsay, France

The phase diagram of High Tc cuprates versusdoping is commonly depicted as two mutually exclu-sive antiferromagnetic (AF) and superconducting (SC)phases. Yet, there is growing evidence that the actualsituation might be more subtle: in parent compoundssuch as doped Nickelates, a static microsegregation(stripes) of holes in an AF background was recently ob-served [1]. The generalization of this concept to high-Tc cuprates is currently a matter of considerable de-bate. We proposed to test its relevance by investigatingthe 60 K superconductor YBa2Cu306.5-6.7 for whichan antiferromagnetic phase was recently observed byNMR/NQR in "aged" superconducting samples [2]. Ourproposed experiments concentrated on basic ques-tions regarding these local AF states, namely the frac-tion of the sample involved, the Néel temperature, andthe degree of magnetic order.

In our search for preparing both magnetic and su-perconducting samples in a reproducible manner, it ap-peared that a true chemical phase separation occurs inYBaCuOx samples due to an unexpected low T reac-tion between water vapour and pure YBaCuO aroundx = 6.5, where one chain over two is empty. The mag-netic phase could be isolated and turns out to be theinteresting missing magnetic parent compound of theYBa2Cu408 superconductor [3]. The mechanism ofpenetration of H2O in YBCO is related to the chainarrangement and we have studied 4 samples preparedusing 2 different annealing procedures (low T, high T),with 2 different H2O concentrations for each. A sketchof our results is presented in Fig. 1, where the samplesare labeled according to the amount of H2O uptake perYBaCuO mole.

All the samples are 100% magnetic with a transi-tion temperature above room T (41 OK for 1.2H2O, asrecorded in ISIS). The presence of oscillations clearlyindicates that the magnetic phase is ordered, likelyantiferromagnetic and the high Néel temperatures aretypical of undoped CuO2 bilayers. The observed sig-nals, which vary with annealing conditions and/or H2Ocontent clearly are completely different from AF YBCOcompounds and are of a new type compared to othermagnetic cuprates. This indeed confirms the chemicaltransformation of the original material. Finally, the ob-vious presence of several oscillating/relaxing compo-nents will allow us, after further analysis and detailedcomparison with ZFNMR spectra, where a new mag-netic component is growing above 0.5 H2O uptake, to

%

CO

0

YBC06.42(H20)1:1.2

300K .

. VV200K .

15K

YBC06.42(H20)0,0.64

300K* * ^

200K

0 Time (fj,s) 3 0 Time (/zs) 3

Figure 1 : ZF/tSR typical signals for YBCO:H2O. Noticethe evolution with T and sample composition.

assign each component to a typical Cu environment/muon site. Ref. [3].

The /*SR signals differ in these antiferromagnetscompared to YBCO6 for two reasons: the hydrogensof the water compete directly with the muon for sites,and the magnetic structure is not exactly the same (asshown by the difference in the ZFNMR spectra). Inthe fully hydrated sample, the field at one muon site isabout half its value in YBCO6, and the ZF oscillationsare remarkably well developed. In partially hydratedsamples, we find evidence of a muon site change atabout 280 K (see Fig. 1). Detailed analysis is ongoing.

REFERENCES

[1] J.M. Tranquada et al., Phys. Rev. Lett. 73 (1994)1003, Phys. Rev. B 54 (1996) 12318.

[2] A. Dooglav et al., Phys. Rev. B 57 (1998) 11792.

[3] W. Günther and R. Schöllborn, Physica C, 271(1996)241.

77

/iSR INVESTIGATION OF THE MAGNETIC PHASE DIAGRAM OF ORGANICSUPERCONDUCTORS

RA-90-07, ZURICH - ST. ANDREWS - BIRMINGHAM - OXFORD -TOHOKU - OKAZAKI - PARIS

C. M. Aegerter*s, C. Ager*, S. J. Blunder E. M. Forgan*, S. T. Johnsontt, H. Keller*, S. L. Lee*, S. H. Lloyd*,B. W. Lowett§, F. Y. Ogrint, F. L Pratt§, T. M. Riseman*, T. Sasaki, I. M. Savic*, H. Taniguchi**

* Physik-lnstitut der Universitat Zurich, Switzerlandf School of Physics and Astronomy, University of St. Andrews, U.K.t School of Physics and Space Research, University of Birmingham, U.K.§ Physics Departement, University of Oxford, Clarendon Laboratory, U.K.f Institute for Materials Research, Tohoku University, Sendai, Japan.** Institute for Molecular Science, Okazaki, Japan.f t Laboratoire de Physique des Solides, Universite Paris-Sud, Orsay, France.

In 1998, we have concentrated on the behaviourof the thermal transition leading to a more disorderedstate of the vortex lattice at high temperatures. Wehave investigated in detail the field dependence of thethermal disruption temperature Tb, observed from achange of the field distribution measured by yuSR. Attemperatures above Tb, the flux lines are more disor-dered along their length, leading to a more symmetriclineshape. As can be seen in Fig. 1, this temperatureis virtually independent of the applied field. This be-haviour observed in ET-Cu below the crossover field ismore reminiscent of BSCCO above the crossover. Itis however also expected for the thermal disruption ofa stack of purely electromagnetically coupled pancakevortices, the value depending solely on the penetrationdepth [1].

A very promising system with which to study thegeneric behaviour of extremely anisotropic supercon-ductors is a-(ET)2NH4Hg(SCN)4 (ET-Hg), which has avery long penetration depth (Xab ~ 1,um) and is thoughtto be even more anisotropic than ET-Cu (7 ~ 1000[2]). Due to the very long penetration depth, the ob-served signal is very small. We have neverthelessbeen able to observe even in this material the pres-ence of a crossover field at an applied field of ~1 mT(see Fig. 2). This value is in fair agreement with the ex-pectation of Bcr = $o/A2, using A determined from thewidth of the field distribution.

0.120

0.100

0.080

S 0.060O

0.040

0.020

0.000

• I

•

Bcr

tI

•

T

mT)

Figure 2: The field dependence of the fiSR depolar-isation rate <r in ET-Hg, the contribution from nuclearmoments and instrumental resolution have been sub-tracted. At very low fields, below ~1 mT the depolari-sation increases. This is interpreted, as in BSCCO, asa crossover to a state of more straight flux lines.

The year 1998 has also seen the publication ofour initial studies of the angular dependence of thefiSR lineshape in the highly anisotropic model organicsuperconductor K-(ET)2CU(SCN)2 (ET-CU) [3]. Thesefindings were already discussed in last year's progressreport [4].

Work supported by the Swiss National ScienceFoundation, the Science and Engineering ResearchCouncil of the United Kingdom, and by a special grantfrom the British/Swiss Joint Research Programme.

REFERENCES

[1] J.R. Clem, Phys. Rev. B 43 (1991) 7837.

[2] H. Taniguchi etal., Phys. Rev. B 57 (1998) 3623.

[3] C. Ager etal., J. Mag. Mag. Mat. 177-181 (1998)561.

[4] PSI Progress Report RA-90-07 1997.

Figure 1: The field dependence of the thermal disrup-tion temperature Tb.

78

//SR STUDIES OF SUPERCONDUCTIVITY AND MAGNETISM IN ORGANIC MOLECULARCRYSTAL SYSTEMS

RA-93-06, OXFORD - RIKEN-RAL- STRASBOURG

S.J. Blundell*s, F.L. Pratt*, B.W. Lovett*, Th. Jestadt*, M. Kurmoo*, W. Hayes*

* University of Oxford, Department of Physics, Clarendon Laboratory, Parks Road, Oxford 0X1 3PU, U.K.| The Institute for Physical and Chemical Research (RIKEN), RIKEN-RAL, Rutherford Appleton Laboratory,

Chilton, Didcot 0X11 OQX, U.K.t IPCMS, 23 rue du Loess, BP 20/CR, 67037 Strasbourg Cedex, France

Existing applications of solid state physics are allbased upon inorganic materials; biological systemshowever utilise the flexibility of carbon chemistry andare nearly all organic. Our research has been moti-vated by the desire to utilise this underlying flexibilityin studying the physics of organic metals, magnets andsuperconductors. These new materials have only be-gun to be discovered by chemists rather recently andare typically rather chemically complex. Neverthelessmuon-spin rotation has proved to be a useful techniquein extracting important information in these systems [1].In particular, experiments have been performed at PSIand ISIS on nitronyl nitroxides [2, 3, 4, 5, 6], Cu-DCNQIsalts [7], organic superconductors based on the mole-cule BEDT-TTF [8, 9] spin-Peierls and gapped systems[10, 11] and molecular magnets based on transition-metal ions [12]

The parent materials of many organic supercon-ductors and metals are shown in Figure 1. Thesemolecules act as very efficient donors or acceptorsof charge. Thus they can be used in salts in whichthe band-filling can be tuned to produce partially filledbands. Muons have been used very successfully tostudy salts based upon these materials [8, 10] but itis very useful to probe the nature of the muon stateformed. This is possible using transverse-field /JSRand ALC experiments. We find temperature depen-dent ALC resonances in all three molecules shown inFigure 1. In particular in BEDT-TTF we find an asym-metric resonance just above 1T corresponding to A «300 MHz and D « 18 MHz. A second resonance devel-ops above 200 K and becomes the dominant feature athigh temperature. Similar behaviour is found in TTF,

but the resonances are shifted to higher frequenciesand the lower resonance is only visible above ~350Kand then only weakly. TCNQ shows a strong ALC res-onace at very low-field. Further studies and calcula-tions are in progress to determine the muon-site fromthese data.

REFERENCES

[I] S.J. Blundell, Appl. Mag. Res. 13 (1997) 155.

[2] T. Sugano etal., Mol. Cryst. Liq. Cryst. 305 (1997)435.

[3] T. Sugano etal., Mol. Cryst. Liq. Cryst. in press (toappear 1999).

[4] R.M. Valladares et al., J. Phys.: Condens. Matter10(1998) 10701.

[5] S.J. Blundell etal., Synth. Met. 85 (1997) 1745.

[6] S.J. Blundell etal., Hyp. Int. 104 (1997) 251.

[7] F.L. Pratt etal., Synth. Met. 85 (1997) 1747.

[8] S.L. Lee etal., Phys. Rev. Lett. 79 (1997) 1563.

[9] S.J. Blundell etal., Synth. Met. in press, to appear1999.

[10] S.J. Blundell etal., J. Phys.: Condens. Matter 9(1997) 119.

[II] Th. Jestadt et al., J. Phys.: Condens. Matter 10(1998)259.

[12] Th. Jestadt etal., Synth. Met. (toappear 1999).

S ' S CN

CN ' — ' 'CN

BEDT-TTF TCNQ TTF

Figure 1: The organic molecules BEDT-TTF, TCNQand TTF.

79

SYSTEMATIC STUDY OF THE MAGNETIC PENETRATION DEPTH IN CHEVREL PHASECOMPOUNDS

RA-94-01, ETH ZURICH - GENEVE

F.N. Gygax*, M. Pinkpank*, A. Schenck*5, M. Decroux*, 0 . Fischer*

* Institute for Particle Physics, ETH-Zurich, CH-5232 Villigen PSI, SwitzerlandI Departement de Physique de la Matiere Condensee, Universite de Geneve, 32 bd. d'Yvoy,

1211 Geneve 4, Switzerland

Since 1994 we are studying with yearly beam timeallocations of a few shifts stoichiometric Chevrel phasesuperconductors with the goal to elucidate a possiblecorrelation of Tc and the magnetic penetration depthas has been found in solid solutions of SnMo6S8_.rSe.i;and PbMo6S8_.TSe.r before [1]. Up to now we have in-vestigated LaMo6Ss (Tc ~ 7 K), AgMo6S8 (Tc ~ 8.4 K),YbMo6S8 (Te ~ 9.6 K), YMo6S8 (Tc ~ 2.3 K) and -not complete yet - ScMo6S8 (Tc ~ 3.6 K). In all in-stances transverse fields of 100 G and 500 G, and insome instances for consistency checks also 1 kG and6 kG have been used. Fig. 1 displays the 500 G -results on the relaxation rates, using for conveniencean exponential relaxation function which led to betterx2 than a Gaussian function. In the end, however,the penetration depth A will be determined from a lineshape analysis of the TF-signals. This log-log plot re-veals that it would be most desirable to have data fromsamples with Tc ~ 4 - 6 K. A candidate compoundwould be NaMo6S8 with Tc = 4.2 K. Grosso modothe present results follow a similar behaviour as thesolid solution compounds: i.e. the penetration depthbecomes smaller with rising Tc. There is one notableexception, namely YbMo6S8 which, although it displaysthe largest Tc among the present samples, does notshow the smallest A, as indicated by the smaller lowtemperature damping rate in Fig. 1. Another puzzle isrepresented by AgMo6S8. While first measurements inthe 1994 and 1995, as reported before, could not de-tect any signature forthe presence of the flux line latticein AgMo6S8 below Tc, although Tc was clearly iden-tified by ac-susceptibility measurements, a new mea-surement on the same sample in 1997 produced theresults shown in Fig. 1. The earlier results would implya penetration depth larger than 10000 A, the new resultimplies A ~ 2000 A. We have no explanation for theseseemingly inconsistent results.

Concerning YbMo6S8 the present data also prove thatno magnetic order down to 2 K develops in contrast toother reports which claim an onset of AF-order around2.7 K. This settles the open question of the valency ofYb in this compound: it is Yb2+.

in

1

ICDC

1CT-

o La Mq, §D YbMOeS,A AgMo6 %O Y Mo6 %

A---—A-.-.

D

10u

Temperature (K)101

Figure 1: Temperature dependence of relaxation ratesfrom exponential fits to /iSR-spectra collected at 500 Gand for various compounds.

REFERENCES

[1] P. Birrer et al., Phys. Rev. B48 (1993) 16589.

80

EVIDENCE FOR A NOVEL MUON SPECIES IN CRYSTALLINE SILICON

RA-93-10, STUTTGART - P S I

M. Schefzik*, A. Seeger**, D. Herlach*, O. Kormannt J. Major* ts, R. Scheuermannt§, A. Rock*, L Schimmele^

* Universitat Stuttgart, InstitutfurTheoretische und Angewandte Physik, D-70569 Stuttgart, Germanyf Max-Planck-lnstitutfur Metallforschung, Heisenbergstr. 1, D-70569 Stuttgart, GermanyI Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland§ Present address: Muon Science Laboratory, Institute of Physical and Chemical Research (RIKEN), Japan

In order to study the dynamical processes of theparamagnetic (normal (MUT) , anomalous (MUBC) muo-nium) and diamagnetic (fid) muon species in crys-talline silicon, undoped and differently doped (1015-1016 cm~3, P or B) floatzone-grown silicon was investi-gated by means of longitudinal-field quenching (LFQ).

During earlier investigations a significant differencebetween the doped and intrinsic samples was ob-served at temperatures below 100 K. Whereas the fitof the most general model function to the results ob-tained at the doped sample is very good in the wholemagnetic-field range, in the case of the undoped sam-ple a significant deviation from the theoretical curveat small magnetic fields is observed [1] as shown inFig. 1(a).

This deviation can be removed by introducing afurther, hitherto unknown, muon species termed Muvwith rather weak hyperfine interaction [2], as shown inFig. 1(b). Although the initial formation probability ofMuv is vanishing, a transition MUT -»• Mu v with a rateof (1.0±0.4)-106 s"1 is obtained. This means, that Mu v

is formed through its precursor MUT- MUV is tentativelyinterpreted as a muonium-vacancy complex. The va-cancy is presumably generated by the muon itself dur-ing the last stage of its slowing-down. This interpreta-tion is supported by recent results of Bech Nielsen etal.[3], in which a hydrogen-vacancy defect pair with weakhyperfine interaction was identified in proton-irradiatedcrystalline silicon by means of EPR.

Zero-field measurements were performed for directobservation of the hyperfine precession frequencies ofMu v . The data evaluation is still in progress, how-ever, preliminary results for the hyperfine parametersof Mu v , Ax - Ay = -12.06 MHz and Az = 25.74 MHzcoincide very well with the rescaled EPR results fromBech Nielsen ef al. [3].

The samples were kindly supplied by Dr.W. Zulehner, Wacker Siltronic AG, Burghausen, Ger-many. The work was funded by the BundesministeriumfurBildung und Forschung, Bonn, Germany under con-tract nos. 03-MA5ST1 and 03-MA5ST2.

REFERENCES

[1] J. Schmidl, Dr. rer. nat. Thesis, UniversitatStuttgart, Cuvillier Verlag, Gottingen (ISBN 3-89712-039-9), 1997.

[2] M. Schefzik, J. Schmidl, R. Scheuermann, LSchimmele, A. Seeger, D. Herlach, O. Kormann, J.

1.0 F

.2 0 8

•g 0.6"5g; 0.4

"Q.

°? 0.2 -

0.0

• ]

C 0.6-

0.4-

F 0 , 2 :

/ •

•'

0.000 0.004

- ^

0.008

PI11

•

-

-

0.012

0.0 0.1 0.2 0.3

B [ T ]0.4 0.5 0.6

0.2 0.3

B[T]

Figure 1: LFQ results obtained on an intrinsic siliconsample at T = 10 K (B | | ( l l l ) ) . (a) The solid line rep-resents the results of a fit with a model including MuT ,MuBc> and fid. Inset: Magnification of low-field rangeB < 0.01 T. (b) The solid line represents the results ofa fit with a model including MuT , MuBc> pd. and Mu v .

Major, and A. Rock, Solid State Comm. 107 (1998)395.

[3] B. Bech Nielsen, P. Johannesen, P. Stallinga, K.Bonde Nielsen, Phys. Rev. Lett. 79 (1997) 1507.

81

RELAXATION RATE AND PARAMAGNETIC FREQUENCY SHIFT OF NEGATIVE-MUONSPIN PRECESSION IN SILICON

RA-97-25, DUBNA - MOSCOW - PSI - STUTTGART

T.N.Mamedov*5, D.G.Andrianovt, I.LChaplygin*, V.N.Duginov*, V.N.Gorelkin*, D.Herlach§, O.Kormann11, J.Major1',A.V.Stoykov*, M.Schefzik1', U.Zimmermann§

* Joint Institute for Nuclear Research, 141980 Dubna, Moscow region, Russiat Institute for Rare Metals "Giredmet", 109017 Moscow, RussiaI Moscow Institute of Physics and Technology, 141700 Dolgoprudny, Moscow region, Russia§ Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandf Max-Planck-lnstitutfur Metallforschung, D-70569 Stuttgart, Germany

As has been shown in theoretical [1] and experi-mental [2, 3, 4, 5] work, the //~SR method makes itpossible to study the behaviour of acceptor centres insemiconductors, substantially supplementing the dataobtained by other nuclear methods. In particular, theundamped component of residual polarization of neg-ative muons in n-type silicon with the highest dopantconcentration at low temperatures was first observedby (iSR [5]. This effect was explained by a transitionof the acceptor centre from the neutral (paramagnetic)to an ionized (diamagnetic) state, and the rate of thetransition was obtained from the experimental data.

The main goal of our experiments in 1998 was to in-vestigate the role of acceptor-donor pairs in the forma-tion of an acceptor centre in the ionized state in n-typesilicon. The measurements were carried out on twosamples with aluminium impurity (2.4-1018 cm"3 and2-10 l 4crrr3) and on three phosphorus-doped samples(3.2-1012 cm- 3 , 2.3-1015 cm- 3 and 4.5-1018 cm-3).

Figure 1: Temperature dependence of the relaxationrate A and transition rate (paramagnetic -> diamagneticstate) v on the highly phosphorus-doped Si sample.

As in the silicon sample with a high (2-10 l8cm-3)concentration of antimony impurity [5], both the dam-ped and undamped components of the residual polar-ization of negative muons are observed in the samplewith phosphorus impurity (4.5-10 l8cm-3) at tempera-tures below 30 K. The rates of muon spin relaxationin the paramagnetic (neutral) state and the transitionfrom trie neutral to ionized state are found within theframework of the previously developed formalism [5]

(see Fig. 1). Unlike the samples highly doped with an-timony and phosphorus, the highly aluminium-doped(2.4-10 l8cm-3) silicon crystal shows only the dampedcomponent of the polarization at temperatures below30 K. The present experimental results are fully consis-tent with our assumption [5] that at high concentrationof donors in silicon the acceptor centres form acceptor-donor pairs. The recombination of charge carriers onthe pairs results in the ionized state of the acceptor.

On the investigated samples relaxation and frequen-cy shift of the muon spin precession were observed attemperatures below 50 K. The frequency shift was aslarge as Sw/w = (7+8)-10-3 at 15K. It is likely that onsome samples the shift is higher at lower temperatures.

On all the samples the temperature dependence ofthe relaxation rate of the muon spin is well approxi-mated by the power function A = C • T~q. It is alsofound that the relaxation rate depends on the concen-tration of impurities: the higher the concentration, theweaker the dependence of the relaxation rate on tem-perature. Evidently, the concentration dependence ofthe relaxation rate is universal and does not depend onthe type of the impurity.

Transverse field measurements on intrinsic siliconsample (p «104 Ohm-cm) show that it can be used tostudy the influence of the applied electric fields on theprocess of acceptor centre formation. In contrast to theresults of the measurements at B = 4mT [6], in theintrinsic silicon sample we observed muon spin relax-ation at temperatures below 15 K, the amplitude of theprecession at zero time was close to those observedat room temperature (in this experiment the transversemagnetic field was B = 0.2T).

REFERENCES

[1] V.N. Gorelkin etal., JETP 66 (1974) 1201.

[2] T.N. Mamedov etal., Hyp. Int. 86 (1994) 717.

[3] W. Beez et al., PSI Annual Report / Annex I (1993)125.

[4] T.N. Mamedov etal., Hyp. Int. 105 (1997) 345.

[5] T.N. Mamedov etal., JETP Lett. 68 (1998) 64.

[6] R. Kadono etal., RIKEN-RAL Muon Facility Report1 (1997)41.

82

MUONS IN SULPHUR

RA-97-28, RAL-PSI

S.F.J. Cox*s, I.D. Reid*

* ISIS Facility, Rutherford Appleton Laboratory, Chilton, Oxfordshire 0X11 OQX, U.K.f Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland

We reported last year the observation of two com-ponents at the muon Larmor precession signal in sul-phur below room temperature - one with very fast andthe other with very slow relaxation - in addition to a aweak ALC radical signal. Further experiments in 1998using the GPD spectrometer and "chemistry" cryostatreached higher temperatures, covering the a-p phasetransition at 368K and the melting point at 392K. Thecombined data (Figures 1 and 2) reveal a broad max-imum in the fast relaxation rate. The correspondingamplitude remains constant above room temperaturewithin the alpha phase but decreases sharply in the /?phase towards the melting point. In molten sulphur,most of the polarization is recovered in the slow com-ponent.

^ 120

\~ 100g(0

a 80CDrr•g 60CDCo| 40oOV! 20LL

n

~0.30

"c 0.25.e(0 0.20"5rr| 0 - 1 5

|| 0 . 1 0oO

I 005w

nnn

o Heating7 Cooling• QuenchedO GPS

• Phase Transition— - Melting Point

<

10

A •

<

II

W 1

10

<

<

<

*

D • ° % C5

e

1

e

150 200 250 300

T(K)350 400

Figure 1: Relaxation rates of the two TF-/iSR diamag-netic signals in S. Earlier GPS data are also shown.

0.25

= 0.20"5.

o Heatingv Coolingn QuenchedO GPS

Phase Transition— • Melting Point

0.15tu

I" 0.10O•a,« 0.05

n nn0.25

<•g 0.15CDtzoa.E 0.10oO

^ 0 . 0 5

0.00150 200 250 300

T(K)350 400

Figure 2: Amplitudes of the two TF-//SR diamagneticsignals in S. Earlier GPS data are also shown.

tation itself, the nuclear magnetism in sulphur beingtoo weak and there being no other electronic momentspresent in sulphur in sufficient concentration. It seemslikely that the variation of relaxation rate with temper-ature, and of the partition between the two fractionswithin the partially disordered beta phase, reflects somecombination of electron and muon mobility but so farthe data are too novel for us to have reached any firmhypothesis.

As expected, the metastable nature of the /? phasewas seen, the data showing a dependence on thermalhistory and retaining some characteristics of the betaphase upon cooling or quenching.

The fast relaxation must be ascribed to encounterwith radiolytic electrons generated by the muon implan-

83

//SR IN II-VI SOLAR CELL MATERIALS

RA-97-23, COIMBRA- HMI BERLIN - KONSTANZ - ISIS

N. Ayres de Campos*-9, J.M. Gil*, H.V. Alberto*, R.C. Viläo*, RJ. Mendes*, L.P. Ferreira*, A. Weidinger1,Ch. Niedermayer*, S.F.J. Cox§

* Physics Department, University of Coimbra, P-3000 Coimbra, Portugalf Hahn-Meitner Institut Berlin, Abteilung FH, Glienicker Str. 100, D-14109 Berlin, Germany| Fakultät für Physik, Universität Konstanz, Universitätstrasse 10, D-78434 Konstanz, Germany§ ISIS Facility, Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire, OX11 OQX, United Kingdom

In elemental (Group IV) and compound (Ill-V) semi-conductors, the majority of what is known of isolatedhydrogen defect centres comes from studies of theirmuonium counterparts [1]. This project aims at extend-ing this study to II-VI compound semiconductors andchalcopyrites. The presence of hydrogen in these ma-terials is receiving much attention in order to study itsfundamental aspects and relations to performances insolar cells and other devices [2].

We started //.SR experiments in CdS in 1997 at PSIin transverse field at low temperatures. An unexpectedsmall splitting around the Larmor frequency was iden-tified. This splitting was studied in detail in 1998 by ori-entational and temperature dependent transverse fieldmeasurements. We were able to resolve these split-tings of extremely low hyperfine constant thanks to thenew MORE (Muons On REquest) set-up. An undopedsingle crystal of Wurtzite structure was used, in a tem-perature range of 2 K to 300 K. The sample was ori-ented so that the most favourable positions were im-posed to the Cd-S bond directions according to thecrystal structure. Spectra were obtained with the sam-ple oriented in two ways: a) c axis parallel to the ex-ternal field (0°) - in this way one bond is parallel tothe field and three bonds are at 70.6°, b) c axis rotatedabout the (110) axis by 54.7° - five bonds at 54.7°,two bonds at 79°, one bond at 164°. In the time spec-trum obtained at 0° with an applied magnetic field of10mT (Fig. 1) two pairs of lines were found positionedsymmetrically around the central diamagnetic Larmorfrequency. The respective Fourier transform is shownin Fig. 2. The splittings found are Avout = 335(1) kHzand Avin = 214 (1) kHz. At 54.7° (at the same field) asingle pair of symmetrically positioned lines was foundwith a splitting of Av = 243(1) kHz.

This is consistent with an anisotropic axially sym-metric muonium with symmetry axis along the Cd-Sbond direction. We assign Avout to the muonium frac-tion at the 9 = 0° bond, Avin to the ô = 70.6° bonds, andAv to the 9 = 54.7° bonds. The splittings due to the 9= 79° and 9 = 164° bonds were not observed by lackof resolution in the spectra taken at 54.7 ° (the MOREset-up was not used in this case). External fields of2 mT, 5 mT and 10 mT didn't change the splittings. Weconsider that the high-field limit is already reached wellbelow 2mT for this muonium state, and calculate thehyperiine parameters to be A\\ = 335 (1) kHz and A± =198(1) kHz [3].

Temperature dependence measurements were per-

10Time (jxs)

15

Figure 1: Muon-spin precession in CdS obtained withthe MORE set-up (c axis parallel to the applied field).

CdS2K 100 G

Frequency (MHz)

Figure 2: Fourier transform of the spectrum in Fig. 1.

formed at 54.7° and 1 mT. The splitting decreases withincreasing temperature and vanishes around 24 K. Thefraction of muons showing the splitting is also decreas-ing, the diamagnetic fraction reaching 100% above thattemperature. A possible scenario is that the paramag-netic electron is distributed over a large complex ofatoms in the neighborhood, the spin density at the muonbeing very weak but not zero. The temperature depen-dence corresponds possibly to an ionisation process ofthe paramagnetic electron.

REFERENCES

[1] S.F.J. Cox, Phil. Trans. R. Soc. Lond. A350 (1995)171.

[2] D. Albin et al., AIP Conf. Proc. 394 (1997) 665.

[3] B.D. Patterson, Rev. Mod. Phys. 60 (1988) 69.

84

THE fx+ IN METAL-HYDROGEN (DEUTERIUM) SYSTEMS

RA-97-17, ETH ZURICH - PALAISEAU - RIKEN - EKATERINBURG - PSI

F.N. Gygax*5, P. Vajda*, D. Andreica*, M. Pinkpank*, A. Schenck*, E. Yagi{, A.V. Skripov§, P. Fischer11

* Institute for Particle Physics, ETH Zurich, CH-5232 Villigen PSI, Switzerlandf Laboratoire des Solides Irradies, Ecole Polytechnique, F-91128 Palaiseau Cedex, France% The Institute of Physical and Chemical Research (RIKEN), Wako, Saitama, 351-01, Japan§ Institute of Metal Physics, Urals Branch of the Academy of Sciences, Ekaterinburg 620219, Russiaf Lab. of Neutron Scattering, ETH Zurich and Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland

/3-RD2+a;(R = rare-earth)The temperature dependences of the ZF-relaxation

data in the compounds HoD212, HoD20 and DyD20

show similar behaviour, and the LF data in HoD2.i2 andHoD20 are identical to the corresponding ZF data - seeFig. 1 [1]. At the critical temperatures Tci (about 6.5KinHoD20, 6.8K in HoD212 and 5.7K in DyD20) the expo-nential relaxation rates A, are largest and hit theresolution limit in the Ho compounds.

102

10'-

10°

i

s

w

¥

hnK

/ HoD2.0

/ r- HoDj.^

^ / HoD2 n

>^v2^-

D2.o - ZF

- ZF/LF- ZF- LF

H

10° 101 102

TEMPERATURE (K)

Figure 1: Temperature dependence of the ZF and LFrelaxation rates A in three p-RD2+x compounds.

0.000-0.002-0.004

t -0.006

|-0.012^-0.014

-0.016-0.018

-1 1 2 31000/T (1/K)

Figure 2: Knight shift and relaxation rate of the TF /uSRsignal in HoD2.i2 versus inverse temperature.

On the T > Tci side the relaxation can well be fittedto A; = ai\T - Tci\-

Pi, with pi = 0.5 for the Ho com-pounds and fa = 0.3 for DyD20. In this temperatureregion most of the muons contribute to the //SR signal(asymmetries At = 20-25%).

On the low-temperature side of Tci a critical behav-iour is also observed, but the exact dependence is noteasy to determine, because of (i) only a reduced frac-tion of the muons visible in the signal (A{ = 3-8%), and

(ii) the observed hysteretic behaviour, possibly relatedto the simultaneous presence of two AF phases.

In a transverse field of 0.4T the Knight shift A'p ofthe /^SR signal in HoD2 i 2 varies linearly with l/T for200 K < T < 300 K (Fig. 2). It can be discussed interms of the magnetic susceptibility of the compound.The exponential relaxation A is also displayed.

Laves-phase compounds TaV2Dj;

§ 10~<x<

10"

B D B D—B—B-B

- e - © - « - e - - - e - - e - ©

TaV2D,65

10° 10' 102

TEMPERATURE (K)

Figure 3: Temperature dependence of the ZF fiSR re-laxation rates A (•) and A (o) in TaV2 and TaV2Di 65.

TaV2 and TaV2Di.65 have been measured in ZF andTF mode [1]. Previous ZF measurements performed atISIS [2] in TaV2 and TaV2Hi were interpreted in termsof a dynamic Kubo-Toyabe (KT) relaxation function. Wefound in our study that it was not possible to performsuch an analysis, the /iSR signal being more complex.We present in Fig. 3 the result of a phenomenologicfit of the data, using a damped KT function, with theKT parameter A and the damping A. It shows that theeffect of dynamics is observed at lower temperature inTaV2 compared to TaV2Di.65, but also that a noticeablechange in the relaxation regime occurs in TaV2 around40 K (trapping?).

REFERENCES

[1] For the 1997 measurements see F.N. Gygax et al.,PSI Annual Report 1997, Annex I, p. 43 and 69.

[2] M. Kemali et al., J. Alloys Comp. 231 (1995) 259.

85

REORIENTATION DYNAMICS AND CHEMICAL REACTION OF CYCLOHEXADIENYLRADICALS IN ZSM-5 ZEOLITES

RA-86-07, STUTTGART - PS I

E. Roduner*s, H. Dilger, U. Himmer*, R. Scheuermann**, I. D. Reid*

* Institut fur Physikalische Chemie, Universitat Stuttgart, Pfaffenwaldring 55, D-70569 Stuttgart, Germanyf Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandt Present address: Muon Science Laboratory, Institute of Physical and Chemical Research (RIKEN),

Wako-shi, Saitama, Japan

Crystalline microporous aluminosilicates, called ze-olites, are of high current interest for a variety of impor-tant applications. They are used as molecular sieves,adsorbants and heterogenous catalysts with shape se-lective activity. One of the most prominent structuresis that of ZSM-5 with its three dimensional network. Inparticular the highly active H-form is used on a largescale for the conversion of methanol to gasoline, butother variations of the material have found numerousother applications.

HZSM-5 was loaded with one benzene moleculeper unit cell, which is about 1.3% of the total weight,and then investigated using the Avoided Level Cross-ing muon spin resonance (ALC-^.SR)[1]. This tech-nique gives rise to resonances in plots of time inte-gral muon spin polarization as a function of the appliedmagnetic field. The resonance fields are related to thehyperfine coupling constants in the adduct of muonium(Mu= fj,+e~, chemically a light hydrogen isotope witha polarized muon as nucleus). In the present case theadduct is the cyclohexadienyl radical, C6H6Mu. If theradical undergoes chemical reactions, in the presentcase a protonation is expected, the ALC-spectrum willchange.

In previous ALC experiments of copper exchangedNa-ZSM-5 we found, that the cyclohexadienyl radicalformes a weak complex with the copper cation [2]. Inthe proton exchanged form, HZSM-5, there was no in-dication for such an association of the proton with theradical above 300 K, but below there was a dramaticchange of the spectrum, indicating protonation of theradical, C6H6Mu+H+ -» C6H7Mu+, to give the cyclo-hexadiene radical cation. There are two possible iso-mers of C6H7Mu+, afforded by protonation in the orthoor para position to Mu. At low temperature, at least theortho adduct is nonplanar and is expected to freeze ina conformation that leaves Mu in two inequivalent po-sitions with Mu hyperfine couplings of about 460 MHz(axial) and 140 MHz (equatorial), while about 730 MHzare expected for the para adduct.

New ALC experiments with DZSM-5 confirm theresults obtained with HZSM-5 (Figure 1). The situ-ation is most clearly seen at 225 K where three newresonances (3,4,5) appear while the two cyclohexadi-enyl resonances observed at higher temperatures (1,2)loose amplitude. The new lines are compatible withprotonation of C6H6Mu in the ortho position, leaving Muin the axial position (Ai at 1.82 T=^4M = 495 MHz, line

J 3; lines 4,5 correspond to Ao lines of the other protonsat saturated carbons). This shows that the catalytic ac-tivity of HZSM-5 which is ascribed mostly to H+, isbased on a delicate, temperature dependent balanceof protonation.

1.3% wt C.K DZSM-5/50

/ V

0.4 0.B 1.2 1.6 2.0 2.4 2.8

B[T]1.4 1.6 1.B 2.0 2.2 2.4 2.6

B[T]

Figure 1: ALC-//SR spectra obtained with benzene inDZSM-5 and HZSM-5 at various temperatures and aloading of 1.3% wt. Note the different scales.

REFERENCES

[1] E. Roduner, Chem. Soc. Reviews 22 (1993) 337.

[2] M. Stolmar, E. Roduner, J. Am. Chem. Soc. 120(1998)583.

86

HETEROGENEOUS PROCESSES OF ENVIRONMENTAL FREE RADICALS

RA-93-02, LIVERPOOL - PSI

Ch.J. Rhodes*5, H. Morris*, T.C. Dintinger*, I.D. Reid*

* School of Pharmacy and Chemistry, Liverpool John Moores University, Byrom St., Liverpool L3 3AF, U.K.t Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland.

To date, the majority of atmospheric processeshave been regarded in terms of largely gas-phase re-actions, and there have been many studies made [1]of model gas systems in an overall effort to under-stand the atmospheric phenomena themselves. Morerecently, however, has emerged the view that aerosols,including solid particulate matter, may participate in theoverall chemistry of the atmosphere, by virtue of reac-tions of either dissolved or surface adsorbed compo-nents [2].

In any surface reacting process, it is important tounderstand the details of surface adsorption and dif-fusion of the reacting components; in the chemistryof the stratosphere direct photodecomposition of halo-carbons plays a major role, for instance as used asrefrigerants (CFC's) and organic solvents, which arethus converted to corresponding halocarbon radicals.Though there is the suggestion [1] that an involvementmay exist of particulate matter in this type of chem-istry, there is no direct study to date, probably in con-sequence to the perennial difficulty of studying surfaceprocesses involving radicals, where conventional meth-ods such as ESR spectroscopy lack the sensitivity re-quired under conditions of fast surface radical diffusionand combination.

As demonstrated admirably by transverse-field [3],ALC [4] and longitudinal-field relaxation [5] measure-ments, radicals may, however, be studied in a range ofsurface environments using muons.

In this preliminary study, we have managed to de-tect MuCH2CCI2- radicals adsorbed on a silica surfaceusing both transverse-field (at PSI) and longitudinal-field muon relaxation (at RAL) measurements. Thetransverse-field measurement provides a direct char-acterisation of the nature of the radical ( A M « = 2 3 8 M H Zat 298 K), and also an indication of the activation en-ergy to its motion over the surface at which it is ad-sorbed, from the temperature dependence of the line-widths; more directly, the longitudinal-field (LF) relax-ation measurements provided an almost identical ac-tivation energy of ~10 kJ/mol for its motion over thesilica surface, but also showed the presence of a moreweakly adsorbed fraction (Ea=0.4 kJ/mol) which wasnot present at lower surface coverages. [Of course,far lower coverages can be studied using a longitudinalfield, where radicals formed at least within the muonlifetime are still detectable; this is not true in the TFmeasurement which relies on a fast rate of formation].

We also studied the MuCH2CCI2- radicals adsorbedon kaolin (an atmospheric particulate constituent stem-ming from dust, wind-blown from deserts), with actuallysimilar results: both TF and LF methods showing an

activation energy of ~9 kJ/mol for the main fraction,with an additional more weakly adsorbed component(Ea=2 kJ/mol) being indicated by the LF study.

We intend to extend this work by studying a rangeof radicals formed by the degradation of atmosphericpollutants on particulate surfaces.

We thank the Paul Scherrer Institute, the EPSRC,the European Union, the Leverhulme Trust, UnileverResearch and John Moores University for financial sup-port of this overall programme of work.

REFERENCES

[1] R.P. Wayne, The Chemistry of Atmospheres,Clarendon Press, Oxford (1985).

[2] M. Kalberer et a/., J. Phys. Chem. 100 (1996)15487.

[3] C.J. Rhodes etal., Magn. Reson. Chem., 33 (1995)S134.

[4] E.Roduner et al. in Radicals on Surfaces, eds.A. Lund and C.J. Rhodes, Kluwer, Dordrecht(1995).

[5] C.J. Rhodes etal., submitted to J. Phys. Chem.

87

RADICAL CHEMISTRY WITHOUT WATER - INSIGHTS INTO CELL MEMBRANES

RA-96-09, LIVERPOOL - PSI

C.J. Rhodes*s, C.S. Hinds*, T.C.Dintinger*, I.D. Reid1

* School of Pharmacy and Chemistry, Liverpool John Moores University, Byrom St., Liverpool L3 3AF, U.K.f Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland.

Pulse-radiolysis is the classical method for the de-termination of the rate constants of free radical reac-tions, but generally necessitates using aqueous media,since the radicals are generated via the radiolysis prod-ucts of water, e.g. hydroxyl radicals. This should pro-vide no difficulty when it is desired to model processesthat occur in aqueous regions of cells, but is surelya great limitation in the understanding of radical reac-tions in cell membranes; these are non-aqueous, andit is known that the rates of radical reactions may beprofoundly altered according to the hydrophilic or hy-drophobic character of their medium.

The muon spin rotation method does not require amedium of any particular kind, beyond it containing anappropriate substrate for muonium addition, since themuons may be implanted into any medium and bothaqueous and non-aqueous phases could, in principle,be studied.

Our current work concerns the kind of radical chem-istry believed to be significant in damage to cell mem-branes: firstly, we consider the duplicity of thiols andthiyl radicals [1], where the thiol in its act of protectiongenerates a thiyl radical, which is reactive and itselfdamages lipid membrane components. We find thatthe thiyl radicals are indeed reactive, and abstract H-atoms from "allylic" groups 3-4 times faster (in THF)than in aqueous solution; however, they are readily in-tercepted by beta-carotene and glutathione, for whichthe rates are also faster by a factor of 3-4 comparedwith aqueous media.

These studies suggest that thiyl radicals are evenmore destructive to membrane lipids than previouspulse-radiolysis data would indicate. Significantly,their reaction rates with antioxidants such as glu-tathione(SH) and beta-carotene are also enhanced(and have anyway far greater rate constants).

It might appear as though Nature has compensatedfor this increased rate of damage by allowing an ac-cording increase in the rate of repair, but such repaircannot of course be perfect since during the course ofour lives we accumulate "free radical damage", arisingfrom lipid peroxidation of cell membranes.

We thought, then, to extend our study to radicalsof the type formed when membrane lipids are attackedby thiyl and other membrane free radicals, namely al-lylic and alkoxyalkyl radicals, arising respectively fromhydrogen atom abstraction at main chain CH2-C=Cand phosphoglyceryl functional groups: these we havemodelled by 1,1,2-trimethylallyl radicals and 1-acetoxy-iso-propyl radicals [formed by muonium addition to 2,3-dimethyl-1,3-butadiene (DMBD) and allyl acetate (Al-lAc), respectively], and have measured the rate con-

stants (Table) for their reactions with vitamin E and vi-tamin K, since these are antioxidants which locate pref-erentially in cell membranes.

Table 1: Rate Constants (M"1 s"1) for for Reactionsof Radicals from BMBD and AIIAc with Vitamin E andVitamin K

vitamin K vitamin E

DMBDAIIAc

3.5 x107 8.4 x106

2.3 x 10s 1.9 x 10s

The reactions with these 2 vitamins are almost cer-tainly of entirely different characters: that with vitaminK involving electron transfer from the radical to the1,4-naphthaquinone unit and forming the correspond-ing carbocation of the initial radical moiety, while thatwith vitamin E involves abstraction of the phenolic hy-drogen atom. The increased stabilisation of the result-ing carbocation in the DMBD system by both allylic de-localisation and the electron releasing effect of the 3methyl groups explains the much higher rate than forthe analogous reaction of the essentially localised AI-IAc system.

For hydrogen abstraction reactions, delocalisedradicals are always slower, so the rate is reduced forthe allylic DMBD radical compared with the AIIAc de-rived radical.

We thank the Leverhulme Trust, the EuropeanUnion, Unilever Research and John Moores Universityfor financial support.

REFERENCES

[1] C.J. Rhodes, in: Toxicology of the Human Envi-ronment - the Critical Role of Free Radicals, ed.C.J. Rhodes, Taylor and Francis, London (1999).

88

WEAK ORGANIC CHARGE TRANSFER SYSTEMS

RA-97-15, RIKEN-PSI-Stuttgart

R.M. Macrae*5, I.D. Reid1, J.U. von Schtitz*, K. Nagamine*

* Institute of Physical and Chemical Research (RIKEN), Wako, Saitama 351-01, Japan| Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandt 3. Physikalisches Institut, Universitat Stuttgart, D-70550 Stuttgart, Germany

Weak organic charge transfer compounds formedby 1:1 co-crystallisation of planar aromatic moleculesare of interest from the perspective of molecular dy-namics and, through the behaviour of triplet excitonsin these systems, as models of low-dimensional excita-tion transport [1]. Muon techniques offer a unique "han-dle" on these properties in that a single ^-generatedfree radical probe can be utilised to study both ther-mal excitations and molecular reorientational dynam-ics (using conventional ^SR techniques) and photoin-duced behaviour and excitonic dynamics (using syn-chronous excitation /*SR [2]), separating two influencesthe effects of which are usually superposed. Initial testswere carried out on the system anthracene/tetracyano-benzene (A/TCNB) at PSI and RIKEN, and preliminaryreports of the results are available [3]. Synchronousexcitation studies await the development of a suitablelaser setup at RIKEN/RAL.

Hexamethylbenzene (HMB)

NC.

Anthracene (A)

Naphthalene (N) 1,2,4,5-Tetracyanobenzene CTCNB)

Figure 1: Donors and acceptors.

Subsequent studies on A/TCNB were extended toinclude temperature-dependent high-resoiution TF y«SRon a single crystal sample, and ALC-^SR measure-ments were carried out on polycrystalline samples oftwo other systems in which the "donor" molecule is re-placed by naphthalene (N/TCNB) and hexamethylben-zene (HMB/TCNB) respectively. (Through systematicchanges of this kind, the properties can to a certain ex-tent be "tuned".) The free radical species formed wereidentified through their characteristic values of Bo, theA M = 1 lineshape envelope origin, by comparison withvalues of AM, the isotropic component of the muon-electron hyperfine coupling, to which it is proportional.Values were taken both from the experimental litera-ture [4, 5] and from the results of first principles calcu-lations using the density-functional-based method B-PW91 [6].

The results were striking in several ways. Firstly, inno case was the adduct to the TCNB moiety observed.This can be only partially due to the somewhat greatersteric freedom of the donor species and the accom-panying stabilisation of radical adducts to it, and must

have implications with regard to the formation mecha-nism of the radical, suggesting that in systems of thistype, at least in the solid state, ionic intermediates arenot involved. Secondly, the relatively open crystal lat-tice of HMB/TCNB leads to considerably more dynam-ical averaging than in pure hexamethylbenzene at cor-responding temperatures [5]. Thirdly, the complex andsubtly T-dependent lineshape in N/TCNB in the lowerpart of the temperature range studied (see Fig. 2) in-dicates that ALC-^SR is particularly effective here indistinguishing changes in static order, implying that in-termolecular interactions have a considerable influenceon the radical hyperfine tensor.

111111111111111111111111 111 i

10000 12000 14000 16000 18000 20000 22000

Field/G

Figure 2: N/TCNB powder: ALC-juSR spectra.

REFERENCES

[1] J. Krzystekand J. U. von Schiitz: Adv. Chem. Phys.LXXVI (1993) 167, and refs. within.

[2] R. Kadono, A. Matsushita, R.M. Macrae, K. Nishi-yama, K. Nagamine, Phys. Rev. Letters 73 (1994)2724.

[3] R.M. Macrae, T.M. Briere, C.J. Rhodes, I.D. Reid,J.U. von Schutz, Proc. Intl. Workshop on JHF Sci-ence (JHF98) III 335.

[4] I.D. Reid, E. Roduner, Struct. Chem. 2 (1991) 419.

[5] E. Roduner, Hyp. Int. 65 (1990) 857.

[6] LA. Eriksson, O.L. Malkina, V.G. Malkin, D.R.Salahub, J. Chem. Phys. 100 (1994) 5066.

89

//SR DYNAMICS OF LIQUID CRYSTALS STUDIED BY ALC

RA-97-22, OXFORD - RIKEN-RAL-PSI

S.J. Blundell*s, B.W. Lovett*, F.L Pratt*, Th. Jestadf, I.D. ReidJ, W. Hayes*,

* University of Oxford, Department of Physics, Clarendon Laboratory,Parks Road, Oxford 0X1 3PU, U.K.

| The Institute for Physical and Chemical Research (RIKEN), RIKEN-RAL, Rutherford Appleton Laboratory,Chilton, Didcot 0X11 OQX, U.K.

t Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland.

Liquid crystals have found many technological usesin displays and sensors and are found in naturally oc-curring systems e.g. biological membranes. They arealso fascinating systems in which novel forms of order-ing can be observed [1, 2]. They are a state of matterwhich is intermediate between homogenous isotropicliquids, which have an average structure which is invari-ant under arbitrary rotations and translations, and crys-talline solids, which are invariant under certain discretetranslations and symmetry operations. Thermotropicliquid crystals (those whose phases are controlled bytemperature) are typically rod-like molecules, usuallycontaining two or more benzene rings. This molecularstructure is crucial to their properties, and the positionsof the various phase transitions which occur in thesesystems is strongly dependent on the dynamical prop-erties of the molecules. Understanding the dynami-cal properties is also of great practical interest as thisaffects, for example, the frequency up to which liquidcrystal displays can be operated, and thus their useful-ness in computer lap-top applications.

Avoided level-crossing (ALC) /iSR [3, 4] is an ef-fective method of studying the dynamics of moleculesand radical states. The electron-nuclear hyperfine in-teraction of organic radicals is generally anisotropic; itpossesses a significant component associated with themagnetic dipolar interaction. Dynamical effects, if suf-ficiently rapid, lead to an averaging of this anisotropy.If the rotation is fast about a single axis, as occurs insome liquid cyrstal phases, the hyperfine interactionpossesses axial symmetry. In the liquid phase the fasttumbling processes will, for normal viscosities, aver-age the anisotropic component to zero, leading to anisotropic hyperfine tensor. The nature of the dynamics(and hence the symmetry of the hyperfine tensor) andthe nature of the radical state (and hence the value ofthe components of the hyperfine tensor) determine theposition and shape of the level-crossing resonances.

In our recent ALC experiments we have studied levelcrossings in the liquid crystal 5CB, which is made ofrod-shaped molecules which contain two benzene rings.This makes it ideal for muon addition, thus allowing theformation of radical states and the observation of levelcrossings. It has a nematic phase, where the mole-cules prefer to point along a particular spatial direction,but where their positions are as would be found in a liq-uid. We have observed four level crossings in this ma-terial, which I show in Fig. 1. These resonances show adependence on temperature as the solid-nematic and

nematic-liquid phase boundaries are crossed, whichwe may attribute to the different dynamic and symmetryproperties of each phase [5].

18000 18500 19000 19500 20000

Field (G)

Figure 1: ALC data for the liquid crystal 5CB in thesolid(S), nematic(N) and liquid(L) phases.

REFERENCES

[1] P.G. de Gennes and J. Prost, The Physics of Liq-uid Crystals, 2nd edition, Clarendon Press, Oxford(1993).

[2] S. Chandrasekhar, Liquid Crystals, Cambridge(1992).

[3] E. Roduner, Chem. Soc. Rev. 22 (1993) 337.

[4] E. Roduner and H. Fischer, Chem. Phys. 54 (1981)261; M. Heming et al., Chem. Phys. Lett. 128(1986) 100; R. Kiefl, Hyperfine Interactions 32(1986) 707. E. Roduner, Hyperfine Interactions 65(1990) 857.

[5] B.W. Lovett et al., in preparation.

90

TEMPERATURE DEPENDENCE OF THE HYPERFINE COUPLING CONSTANT OFMUONIUM IN LIQUID H2O AND D2O AND IN SOLID Si8O12-CAGES

RA-98-18, STUTTGART - PSI - BERLIN - STUTTGART

R. Scheuermann**5, E. Roduner*s, H. Dilger*, I.D. Reid*, D. Herlach*, R. Std(3er§, M. Pach§, J. Major^,M. Schefzik**

* Universitat Stuttgart, Institut fur Physikalische Chemie, Pfaffenwaldring 55, D-70569 Stuttgart, Germany| now: Muon Science Lab., Inst. of Physical and Chemical Research (RIKEN), Wako-shi, Saitama, Japan\ Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland§ Humboldt-Universitat zu Berlin, Institut f. Physikal. u. Theoret. Chemie, D-10115 Berlin, Germanyf Max-Planck-lnstitut fur Metallforschung, Heisenbergstr. 1, D-70569 Stuttgart, Germany** Universitat Stuttgart, InstitutfurTheoretische und Angewandte Physik, D-70550 Stuttgart, Germany

The direct measurement of the triplet-singlet tran-sition frequency of muonium (Mu) in zero applied mag-netic field yields the highest accuracy for the value ofthe hyperfine coupling constant^.,,. This kind of experi-ment requires an experimental time resolution of betterthan 150ps to detect oscillation frequencies of about4500 MHz. The feasibility of such fast-timing experi-ments has been demonstrated in measurements of A^in quartz [1]. Our interest is to study the variation of A^due to the dynamics of Mu trapped inside a cage.

For the case of atomic H and D in cages formed byH2O and D2O molecules, electron paramagnetic res-onance (EPR) experiments revealed a hyperfine cou-pling constant smaller than the vacuum value, a massdependence and a negative temperature coefficient [2].High-resolution EPR studies on H and D in solid com-pounds containing the cube-shaped Si8Oi2-units (T8-units, octasilsesquioxanes) were performed at the Uni-versity of Stuttgart. The hyperfine coupling constantsof atomic H and D in different T8-cages decreases bysaO.2% on lowering the temperature from 194 K to 40 Kand the temperature behaviour fits a single-frequencyoscillator model [3].

With a newly developed zero-field /<SR spectrome-ter an effective time-resolution of 140 ps at a muon mo-mentum of 60 MeV/c was achieved by mounting ultra-fast plastic scintillators directly onto the photomultipliertubes and using low-loss cables for the signal trans-mission to a constant fraction discriminator EG&G mod-el CFD935. The direct output of a logic unit SIN OR101provided the input signals for an EG&G ORTEC pi-cosecond time analyzer (pTA) model 9308. With thepresent version of the GUSTAV DAQ system the effec-tive histogram length was 260 ns at a time resolution of20 ps/channel) {http://musr.mpi-stuttgart.mpg.de).

Using the set-up described above, it was possibleto measure the triplet-singlet transition frequency ofMu in fused quartz (SUPRASIL, 4435.7 MHz, Fig. 1), inliquid H2O (4434.3 MHz) and in liquid D2O (4432.9 MHz)at room temperature. The use of a standard chemistrycryostat drastically reduced the count rate due to thesmall solid angle for the stop-detector. The effectivetime-resolution, however, did not change.

In order to determine the muonium formation prob-ability in different T8 samples TFjuSR experiments have

1800

1600-

1400:

1200

1000-

800'

iW '•400-

200-

o:

2000 3000

FREQUENCY (MHZ)

S000

Figure 1: FFT spectrum of zero-field ^SR on fusedquartz at room temperature.

been performed at an applied magnetic field of 15mTwith the GPD spectrometer and GUSTAV DAQ sys-tem. At 300 K a large Mu fraction («25%) was found inmethyl-T8 (MeT8), whereas no Mu precession signalswere detected in the polymerized form MeTn. In ethyl-T8 and in HT8 the Mu fraction was found to be smallerthan 20% at 300 K. In all three cases A^ is close to thevacuum value. An accurate determination, however,is not possible from these TF/iSR experiments. Theremay even exist two different Mu species (in-cage orintra-cage) with very similar A^, which would be distin-guishable in zero applied field only. MeT8 was inves-tigated in the temperature range 100K - 350 K. Withincreasing temperature the relaxation rates of the Musignals decrease, whereas the asymmetry of the dia-magnetic species (^d) increases and the correspond-ing relaxation rate decreases. The large missing frac-tion («25% at 300 K) clearly indicates some dynamicprocesses. The analysis of these data is still in progress.

Work partially funded by the Ministerium fur Bildungund Forschung, Bonn, Germany under contracts Nos03-MA5ST1 and 03-MA5ST2.

REFERENCES

[1] E. Holzschuh etal., Helv. Phys. Acta54 (1981) 552.

[2] E. Roduner etal., J. Chem. Phys. 102 (1995) 5989.

[3] E. Roduner et al., J. Phys. Chem. A 102 (1998)7591.

91

RANGE AND RANGE DISTRIBUTION OF LOW ENERGETIC MUONS IN MATTER

RA-94-07, PSI - BRAUNSCHWEIG - KONSTANZ - BIRMINGHAM

M. Birke*'f, E.M. Forgan*, H. Glückler*, A. Hofer*>§, T.J. Jackson*, F.J. Litters^, H. Luetkens**, E. Morenzoni*,Ch. Niedermayer§, M. Pleines*^, Th. Prokscha*, T.M. Riseman*, A. Schatz*, G. Schätzt and H.-P. Weber*

* Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandf IMNF der Technischen Universität Braunschweig, D-38106 Braunschweig, Germanyt School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, United Kingdom§ Fakultät für Physik, Universität Konstanz, D-78434 Konstanz, Germany

To study physical properties of thin films or multi-layer systems by the low energetic //.SR technique (Eß

< 30 keV) it is necessary to know the range and therange distribution of muons in matter. In 1998 we haveextended our measurements to thin AI and thin Au filmssputtered on quartz glass substrates. Together with ourdata from 1997 where thin Cu films were investigatedwe have now a series of range and range distributionmeasurements for metals which cover the interval fromZ=13(A I ) toZ = 79(Au).

The investigated films have thicknesses of 100nmand 400 nm for AI and 40 nm and 400 nm for Au. Eachmetal was deposited by sputtering [1] onto a quartzglass disc (SUPRASIL, 50 mm diameter, 1 mm thick-ness). The thickness of the deposited metal layer wasmonitored by micro balances during the preparationand afterwards determined by a RBS analysis [2].

Low energy p+ with adjustable implantation ener-gies Ep between 2.5 keV and 28.5 keV were implantedinto the samples. TF-//SR measurements were per-formed in a 10 mT magnetic field at a temperature of20 K (to minimize the influence of ^-diffusion). Muonscoming to rest in the metal precess at a frequency z^+of « 136 kHz/mT. If the p+ pass the metal layer andstop in the quartz glass, however, muonium (Mu) for-mation occurs. The Mu precesses at a frequency vMu

of « 13.9MHz/mT. Thus, the fi+ and the Mu preces-sion signals can be easily distinguished. The asymme-tries of the two signals are a direct measure of the frac-tions of fi+ stopped in the metal or in the quartz glass.By increasing Eß the asymmetry of the /i+ signal de-creases indicating that the fraction of //+ stopped in themetal layer decreases. To calibrate the ß+ asymmetryas a function of energy, thick AI and Au metal films (d= 400 nm) where also measured. In these samples allfi.+ come to rest inside the metal layer. These data arecompared with predictions of the ß+ ranges and pt+

range distributions for the different metals obtained byMonte-Carlo computer simulations.

In Fig. 1 the observed / i + asymmetry for the thin AI(upper part) and Au sample (lower part) as a functionof Eß is shown. The solid lines represent the predic-tions for the energy dependence of the fx+ asymmetryobtained by computer simulations using the computercode TRVMC95 [3].

The lower asymmetry found in the Au sample is dueto n+ which are backscattered from the sample andtherefore lost. Even at the highest Eß a diamagneticasymmetry of 5 % is observed. This effect is due to

30100 nm AI on SiCX

40 nm Au on SiCX

0 5 10 15 20 25 30

Implantation Energy E [keV]

Figure 1 : Asymmetry of fi+ stopped in thin metal filmson SiO2 as a function of the implantation energy Eß.The solid lines describe simulated dependencies of thefi+ asymmetry as a function of Eß.

//+ which are stopped in the quartz glass but do notform muonium. //SR measurements using the surface-n+ beam in ?rM3 performed on suprasil glass showedthat the asymmetry of fi+ which do not create muoniumamounts to « 3%. The remaining 2% are backgroundsignal, due to p+ missing the sample, as confirmed bytransport simulations.

REFERENCES

[1] the samples were prepared by M. Horisberger, PSI.

[2] the RBS analysis was done by M. Döbeli, PSI.

[3] W. Eckstein, Computer Simulation of Ion-Solid In-teractions, Springer Verlag Berlin 1991.

92

THE USE OF SLOW MUONS FOR PRECISION MEASUREMENTS OF MAGNETIC FIELDSON A 10 nm SCALE AT A YBa2Cu3O7-* SURFACE

RA-94-15, BIRMINGHAM - PSI - BRAUNSCHWEIG - KONSTANZ

M. Birke1*, E.M. Forgan*, H. Glückler*, A. Hofer*§ T.J. Jackson*, F.J. Litterst*, H. Luetkens**, E. Morenzoni*,Ch. Niedermayer§, M. Pleines^, T. Prokscha*, T.M. Riseman*, A. Schatz*, G. Schatz§, H.P. Weber*

* School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, United Kingdomf Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland| IMNF der Technischen Universität Braunschweig, D-38106 Braunschweig, Germany§ Fakultät für Physik, Universität Konstanz, D-78457 Konstanz, Germany

The slow muon source at PSI [1] represents aunique possibility to investigate magnetic properties ofthin films and surfaces. By accelerating or deceleratingthe muons just above the sample surface, we can tunethe average implantation depth by a factor of ~10, andthereby measure variations of magnetic field over dis-tances of order 15-130nm, corresponding to p+ im-plantation energies between 2.5 and 28.5 keV, respec-tively. We chose the surface of a high Tc superconduc-tor in the Meissner state as a testbed for the technique,and as a means of measuring the properties of the su-perconductor, and comparing Monte-Carlo codes formuon stopping in solids with experiment.

It is well known that superconductors in the Meiss-ner state do not completely exclude magnetic flux, butthat the field penetrates a short distance into the sur-face. In the simplest case, it is expected that an exter-nal field Bo, applied parallel to the surface, penetratesaccording the equation:

B(z) = (1)where z is the depth perpendicular to the surface, andA is the London magnetic penetration depth. Directconfirmation of the functional form of Eqn. 1 is difficultto obtain and is of some interest, since it will not applyin all cases.

The measurements were performed on a c-axis ori-ented, 700 nm thick epitaxial YBa2Cu307_,5 film grownat the TU München (group of H. Kinder) by thermal co-evaporation on a LaAIO3 2 inch substrate. The qual-ity of the films can be considered homogeneous overthe entire film area. A magnetic field of 10mT paral-lel to the sample surface was applied after cooling thesample to 20,K so that no flux lines were trapped in-side, and the field below the sample surface was dueto the finite value of penetration depth. The distributionof precession fields of implanted muons was analysedby a Maximum Entropy procedure. Typical results areshown in Fig. 1.

The average depth of muon implantation was de-termined using a version of TRIM code provided by W.Eckstein. The average field (derived from the muonprecession frequency) at each depth is plotted for sev-eral temperatures in Fig. 2. It can be seen that there isgood agreement with Eqn. 1, and from the results thevariation of penetration depth with temperature may bedetermined. These results are in preparation for publi-cation.

10 12B - field (mT)

Figure 1: Typical field distribution obtained by maxi-mum entropy Fourier analysis of the muon decay spec-tra. This field distribution reflects the muon stoppingdistribution.

eldB

o••£cD)CO

Muonc

100,

80

60

4 0

) 5

• 20«.50^,70• 80

Implantation Energy1C) 20

KKKK

A.[nm] B -135(5)162(5)22-3(8)453(1 o;>

[keV]3 0

'

—w-

\

0 30 60 90 120 150 180

Muon Implantation Depth z [nm]

Figure 2: Field values versus depth of implantation, forvarious values of sample temperature. The solid linesrepresent exponential fits, with the value of A as theonly free parameter.

REFERENCES

[1] E. Morenzoni et al., J. Appl. Phys. 81 (1997) 3340.

93

LOW ENERGY /xSR STUDY OF THE FIELD DISTRIBUTION OF A FLUX LINE LATTICECROSSING A YBa2Cu ;O-_, SURFACE

RA-94-15, KONSTANZ - PSI - BRAUNSCHWEIG - BIRMINGHAM

M. Birke™, E.M. Forgan§, H. Glückler*. A. Hofer*t, j . J . Jackson§, K. Kupfer*, F.J. Litterst*. H. Luetkens™,E. Morenzonit, Ch. Niedermayer*, M. Pleines**, T. Prokschat, T.M. Riseman«, A. Schatz*, G. Schatz*

* Fakultät für Physik, Universität Konstanz, D-78457 Konstanz, Germanyt Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandi IMNF der Technischen Universität Braunschweig, D-38106 Braunschweig, Germany§ School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, United Kingdom

We have used low energy p+ to study the evolu-tion of the magnetic field distribution as the flux linesemerge through the surface of a superconducting film.

The measurements were performed on c-axis ori-ented, 700 nm thick epitaxial YBa2Cu307_,5 films withTc = 87.5 K grown at the TU München (group of H.Kinder) by thermal coevaporation on LaAIO3 (2 inches)substrates [1]. The quality of the films can be consid-ered homogeneous over the entire film area [2]. To as-sess the region outside the superconductor a secondYBa2Cu3O7_,5 film was grown under exactly the sameconditions and covered ex situ with a 70 nm thick Aglayer.

Fig. 1 shows three field distributions, which wereobtained by implanting muons with energies of 28.5 keVand 2.5 keV into the uncovered YBa2Cu3O7_,5 film andwith an energy of 2.5 eV into the Ag covered film. Ac-cording to TRIM simulations [3] muons with these ener-gies come to rest at a mean depth of 130 nm and 15 nmbelow the YBa2Cu307_5 surface in case of the uncov-ered film and stop 55 nm in front of the superconductinglayer within the Ag in case of the Ag covered film. Thedata were taken after cooling the sample in an externalfield of 1.04mT to T = 20 K. For z = 130 nm the fielddistribution shows the typical features characteristic fora 3D flux line lattice, namely a pronounced tail towardshigh fields arising from regions within the vortex latticeclose to the vortex cores, a cusp which correspondsto the most probable field Bsad at the saddle point be-tween adjacent vortices and a cutoff on the low fieldside corresponding to the field minimum at the pointwhich is most remote from the vortex cores. As theflux lines emerge towards and through the surface ofthe YBa2Cu307_i film, the saddle point field moves to-wards the applied field and the overall width of the fielddistribution diminishes. On the outside of the super-conducting film the field distribution is still asymmetricbut the tail towards high fields is drastically reduced.

For a quantitative analysis of this behaviour we plotin Fig. 2 the measured depth dependence of Bext - Bsad

and compare it to the results of calculations for threedifferent values of the London penetration depth \ab

(dashed lines). These calculations simulate the depthdependence of the field distribution by solving the Lon-don equations inside the superconductor, whereas theLaplace equation applies on the outside. For a pen-etration depth of about 150 nm, excellent agreement

0.010 0.015Magnetic Field [T]

-100-50 0 50 100 150 200

depth [nm]

Figure 1 : Left: Measured field distributions at T = 20 Kand Bext = 10.4mT at distances of z = 130 nm and z =15 nm below and z = -55 nm in front of the surface of theYBa2Cu307_,5 film. As the flux lines emerge throughthe surface of the film both the cusp shift and the tailon the high field side are drastically reduced. Right:Depth dependence of Bext-Bsad obtained at T = 20 Kand Bext = 10.4mT. The solid curves are predictionsof the London model for different values of the Lon-don penetration depth. Excellent agreement betweenthe experimental and calculated values is obtained forAa6=150nm.

between the measured and calculated cusp shift is ob-tained. These results clearly demonstrate the potentialof the low energy ^SR technique for future thin film andsurface studies.

REFERENCES

[1] B. Utz et ai, IEEE Transactions on Applied Super-conductivity? (1997) 1272.

[2] P. Berberich et ai, Physica C 219 (1994) 497.

[3] W. Eckstein, Computer Simulation of Ion-Solid In-teractions Springer-Verlag Berlin 1991

94

MAGNETISM OFTHIN CHROMIUM LAYERS STUDIED WITH LOW ENERGY //+SR

RA-94-15, BRAUNSCHWEIG - PSi - BIRMINGHAM - KONSTANZ - KRAKOW

M. Birke', E.M. Forgan*, H. Glückler§, B. Handke11, A. Hofer*, T.J. Jackson*, J. Korecki^, M. Kubik^, F.J. Littersf,H. Luetkens*§, E. Morenzoni§, Ch. Niedermayer*, M. Pleines*5, T. Prokscha§, M. Przybylski11, T.M. Riseman*, A.Schatz*, G. Schatz*, T. Slezak^ and H.R Weber§

* IMNF der Technischen Universität Braunschweig, D-38106 Braunschweig, Germanyt School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, United Kingdomt Universität Konstanz, D-78434 Konstanz, Germany§ Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandf Department of Solid State Physics, Academy of Mining and Metallurgy, Krakow, Poland

Bulk Cr is an antiferromagnet which forms an in-commensurate spin density wave (ISDW) below TN

= 311 K. As mediating spacer layers thin epitaxial Crfilms play an important role in exchange coupled mag-netic superlattices. The dependence of the Cr mag-netism on the layer thickness in Fe/Cr multilayers wereinvestigated with neutron scattering and PAC [1], [2]and [3]. The Néel temperature of the ISDW decreaseswith decreasing Cr thickness. Finally at a critical thick-ness of « 42 Â, which is close to the modulation pe-riod of the SDW, Tjv drops to zero. A still open ques-tion is the magnetism of Cr below this critical thick-ness. Both commensurate antiferromagnetic and para-magnetic phases are suggested. Also for Cr layerslarger than 42 Â and temperatures from TN up to 500 K(which is well above the bulk TN) a commensurateSDW structure has been reported [1].

To study the magnetic properties of thin Cr(001)films two different Cr-multilayers were grown by molec-ular beam epitaxy (MBE) on MgO(001) substrates:A: 340Â AI/100Â Cr/200Â Au/40À Cr/MgO(001)B: 300Â AI/40Â Cr/100Â Au/40Â Cr/200Â Au/40ÂCr/MgO(001). These samples were investigated bylow energy ^+SR (LE-/i+SR [4]), now available atPSI. The muons were implanted into the sample withenergies between 1.5 keV and 28.5 keV. From theseimplantation energies the stopping distribution of themuons in the sample can be calculated using a mod-ified version of the Monte Carlo simulation programTRIM (Transport of Ions in Matter) [5]. These simu-lations are in agreement with range experiments per-formed at the LE-^+SR apparatus in 1998 [6]. ZF andTF /j,+SR measurements were performed at tempera-tures of 20 K to 320 K with external fields of 5 mT and10mT.

ZF experiments using 6.5-keV muons (most of themuons stop in the first Cr layer) show high depolari-sation rates of Ar « ( lO .ö iS .ö ) / ^ " 1 at low tempera-tures.

The TF spectra can be satisfactorily fitted to a sin-gle exponentially damped precession signal. For sam-ple A, Fig. 1 shows the energy dependence of the to-tal asymmetry at 20 K and 320 K. A significant loss ofasymmetry can be seen for muons penetrating the Crlayer. This indicates a magnetically ordered state forboth temperatures. For 6.5-keV muons a temperaturescan has been performed. It shows a strong increase

1 8 -

1 7 -

1 6 -

1b-

14-

13-

12-

1 1 -

10-

9-

8-

7 -

Al

ÏÏ

Ï

I

I

Ï

Cr |

I

I

Au

i

ï

| c r |

o

•

MgO

320 K20 K

I

I5 10 15

Implantation energy (keV)

Figure 1: Total asymmetry in TF measurements onsample A at 20 K and 320 K. A calculated energy toimplantation depth scaling is given at the top of the fig-ure.

of the asymmetry with increasing temperature begin-ning at 270 K, which is approximately the expectedNéel temperature for a 100Â Cr layer. Surprisingly,the total asymmetry is still reduced for muons stoppingin the Au layer. This will be the subject of further sys-tematic investigations.

The 40 Â Cr film (sample B) is found to be magneticup to 320 K, without the increase of the total asymmetryat high temperatures.

REFERENCES

[1] H. Zabel et al., J. Phys. D: Appl. Phys. 31 (1998)656.

[2] E.E. Fullerton etal., Physica B 221 (1996) 370.

[3] J. Meerschaut et al., Phys. Rev. Lett. 75 (1995)1638.

[4] E. Morenzoni, Appl. Magn. Reson. 13 (1997) 219.

[5] W. Eckstein, Computer Simulation of Ion-Solid In-teractions, Springer, Berlin, 1991.

[6] M: Birke etal., this Annex.

95

SLOW MUONS TO STUDY CMR AND SPIN-GLASS FILMS

RA-94-15, ETH ZURICH - LEIDEN - PSI - BRAUNSCHWEIG- BIRMINGHAM- KONSTANZ

A.Schenck*5, F.N.Gygax*, J.Aartsî, A.Amato*, D.Andreica*, M.Birke§, E.Forgan1', H.Glückler*, A. Hofer**,T. Jackson11, J. Litterst5, H. Luetkens§, E. Morenzoni*5, J.A. Mydosh*, Ch. Niedermayer**, M. Pinkpank*, M. Pleines*,T. Prokscha*, T. Riseman11, A. Schatz§, G. Schatz**

* Institute for Particle Physics, ETH Zürich, CH-5232 Villigen PSI, Switzerlandt Kamerlingh Onnes Laboratory, Leiden University, P.O. Box 9504, 2300 RA Leiden, The Netherlands% Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland§ Institut für Nukl. Festkörperphysik, Technische Universität Braunschweig, D-38108 Braunschweigf School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, U.K.** Fakultät für Physik, Universität Konstanz, D-78434 Konstanz, Germany

We have used the slow //+-beam set-up developedat PSI [1] to perform exploratory yuSR measurements inthin films of colossal-magnetoresistance (CMR) mater-ial and AuFe spin glass. The epithermal muons emerg-ing from the nitrogen moderator were re-acceleratedand selected with energy of 6 keV to reach implantationdepths of 35 or 25 nm, respectively, in the two samples.

Lao.7Cao.3Mn03: The temperature dependences ofthe TF-relaxation has been measured in CMR films ofLao.7Cao.3Mn03. The epitaxial film of 60 nm thicknessare sputtered on a SrTiO3 substrate [2]. The data werefitted by an exponential relaxation function - see Fig. 1.

v>

a.o<x<

ÜJQ

<

100 150 200 250 300

100 150 200 250 300TEMPERATURE (K)

Figure 1 : Temperature dependence of the ^SR relax-ation rate A and the signal amplitude - TF 0.01 T.

The relaxation shows a critical behaviour, here ten-tatively fitted with A oc \T - Tc\-

X, indicating a Tc ofabout 110K. This is surprising since the conventionalmeasurements indicate that the films order magneti-cally around 175 K [2]. The amplitude of the observed^SR signal shows a decrease forT < 160 K.

Spin-glass system: A AuFe(3at.%) film of 50 nmthickness has been studied in the temperature range15 - 150 K. The data were fitted by a stretched expo-nential relaxation function, a exp{-(At)ß}. The tem-perature dependences of A and ß are shown in Fig. 2;the /xSR signal shows full and constant amplitude overthe studied T range.

10

g

RE

LAX

AI

ND

EX

ß

101

10°

10"1

101

0.8

0.6

0.4

0.2

10'

%

Ä

*

TEMPERATURE

s

102

I

102

(K)

Figure 2: T dependence of the relaxation A and theexponent ß obtained from a stretched exponential fit tothe TF /iSR spectra (0.01 T) in a AuFe(3at.%) film.

As has become evident in recent years, for bulkspin glasses of various kinds ß approaches 1/3 on ap-proaching the freezing temperature Tg, which is notreached yet at the lowest applied temperature of 15 K.For bulk AwFe(3at.%) Tg ~17K; in a 50 nm film Tg

could be reduced by 80% [3]. Extrapolating ZF-resultson bulk ,4«Fe(1 at.%) [4] one would expect a damp-ing rate of ~ 15-18 ßS~l for T < Tg in a AuFe(3at %)sample, our measurements show ~ 9 / /s " 1 at 15 K.

REFERENCES

[1] E. Morenzoni et al., J. Appl. Phys. 81 (1997) 3340.

[2] J. Aarts et al., Appl. Phys. Lett. 72 (1998) 2975.

[3] J.A. Mydosh, Spin glasses: an experimental intro-duction, Taylor & Francis, London (1993).

[4] Y.J. Uemura et al., Phys. Rev. B 31 (1985) 546.

96

LOW ENERGY n+ BEAM DEVELOPMENTS

PSI - BIRMINGHAM - BRAUNSCHWEIG - KONSTANZ

M. Birke**, E.M. Forgan§, H. Glückler*, A. Hofer*t, T.J. Jackson§, F.J. Litterst*, H. Luetkens**, E. Morenzoni*Ch. Niedermayert, M. Pleines*t, T. Prokscha*, T.M. Riseman§, A. Schatz*, G. Schätzt, H.P. Weber*

* Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerlandt Fakultät für Physik, Universität Konstanz, D-78434 Konstanz, Germanyt IMNF der Technischen Universität Braunschweig, D-38106 Braunschweig, Germany§ School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, United Kingdom

Several improvements of the low energy fi+ beam(LE-jU+, see Fig. 1) have been implemented and suc-cessfully tested during the last year. The improvementstranslate into an increase of flux of the very slow n+ (Eß

~ 10 eV) at the moderator, an optimized transport effi-ciency onto the sample, a smaller beam spot size, andthe extension of the available implantation energies to0 < E^P < 30 keV.

The flux of the epithermal p+ at the source hasbeen optimized by a better transport of the incomingsurface muon beam onto the moderator and by sub-stituting the patterned 380-jum-thick Si substrate by apatterned 125-/jm-thick Ag substrate [C. David et al.,this Annex]. Besides easier handling and a better heatconductivity, the Ag substrate has a higher stop densityfor the surface muons compared to the Si substrate.This causes a 10 - 15% higher stop density in the solidvan der Waals moderator layer (N2) condensed on thesubstrate, giving rise to an equal increase of the ep-ithermal fj,+ flux. Both improvements result in a gain of~ 30%.

The transport efficiency of the low energy fi+ beamonto the sample was increased up to 42% (for themaximum high voltage settings of the transport sys-tem, i.e. +20 kV at the moderator) by optimizing the po-sitions of the "einzel lenses" and by installing a newconical shaped electrode (conical lens, see Fig. 1).The conical lense also yields a small round beam spotsize of < 10mm FWHM (see Fig. 2). This allowsto study small samples with low background in LE-H+SR experiments.

The range of implantation energies was extendedby introducing an acceleration - deceleration section infront of the sample and by floating the sample cryostatbetween -12.5 kV and +12.5 kV. With only two high volt-age settings of the transport system (+15 and +20 kV)all energies between 0 and 30 keV are accessible now.With the LE-/U+ beam setup several new experimentshave been performed. For details, we refer to the cor-responding contributions in this Annex.

In the 7rE3 area the maximum rates obtained withthe improved setup amount to ~ 500/s epithermal p+ atthe moderator and ~ 210/s /i+ hitting the sample at asurface muon rate of ~ 1.7 x 107/s (Ap/p ~ 7% FWHM).The observed p+ decay asymmetry of 25% in a verypure AI sample corresponds to a polarization of ~ 92%.This is in agreement with the measured TTE3 surfacefi+ beam polarization.

MCP1

Spin

Low Energy (0 - 30 keV)muon beam

Gate Valve

Conicallens

V

Heimholte Coils

Sample or MCP2position

Sample Cryostat

Positron counters

Figure 1 : Present setup of the low energy /*+ beam.

•jj 40

- 30

20

10

0

-10

-20

-30

-40 -30 -20 -10 0 10 20 30 40x[mm]

Figure 2: Contour plot of the beam spot of the low en-ergy n+ at the sample position, measured with a posi-tion sensitive MCP detector with delay-line readout.

97

DEVELOPMENT OF A NOVEL TIME-DIFFERENTIAL A*SR DATA-ACQUISITION SYSTEMAND ITS APPLICATION TO RADIO-FREQUENCY //SR

RA-97-26, STUTTGART-PSI

O. Kormannt, J. Major**5, I. Reid*, A. Rockt, M. Schefzikt, D. Herlach*

* Universitat Stuttgart, InstitutfurTheoretische und Angewandte Physik, D-70550 Stuttgart, GermanyI Max-Planck-lnstitutfur Metallforschung, Heisenbergstr. 1, D-70569 Stuttgart, Germanyj Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland

The data acquisition hardware [1], based on the Pi-cosecond Time Analyzer (EG&G Ortec model 9308)and the data-acquisition software Gustav [2], was fur-ther developed and improved [3]. Long-time mea-surements with the data-acquisition system also innegative-muon experiments (see RA-97-25) showedreliability and easy-to-use features.

A new radio-frequency (rf) control unit, which de-tects electric-field break-downs, was developed for theRF/iSR spectrometer. The rf control unit consists of ademodulator, which rectifies the rf signal from the de-tection coil, and a fast 8-bit transient-recorder card inthe data-acquisition computer, which samples the rec-tified signal.

Hitherto, RF/iSR experiments on isotrop ("normal")muonium (MUT) were performed at rf frequencies (uT{)of 120 MHz to 180 MHz [4]. In this frequency range itis possible to resolve the various rf transitions whichare characteristic for MuT. However, experiments inthe frequency range 20 MHz to 80 MHz demonstratedthat various muonium species can be better detectedthan at the higher frequencies. Around 20 MHz threeMUT transitions contribute to a single resonance line(Fig. 1a) and around 80 MHz the detection of anom-alous (anisotrop) muonium (MUBC) becomes feasible

0.040-

0.030-

0.020-

0.010-

0.000-

0.020-

0.015-

0.010-

0.005-

0.000-

M U B C

I I

&

I

I

II

(a)

I(b)

0 2 4 6 8

5[mT]

Figure 1: RFjuSR resonance lines of MuT and MuBcobtained on Si:B-CZ: (a) z/rf = 20.45 MHz, rf power Prf

= 50 W; (b) i^rf = 80.93 MHz, Pvi = 15W.

0.015

0.010-

0.005 -

0.000 -

100 150T[K]

Figure 2: Temperature dependence of the on-reso-nance rf asymmetry of MUT and MUBC obtained onSi:B-CZ; vrf = 81.1 MHz, Prf = 15W.

(Fig. 1b). Species that are formed with delay may notbe observable in TFjuSR experiments (e.g. MuT in Si:B-CZ) [5,6], whereas in RF^SR the MuT signal is re-vealed (Fig. 2).

This work is supported by the BundesministeriumfurBildung und Forschung, Bonn, Germany under con-tract no. 03-SE4ST2.

REFERENCES

[1] J. Schmidl, Dr. rer. nat. Thesis, Universitat Stutt-gart, 1997, (ISBN 3-89712-039-9)

[2] http://musr.mpi-stuttgart.mpg.de/www/musr/gustav/gustav. htm

[3] O. Kormann, J. Major, I.D. Reid, A. Rock,M. Schefzik, R. Scheuermann, J. Schmidl, D. Her-lach: PSI Annual Report 1997, Annex I, p. 87

[4] S.R. Kreitzman, B. Hitti, R.L Lichti, T.L Estle, andK.H. Chow: Phys. Rev. B 51 (1995) 13117

[5] R. Scheuermann, Dr. rer. nat. Thesis, UniversitatStuttgart, 1997 (ISBN 3-89588-979-2)

[6] B.D. Patterson, Rev. Mod. Phys. 60 (1988) 69.

NEXTleft BLAMK

99

M ic ro a n dNano T e c h n o l o g y

100

REPLICATION OF NANOSTRUCTURES WITH POLYMER INJECTION MOLDING

H. Schift, D. Bachle, C. David, J. Gobrecht, R.W. Jaszewski, T. Mezzacasa, L. Scandella (PSI),S. Koppel, W. Kaiser (FHAargau, Windisch); M. Gabriel, M. Wey (AWM Werkzeugbau AG, Muri)

Polymer nanostructures have been fabricated using a thermal injection molding process. The master struc-ture was fabricated with e-beam lithography and reactive ion etching, and was replicated in polycarbonatediscs. A replication fidelity of 25 nm was achieved, which is an order of magnitude below the structuring ofadvanced Compact Disc fabrication processes.

Polymer replication offers the possibility to fabricatehigh precision micro-optical elements at low cost andhigh throughput. Once a master structure is gener-ated, it can be used as a mold insert and identicalstructures can be replicated by thermal injectionmolding. The most prominent example is the CompactDisc (CD) fabrication, in which discs with sub-micronsurface structures are replicated with cycle times ofless than 4 sec. The recent development of the DigitalVersatile Disc (DVD) involves the structuring of400 nm wide and 100 nm deep features over the sur-face of a 0.6 mm thick and 120 mm diameter disc.This is the result of the continuous optimization ofmaterials, machine and process technology.

on a NETSTAL Discjet 600 machine using a polycar-bonat (PC) Makrolon CD 2005 from BAYER. Duringthe injection process, the mold cavity and the masterare filled with the polymer melt, which solidifies in thecavity. Then the cavity is opened and the polymerparts are demolded automatically. Best replicationfidelity was achieved using mold temperatures slightlybelow the glass transition point. It was possible to fillgrating structures down to a width of 25 nm.

A486-,243-1

r2

•r «* >• » I

» « «• rti H* \

* " *« 1* <w •„ ^

' S ^ , »* • - • •#. « \

| Njj, « ^« .» r

•- *;

• . . : . i ;

Currently the limits of miniaturization are determinedmore by the fabrication of the master and the signalread-out than by the replication process itself. It hasbeen known for a long time that polymers exhibitresolutions below 10 nm, and recently it has beenshown that thermoplasts can be structured in the nm-range by hot embossing processes. Our aim was tosee if these limits can also be achieved with modeminjection molding processes.

Master structures were fabricated by e-beam exposureof thin PMMA films. The interferometric stage of ourLEICA LION LV1 e-beam machine makes it possibleto achieve a lateral positioning accuracy of 30 nm. Thedeveloped resist pattern was subsequently transferredinto the underlying SiO2 substrate using reactive ionetching (see SEM micrograph). Structures with depthsdown to 60 nm and aspect ratios up to 2:1 wereobtained. The master was then inserted into a CD-molding tool of AWM. Injection molding was performed

• v.

o W o

o:

Characterization of the surface quality of the structureswas done using scanning electron (SEM) and force(AFM, see micrograph) microscopies. It can be shownthat the polymer exactly conforms to the surface of themaster structure.

The combination of high precision e-beam writing ofthe master structure and the well defined replicationconditions of CD injection molding makes it possible toeasily fabricate hundreds of identical nanostructuredelements. The potential of this technique for the fabri-cation of high density storage devices is obvious.Nanogratings with sub-|im periods can also be usedas simple calibration standards for scanning probemicroscopies, where reference patterns are neededfor the calibration of the piezo scanners.

FUNDING: SNF (NFP36), PSI, FHA, AWM.

REFERENCE

R.W. Jaszewski, H. Schift, J. Gobrecht, P. Smith, Mi-croelectronic Eng. 41/42, 575-578 (1998).

101

NANOREPLICATION IN THIN POLYMER FILMS

Ft.W. Jaszewski, D. Aliiata, J. Gobrecht, T. Mezzacasa, H. Schift (PSI)

Thin polymer films used for lithographic applications can be structured in the nanometer range with moldingprocesses. In order to explore the limits of replication, we fabricated master structures and used them in ahot embossing process. It was possible to obtain nanostructures in polymers with feature sizes down to10 nm.

The excellent replication fidelity obtained with poly-mers makes it possible to use them in a variety oflithographic applications where a resist has to bestructured. The processed polymer can then be usedas a mask in a subsequent pattern transfer process ordirectly as a functional surface relief structure. Largesurfaces can be patterned with nanostructures in ahighly parallel manner, and the main limitations oflithography using conventional exposure and de-velopment can be avoided.

:*' . : . '• .: i I: - \-' ". . " ;v- ",

Hot embossing uses the difference in the thermome-chanical properties between a hard stamp and a ther-moplast during temporary intimate physical contact.First, the polymer is heated beyond its glass transitiontemperature, and it assumes a viscous state. Thisallows it to be shaped under pressure by imprintingwith a hard master. During the embossing, the viscouspolymer flows into the cavities of the mold. Once thepolymer has exactly conformed to the shape of thestamp, it is hardened by cooling and then the master isdemolded.

In order to advance our knowledge of this technique, itis important to understand the behaviour of polymerswhen they are structured in the nanometer range. Thelower limit to which a polymer can be structured hasbeen shown to be at least 10 nm; therefore, one needsembossing stamps featuring such small structures. Weused pieces of a (100) silicon wafer consisting ofnanogrooves obtained by anisotropical wet etching(see SEM micrograph). This lithographic processyields grooves with very smooth sidewalls and a verysharp V-shaped bottom. Thus, both the volume andthe form of the cavity can be precisely determined.The end radius of the groove is estimated to besmaller than 5 nm. These features have been em-bossed into a 2 |im thick film of low molecular weight

(Mw=75k) PMMA at a temperature of 160°C and at apressure of 10 MPa. In order to prevent air from beingtrapped in the lower part of the grooves, the emboss-ing was performed under vacuum. Scanning electronand atomic force micrographs of the polymer struc-tures show that the polymer precisely conformed to themaster (see AFM image and profile plot below). TheAFM analysis was performed in the Intermittent Con-tact (IC) mode, thus minimizing the danger of a de-structive interaction between the tip and the polymer.The radius at the top of the polymer ridge was found tobe smaller than 10 nm. This is on the order of the sizeof the AFM tip radius, which means that the actualradius of the ridge may be even smaller.

'; '••• r:

: ' r ; . '1

; : - - ' :

mm

s||i;;::;;;:i

-

: /

: J

/ \\

1

Jim

jam0.92-

0.46

0.00

2- \X,, V

These results lead us to think that the limit of replica-tion in polymers is below 10 nm. Special efforts haveto go into the fabrication of a high quality nanomaster,the optimization of the polymer properties, and theimplementation of suitable sample characterizationtechniques.

FUNDING: SNF (NFP36), PSI.

REFERENCE

R.W. Jaszewski, H. Schift, J. Gobrecht, P. Smith,Microelectronic Eng. 41/42, 575-578 (1998).

102

MANUFACTURE OF MODERATOR GRATINGS FOR EPITHERMAL MUONS

C. David, L Scandella, T. Mezzacasa, T. Prokscha, A. Hofer, E. Morenzoni, H. Gluckler (PSI)

In order to improve the yield of epithermal muons generated by moderation of surface muons, silicon and silvermoderator gratings have been manufactured using lithography, anisotropic wet etching, and replication bymetal embossing. A comparison between different moderators shows, that microstructured Ag moderatorshave the highest efficiency and the best handling properties.

The low energy muon beam (LEM) for jxSRinvestigations on thin films and multi-layers isgenerated by moderation of polarised surface muons(4MeV) in a v.d.Waals gas layer (e.g. Ar, N2)condensed on a cold substrate (T=10K). The exitingepithermal muons with energies of about 10eV areused as the source of the LEM beam [1].

-4 MeV

Cold Substrate

epithermal

fast H

Condensed van der Waalsgas layer ( 200 - 300 nm)

Fig. 1: Principle of muon moderation.

The effective surface of the v.d. Waals gas layer canbe increased by patterning the moderator surface. Weproduced line patterns of approx. 30|j.m pitch on380jxm thick S/<100> wafers using photolithography.By an anisotropic wet etching process in a KOHsolution we generated V-shaped grooves defined bythe <111> crystal planes with an angle of 54° to thewafer surface (see fig. 2).Flux measurements with several types of moderatorsshowed, that a gain in flux of epithermal muons ofapprox. 70% can be obtained when using Si gratingscompared to flat Si moderators. This corresponds wellto the geometric gain in surface by the V-groovepattern. However, the Si moderators turned out to bedifficult to handle when cooled down in a He cryostat.Firstly, Si has a poor heat conductivity and secondly,the moderators tend to crack during cooling whenclamped tightly to its holder for good thermal contact.

To circumvent these problems, we developed themethod shown in fig 2. The V-grooves of a Si gratingare replicated into 125p.m thick foils of high purity,annealed silver by embossing it with forces up to 106kgwhich corresponds to 10 tons/cm2. Surprisingly, themasters made of 2.8 mm thick Si wafers withstood thisforce enabling us to fabricate larger numbers of 3cm x

<100> Si wafer with )200nmSi3N4

e-beam or photo-lithography

Cr lift-off

Reactive ion etchingofSi3N4

KOH wet etch

Si3N4 removal in HF J

Embossing into0.12 mm Ag foil

replicated Agmoderator .*

Fabricationof Si master

I Replicationf by embossing

Fig. 2: Steps for fabrication of an Ag moderator grating

3cm Ag replicas. Figure 3 shows SEM micrographs ofthe Si master structures and the replicated Aggratings. Although the ragged ridges indicate that theAg has not been completely pressed into the Sigrooves, this does not reduce the moderationefficiency. Instead, the Ag moderators yield another10-15% increase in muon flux which can be attributedto the higher stop density in Ag compared to Si.

• • V *

Fig. 3: SEM micrographs of V-grooves in a Si master(left) and an Ag replica made by embossing(right).

FUNDING: PSI.

REFERENCE

[1] See the Annual Report of the Laboratory of MuonSpin Spectroscopy.

103

DIFFRACTIVE OPTICAL ELEMENTS GENERATED BY E-BEAM LITHOGRAPHY

C. David, P. Haberling (PSI), J. Sochtig (CSEM, Zurich), H.J. Heimbeck (FISBA OptikAG, St. Gallen),G. Scherer (3D Ltd., Unterageri) ) , S. Schoch (Baumer electric AG, Frauenfeld), L. Schweri (Oerlikon

Contraves, Zurich), L. Stauffer (Leica AG, Heerbrugg), W. ft Tompkin (OVD Kinegram, Zug)

Various diffractive optical elements have been generated for industrial partners participating in the KTIproject "Swiss e-beam generated optical elements and moulds (SEGOEM)". The project gives industryaccess to the microstructuring facilities at LMN for the development of prototypes and the fabrication ofmasters for production by replication techniques.

The electron beam lithography system LION installedin the LMN clean room has some unique features thatmake it especially suited for the generation of diffrac-tive optical devices. Besides its high resolution (betterthan 50 nm feature size) it is capable of drawing largenumbers of straight or curved lines across long dis-tances without the need of field stitching. This isachieved by moving the sample continuously duringexposure instead of deflecting the beam. Furthermore,the line width can be set very accurately by defineddefocusing of the beam.

Fig. 1: Curved line structures generated by e-beam lithography.

Within the SEGOEM project funded by the KTI, thistechnology is available to the six industrial partners forthe development of optical devices. Among the appli-cations are optical encoders, security features, Dam-mann-gratings, and anti-reflective gratings.As the exposure times for such devices are typically inthe order of many hours, e-beam technology is notsuited for mass production. It can, however, provideprototypes for research and development or mastersfor replication techniques.The complicated technological steps require a closeco-operation of all project partners: The generation ofexposure data, the e-beam exposures and parts of thepattern transfer are carried through at LMN, while thedesign, numerical calculations, the replication andcharacterisation of the devices is performed mainly atthe CSEM-Z laboratories in close contact with theindustrial partners.As an example for such a device, the structures of anangular encoder are shown. The diameter is about

i.:'.

Fig. 2: Structures of a master for the productionof an angular encoder.

11mm. The design consists of circular tracks with 512sectors where each sector is filled with 720 nm wideradial lines. The difficulty to realize this device withconventional e-beam tools lies in the fact, that all45.000 lines have a different angle, and that a dutycycle of close to 0.5 has to be maintained for eachangle. The e-beam generated PMMA master is repli-cated by galvanoforming and injection moulding formass production of an encoder with 2048 increments.

FUNDING: KTI (SEGOEM), PSI.

104

NANOSTRUCTURED ELECTRODES FOR ELECTROCHEMICAL INVESTIGATIONS

C.R. Musil, D. Bachle (PSI), H. Siegenthaler (Univ. Bern)

Electrochemical reactions are greatly influenced by the surface morphology of the electrode material in themicro and nanometer range. Most "real world" instances of electrochemistry occur at highly heterogeneouselectrodes, where an understanding of the mechanisms involved demands high resolution analysis. In thisregard, nanostructured electrodes provide a powerful approach by which the processes that occur duringelectrochemistry can be investigated in a systematic and controlled way.

When investigating electrochemical processes byscanning probe microscopy, electrode material nano-structured into electrically active and inactive sectionsprovides the surface contrast necessary for an analy-sis of the evolution of the reaction products. A particu-larly simple method to fabricate such structures con-sists of defining openings in a PMMA resist layer ontop of a conductive film by e-beam lithography. Inmany electrolytes, PMMA acts as an inert and stableinsulating layer, thus avoiding a pattern transfer step.

In connection with the investigation of electrochemicalintercalation of oxygen in MBE-grown La2CuO4 thinfilms, substrates containing such films were coveredwith a 100 nm thick layer of PMMA. Alignment featureswere exposed in the PMMA resist by e-beam writing,developed, and transferred into the La2Cu04 film byetching in 0.5 mM H2SO4. After stripping the unex-posed PMMA mask in acetone, another layer ofPMMA was applied and a pattern consisting of 5 x 5repetitions of a basic cell containing squares of size 3jxm to 100 nm and lines of width 200 nm and variablespacing was written in the PMMA resist after aligningthe second exposure to the marks previously defined.The customized electrode assembly is depicted in theoptical image below. The time evolution and stability ofthe oxygen intercalation at the openings in PMMAcould be monitored at each individual nanoscale elec-trode by aligning the scanning probe microscope to thetopographically defined etched registration marks.

In a related experiment concerning the electropoly-merization of pOPh film at nanostructured electrodes,p-doped silicon wafers were evaporated on the frontand back sides with 20 nm of titanium followed by 100nm of gold and sawn into octagons of width 7.5 mm.As in the La2Cu04 investigations, the gold surface wascovered with PMMA and openings consisting of linesand squares were defined by e-beam exposure anddevelopment, thereby creating a nanoscale pattern ofgold electrodes. The dual-sided metal deposition pro-vided a convenient low impedance (ca. 100 Q) con-nection between the back side and the nanostructuredsurface, and the geometry of the substrate was opti-mized in order for the sample to fit within the electro-chemical cell (see photograph below). In situ analysisof the film during electropolymerization revealed en-hanced nucleation and growth of the polymer alongthe sides and neighboring surface of the PMMA layer.

The technique of e-beam lithography along with pat-tern transfer by lift-off can also be applied to simulatethe surface structure of non-ideal electrodes. In thiscase, a thin metallic layer is deposited on an insulatingsubstrate and structured to imitate the natural topo-graphy and crystal structure found in macroscopicelectrodes, which controls the early stages of electro-chemical metal deposition.

FUNDING: SPP MINAST, PSI.

105

NANOSTRUCTURED SURFACES FOR BIOMOLECULAR IMMOBILIZATION

C.R. Musil, D. Bachle (PSI), M. Pawlak (Novartis), C. Duschl (EPFL)

The selective and spatially defined immobilization of individual biomolecules has great potential in thedevelopment of bioanalytic methods. Nanostructuring represents a key technology in this area through thedefinition of surfaces which provides the desired physical and chemical attributes on a scale commensuratewith the size of biomolecules. In this context, openings in gold film on glass and patterned self- assembledmonolayers on gold down to 50 nm in lateral size have been achieved by e-beam lithography.

Through the contrast of material properties, nano-structures can provide a tailored environment of sur-face chemistry and as well as topographic relief. Thecontrolled interaction of specific biomolecules withthese surroundings constitutes a form of "moleculararchitecture". In the following, two examples are givenof specific structures which utilize this principle.

Samples exhibiting the orthogonal surface chemistry ofgold (thiol) and glass (silane) and steric constraintswere prepared by e-beam lithography with patterntransfer by argon ion etching. Glass wafers 4 inches indiameter were covered with a 30 nm thick film of goldfollowed by a 100 nm thick layer of PMMA resist. Afterexposure by e-beam writing and development, thepattern in the resist was transferred through the goldlayer by dry etching in argon at a pressure of 75 mtorr,a RF power of 26 W, and a self bias of -310 V. Despitethe poor etch selectivity (the PMMA etching at a ratefive times that of gold), fine feature definition was stillpossible through the careful optimization of the PMMAexposure dose and of the gold etching time. Aftersubjecting the etched samples to an oxygen plasma inorder to remove any residual organic contaminantsfrom the surface, the structures were analyzed by opti-cal, scanning electron, and scanning force microsco-pies. From these observations, it was determined thatopenings 20 nm in depth and down to 50 nm in linewidth were achieved through the gold layer on glass(see scanning electron micrograph below).

50 nm

In an extension of nanostructured surfaces with spe-cific surface chemistry, the feasibility of laterally struc-turing a self-assembled monolayer (SAM) using e-beam lithography was demonstrated. A process iden-tical to that above was employed up to and includingthe development step. The samples were then ashedin an oxygen plasma for 30 seconds and immediatelyafterwards submerged in a 1 mM solution of octade-cane thiol (ODT) in heptane for 18 hours, during whichtime a SAM formed on the areas uncovered by e-beam exposure. The remaining unexposed PMMAwas removed by soaking the samples in acetone. Theselective deposition of the SAM on gold was verifiedby using the ODT monolayer as a mask against a fer-ricyanide gold etch. Dot features down to 50 nm indiameter were achieved by this process, as indicatedin scanning electron micrograph below. Despite theirdisparate chemistry, the patterning of PMMA resist bye-beam lithography was found compatible with theformation of an alkanethiol SAM on the gold surfaceuncovered. Furthermore, the SAM was retained afterremoving the PMMA in acetone and of sufficient qual-ity to protect the underlying Au from a ferricyanideetch.

These nanostructured surfaces offer functional mediafor subsequent biomolecular immobilization and analy-sis by optical or amperometric detection.

FUNDING: SPP MINAST, PSI.

106

THERMOGRAVIMETRIC ANALYSIS OF MICRO CRYSTALS : EXPERIMENT ANDFINITE ELEMENT SIMULATIONS

J.-H. Fabian, J. Gobrecht, T. Mezzacasa, L Scandella (PSI), R. Berger, H.P. Lang, Ch. Gerber,J.K. Gimzewski (IBM), E. Meyer (Univ. Basel)

A new method for thermogravimetric analysis of nanogram quantities of materials has been developed. Forthat purpose micro crystals were attached on a heatable cantilever-type microbalance. Finite-elementsimulations of the miniaturized chemical analytical system were performed to determine the mass of themicro-crystal and the mass change by comparing the experiments.

Thermal gravimetric analysis is routinely used in ana-lytical chemistry and physics to characterize sub-stances and composite materials. Standard thermo-gravimetry (TG) requires material quantities in themilligram range. Consequently, analysis of micro-crystals having a mass of only a few hundred ofnanograms requires tens of thousands of crystals.We have developed a new micromechanical thermo-gravimetric (MMTG) technique to investigate singlemicro-crystals of nanogram sample mass [1]. Bothnew experimental measurements of individual singlezeolite crystals (ZSM-5) as well as finite-element (FE)simulations of the miniaturized chemical analyticalsystem were performed. The method consists of acantilever-type microbalance, which is used to heatand weigh nanogram quantities.

A piezoresistive cantilever is mounted on a piezoelec-tric actuator and vibrated at its first resonance fre-quency. The resonance frequency can be monitoredby measuring the change of resistivity of the piezore-sistive layer upon cantilever bending. This is achievedusing a Wheatstone bridge electrical circuit. For ourMMTG applications, the applied bridge voltage is in-creased, resulting in an ohmic heating of the microme-chanical piezoresistive layer. Temperatures up to sev-eral hundreds °C can be obtained. The mechanicalproperties as well as the heating and temperaturedistribution of such a device are carefully simulated byFE calculations. These simulations allow the sensor tobe calibrated accurately.

To demonstrate the high potential of the new tech-nique, MMTG analysis of the desorption behavior of p-nitroaniline dye molecules (PNA) emanating from asingle zeolite crystal were performed [2]. The figuresabove show an optical micrograph and a mesh sizedistribution used for FE simulations of a ZSM-5-typecrystal attached to the apex of a cantilever. Thedesorption of PNA molecules from the zeolite wasobserved to occur at ~ 215°C and caused a weightloss of = 10.5 ng. The absolute sample mass andmass change are determined by comparing the meas-ured and the FE-simulated resonance frequency.

- 100

50 100 150 200

temperature (°C)250 300

- 92

350

Furthermore, FE calculations help to optimize the de-sign, in term of response time and resolution, for mi-cromechanical thermogravimetric sensor applications.

FUNDING: PSI, MINAST (NOSE).

REFERENCES[1] R. Berger et al., Chem. Phys. Lett. 294, 363 (1998).[2] J.-H. Fabian et al., in Micro Total Analysis Systems

'98, eds. D.J. Harrison and A. van den Berg, Klu-wer Academic Publishers 1998, p. 117.

107

HIGHLY EFFICIENT SOLAR CELLS ON LOW COST SHEET RIBBON GROWNSILICON

H. Elgamel, J. Gobrecht, D. Bachle, T. Mezzacasa (PSI), J. Hanoka (Evergreen Solar)

The aim of this project was to clarify the potential of low cost polycrystalline silicon substrate material forthe fabrication of photovoltaic solar cells with increased efficiency by applying high-quality cell processing.We applied the PSI high-efficiency solar cell process on this material and introduced several processvariations which account for the polycrystalline nature of the material. A cell efficiency above 15% underAM 1.5 standard conditions has been achieved.

In spite of many other materials under investigation,silicon is considered the material of choice for largescale photovoltaic energy conversion because of itsabundance, stability, low toxicity etc. A large fraction ofthe cost of a silicon solar cell originates from the costof the single- or polycrystalline substrate because it isnecessary to saw the wafers or slices from largerpieces of bulk silicon. This process also increases the"grey energy" i.e. the energy necessary to produce thecell.

Evergreen Solar of Waltham, USA has developed aproprietary process which allows to pull thin, poly-crystalline Si ribbons directly from the melt. Theseribbons can be used for solar cell processing withoutfurther treatment. A low cost cell manufacturing proc-ess also developed by Evergreen leads to cell efficien-cies around 10%. This value may be too low in futuresince even cheaper Si thin film cells are expected toachieve these conversion efficiencies soon.

PSI's task in this project was to show that the SR ma-terial has a potential for achieving much higher effi-ciencies if a careful and clean cell processing isapplied. In an earlier research project on Si solar cellsPSI had developed a process [1] which led on singlecrystalline float-zone substrates to solar cells achievingup to 21.2% conversion efficiency under air-mass (AM)1.5 standard conditions, confirmed by the Institut furSolare Energiesysteme, Freiburg.

Algetter

no

no

yes

H-pass

no

yes

yes

I .

B:31.3

A: 30.7

B: 33.0

A: 32.6

B: 33.8

A: 34.8

B:552

A: 545

B:569

A:565

B:581

A: 584

FF (%)

B:71.5

A: 70.3

B: 74.1

A: 71.9

B: 77.4

A: 73.0

11 (%)

B: 12.4

A: 11.7

B: 13.9

A: 13.3

B: 15.2

A: 14.8

l^ is given in mA/cm2, V^ in mV. B= best cell, A= aver-age of cells in lot. In the table results are given for avariety of cells, each processed differently.Generally the processing was done under standardCMOS conditions i.e. class 10 clean room environ-ment etc. Particularly, three steps were studied inmore detail:

• hydrogen passivation of the solar cell structure• Al gettering• antireflection (AR) coating.

The effect of plasma hydrogen passivation on the mi-nority carrier diffusion length is given below.

Because of a significant scatter in the structure of thestarting material the effect of plasma passivation ismore pronounced in some cases than in others.

Wafer

SR1

SR2

SR3

Ln before

75.2±6.3 u.m

95±13.1 u.m

97.3+4 u.m

after plasma passiv.

106.1 ±5.7 urn

141.5+2.7 |im

171.6±7.2u.m

Optimum AR coating for the solar spectrum wasachieved by evaporating a double layer consisting of43 nm ZnS and 103 nm MgF2. Al-gettering was donesimultaneously with the phosphorous emitter diffusion.

The output IV-curve under standard AM1.5 illuminationconditions of the best cell (15.2%) is given below. Thecell's area is 4 cm2. An efficiency above 16% seemsrealistic if substrate material with an optimum dopinglevel can be used.

FUNDING: PSI, Swiss Federal Office of Energy; EWZZurich; alpha real ag, Zurich; Evergreen Solar, USA.

REFERENCE

[1] H. Kiess et al., Proc. 12th Europ. PV Conf., Am-sterdam (1994), pp 763.

108

EFFECTS OF SPONTANEOUS ORDERING ON THE ELASTIC PROPERTIES OFGalnP2

S. Gehrsitz, H. Sigg (PSI), C. Geng (Univ. Stuttgart), F.K. Reinhart (EPFL)

Near infrared Brillouin scattering is used for a precise investigation of the elastic properties of GalnP2

epitaxial layers with different degrees of ordering. Measurements of the angular dispersion of the soundvelocity reveal a cubic symmetry for the disordered crystal and a trigonal one for the single-variantordered structure. This provides the first experimental evidence for an ordering induced lowering of theelastic symmetry in semiconductor compounds. Ordering induced changes of the trigonal elasticconstants are attributed to a relaxation of internal strains along the ordering direction.

Spontaneous long-range ordering has been found inseveral ternary Ill-V semiconductor compounds. Inparticular the GalnP2 system is an attractive materialfor the experimental investigation of ordering effects,since it can be grown lattice matched on GaAs sub-strates with a high degree of ordering, when grownunder specific conditions by metalorganic vapor phaseepitaxy (MOVPE). Ordering of the cation (group III)sublattice occurs along the [111]B directions (CuPtgtype), and results in {111}B monolayer superlattices ofalternating Ga-rich and In-rich planes. The degree ofordering depends on the growth temperature and sub-strate misorientation and is characterized by the or-dering parameter T\ (0<n<l). The Brillouin light scat-tering experiments where performed on a disorderd(crystal structure of the ZnS type) and an orderedGalnP2 sample with T\= 0 and 0.51 respectively.

ZnS CuPtp

[OOlJf[100]

By making use of various scattering geometries and byscanning the angle between the crystallographic direc-tions and the incident light, we systematically analyzethe angular dispersion of the sound velocity, which isrelated to the symmetry properties of the crystals. Tocheck the cubic symmetry of the disordered (r[=0)sample, we first determine the complete set of cubicelastic constants (C 0 ^ , C c i 2 and Cc

44) and comparedthe as defined angular dispersion of the sound velocitywith the one observed experimentally. All the datapoints are found to be in good agreement with thecalculated dispersion, which confirms the assumedcubic symmetry. For the comparison in the [001] -[110] cubic plane, see Figure.

1S0-

150-

a.P

110-10 20 se in so no

0 {degree)70 80 98

The same procedure is performed for the orderedsample. Perfect agreement between ordered and dis-ordered GalnP2 is found for the angular dispersionnear the cubic [001] crystal axis. But a clear discrep-ancy is found for phonons propagating along the or-dering direction [111]. For this direction the squaredsound velocity (pv2, p is the crystal density) is ap-proximately 1.5 % higher for the LA phonon than thatpredicted by the cubic dispersion curve. Although theobserved deviations are small, they demonstrate thereduced elastic symmetry of the partially orderedcrystal. We could show that the angular dispersion isconsistent with the trigonal symmetry with point groupC3v, expected for CuPts type ordered crystals. For thefull description of the elastic properties six independentelastic constants had to be determined.

To compare these trigonal elastic constants of theordered compound with those of the cubic (disordered)crystal, we transform the cubic stiffness tensor to thetrigonal coordinate system. In this way we can showthat the trigonal elastic constants of ordered anddisordered GalnP2 have very similar values, except forC ^ and C\3. Since Cl

33 and C1^ are closely relatedto trigonal shear strains producing internal strains, weattribute the ordering induced change of the trigonalelastic constants to relaxation of internal strain alongthe ordering direction.

FUNDING: SNF and PSI.

109

HIGH FREQUENCY OPTICAL RECTIFICATION IN BULK GaAs

S. Graf, H.C. Sigg, H. Siegwart (PSI), W. Bachtold (ETHZ)

We have investigated high frequency optical rectification in bulk GaAs. The traveling wave scheme allowsan efficient coupling of the optical rectification induced signal of an optical pulse propagating in the GaAsslab. The functionality in the far infrared has been demonstrated with short (< 6 ps) pulses of the freeelectron laser FELIX.

We present the first realization of a traveling wavedetector based on optical rectification in bulk GaAs.Optical rectification (OR) is the generation of a quasi-static polarization co-propagating with the opticalexcitation. The created polarization pulse in the trans-mission line is converted into an electrical microwave.With the traveling wave principle, the bandwidth ver-sus sensitivity tradeoffs of classical detectors can becircumvented [1]. This means that the pulse length isnot limited by the interaction length with the medium.Although low in sensitivity the detector has theadvantage of it's simple structure, is working at roomtemperature and should be useful in the large wave-length range from 0.8 to15jxm.

microstrip

50£2 coax

semi- insulatingGaAs

• < . . - -

[001][110]

[110]

= 4 JJJ (Tpuis < 6 ps)

The light coupling in the GaAs detector is from the[110]-cleaved edge. The microstrip structure is formedby gold layers evaporated on the top and the bottomside of the semi-insulating substrate material. Intensefar infrared light pulses are provided by the free elec-tron laser of the FOM institute for plasma physics(FELIX). Pulse transients and radiation wavelengthused are in the range of 1-5ps and 8-15|j.m, respec-tively.

The measured OR signal obtained by exciting theGaAs slab with light pulses of 4 p.J shows the temporalsignature predicted by the equivalent circuit model ofthe transmission line. The signal shape closely followsthe time derivative (TD) of a Gaussian pulse with awidth corresponding to the 36 GHz bandwidth of thesampling scope. The measured pulse width is in ourcase not affected by velocity mismatch.

[001] [111] [110]

-30 0 30 60Angle of Polarization (deg)

The angular dependence of the signal on the directionof the incident, linearly polarized radiation field followsalmost exactly the theoretical amplitude of the ORinduced polarization component along [001]. Since[001] is the E field direction of the fundamental trans-mission mode HEo, we conclude that the appliedmode-coupling concept is well understood. Whenscaled to the optical pulse length of 5ps signal maximaof 0.2V were measured with pulses of 2uJ energy.This value is in good agreement with the theoreticalestimation of 0.8V.

Asymmetric GaAs/AIGaAs quantum well structureswhich are known to exhibit large nonlinear opticaleffects are promising candidates to improve the per-formance of detectors based on OR. Speciallydesigned structures achieve more than six orders ofmagnitude higher OR coefficients compared to bulkGaAs values.

FUNDING: PSI, SNF.

REFERENCE[1] M. Kwakernaak, H. Sigg, D. Erni, New Directions

in Terahertz Technology, NATO ASI Series E, Vol.334(1997).

25 50 75Time (ps)

25 50 75

Time (ps)

100

110

ATOMIC STRUCTURE OF THE CARBON-INDUCED c(4 x 4) RECONSTRUCTIONON Si(001)

O. Leifeld, D. Grutzmacher (PSI), E. Kaxiras (Harvard University), K. Kern (EPFL)

The initial stages of carbon alloying into the Si(001) surface have been studied by scanning tunnelingmicroscopy (STM) and density functional theory (DFT). To explain the experimentally observed c(4 x 4)surface reconstruction, we propose a novel model that involves pairing of Si dimers mediated by the pres-ence of a complex of a C dimer and four nearest neighbour subsurface C atoms. The model is backed bytotal energy and thermal stability simulations. Its calculated surface charge density agrees well with thefilled state STM images.

Epitaxial alloys of column IV elements grown on Sivastly extend the range of electronic properties thatcan be obtained using Si alone. In particular the ter-nary SiGeC alloy system has attracted much attentionrecently, as the bandgap and the strain can be con-trolled independently. So far very little is known aboutC incorporation on Si substrates, which is very difficultdue to the very large lattice mismatch between the twobulk materials (-35%). This fact makes detailed stud-ies of the structure and properties of the C/Si(001)system of particular interest.Evaporating carbon from a graphite sublimation sourceonto Si(001) at 550°C forms areas having an unusualc(4 x 4) surface reconstruction. This is clearly seenfrom ultra-high vacuum scanning tunneling microscopymeasurements, which reflect a convolution of theatomically resolved surface morphology with the localelectronic density of states at the surface.

i*V +2V, |' :

G.2nA'' * + >•

The basic feature is the presence of two brightspots per c(4 x 4) surface unit cell, at an apparentheight of about one ML. The two spots appear inpairs with an intrapair distance equal to 5.9 A andan interpair distance of 9.6 A. The distance be-tween successive pair centres is 15.5 A, which isthe normal length of four times the basic repeatdistance on the unreconstructed (001) surface. Theintrapair distance is far too long to correspond toany kind of interatomic bond. So the bright spotscannot be associated with single atoms. We pro-pose that the paired spot structure corresponds totwo Si dimers, which would normally be at a secondneighbour distance, i.e., 7.7 A apart, and havecome closer together at a distance of 5.9 A due tostrain induced by incorporation of substitutional Catoms in the Si lattice. We have evaluated eight dif-

ferent models involving different amounts of C on lat-tice sites, that could potentially account for the largedisplacement, by investigating their total energies andthe thermodynamic stability [1]. We also evaluated theelectronic charge densities on the surface, based onDFT calculations. We find that only in the modelshown below are the surface Si dimers displacedsufficiently towards each other to form effectively a

pair of dimers, as it is seen with STM. This structureconsists of two Si dimers in the surface c(4 x 4) unitcell, brought closer together by 1.8 A due to thepresence of a complex of six C atoms between theSi dimers, arranged in a surface carbon-carbondimer and its first four subsurface neighbours. Thisstructure also is the energetically most favourable.Furthermore, it is predicted by Monte Carlosimulations, that it is stable against interchanges ofC and Si at high temperatures .

FUNDING: PSI, EPFL.REFERENCE

[1] O. Leifeld, D. Grutzmacher, B. Miiller, K. Kern,E. Kaxiras and P.C. Kelires, accepted for publica-tion in Phys. Rev. Lett., 1999.

111

TEM STUDY OF AN ANTICORRELATION IN CORRUGATED LAYERS OFSii.xyGexCy/Si SUPERLATTICES

E. Muller, R. Hartmann, D. Grutzmacher (PSI)

SiC and SiGeC layers were grown as a superlattice on Si(001) substrate. Exceeding the relatively narrowrange of growth temperature and C-concentration allowing for high quality layers, a pronounced undulationof the layers was observed. This undulation has been found to locally exhibit an anticorrelation relationfrom one layer to the next one. This is in clear contrast to the vertical alignment of corrugated layers knownfrom the pure SiGe system.

The lattice constant of Ge is about 4.2% larger thanthat of Si. This results in a low critical thickness ofSii.xGex layers for the onset of defect formation.Recently, it has been shown, that the high lattice mis-match of a SiGe layer on Si can be partly or even fullycompensated by the incorporation of C-atoms [1].Growing a SiC layer on Si even allows the growth of alayer having tensile strain on Si. In spite of their lowsolubility in Si, carbon atoms can be incorporated intoSi or Sii_xGex up to a concentration of several percenton substitutional sites. The growth of such layers,however, is rather difficult. As soon as either thegrowth temperature, the C-concentration, the layerthickness or the growth rate are inappropriate, a pro-nounced corrugation of the layers is observed.

#** .•*

The above figure shows a transmission electron mi-crograph of a Si0.985C0.015/Si superlattice (SL) grownat 530°C. Faint ellipsoidal contrast features beingperiodically aligned can be recognised. If the carboncontent is increased to a concentration of 2.5%, apronounced undulation of the SiC layer is visibleinstead (see the figure below). On both micrographesthe corrugation of subsequent layers appear to becoupled via an anticorrelation: The features observedin the 2nd, 4 th, 6th etc. layer are shifted by one half oftheir lateral period compared to those of the 1 s t , 3rd,5th etc. layer.

SEM images show that the surface roughening is not areal long range order phenomenon. Possibly due tothe presence of two equivalent orientations a compli-cated system of "hills" each being elongated along oneof the <110> directions is observed. Usually only 3 to 5

hills are lying parallel to each other. On TEM imagesthis results in regions showing the above discussedanticorrelation alternating with areas being character-ised by flat SiC-layers. The latter represent sectionsimaged across the hills seen by SEM, while thoseshowing surface roughening are seen in a projectionalong the hills.

Since the presence of such an anticorrelation is notrestricted to SiC/Si SLs but is also observed inSi-|.xGexC/Si SLs as well as in Si-|.xGexC/Ge SLs (ascan be concluded from ref. [2]), it appears to be amore general phenomenon caused by the presence ofthe carbon atoms. It clearly differs from the pure SiGesystem where a direct alignment of surface corruga-tions is observed. This can easily be illustrated by thevertical alignment of Ge-dots (see the figure below).

FUNDING: Swiss National Science Foundation.

REFERENCES[1] R. Hartmann, U. Gennser, H. Sigg, D. Grutz-

macher, K. Ensslin, Appl. Phys. Lett. 73, 1257(1998).

[2] B.-K. Yang, M. Krishnamurthy and W.H. Weber, J.Appl. Phys. 82, 3287 (1997).

112

PHOTOLUMINESCENCE OF C INDUCED Ge QUANTUM DOTS EMBEDDED IN Si

Ft. Hartmann, H. Sigg, U. Gennser, H. Siegwart, K. Ensslin, D. Grutzmacher (PSI)

C-induced Ge quantum dots grown by molecular beam epitaxy (MBE) at deposition temperatures rangingfrom 450-550°C exhibit pronounced photoluminescence (PL). This luminescence exceeds in intensity thatof single crystal Si by several orders of magnitude and is about 10-20 times more intense thanluminescence observed from 2-dimensional SiGe quantum structures. To clarify the origins of this PL, westudied the dependence of the optical properties on the excitation density and post-epitaxiai annealing.

The reduction of the dimensions from a quantum wellconfinement to a O-dimensional confinement in quan-tum boxes is considered as one promising approach toenhance the non-phonon (NP) assisted PL in columnIV alloys. Self assembly proves to be a powerfulmethod for the fabrication of low dimensional Geislands in Si. It has been found that the coating of Si(100) surfaces with fractions of a monolayer (ML) of Cleads to a dramatic increase in density of Ge islandsand decrease in size. Island formation on C coated Sisurfaces has been obtained at Ge coverages waybelow the critical thickness of 4 ML Ge for island for-mation on bare Si surfaces.

The first figure shows a typical PL spectrum of a sam-ple containing Ge dots formed after the deposition at450°C of 0.1 ML C and 2.5 ML Ge. The dots are over-grown by 30 nm of Si. Note that the NP signal of the Sibulk material at 1050 meV is barely visible. The narrowline at 1094 meV is the corresponding TO phonon lineof Si. The additional PL at 1.09 eV and its phononreplica at 1.04eV is assigned to the quantum dot. Incontrast to the Si PL the dot PL exhibits a stronglyenhanced NP line. A careful peak fitting of the over-lapping PL lines in the spectra yields a linewidth of theNP line of the dots of about 20 meV, which most likelyis due to the inhomogeneous size distribution of theGe islands.

Dot-PL

TO

simulated

1.0 1.1energy [eV]

In the second graph we show a series of PL spectrataken from the same sample after a 4 min anneal attemperatures varying from 550 to 1100°C. The postepitaxial temperature treatment leads to an interdiffu-sion of the Si and Ge atoms, which will lead to biggerislands and at higher temperatures eventually to theformation of a 2-dimensioal alloyed layer. The increasein size of the dots leads to the blue-shift of the spectrain the temperature regime from 550-750°C. At higheranneal temperature the transition into a 2-dimensionalalloyed layer becomes visible by the decrease in NP

PL intensity and simultaneous increase of the TOphonon PL. After a 950°C anneal the spectrum showsthe typical signature of a 2-dimensional SiGe quantumwell.In the third graph the dependence of the PL on the

T =450°CGrowth

RTA (4min):1100°C

1050-C

950°C

850°C

750°C

650'C

550°C

as grown

0.9 1.0energy [eV]

excitation density is shown. A blue-shift of the dot re-lated PL is detected with increasing excitation den-sities. This is explained by the consecutive filling of low

3

CD

1.D84

f [ 1.DB0'

P-c

norg

1.072

{\0 20 40 BO BO 100

intensily [%]

^ J A

intensity

0.25%

l^^j\ 2.5%. , . i ,. . I-P—*"^ / ^

^i 1 , i i i r i • • •

[ 5 %

25%

^ joo% |

0.9 1.21.0 1.1energy [eV]

energy states in big dots towards states at higherenergies in small dots. In conclusion, the lineshape ofthe PL, the observed blueshift in dependence on thetemperature treatment and on the excitation densitystrongly suggests that the PL originates from carriersconfined in low dimensional structures.

FUNDING: NF, PSI.

113

INTERSUBBAND ABSORPTION IN Si/SiGe MQW

G. Dehlinger, U. Gennser, hi. Sigg, K. Ensslin, D. Bachle, R. Mezzacasa, D. Gruetzmacher (PSI)

Silicon based detectors for the infrared regime may offer some advantages over other material systems.Modulation doped multi quantum wells (MQW) in Si/SiGe grown by MBE show sharp absorption peaks andtherefore their high quality. We use this to study the physics ofintersubband transitions.

The Si/SiGe material system offers the possibility toapply the methods of bandgap engineering to Si tostudy fundamental physics and develop new deviceconcepts and combine them with the mature Si proc-ess-technology.

Here we study intersubband absorption of Si/SiGeMQW structures to evaluate the possibilities of thismaterial system for the fabrication of optical devices inthe mid infrared regime. With the possibility of tailoringthe resonance-wavelength by adjusting the quantumwell width it is of interest for making infrared-detectors,sensitive at 8-13(j,m, the atmospheric window. In par-ticular Silicon offers advantages over Ill-V semicon-ductors, since here it is possible to absorb the radia-tion vertically and therefore to make planar detectorarrays.

Due to the large band offset in the valence bandbetween silicon and germanium thin layers of SiGeembedded in Si form a quantum well for holes. Usingspecial modulation doping it is possible to ensure thatthe holes occupy the first energy level. Incoming lightwith an energy equal to the energy-separationbetween two levels and with the right polarisation canbe absorbed by exciting an hole from one level to anhigher one.

growth direction

Due to the fact that silicon has heavy and light holes,the polarisation of the incoming light can be parallel orvertical to the growth direction, depending on the in-volved energy states.The samples were grown by MBE at 550 °C. Theyconsists of 5 to 20 wells separated by 20 nm Si. Thecenter of the Si layers were boron doped (2*1018cm'3).

For the optical measurements the samples werecleaved in pieces of ca 8x4 mm2. To optimise the cou-pling of the light into the sample, the front and rearedge were polished in 45° to get a waveguide struc-ture.

Subsequently, the sample was coated with a metalfilm. In this way multiple internal reflection of the lightenhance the optical path. Spectra were taken withdifferent polarisation, which were then divided to dis-tinguish intersubband absorption and other(absorption) effects, such as free carrier absorption.The measurements were made with a Bruker fourrier-transform-spectrometer with a glowbar as light-sourceand a MCT-detector. Spectra were taken at room-temperature as well as at 77K and 4K.

0.1

0.0

sample B517 mod. dop. MQW(37% Ge; 3,4 nm well, 20x)

\ II300K77K4K

1000 1500

wavenumber [cm'1]

2000

The graph compares spectra, obtained for differenttemperatures. With decreasing temperature the peak-absorption changes slightly to higher energies, whilethe FWHM decreases to 20 meV, which compares wellwith the best values reported in the literature. From theintegrated absorption one can obtain the sheet-densityand it fits fairly well with the results obtained from Hall-measurements. Because some of the structural pa-rameters are not well determined, it is difficult to makea good comparison between the theoretical and themeasured values for the peak position.

Further experiments are in progress, where top gatesare used to influence the sheet-density and thus theabsorption.

FUNDING: SNF, PSI.

114

METAL-INSULATOR TRANSITION IN A SiGe 2-DIMENSIONAL HOLE GAS

V. Senz, U. Dotsch, U. Gennser, D. Grutzmacher, D. Bachle, T. Mezzacasa, K. Ensslin (PSI),T. Ihn (ETHZ), Heinzel (ETHZ), D. Weiss (Univ. Regensburg)

The low temperature resistance of Si/SiGe 2-dimensional hole gas has been studied using the gate con-trolled carrier density as a parameter. A metallic phase is found for high carrier densities, and both E-fieldand temperature scaling is demonstrated for the metal- insulator transition.

The recent discovery of a metallic phase at zero mag-netic field in two-dimensional (2D) systems have castaway the two-decade long belief that these systemsare all localised. There have by now been reports onmetal - insulator transitions in several different 2Dsystems; however, in the SiGe 2-dimensional holegases (2DHG) so far only T-scaling has been estab-lished. In the present work we see evidence for bothT- and E-scaling in this system.

The samples employed in this study were grown bymolecular beam epitaxy, and consist of a p-modulationdoped Sio.85Geo.15 quantum well. The transport effec-tive mass of the holes confined in the SiGe quantumwell is m* = O.25mo, as extracted from the temperaturedependence of the magnetotransport measurements.Conventional Hall bar structures were fabricated with aTi/AI Schottky gate to modulate the carrier density.Transport measurements were performed attemperatures down to 35 mK. The carrier density wasmeasured both using classical Hall measurements andfrom the Shubnikov-de Haas oscillations.

The current-voltage characteristics were measured atthe base temperature of the dilution refrigerator (35mK). At high carrier densities (low resistance) the re-sistance behaviour is that of a metal: for large electricfields the resistance increases. For the lowest carrierconcentrations, the behaviour is instead insulating,with a lower resistance at higher E-fields. At a criticaldensity nc an electric field independent resistance isobserved. The transition is found for k?l = 1, where i

is the mean free path as deduced from the mobility,and kp = (2jins)

1/2 is the Fermi wave vector. In thefigure to the left pxx is plotted against 5n/E1/((z+1)v)j

where 5n = (n-nc)/nc, and where a value of the expo-nent 1/((z+1)v) = 0.235 is found to give the best over-lap between the different density curves.

0.011/zvI8J/T

The transition can also be seen in the temperaturedependence of the resistance. In the above figure thescaling plot is presented, with pxx plotted againstSn/T 1 ' ^ ) . The exponent 1/(zv) = 0.433 has beenobtained in similar fashion as the exponent of the E-field scaling. Again, the curves fall on two separatelines, with a reasonable overlap between the curves.

From the E- and T-field scaling we obtain the values z= 1.2 + 0.1 and v = 1.95 ± 0.25, which is in goodagreement with the value z = 1 expected for a stronglyinteracting 2D system, and v = 7/3 derived for quan-tum percolation, it should also be compared to thevalues found in Si MOSFETs: z = 0.8 ± 0.1 and v = 1.5±0.1 .

No effect is seen when a magnetic field is appliedparallel to the 2DHG. The symmetry of the conductingheavy hole state makes the magnetisation behave asa system with an easy axis. In contrast, it has beenreported that the field destroys the metallic phase in SiMOSFETs. Together, these observations indicate thata spin gap at B = 0 due to spin-orbit interaction may beresponsible for the metallic phase.

FUNDING: SNF.

115

AMPEROMETRIC IMMUNOSENSING OF PENICILLINS

C. Padesie, A. Grubelnik, J.J. Hefti, L Tiefenauer (PSI)

The determination of penicllins and related degradation products is of increasing interest for quality controlof food. Amperometric immunosensors combine the high specificity of antigen-antibody interactions withthe high sensitivity of electrochemical detection principles.

Analysis of antibiotics in food has a commercial and alegal aspect. Antibiotics in their active form inhibit fer-mentation processes, e.g. in cheese production.Determination of degradation products on the otherhand may give indications on abuse of antibiotics.

Among the antibiotics used in animal farming, ben-zylpenicillin takes a prominent position. As a majorcomponent of most antibiotic cocktails, it may be usedas an indicator for the presence of further antibiotics infood samples.

A four-membered p-lactame ring is common to penicil-lins in their active form. The (3-lactame ring is openedupon reaction with a key enzyme of bacterial cell wallsynthesis. Hydrolysis in aqueous solutions also leadsto ring opening and deactivation of penicillins.

COOH

In order to determine both the active and deactivatedpencillins by immunochemical methods, antibodiesdirected against the intact and the open ring form areneeded. Two corresponding BSA-conjugates havebeen synthesised and used to immunise a series ofrabbits. The resulting antisera were collected andtested using penicillin—biotin conjugates in an enzyme-

linked immunosorbent assay (ELISA). In buffer solu-tions, detection limits are in the range of 10ppb forintact as well as for hydrolysed benzylpenicillin.

In order to achieve electrochemical detection, it isnecessary to couple electrochemically active moietiesto the antibodies. For this purpose microperoxidase(MP-11) was used [1], which is known to catalyse theelectrochemical reduction of hydrogen peroxide

(HA)-

HoO2O2 2e"MP-11

2 OH"

This reaction was measured at a potential of 180 mV(vs. Ag/AgCI) using electrodes with immobilised MP-11or antibody-MP-11 conjugates.

ELISA calibration curves

hydrolysedbenzylpenicillin

1 10 100 1000

Antibiotic Concentration [ppb]

10000

When MP-11 was directly bound to the electrode sur-face (a), much higher signals were observed com-pared to immobilisation via streptavidin and an anti-gen-biotin conjugate (c). As this effect is most proba-bly due to the limited conductivity of protein layers,present investigations are focusing on improvementsof the electron transfer in the system.

FUNDING: SPP Biotechnology, PSI.

REFERENCE

[1] C. Padeste, A. Grubelnik and L. Tiefenauer, Anal.Chim. Acta, 374, 167(1998).

116

FUNCTIONALISED GLASS CHIPS FOR NEURON CELLS

H. Sorribas, C. Padeste, A. Grubelnik, T. Mezzacassa, L. Tiefenauer (PSI), P. Sonderegger, L Leder(Univ. Zurich), R. Douglas (ETH/Univ. Zurich)

Using photolithographic techniques microchannels suitable in size to host single neuron cells have beengenerated on glass chips. By a directed immobilisation of biomolecules, adhesion of neurons andoutgrowth of axons could be achieved.

Designed networks consisting of dissociated neuronsmay be very helpful to investigate neuronal informationprocesses. The dimensions of sensory neuron cellsand their outgrowing axons are in the range of 20 and100 micrometres, respectively. Guidance of neuroncell outgrowth in a predefined direction can beachieved by suitable microstructures. An array of mi-crochannels on a glass chip has been generated usingphotolithographic techniques. First, gold lanes havebeen produced on glass and covered by an insulatingoxide layer. These conducting gold lanes end in theindividual microchannels and will allow an electricaddressing of axons. A biocompatible polyimide wasused for the structural definition of the microchannels.

Neuron

Col. merr.bra-io

Axonin-1

Axonin-110 nm

Gold

Glass

These axonin constructs can now be stable immobi-lised via their cysteine on protected gold in a sub-monolayer density as demonstrated by radiolabellingtechniques. Neuron cells adhere to the axonin-treatedgold surfaces confirming that the immobilised mole-cules are functional.

In parallel, glass surfaces have been treated in orderto achieve a neuron-compatible surface. The tetrapep-tide RGDC is covalently immobilised on the surfaceusing reactive silane and a heterobifunctionalcrosslinker. Neuron cells which have been isolatedform dorsal root ganglia of chicken adhere on thebiofunctionalized glass and the outgrowth of axonswas similar to cultures on laminin-treated plastic sur-faces.

Specific adhesion proteins present on the gold padswill promote a close neuron-electrode contact and mayalso induce synaptogenesis. The adhesion proteinaxonin-1 essentially consists of Ig- and fibronectin (Fn)domains and naturally occurs in cell membranes ofoutgrowing axons. Recombinant axonin moleculeswhich have a C-terminal Cys at a peptidic linker havebeen expressed in an eucaryotic cell system.

The reported achievements are an important steptowards the realisation of predefined networks of neu-rons on a chip.

FUNDING: SPP Biotech, PSI.

117

SPM FOR FUNCTIONAL IDENTIFICATION OF INDIVIDUAL BIOMOLECULES

R. Ros, C. Padeste, L. Tiefenauer (PSI), F. Schwesinger, A. Piuckthun (Univ. Zurich), D. Anselmetti(Novartis), M. Kubon, R. Schafer, H.-J. Giintherodt (Univ. Basel)

Binding forces of individual antigen-antibody pairs have been determined using a combination of scanningprobe microscopy (SPM) techniques. A new software has been developed which allows automatedmeasurement and data processing.

Precise determination of biochemical binding forcesby using SPM-techniques requires both, a carefuldesign of the molecular arrangement and asophisticated measuring method. We used aconvalently immobilised antigen coupled to an AFM-tip and a recombinant single chain Fv fragment (scFv)of the corresponding antibody molecule which isimmobilised in a correct orientation on an flatprotected gold surface. When the scFvs are presentat a very low surface density on gold, individual, wellseparated molecules in an AFM-image can beselected for the subsequent measurement of bindingforces: The functionalised tip is first approached tothe selected scFv molecule and then retracted from it.At a certain point the antigen-antibody complex willbreak and this rupture force can be extracted from theforce-distance (F/D) curve. Based on a statisticallysufficient number of unbinding force values, 50 + 4picoNewton (pN) was calculated for the wild typeprotein. For the mutant His(H58)Ala which hasroughly a ten times lower affinity constant, anunbinding force value of 40 ± 3 pN has been found. Inthis mutant scFv a single histidine of the binding sitehas been exchanged by alanine. A series of additionalpoint mutant scFvs are now being investigated. Puta-tive correlations of their kinetic and thermodynamicparameters to the measured unbinding force valueswill be investigated aimed at better understanding ofbasic biorecognition processes.

Considering a future application of AFM techniquesfor functional screening we further improved thesoftware required for the measurement and dataacquisition of F/D-curves. First, the instrumental driftis automatically corrected in order to prevent anunwanted application of destructive forces to thesample. Second, a high number of F/D-curve areautomatically acquired and stored. Third, an algorithmhas been introduced allowing the automatedextraction of unbinding force values from the F/D-curves, followed by a statistical analysis. Thus, thetime required for the measurement of one scFv typecould be decreased from one hour to about 20minutes.

*******

»*fhf.

i* • "tuftW »itU» />'-'

I. Imaging: . Force Spectroscopj:

point selection

z - drift correction

analysis

jy i - drift correction

DI. Data Processing: ^ ^

unbinding force value

Furthermore, a sample preparation method was de-veloped for multiple analysis. By using a simplestamping technique, hydrophilic squares have beengenerated on a flat gold surface allowing toimmobilise different scFv-types on the same chip.This method allows to investigate up to nine scFvs inthe same experiment under the same conditions. Thisis an important prerequisite to detect the smalldifferences between the different mutant proteins.

FUNDING: NFP36 Nanoscience, PSI.

REFERENCE

[1] R. Ros, F. Schwesinger, D. Anselmetti,M.Kubon, R. Schafer, A. Piuckthun and LTiefenauer, Proc. Natl. Acad. Sci. USA 95, 7402(1998).

118

ONE-DIMENSIONAL CONFINEMENT OF MOLECULES VIA SELECTIVEADSORPTION

H. Rauscher, J.A. Lin, A. Kirakosian, F.J. Himpsel (Univ. Wisconsin at Madison), T.A. Jung (PSI), U. Rohr,J.D. Brand, K. Muellen,(MPIP, Mainz)

Molecular self assembly and growth has been explored to generate well defined linear (1-d) arrays ofmolecules on a semiconducting/insulating nanopatterned substrate. Self organized CaF/CaF2 stripe pat-terns served as a template for selective adsorption of 3,10-di(propyl)perylene. Beyond the nanopatterningthe selectivity of adsorption can be discussed in terms of the interaction of occupied (HOMO) and unoccu-pied (LUMO) molecular orbitals with the valence and conduction bands ofCaF and Ca F2, respectively

Nanostructure fabrication with characteristic dimen-sions in the nanometer range is essential for futureapplications in nanoelectronics, nanooptics, or towardsthe integration of functional materials on the nanome-ter scale. In this size range, dimensionality effects arebecoming important and these structures exhibitunique fundamental properties such as density ofstates singularities or discrete molecular electronicstates over large distances. In this work a multi stepapproach of substrate and molecular self-assembly ispresented that resulted in the successful creation ofwell defined stripe structures.

First we prepare a self organized step array on theSi(111) surface which is readily decorated with a self-assembled CaF/CaR, stripe pattern (1) CaF is chosenas a growth material because of the lattice constantmatching the Si(111) within one percent. After deposi-tion of CaF2 at 610 C and subsequent annealing at830 C for 10 s a Si-Ca-F interface is formed. If theCaF2 deposition is continued beyond a full monolayercoverage, it nucleates as CaF2 from the step edgesand is imaged in Scanning Tunneling Microscopy STMby protruding wires. A typical result of this process isshown below (imaged surface area: 120nm x 120nm.

trusions on the dark background of the single mono-layer coverage with CaF stoichiometry.

\ \ \ \

Second the 3,10-di(propyl)perylene (DPP) moleculesare deposited onto this substrate by sublimation inUltra High Vacuum (UHV). Third the distinctly differentlocal chemical reactivities of CaF1/CaF2 on siliconSi(111) substrate surface are used to arrange themolecules along the more reactive CaF1 stripes. In theSTM image of a such prepared surface below, 120nmx 120 nm in size, the molecules are imaged as pro-

Sfc

The reason for this chemical specificity of adsorptionlies in the distinctly different electronic structure of theCaF monolayer and the CaF2. This is also reflected bythe local band gap changing by as much as 9.6 eVbetween CaF2 and CaF, as it can be confirmed bySTM chemical imaging (3).

Thus we have presented the successful generation ofvery well defined structures consisting of molecularand atomic layers with monolayer precision. In thefuture, such compound, self assembled nanostructuresmay provide means towards the integration of mo-lecular devices into silicon electronics.

FUNDING: Univ. Wisconsin at Madison, US NSFDMR-9624753 DMR-9632527, PSI, MPIP-Mainz.

REFERENCES

[1] H. Rauscher, T. A. Jung, J.-L Lin, A. Kirakosianand F.J. Himpsel, submitted to Chemical PhysicsLetters.

[2] J. Viernow, J.-L. Lin, D.Y. J. Viernow, J.-L. Lin,D.Y. Petrovykh, F. M. Leibsle, F.K. Men, and F.J.Himpsel, Appl. Phys. Lett. 72 (1998) 948; J.-L.Lin, D.Y. Petrovykh, J. Viernow, F.K. Men, D.J.Seo, F.J. Himpsel, J. Appl. Phys. 84 (1998) 255.

[3] T.A. Jung, F.J. Himpsel, R.R. Schlittler, J.K. Gim-zewski. In: Scanning Probe Microscopy, Analyti-cal Methods Springer (1998) pp11-48.

119

R a d i o - a n d

E n v i r o n m e n t a l C h e m i s t r y

Heavy Elements

Surface Chemistry

Aerosol Chemistry

Analytical Chemistry

Project Radwaste

Cement Chemistry

120

A 252Cf FISSION FRAGMENT SOURCE CALLED MISS PIGGY

D.T. Jost, A. Tiirler, D. Piguet, B. Eichler (PSI), U. Kindler, R. Eichler, Ch. Dullmann (Univ. Bern),H. W. Gaggeler (Univ. Bern, PSI)

A fission fragment source using 4 u.g 252Cf was built as a source for short-lived isotopes. This facility is usedto study the chemistry of carrier-free nuclides. For safety the operation of the device is controlled with aprogrammable controller. The interface with the user is a LabView application running under Windows.

1 INTRODUCTION

Spontaneous fission of 252Cf produces a wide range ofradioactive nuclides which can be used to studychemical reactions in very diluted systems (singleatom chemistry). The alternatives to produce suchisotopes are nuclear reactors or particle accelerators,which are too expensive and too large to operate in auniversity laboratory.

2 DESIGN GOALS

The 252Cf-source has to provide a continuous output offission products which can be transported with a gas-jet to a chemistry experiment. Different carrier gasesand recoil chamber pressures have to be possible Thedevice has to be save, easy to operate and produce aslittle radioactive waste as possible.

3 SOURCE ARRANGEMENT

4|ig 252Cf are electroplated on a Pt disk. The disk ismounted on a movable carrier and covered with a 6p.mAl foil. When the source is not used it is moved to thepark position. With the source in this position the recoilchamber can be replaced. The recoil chamber with themovable stem makes it possible to accommodate thedifferent range of the fission products in the differentjet gases (1.8 cm in N2, 8 cm in He). The entrancewindow of the recoil chamber is again a 6u.m Al foil.This double barrier technique protects the chemistryapparatus from a 252Cf contamination. The sourceassembly is housed in the middle of a pink steel ball of1m diameter (see cover photo of this report). Thesource and the stem can be remotely operated fromthe top of the steel ball. The empty space of the ball isfilled with boron-paraffin and polyethylene granulate asneutron shielding.

4 CONTROL SYSTEM

The parameters position of the source (park or runningposition), pressure in the recoil chamber, gas flow rateand gas composition are controlled by a SELECTRONcontrol and interlock system, which ensures a safeoperation of the source. This control system runsindependently; a LabView interface running on a PC isused to communicate user parameters and to displaythe system status.

5 OPERATION

A LabView application provides an easy to use userinterface. The user selects the desired carrier gasesand flow rates. The recoil chamber pressure and thecarrier gas parameters are continuously displayedtogether with the status and error messages of thecontrol system. These parameters are periodicallylogged. It is planned to include the chemistry devicesin to the same display system.

6 OUTLOOK

A second source called Kermit is under construction. Itwill feature a recoil chamber suited for chemicallyreactive gases and a chromatography system directlycoupled to the chamber without an intermediate gas-jet transport. This will expand the range of availablenuclides to even shorter half-lives.

Jet outlet

Volurne-variation-slide

ZSource park-position ^Source running-position

Fig. 1: 252Cf source delivered by the Institut furKernchemie at Mainz University for use with agas-jet transport system.

121

MISS PIGGY, THE NEW CALIFORNIUM-252 FISSION SOURCE AS A GENERATOROF SHORT-LIVED RADIONUCLIDES

Ch.E. Dullmann, D. Jordi, R. Eichler (Univ. Bern), D.T. Jost, A. Turler (PSI), H.W. Gaggeler(Univ. Bern & PSI)

A 252Cf fission source is a versatile instrument in various research fields. It provides fission products withhalf-lives from a few seconds up to several hours. In connection with a gas-jet and the PSI Tape DetectionSystem or a trap, respectively, more than a hundred different nuclides can be detected. The typical ele-ments of the lighter mass peak are homologues of the heaviest elements currently being chemicallyinvestigated and therefore well suited for modelling the chemical behaviour of these transactinides.

The nuclide 252Cf decays mainly via a-emission, buthas also a spontaneous fission branch of three per-cent. The mass numbers of the fission products aregrouped in two peaks with mass numbers from 95 to120 and from 130 up to 155. The primary fission pro-ducts have very short half-lives and are therefore wellsuited e.g. for kinetic studies of chemical processes.The elements that most of the nuclides of the lightpeak belong to are homologues of transactinides thathave isotopes which are long lived enough to be che-mically investigated like Rf, Db, Sg, Bh or Hs. There-fore these fission products are well suited for model-ling studies of transactinides.

The transport of the fission products, which passthrough the cover foil of the source into the recoilchamber and from there to the chemistry and/or de-tection device is done with a gas-jet system [1].

Various gases like He, N2 or O2 can be used for thispurpose and the jet can either be run with or withoutaerosol particles, depending on the properties of thedesired nuclides. If the gas-jet is run with an aerosol,e.g. carbon particles with He/N2, the PSI TapeDetection System can be used to analyse thedecaying nuclides by means of Y-sPectrosc°Py- Aspectrum acquired with this system is given in [2]. Ifthe jet is run with pure gases, e.g. He/O2, the productscan be collected in a charcoal trap which can also bemeasured y-spectroscopically.

REFERENCES

[1] R.J. Silva et al., Nucl. Inst. Meth. 147, 371 (1977).

[2] D.T. Jost et al., this Annual Report.

46Pd

45Rh

f-tfaj 6.0 h* } t s i .• • ^ • ^ l MO-99^ ^ ^ ^ ^ 1 66)1

Nb-96 fib-BS

•••Tc-10114.2 m

Mo-IGO

Nb-9915 s*

mRu-10338,35 d

:rgfMo-101H,Bm

Nb-1001 5s

Zr-99

2,1 8

•Rh-10535.4 h*•Tc-10354,2 s

MO-10211.2m

Nb-1017.1 s

Zr-100

7.1 s

IHIlRu-1054.44 h

Tc-10418,2 m

Mo-10367,5 s

N i !('

•Rh-10721.7 m

£STc-1O57.6 m

Mo-1041.0m

NO *Ww A

: * -

Rh-10816.6 s*

Ru-1073,8 m

Tc-10636 s

Mo-10535,6 s

\l* '--I

•Rh-109

BOs

Ru-108

4,5 m

Tc-10721.2 B

Mo-106B,7e

J t *

' - . i :

W-11]

Rh-1103.3 s-

Ru-10934.5 s

Tc-1085 17 s

Mo-1073.5 s

NEi Oft

' JH

Rh-11111 s

Ru-11011,6 s

TtS-1090,8S«

*fo-i»a

VJ *r*

3 S " - -

Rh-1122,1 s*

»1-11»

til"•MO-IOS

•: i •••t

Pd.114

Rh-1132.80 s

Ru-118

Tc-m

tb-no

P t1 fRh-114

1.85 s*

n«-tt3

Tc-112

Mo-111t

Pd-116

11.8 s

wf

" . ; " •

4,5 ••

'ft68s

fl«-11Sr .'•

tt,*!3

Fig. 1: Cut out of the chart of nuclides that shows the light mass peak of 2S2Cf fission products.• : stable nuclides; • : nuclides detected with Miss Piggy.

56

Ba

55

Cs

53 ^ ^ ^ H i-JSS

1 i^i^B JSW52 ^ ^ H T 8 - 1 8 ?

51 |SiJ-1SS|SI>-^

Sb I •w*! «M«'

f-128

rT^3,85*

'£?!,

Sb-126

9h-

••

1-131

8,02 d

T8-;I3{>

•Sb*f29

•Cs •:«!

Xe-133

5,S5d

1-132

2.3 h

Te-131

25 m*

•".•.*?)

/* J1'•1-133

20.8 tT

Te-132

76.3 h

Sb-131

23m

59 ^ HPr ^ H58 ^ ^ H C e - 1 4 1

Ce ^ ^ B 32.50 d

57 ^ ^ H L a - 1 4 0

La ^i^i^i|40.272 tl

^ ^ ^ ^ ^ H Ba-139

^ ^ ^ ^ H ^ ^ 83.06 m

^""••SlO-,-- <?|Cs-138

> F a j '~ . f j 32,2m

Xe-135M^IXe-137

9,1 h* ^ i ^ H 3,83 m

1-134 I 1-135

52,0 m I 6,81 h

1-136

84s

Sb-132js^134s&"1i34

2.8 m* 1 ^ Jn § tpSS*

Pr-143

13.57 d•Ba-140

12,75 d

Cs-139

9,3 m

Xe-13814,1 m

1-137

24,2 s

W *

Ce-143

33,0

La-142

92.5 m

Ba-141

18.3 m

Cs-140

63.73

Xe-139

39,7 s

1-138

6.4 s

•-'• ! '.5

Co-141

La-143

14.23 m

Ba-142

10,7 m

Cs-141

24,94 3

Xe-140

63,7 s

Pr-146

24.0 m

Ce-1452,98 m

La-144

40,9 s

Ba-143

14,5 s

Cs-142

1.70 s

Xe-141

1.72 s

1-146.

0.68S

Pr-147

13.6 m

Ce-146

13,5 m

La-145

24.BS

Ba-144

11,5s

Cs-143

Pr-148

2,27 m-

Ce-147

57 s

La-146

6,3 s

Ba-145

4.3 s

'Oa-144

1,0*

X9-143

l-UH

Ce-148

48 s

La-147

4,0 s

Ba-146

2,2 s

Pr-150

6,1s*

Ce-149

5s

U-ttft1,4^3.

Ba-1-»7

C<i-UB

xc • •••

P"l 1

Pr-151

18,9 s

Ce-150

4,1 S

Bg-i-iii

Pr-152

3,8 s

*TU-1S0

Pr-t«S

.OMK

ia-isi.

:• -> -i

Cs '4/

" • " •

Fig. 2: Cut out of the chart of nuclides that shows the heavy mass peak of 2

• : stable nuclides; • : nuclides detected with Miss Piggy.

2Cf fission products.

122

THERMOCHROMATOGRAPHIC INVESTIGATION OF RUTHENIUM WITH OXYGENAS CARRIER GAS

Ch.E. Dullmann (Univ. Bern), B. Eichier (PSI), H.W. Gaggeler (Univ. Bern&PSI), A. Turler(PSI)

The behaviour of long-lived radioactive ruthenium in a temperature gradient tube with oxygen as carriergas was investigated. Two deposition zones were observed and assigned to the compounds RuO3 andRuO4. RuO3 was transported along the column in a transport reaction of the type RuO3 (ads) <-» RuOs (g) +1A O2 (g) whereas RuO4 was transported by mobile adsorption RuO4 (g) <H> RUO4 (ads). The adsorptionenthalpy of RuO4 was determined by simulating the process with a Monte Carlo model and by the quasithird law method. The results of both methods were in good agreement.

1 INTRODUCTION

The chemical properties of the transactinide elementHs (element 108) have not yet been investigated ex-perimentally, but it is expected to be a member ofgroup 8 and thus homologous to Ru and Os. Thelongest-lived known isotope of Hs is 269Hs [1] with ahalf-life of about 10 s. Due to the expected very lowproduction cross section of 269Hs and the short half-life,model studies with lighter homologues are essentialbefore a transactinide experiment can be performed.Since Ru and Os are known to form highly volatileoxides, a gas chromatographic separation of Hs asHsO4 is highly desirable.

2 EXPERIMENTAL

In a first step the behaviour of Ru oxides was investi-gated in thermochromatography experiments in orderto determine which species are formed and tomeasure their enthalpies (AHads) and entropies (ASads)of adsorption. Until now, only calculated values areavailable [2]. For these studies commercially available106Ru (T1/2=373.6 d) in 4M HCI was used. Becauseoxychlorides of Ru are supposed to be also volatile atlow temperatures, it was necessary to prepare Rusamples free of chloride. Therefore, Ru was depositedin metallic form from the HCI solution on Ni foils. Thechlorides were removed by washing the foils withNaOH solution. However, the volatilisation of Ru fromthese foils was slow and proceeded with low yield. TheNi foils were therefore not suitable as starting materialfor thermochromatographic experiments. Athermochromatography with SiO2 filled columns wascarried out to transfer the Ru onto quartz powder. TheNi foils were used as starting material in theseexperiments which lasted several hours in order toincrease the yield of the transfer. By applying a steeptemperature gradient, the deposition zone of Ru waskept short. The quartz powder containing the Ru wasthen used in the actual thermochromatographyexperiments. Ten experiments were carried out at gasflow rates between 20 ml/min and 200 ml/min O2

lasting between 15 min and 4h.

3 RESULTS AND DISCUSSION

In Fig. 1 a typical thermochromatogram is shown. Afterthe experiment, 3 Ru peaks were observed. A small

fraction of the Ru remained in the starting zone andwas not mobilised e.g. due to diffusion into the quartzpowder. A deposition peak in the middle of thechromatogram originated from Ru that was depositedon the column surface in the form of themacroscopically unstable RuO3. It was transportedalong the column in a transport reaction of the typeRuO3 (ads) <-> RuO4 (g) + Vz O2 (g). Therefore, thedeposition temperatures varied over a wide range inthe different experiments. In experiments which lastedmore than one hour, the peak disappeared. Under theassumption of a dissociative adsorption withASdiss ads=-73.9 J/K-mol, A H ^ ads=-77±14 kJ/mol wasdetermined.

10

5 - •

0

—Temp— -Monte-Carlo Hads=-54kJ/mol

950

750

• 5 5 0 c?

j %£ RuO3(.d l )^IlllllllllllllllllrttffllllllllllllllllllllS

Distance [cm]

Fig. 1: Thermochromatogram of a Ru experimentlasting 1 h with an O2 flow rate of 50 ml/min.

The Ru deposited at low temperature was transportedby mobile adsorption in the form of the very volatilecompound RuO4. Using the quasi third law method [3]AHads=-55±4 kJ/mol and ASads=-168.7±0.5 J/K-molresulted. The result of a Monte Carlo simulationprocedure [4] is also shown in Fig. 1. The width of thepeak is in good agreement with experimental data, andAHads=-55.0±4.2 kJ/mol was determined, in goodagreement with the quasi third law method.

REFERENCES

[1] S. Hofmann et al., Z. Phys. A354, 229 (1995).

[2] B. Eichier et al., Radiochim. Acta 56, 133 (1992).

[3] B. Eichier et al., Radiochim. Acta 30, 233 (1982).

[4] I. Zvara, Radiochim. Acta 38, 95 (1985).

123

STUDIES OF GASCHROMATOGRAPHIC BEHAVIOUR OF 1691M1831MRe

IN THE GAS SYSTEM He/O^HCI

ft Eichler (Univ. Bern), H. W. Gaggeler (Univ. Bern & PSI), B. Eichler, A. Turler (PSI)

Using thermochromatography with 1S3/1S4Re and a modified low temperature OLGA III technique with 169~176Rethe formation of only a single rhenium compound in the gas phase of He/O/HCI was observed. We deter-mined the adsorption behaviour of this compound assuming the compound to be ReO3CI. The high forma-tion rate, the high volatility and the insensitivity of the formation reaction to different mixtures of the reactivegases makes this gas chemical system an ideal candidate for an experiment withBh (element 107).

The production rates of bohrium nuclides are ex-pected to be very low and the predicted half-lives ofthe longest-lived currently accessible isotopes are ofthe order of 10 s or less. Therefore, we searched fora gaschemical system where high reaction rates andshort separation times are achievable. High tempera-ture gaschromatographic studies of rhenium oxidesand oxyhydroxides in different oxidising gas systems[1] showed rather low overall yields for short- livedrhenium isotopes (169|170Re) in comparison to longer-lived isotopes (m>176Re). This fact was attributed to akinetic hindrance of the formation reaction of HReO4

or ReO3. The chlorination kinetics seemed to befaster. So we decided to use an O./HCI mixture asreactive gas.Four different rhenium containing compounds that arestable in the gas phase are described in the literatureabout the chemistry of the chemical systemRe/HCI(CI2)/O2>: ReCI3, ReCI4J ReOCI4 and ReO3CI [2].Using the selected reactive gas mixture we expectedto produce all of them.Thermochromatographic experiments with 183184Rewere carried out using different mixtures ofHe/O/HCI. In all experiments the formation of only asingle volatile compound was observed (Fig. 1). Weobtained the same surprising result with rheniumsamples that were prepared by drenching a graphitewool plug instead of a quartz filter stripe withNH4

183l184ReO4 followed and then reduced in an Ar:H2

(Vol-% 93:6) mixture [1].Comparing the boiling and melting points of the com-pounds described in [2] with the deposition tempera-ture of the rhenium compound, it seems clear thatonly ReOgCI could be so volatile. This plausibilitystatement is needed because of the scarce thermo-chemical or mass spectrometric data for these com-pounds. The adsorption enthalpy AHads= - 63 kJ/molwas calculated from the thermochromatograms.

60 -

40 -

20 -

0 •J U . , . . . . . . .

ReO3Cl Ti = - 7 °C "

i :

\ ^

i i i J

, * • • • * • * " * * •

1 100

900

700 —U

500 "

300

100

10 20 30 40 50 60 70 B0 90 100I [cm I

Fig. 1: Thermochromatogram of ReO3CI.

At the PSI Philips Cyclotron, on-line experimentswere carried out. We produced rhenium isotopes in ahot fusion reaction 156enrDy[19F,6n]169Re at a beamenergy of 127 MeV.With regard to the high volatility of ReO3CI we built amodified OLGA III system to reach isothermal tem-peratures between -24 and 200°C (Fig. 2).

HCI/02

C/He-jet

Fig. 2: Modified OLGA III system:1 reaction oven; 2 heating/cooling jacket;3 quartz column with the "reclustef chamber.

In this temperature range gaschromatographic reten-tion curves were measured for long-lived 174i176Re andfor short-lived 169170Re compounds as well as for2i8po/2i4Bj c o m p o u n c | s (Fig. 3). Using a Monte-Carlomodel (dotted lines) the adsorption enthalpy ofReO3CI was evaluated to AHads= -61 kJ/mol. This is ingood agreement with the thermochromatographicresults. The formation of the same compound in off-line and on-line experiments can be assumed.

100

B0

60 ••

Z 408

20

0

ReO3CI

4 *'BiCyBifJCI !•

5 25 45 65 85 200T[°C]

300 400

Fig. 3: Retention curves in the O/HCI/He system.

The overall chemical yield for the 16s-169Re is about50% and the separarion time is 7s. The achieved highseparation factors for actinides, Bi and Po are essen-tial for an unambiguous detection of bohriumnuclides.

REFERENCES[1] R. Eichler et al., PSI-Ann. Rep., Annex MIA, 59

(1997).[2] Gmelin, Masurium Rhenium 70, 107 (1941).

124

PHYSICOCHEMICAL CHARACTERIZATION OF SEABORGIUM AS OXIDEHYDROXIDE

S. Hubener, A. Vahle, St. Taut, H. Nitsche for a Bern University - FLNR Dubna - GSI Darmstadt - TUDresden - GH Kassel - University Mainz - Forschungszentrum Rossendorf - PSI Villigen collaboration

High-temperature on-line gas chromatography of oxide hydroxides was used to characterize the physico-chemical properties of seaborgium. The results indicate the formation of a low volatile seaborgium oxidehydroxide.

In continuation of experiments with seaborgium oxy-chlorides [1], the goal of the present gas chemistryexperiment of the seaborgium collaboration was toshow that Sg forms low volatile oxide hydroxides.

The High-Temperature on-line Gas chromatographyApparatus HITGAS developed for studying low volatileoxide hydroxides is shown schematically in Fig. 1.Considering the low rate of the basic reactions of thegas chromatography of group 6 oxide hydroxides inquartz glass columns [2], reasonably short open tubu-lar chromatography columns were used and directlycoupled with the GSI Rotating Wheel MultidetectorAnalyzer ROMA: The chromatography furnace wasflanged onto the ROMA and a separate depositionchamber was placed between column and rotatingwheel, as described earlier in detail [3]. In test experi-ments with short-lived Mo and W isotopes retentiontimes as short as 8 s and chemical yields of about 60% were achieved [3,4]. On-line alpha spectroscopywas carried out with a resolution of 25 keV. However,1 urn collection foils, to be applied for 4 JC alpha-spec-troscopy, which were stable in pure He carrier gas,were immediately destroyed when the reactive gasH2O/O2 was added. Considering the results of the testexperiments and potential interferences of short-livedPo isotopes, we decided to study 266Sg (T1/2 = 21 s)instead of the shorter-lived 265Sg (T1/2 = 7.4 s) and de-tect it in 2TC geometry by registration of 266Sg a-decayand time-correlated spontaneous fission events of the262Rf daughter.

The seaborgium experiments were performed at theGSI UNILAC accelerator with a mixed 248Crn/152Gdtarget (GSI, 820 ug/cm2 248Cm, 85 ug/cm2 162Gd) at a^Ne beam energy of 118 - 120 MeV. Assuming across section of 80 pb for the 248Cm (22Ne,4n) 266Sgreaction the production rate was about one 266Sg atomper hour. A He/MoO3 jet was used to transport thenuclear reaction products with a gas flow rate of 2.0l/min to the HITGAS. The temperature of the chroma-tography columns was 1325 K in the reaction and1300 K in the isothermal zone. At the column en-trance, 0.5 l/min O2, moistened with H2O at 323 K,were added as reactive gas. 25 urn Al-foils were usedto collect the species under study. The ROMA wasoperated with a cycle-time for collection and detectionof 10 s. 15 equidistantly positioned PIPS detectorswere used to detect spontaneous fission and a-decayevents [5]. Short-lived W isotopes were monitored witha HPGe detector. The chemical yield of W oxide hy-

droxides and the separation of interferingspontaneously fissioning actinides were watched con-tinuously. At spontaneous fission rates above 2 cphand/or W yields lower than 40 % the quartz glass col-umns were replaced.

quartz wool quartz glass column

gas jet

cyH2O(gl

I Lcarrier gas outlet -

cooling block

•*— cooling gas inlet

-!=Y —* cooling gas outlet

0 10 20 30 40s[cm]

ROMA wheel

PIPS detector

HPGe detector

vacuum pump

Fig 1: On-line gas chromatography apparatus,schematically.

We detected two correlated 266Sg - 262Rf decay chains:

Table 1: Correlated decay events.

chain

1

2

E«[MeV]

8.66

8.70

life-time 266Sg [s]

84.9

4.8

life- time 262Rf

7.0

3.7

[s]

We interpret these results with the formation of a lowvolatile seaborgium oxide hydroxide typical for group 6elements.

ACKNOWLEDGMENT

The support by the BMBF and the GSI under contracts06 DR 824 and DRNITK is gratefully acknowledged.

REFERENCES

[1] M. Schadel et al., Nature 388, 55 (1997).

[2] A. Vahle et al., Radiochim. Acta 78, 53 (1997).

[3] S. Hubener et al., FZR-218, 87 (1998).

[4] A. Vahle et al., FZR-218, 85 (1998).

[5] St. Taut, this Annual Report.

125

ON-LINE CHROMATOGRAPHY OF SHORT-LIVED W ISOTOPES WITH THE MULTI-COLUMN TECHNIQUE

A. Kronenberg, J. V. Kratz, A. Nahler (Universitat Mainz), W. Bruchle, E. Jager, M. Schadel, B. Schausten,E. Schimpf (GSIDarmstadt), D.TJost, A. Turler(PSl), H.W. Gaggeler (Univ. & PSI), G. Pfrepper (Leipzig)

For on-line chromatographic studies of transactinide elements in aqueous solutions, a separation schemehas been proposed [1] that uses the nuclear half-life of the studied isotope as an internal clock. To preparesuch experiments with seaborgium, we have studied the chemical behavior of its homologs on anion andcation exchangers in HNO3/HF solutions using the multi-column technique.

The first aqueous chemistry of Seaborgium requiredaccumulation of 5000 single experiments with theAutomated Rapid Chemistry Apparatus ARCA. Thisexperiment showed that seaborgium, in 0.1 M HNO/5*10"" M HF, forms neutral or anionic oxyfluorides [2],but the experiment did not determine a distributioncoefficient. In future experiments with seaborgium andother transactinides, we plan to determine Kd-valueson anion and cation exchangers in HNO3/HF solutionsusing the multi-column technique [1] via activities of along-lived descendent.In this technique, the KCI aerosol particles transportingthe activities in a gas-jet are continuously dissolved ina degasser unit in an aqueous solution which is contin-uously pumped through three consecutive columns.The first column (F) acts as a filter for all decay prod-ucts of the interesting element and is passed by thelatter. The second column (C) is the true chroma-tographic column where the element of interest expe-riences a retention time, tR, being of the order of itsnuclear half-life, T1/2. The daughter nuclides that areformed during the retention time of the element ofinterest on column C are strongly retained in C(daughter activity Ac). The part of the element of inter-est that passes C is sorbed on the third column Dwhere it decays into its daughter which is stronglyretained in D (daughter activity AD). After the experi-ment, the activities of the decay products on column Cand D are measured and the retention time is obtainedaccording to:

In2

The distribution coefficient, Kd, of the element of inter-est in column C may then be obtained as:

where t0 = column hold-up timeV = flow rate of the mobile phaseM = mass of the stationary phase.

The degasser unit has been developed and tested atthe Mainz TRIGA reactor. It has been shown that thedissolution of the aerosol can be performed quantita-tively, but only at the expense of a very high pressurein the target chamber. As a result of many experi-ments, the dissolution yield in the degasser unit is nowover 80% at a tolerable pressure in the target chamberof 1.7 bar at a gas-flow rate of 2 L/min.

We describe here first results of an on-line test of thesystem with 145-s 170W produced in the Dy(18O, xn)reaction at the Philips cyclotron of the PSI. In order toseparate 17°W and 170Ta, we used a combination of theanion exchangers Wofatit HS36 (columns F,C;logKd(W) >2), logKd(Ta) >4) and Dowex 1X8 (columnD; logKd(W) >3, logKd >4). Because only W passescolumn F, W enters column C free of Ta. The part ofTa that grows in during the retention time of W oncolumn C is strongly retained in C (Ac). The part of Wthat passes C is sorbed on column D for a time longenough to decay completely into Ta (AD). The resultsof the on-line experiments are shown in Table 1.

Table 1: Kd-values of W on Wofatit HS36 at variousHF/HNO3- concentrations.

Solution

0.01 M HNO/O.I M HF

0.05 M HNO3/O.I M HF

0.05 M HNO3/O.5 M HF

logKd fromexperiment2.73 ±0.102.85 ± 0.092.00 ± 0.082.18 + 0.072.35 ± 0.052.41 + 0.07

The errors of the Kd values are mainly due to theerrors in the weight of the ion exchange resin. Thevalues are roughly in agreement with the Ka valuesfrom batch experiments. The scatter of the values fromindividual runs is still too large and is attributed tofluctuations in the flow rate of the mobile phase whichwas not ideally continuous. It is foreseen to replace theperistaltic pump used in the present experiments by aHPLC pump to ensure a pulse-free flow of the solutionthrough the three columns in future experiments. Also,the hold-up time of the activity in the degasser must bereduced.

From the variation of the Kd with varying counter ionconcentration (NO3), a slope of -0.9 ± 0.45 results in alogKd vs. logcHNO3 plot indicating that the anionic com-plex is likely to be monovalent, in agreement with [3].Its stoichiometry might be WOF5" or WO2F3.

REFERENCES

[1] G. Pfrepper et al., Radiochim.Acta 77, 201 (1997).

[2] M. Schadel et al., Radiochim.Acta 77, 149 (1997).

[3] G. Pfrepper et al., FLNR Scientific Report 1993-1994 (1995) p. 161.

126

PRODUCTION OF 262Db IN THE REACTION 248cm(i9F, 5n)

A. Türler, ft. Dressler, B. Hehler, D.TJost, D. Piguet (PSI), Ch. Düllmann, ft Eichler, M. Gärtner,H.W. Gäggeler (Univ. Bern & PSI), M. Schädel (GSI Darmstadt), St. Taut (FZ Rossendorf),

A.B. Yakushev (FLNR Dubna)

The nuclide 262Db was produced in the reaction 24BCm(9F, 5n) at the PSI Philips cyclotron at a beam energyof 106.5 MeV. Four time correlated a-a decay chains to its daughter 2S8Lr were observed. A productioncross section of 0.26*_°0^ nb was deduced, which is more than one order of magnitude smaller than thepeak cross section in the reaction249Bk(aO, 5nf62Db.

1 INTRODUCTION

In the previous Annual Report [1] we reported the pro-duction of 262Db in the reaction 248Crn(19F, 5n) at the PSIPhilips cyclotron. Two time correlated a-a decaychains to its daughter 258Lr were observed. In the meantime the 19F beam intensity could be enhanced byabout a factor of four by using CF4 instead of SF6 inthe ECRIS source. We therefore continued our studieson the production of 262Db in the reaction 248Crn(19F, 5n)in order to measure the production cross section.

2 EXPERIMENTAL

The experimental set-up was similar to the one alreadydescribed in [1]. In the new series of experiments sixconsecutive pairs of PIPS detectors were used.

3 RESULTS AND DISCUSSION

Despite of the higher beam intensity only two furthera-a decay chains were registered. The decay timesand a-decay energies of all observed decay chains aresummarized in Table 1. A careful analysis of all eventsin the energy region 8.40-8.70 MeV revealed that only0.13 a-a decay chains unrelated to the decay of 262Dband 258Lr have to be expected. From the measureddecay times, half-lives of 27^° s and 3.8!, g s were

estimated for 262Db and 258Lr, in good agreement withliterature data. Assuming an a-decay branch in 262Db of67% the 4 correlated a-a decay chains correspond toa production cross section of 0 . 2 6 ! ^ nb at 106.5 MeVbeam energy. The production cross section in the

Table 1: a-decay energies and decay times of allobserved decay chains attributed to the decay of262Db and its daughter 258Lr.

Ea(262Db)

(MeV)

8.451

8.684

8.414

8.427

tdecay

(s)

0.88

36.66

55.68

56.06

Ea(258Lr)

(MeV)

8.556

8.596

8.532

8.572

At

(s)

8.82

8.40

0.60

0.80

248Crn(19F, 5n)282Db reaction is more than one order ofmagnitude smaller compared to the peak cross sectionin the 249Bk(18O; 5n)262Db reaction and already as smallas the production cross section for the ^ C m ^ N e ,5n)265Sg reaction. In Fig. 1 the excitation functions forthe 4n and 5n evaporation channel in the 19F+248Cmand the 1BO+249Bk reactions calculated with HIVAPSI[2] are shown and compared with experimental data[3]. The measured production cross section for 282Db inthe 24SCm(19F, 5n) reaction agrees very well with thepredicted cross section whereas the production crosssections for 263Db and 262Db in the 249Bk(18O; 4,5n)reaction are underestimated by about one order ofmagnitude.

D10 i

18O + 249Bk I 4n5n

/ , - " v ' - > , 4 n '• "~

Fig. 1:

80 85 90 95 100 105 110 115 120

Energy [ MeV ]

Calculated excitation functions for the 4n and5n evaporation channels in the reactions18O+249Bk (dotted lines) and 19F+248Cm (solidlines) using the HIVAPSI code. For compari-son the experimentally measured cross sec-tions are also shown.

REFERENCES

[1] A. Türler et al., PSI Annual Report 1997, AnnexMA, p. 48 (1998).

[2] R. Dressier, this Annual Report.

[3] J.V. Kratz et al., Phys Rev. C 45, 1064 (1992).

127

FLUORIDE COMPLEXATION OF RUTHERFORDIUM (Rf, ELEMENT 104)

E. Strub, J. V. Kratz, A. Kronenberg, A. Nähler, P. Thörle, S. Zauner (Univ. Mainz), W. Brüchle, E. Jäger,M. Schädel, B. Schausten, E. Schimpf (GSI Darmstadt), D. Schumann, U. Kirbach (TU Dresden),D.T. Jost, A. Türler (PSI), H.W. Gäggeler, R. Eichler, A. Eichler, M. Gärtner (Univ. Bern & PSI),

Y. Nagame, K. Tsukada (JAERI, Tokai), J.P. Glatz (Transuraninstitut Karlsruhe)

Previous studies had shown that Rf behaves differently from Hf and Zr and resembles Th on a cationexchanger in HNO/HF solutions. These studies were extended to higher HF concentrations. Furtherexperiments using an anion exchanger indicate that Rf forms neutral complexes at HF concentrations ofabout 0.05 M at which it starts to be eluted from the cation exchanger.

It had been shown experimentally [1] that 261Rf is noteluted from the cation exchange (CIX) column underthe conditions of experiments performed previouslywith seaborgium (Sg, element 106) [2,3] and can onlyappear in the Sg sample as a result of the decay of

Sg.

Further, the distribution coefficient (Kd) of 261Rf on acation exchange resin at a higher HF concentration(0.1 M HNO/O.OI M HF) had been studied [1]. At thisconcentration, despite Zr and Hf being eluted, Rf is stillretained on the column.

Now, the Kd values of Rf at HF concentrations of 0.03M and 0.05 M (both in 0.1 M HNO3) have been deter-mined to establish at which concentration the K,, dropsdown indicating that neutral or anionic complexes of Rfare being formed.

In addition, experiments using an anion exchanger(AIX) have been performed. It is expected that Rf willform anionic complexes resulting in high Kd values athigh HF concentrations, while Th does not form ani-onic fluoride complexes.261Rf was produced in the 248Crn(18O,5n) reaction at thePSI Philips Cyclotron. A 730 (xg/cm2 248Cm target wasbombarded with a 0.5 u A ^ ™O** beam. The targetcontained 10% Gd thus producing simultaneouslyshort-lived Hf isotopes. Rf and Hf were transported bya He(KCI) gas-jet and collected for 90 s by impactionon a slider in the Automatic Rapid Chemistry Appara-tus ARCA II. The residue was dissolved in 200 jil 0.1M HNO./X M HF (x variable) and fed onto the ion ex-change column. The effluent was evaporated to dry-ness as sample 1. In order to elute remaining Rf fromthe column, a second fraction (200 u.l) was collectedwhich is known to elute group 4 elements from thecolumn (i.e. 0.1 M HNCyo.1 M HF (CIX) or 5 MHNO/0.01 M HF (AIX)). This fraction was prepared assample 2. 78-s 261Rf was detected by oc-spectroscopy.The counting time was 12 min. Every 8th pair of sam-ples was monitored by additional y-spectroscopy todetermine the distribution of Hf.

The data (see figures) show that Rf behaves differ-ently from Zr and Hf and seems to resemble Th onboth AIX and CIX (dotted lines). On the CIX, the fall ofthe Kd values for Rf occurs between 0.01 M and 0.1 MHF, i.e. at one order of magnitude higher HF concen-trations than for Zr and Hf. It is remarkable that the K,,values of Zr and Hf on the AIX rise in the same HFconcentration range where they fall on the CIX indi

eating that the formation of anionic complexes takesplace simultaneously with the decrease of cationiccomplexes. Conversely, the Rf data suggest that Rfforms neutral complexes (type [RfFJ).

A first attempt of interpretation could be that the ionicradii of Zr"* and Hf* are very similar due to the lantha-nide contraction resulting in a similar complexationbehaviour, while in Rf, the actinide contraction is sur-passed by the relativistic expansion of the d orbitals. Athird beam time at the PSI will be used to verify andcomplete the AIX data.

3 -

?2 -

1-

0 -

- 1 -

Zr offlineHf offline

.... Th offline* Hf online (PSI)O Rf online (PSI)

frft O '"

CDC (Aminex A6), 0.1 M HN03

- 5 - 4log c.(HF)

- 1

Fig. 1 : Sorption of Zr, Hf, Th and Rf on a CIX resin2(Aminex A6) at various HF concentrations.

3-

2-

Zr offlineHf offlineTh offline

•fr Hf online (PSI)O Rf online (PSI)

0 -

- ! •

AIX (Riedel de Haen), 0.1 M HNO,

- 5 - 4 - 3 - 2log co(HF)

Fig. 2: Sorption of Zr, Hf, Th and Rf on a AIX resin(Riedel de Haen) at various HF concentrations.

REFERENCES

[1] E. Strub et al., PSI Annual Report 1997, AnnexIIA, p.6.

[2] M. Schädel et al., Nature 388, 55 (1997).[3] M. Schädel et al., Radiochim.Acta 77,149 (1997).

128

PRELIMINARY STUDIES FOR THE SEPARATION OF 263Rf IN AQUEOUSSOLUTION PART I: TRACER EXPERIMENTS

D. Schumann (TU Dresden), H. Bruchertseifer, R. Dressier, B. Eichler, D.T. Jost, A. Turler,P. Zimmermann (PSI), R. Eichler, H. W. Gaggeler (Univ. Bern & PSI)

Ion exchange on strongly acidic cation exchangers from 0.5 M HF solution should be a suitable tool forthe separation of long-lived Z63Rf from contamination of Pb, Bi and actinide isotopes and their decayproducts. Determination of the No fraction is possible after elution with 1.5M HNO3.

Recent studies of the nuclear reaction 22Ne+248Cm [1]showed evidence for the format ion of 263Rf in the a 3 n -reaction channel . Accord ing to calculat ions based onthe Viola- Seaborg fo rmula this nucl ide is expected tohave a half-life between 10 min and 1 h. Several con-taminat ion, among the decay products of long-l ived At,Bi and Pb isotopes, have cc-energies similar to 263Rf. Inorder to get a very pure Rf fract ion, separat ion of theunwanted contaminat ion by use of ion exchange inaqueous solution was chosen as a suitable method.Earlier investigations of the chemical propert ies of 261Rfin aqueous solution showed that this e lement forms -similar to its lighter homologues Zr and Hf - highlystable anionic complexes with f luoride at concentra-t ions of 0.1 M HF and higher [2]. Therefore, in0.5 M HF solution Rf should pass through a cationexchange co lumn, whereas the by-products of thenuclear reaction (mainly heavy actinides such as Fmand Md, Pb, Bi and their decay products) as well as259No, the a-daughter of 263Rf, are retained on thecat ion resin [3].

For the tracer studies the fol lowing nuclides wereused:•89Zr, as a tracer for Rf, produced by bombardment of

metall ic Y with protons and 211At produced by bom-bardment of metall ic 209Bi wi th a-part icles at the U-120 accelerator in Rossendorf . Details of the samplepreparat ion are given in [4,5],

. " R a , 212Bi and 212Pb, eluted f rom a 228Th-224Ra genera-tor [6],

•152Eu and 22Na (Na as a model for K, which comesf rom the KCI gasjet t ransport system into solut ion),

•87mSr, tracer for No, eluated f rom an 87Y-87mSr genera-tor [7]

A plastic co lumn with an inner diameter of 3 m m wasfilled with ca. 300 mg cation exchanger resin( D O W E X W 50x8, 200-400 mesh) . To improve thesorpt ion of rare earth e lements (as model for act i-nides) on the co lumn, it was first loaded with 0.2 ml10'2 M La(NO3)3 (in H2O) and then the co lumn waswashed with 0.5 M H F solut ion in order to producefreshly precipitated LaF3 . 1 ml of 0.5 M HF solutioncontaining the radionucl ides were passed through theco lumn with a f low rate of about 0.3 ml/min. Then theco lumn was eluted with 0.5 ml 0.5 M HF solut ion.These 1.5 ml of elute were measured with a HPGedetector and compared with the content of the initialsolut ion. The sorption behaviour of the elements understudy on the cation exchanger under the descr ibedcondit ions is shown in Tab . 1 . For211 At a decontamina-t ion factor of about 3 w a s obta ined.

Table 1 : Decontaminat ion factors of several e lementson D O W E X W 50x8 with 0.5 M H F solut ion.

^Na224Ra87mSr152Eu212Pb212BiB9Zr

% in the eluate

<0.7<1.0<0.05<0.2<0.15<0.5100

decontaminat ion factor

>1.5x10 2

>1x102

>2x10 3

>5x102

>7x102

2x10 2

—

In several cases upper (for eluate content) or lowerlimits (for decontaminat ion factors) are given. This iscaused by the detect ion limits. It can be seen that allcontaminat ion except At is retained in sufficientquanti ty on the cat ion exchanger co lumn, whereas Zrpasses the co lumn completely wi thout sorpt ion. Theseparat ion of astat ine requires further t reatment .Since No, the a-daughter of Rf, is expected -similar toits tracer e lement Sr - to be strongly retained on thecation exchanger, determinat ion of its content wouldgive addit ional information on the cross sect ion of thenuclear reaction and the half-l ife of 263Rf. Sr can beeluted from the column with 1.5MHNO3, whilelanthanides still remain on the column at this acidconcentration. In Fig. 1 the chromatogram for theelution of Sr is shown.

0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5

elution volume [ml]

Fig. 1 : Elution of Sr from DOWEX W 50x8 with 1.5 MHNO3.

This work was supported by BMBF, Germany.

REFERENCES[1] R. Dressier, Ph.D. thesis, Bern 1999.[2] Z. Szeglowski et al., J. Radioan. Nucl. Chem. Lett.

186,353(1994).[3] H. Bruchertseifer et al., JINR Rap. Comm. 1 [34]

(1989)50.[4] D.Schumann et al., Radiochim. Acta 72 (1996) 137.[5] Gmelin, Astatine, 8th Ed. Berlin, Springer 1985.[6] F.Sebesta et al., J.Radioanal.Chem. 21 (1974) 151.[7] R. Su(3 et al., Zfl Report 81 (1989), 5.

129

PRELIMINARY STUDIES FOR THE SEPARATION OF 263Rf IN AQUEOUSSOLUTION PART II: ON-LINE EXPERIMENTS

D. Schumann (TU Dresden), H. Bruchertseifer, R. Dressier, B. Eichler, D.T. Jost, A. Turler,P. Zimmermann (PSI), R. Eichler, H. W. Gaggeler (Univ. Bern, PSI)

The separation of short lived Hf from lanthanides with strongly acidic cation exchange columns in 0,5HF solution was studied. An average yield of 85 % Hf can be achieved with a separation time of aboutone hour. Hf isotopes were produced in the reaction f5SGd(BO, xn)170"" Hf.

In part I [1] a chemical system was developed toseparate rutherfordium from the by-products of thenuclear reaction, mainly heavy actinides, Pb and Biisotopes and from its a-decay daughter No. In tracerexperiments it could be shown that decontaminationfactors of about 500 and more can be reached, usingion exchange on strongly acid cation exchangercolumns and 0.5 M HF solution as eluent.

In the present work this separation system was appliedto a model experiment with short-lived Hf isotopes.The reaction 158Gd (18O, xn)17Onx Hf was used toproduce Hf isotopes with half-lives from 2-26 min.Their decay products Lu and Yb can serve as tracernuclides fortrivalent actinides.

For the separation columns filled with ca. 300 mgDOWEX W 50x8, 200-400 mesh, were used whichwere treated as described in [1].

The nuclear reaction products were transported fromthe target chamber to the laboratory by a KCI gas-jetsystem with an yield of 55%. There it was collected ona Nucleopore filter for one hour. Then, the reactionproducts were washed from the filter with 0.8-1.0 ml0.5 M HF solution and passed through the cationexchange column with a flow rate of about 75 u,l/min(about 3 drops per minute). The eluate, containingonly the Hf fraction, was evaporated to dryness in aquartz vessel and dissolved in about 150 uJ of aspecial solution (40 % alcohol, 1 M HCI, 0.5 M HF, 5Bq/ml 241Am tracer). To simulate the real experiment,samples suitable for 47t-a-measurements were pre-pared by evaporating this solution on thin polypropyl-ene foils (40 p.g/cmz).

80 100 120 140 160 180 200

Energy [ keV ]

Fig. 1: y-spectrum of the starting solution (afterwashing from Nucleopore filter).

These foils were measured with a y-spectrometer. Thewhole procedure from washing the filter untill thebeginning of measurement took about 25 min.

In Fig. 1 the y-spectrum of the solution after washingfrom the filter is depicted. Fig. 2 shows the y-spectrumof the Hf fraction after separation on the cation ex-change column. The isotopes 167Lu and 167Yb wereused as models for heavy actinides. They are pro-duced as decay products of 2.05 min 167Hf, which hascompletely decayed before the end of the separation.Fig. 2 shows that no evidence for these isotopes canbe found in the Hf fraction. In Tab. 1 the count rates ofsome characteristic isotopes in the starting solutionand the Hf fraction are listed. The values for 16BHf arehalf-life corrected.

Table 1 : Counting rates of some isotopes.

energy filter

solution Hf

fraction

yield

""Yb/^Hf

113 keV

25600

62

0.2%

167Yb/«"Hf

106 keV

11000

...

<0.5%

™Hf

157keV

5778

5125

88.7%

i « H f

117 keV

6527

5400

82.7%

We achieved an average yield of 86% for Hf and andecontamination factor of at least 200 for lanthanideswhich is in good agreement with the results found forZr and Eu in the tracer experiments [1], described inpart I.

This work was supported by BMBF, Germany.

80 100 120 140 160 180 200Energy [ keV ]

Fig. 2: y-spectrum of the Hf fraction (after separationon DOWEX W 50x8)

REFERENCE[1] D. Schumann et al., this Annual Report.

130

THERMOCHROMATOGRAPHIC INVESTIGATIONS OF 254Es BROMIDES

U. Graf (Univ. Bern), B. Eichler (PSI), A. Turler(PSI), H.W. Gaggeler (PSI & Univ. Bern),R.G. Hake (ORNL)

The adsorption of EsBr3 on quartz surfaces was investigated by the thermochromatographic method usingbrominating carrier gases. The enthalpies of adsorption were evaluated from the experimental data.

1 INTRODUCTION

Thermochromatography has been successfully appliedto investigate the volatility and adsorption behavior ofthe relatively low-volatile chlorides of transplutoniumelements. Investigations of Am, Bk and Es trichlorideswere recently carried out [1,2]. Very similar AHR

om

values resulted for all three elements. In experimentswith Bk under strongly oxidizing conditions a secondcompound which deposited at lower temperatures wasobserved. This was attributed to the formation of BkCI4and AHa

0(T)(BkCI4) = -150+5 kJ/mol was determined [2].In this work thermochromatographic investigations ofEsBr3 were conducted for the first time.

2 EXPERIMENTAL

The thermochromatographic device consisted of twosections. The first section served as evaporator. In thesecond section a temperature gradient was realized byheating one end of a metallic tube and cooling theother end with water. The gradient ranged from1000°C to 25°C. As thermochromatographic columnsempty quartz tubes with a length of 100 cm and aninner diameter of 2.8 mm were used. The columnswere pre-treated at 1000°C during 1h with HBr at aflow rate of 25 ml/min. Samples of 254Es (in nitric acidsolution) were prepared by placing two drops (~ 40 \i\,-2-1010 atoms of 254Es) on a strip of quartz fibre paperand dried over H2SO4 in a desiccator. The driedswatches were inserted into the chromatographiccolumn and EsBr3 was formed with HBr at a flow rateof 25 ml/min by inserting the column in a short reactionoven kept at 1000°C. EsBr3 was deposited in a narrowdeposition peak at the exit of the reaction oven. Thecolumn containing EsBr3 at the starting position wasnow introduced into the thermochromatography ovenusing HBr as carrier gas. After completion of theexperiment, the distribution of the activity in thecolumn was measured using an automatic columnscanner equipped with two BGO crystals. By waiting24h after completion of the experiment the depositionzone of 254Es could be determined by measuring the y-rays emitted by the 3.2 h a50Bk daughter nuclide.

3 RESULTS AND DISCUSSION

The experimental conditions and the observeddeposition temperatures are summarized in Table. 1.

Table 1: Experimental conditions and results of thethermochromatographic experiments with EsBr3.

Species Br2 Gas

flow rate

(ml/min)

EsBr3 11.7

11.7

17.6

17.6

17.6

17.6

31.1

31.1

Duration

(min)

60

120

60

120

240

360

60

120

Depositiontemperature

(K)

773

864

773

728

728

728

728

708

Entropy ofadsorption

(J/molK)

-168.1

-167.7

-168.1

-168.4

-168.4

-168.4

-168.4

-168.5

Enthalpy ofadsorption

(kJ/mol)

-181

-20611

-185

-179

-183

-184

-178

-177

-181±3

1) Value omitted for the calculation of the mean AH°m.

The current measurement shows that similar to thetrivalent lanthanide bromides, also the trivalentactinide bromides are more volatile than thecorresponding chlorides. Under the assumption thatthe empirical correlation established for chlorides [3]

-AWaom=(0.600±0.025)AHs

0(298)+(21.5±0.25) [kJ/mol]

is also valid for bromides, a sublimation enthalpyAA7s

0(298)(EsBr3)=266±15 kJ/mol can be estimated.

ACKNOWLEDGMENT

The authors are indebted for the use of 254Es to theOffice of Basic Energy Sciences, U.S. Department ofEnergy, through the transplutonium elementproduction facilities at the Oak Ridge NationalLaboratory, managed by Lockheed Martin EnergyResearch Corporation.

REFERENCES

[1] J. Adams et al., PSI Annual Report 1997, AnnexMIA, p. 62(1998).

[2] A. Yakushev et al., PSI Annual Report 1997,Annex 11IA, p. 63(1998).

[3] B. Eichler et al., Report JINR, P12-9454, Dubna(1976).

131

HIVAPSIA PROGRAM TO ESTIMATE CROSS-SECTIONS IN HEAVY ION REACTIONS

R. Dressier (PSI)

The code of the well known program HIVAP was adapted as a 32-bit application and is now available underMS DOS, Win 3.11, Win 95, and Win NT. It was found, that the original code exhibits a series of numericalerrors leading to inaccurate results.

One of the widely used programs to calculate fusion-evaporation cross-sections in heavy-ion inducednuclear reactions is the HIVAP code. It was developedin 1981 by W. Reisdorf [1,2] at GSI. After the shut-down of the IBM 3090-60J at the end of 1994 the sup-port of MVS/TSO ended 1996. The code was trans-ferred to the UNIX platform at GSI by F.P. HeBberger.Unfortunately, the results are in disagreement with thecalculations of the original HIVAP code.For the design of our heavy element experiments it isimportant to estimate the production cross-sectionswith the highest possible accuracy. In addition, thepossibility to carry out model calculations independ-ently of the GSI infrastructure is highly desirable.

The HIVAP source code was adapted as a 32 bit appli-cation for all MS-PC platforms. During this adaptions aseries of programming errors were corrected, mainly inthe use of single and double precision variables.The new program code, called HIVAPSI, exhibits agood agreement with experimental data (Tab. 1). Thepeak energy and maximum cross sections are com-pared with experimental data taken from [3-8].One reason for this good agreement lies in an im-portant difference between HIVAPSI and the originalHIVAP code. HIVAPSI generates higher overall fusioncross sections and the peak energy for the evapora-tion residues is lower. This behaviour is shown inFig. 1.

75 80 85 90 95 100 105 110 115

75 80 85 90 95 100 105 110 115

Fig. 1: Comparison of fusion and evaporation resid-ual cross-sections in the reaction 18O+248Cmcalculated with the original HIVAP-code(dotted line) and HIVAPSI (solid line).

The discrepancies between experimental and calcu-lated cross-sections for neutron-rich isotopes of heavi-est elements might be caused by the N=162 andZ=108 shell-closure and by higher fission barriers [18]compared to those assumed in HIVAPSI.

Table 1 : Comparison of calculated cross-sections withexperimental data taken from [3-17].

Reaction

232Th(22Ne, 4 n f ° F m238U(16O, 4 n f ° F m

242Pu(12C, 4n)250Fm246Crn(16O, 4n)258Rf248Crn(16O, 4n)260Rf249Bk(15N, 4n)260Rf

248Crn(18O, 5n)261Rf2 4 8 CmfO, 4n)262Rf244Pu(22Ne, 5n)261Rf244Pu(22Ne, 4n)262Rf

249Cf(15N, 4n)260Db249Bk(18O, 5n)262Db

249Bk(18O, 4n)263Db2 4 9 C f f 0 , 4n)263Sg

248Crn(22Ne, 5n)26SSg248Crn(22Ne, 4n)266Sg248Crn(22Ne, 4n)266Sg

' -calo

[MeV]

103

85

6891

89

79

97

88

114

108

8398

90

94121

110121

" H I V A P S I

[nb]

5901100

11000

14

7.3

19

4.47.1

1.8

1.9

3.8

0.89

1.2

0.62

0.0360.0700.006

[nb]

3001500

8000

10+1°3

6 + 1

20 ±1

5

5 ± 11 5 + 2 6

' - ° -0 81 ±1

3

6 + 3

10 + 6

0.90.24....

0.025

The author thanks Dr. F.P. HeBberger, Dr. M. Schadeland E. Stiel for their support at GSI and the helpfuldiscussions.

REFERENCES[I] W. Reisdorf et al., Z.Phys. A300 (1981) 227.[2] W. Reisdorf M. Schadel, Z.Phys. A343 (1992) 4.[3] A. Ghiorso et al., Phys. Rev. Lett. 22 (1969) 1317.[4] A. Ghiorso et al., Phys. Rev. Lett. 24 (1970) 1498.[5] A. Ghiorso et al., Phys. Lett. 32 B (1970) 95.[6] W. Neubert Nucl. Data Tab. 11 (1973) 531.[7] A. Ghiorso et al., Phys. Rev. Lett. 33 (1974) 1490.[8] V.A. Druin et al., Sov. J. Nucl. Phys. 29 (1979) 591.[9] J.M. Nitschke et al., Nucl. Phys. A 352 (1981) 138[10] C.E. Bemis et al., Phys. Rev. C 23 (1981) 555.[II] L.P. Sommervilleetal., Phys. Rev. C 31 (1985) 1801.[12] N. Shinohara et al., Phys. Rev. C 34 (1986) 909.[13] J.V. Kratz et al., Phys. Rev. C 45 (1992) 1064.[14] Y.A. Lazarev, Int. Proc. Int. Workshop, Hirschegg

1996, p 11.[15] Yu.A. Lazarev et al., WS - IGISOL - 6, Dubna

1997, p123.[16] V.K. Utyonkov et al., VI Int. School-Seminar,

Dubna 1997.[17] A. Turler et al., Phys. Rev. C 57 (1998) 1648.[18] S.I. Bastrukov et al., J. Phys. G 24 (1998) L1.

132

CALCULATION OF CROSS SECTIONS IN cc-xn-REACTIONS

R. Dressier (PSI)

The break-up fusion model was used to predict the cross sections in a-xn reactions.

The prediction of the production cross sections in in-complete- or break-up fusion reactions with pure sta-tistical evaporation codes leads to non satisfying re-sults. The previously described HIVAPSI code [1] e.g.underestimates the cross sections in a-xn reactions bya factor of about 200.To design search experiments for the synthesis ofneutron-rich isotopes at the limit of the chart of nu-clides, a more accurate method to predict these pro-duction cross sections is needed.The break-up fusion model [2,3] assumes that theprojectile is broken up by coulomb excitation into an a-particle and a projectile rest. Following the calculationsin [4], the energy distribution for the emitted a-particlesis:

6Pd E « { (m P - ma) • Q a + ma • (Ep - ZPEC) + mp • (Ea - ZJEC))

-Ama • (Ep - ZPEC) • mp • (£„ - ZaEc)

ER=EP-Ea-Qa-ZaEc

(1)

4re • 60 • (rT +rp + 2.7fm)

where: mp - mass of projectile, Zp - atomic number ofprojectile, Ep - energy of projectile, rp - radius of pro-jectile, ZT - atomic number of target atom, rT - radius oftarget atom, ma - mass of the a-particle, Ea - energy ofthe a-particle, £„ - energy of projectile rest, Ec -relative coulomb energy, Qa - binding energy of the a-particle in the projectile.To simplify the calculation it was assumed that thebreak-up takes place at the interaction distance rM

(rinl=rT+rp+2.7im [5]). It was further assumed that com-plete fusion of the target and the projectile rest takesplace only if the a-particle is emitted away from thetarget nucleus. Then the production cross section a ^of an a-xn reaction can be expressed as

a

Er=EP-Qa-ZaEc

where E™ is the maximum energy of the a-particle.The cross section axn of the neutron evaporation

reaction of the target with the projectile rest can becalculated e.g. with the HIVAPSI code [1].In Fig. 1 the results of these calculations for the reac-tion 248Crn(18O; a3n)259No are compared with experi-mental data [6]. The Qa value is 6.23 MeV.If the binding energy of the a-particle is higher, Eq. (2)must be modified to take into consideration that theprojectile must approach closer to the target nucleus.Therefore, the break-up will frequently take place atcloser distances than rlnt. This will increase the prob-ability to overcome the interaction barrier and can betaken into account by calculating the ratio rtan. of the

penetrability of the barrier for the projectile and theprojectile-rest nucleus. Eq. (2) can now be rewrittenas:

(3)dE„-r,barr

In Fig. 2 the cross sections calculated with this modelare compared with experimental data in the reaction24*Pu(22Ne, a3n)259No [6,7]. Here, Qa is 9.67 MeV, about3 MeV higher than in the first example. Obviously, onlya part (about 10 %) of all break-up fusion reactionstook place at distances closer than rlnl. The logarithmicmean of Eq. (2) and (3) leads to a reasonableagreement with experimental data.

10*-

80 85 90 95 100 105 110 115 120E [MeV]

Lab

Fig. 1: 248Crn(18O; a3n)269No: • - experimental data [6],solid line - Eq. (2), dotted line - a 3n evapo-ration calculated with HIVAPSI.

102-

c

D

10-2100 105 110 115 120 125 130 135 140

Fig. 2:

E [ MeV'Lab

•2 5 9No:S44PufNe; a3n)[6,7], dashed upper

• - experimental dataline - Eq. (3), dashed

lower line - Eq. (2), solid line - logarithmicmean of Eq. (2) and Eq. (3), dotted line - a3nevaporation calculated with HIVAPSI.

REFERENCES[1] R. Dressier, this Annual Report.[2] A.K. Kermann, K. McVoy, Ann. Phys. 122 (1979) 197.[3] T. Udagawa, T.Tamura, Phys. Rev. Lett. 45 (1980)1311.[4] N. Matsuoka, et al., Nucl. Phys. A 311 (1978) 173.[5] R. Bass, Nucl. Phys. A 231 (1974) 45.[6] R. Dressier, Ph.D. thesis, Bern 1999.[7] R. Dressier et al., PSI Annual Rep. Annex MIA 1996, 84.

133

CORA - A NEW CONTROL PROGRAM FOR THE ROMA DETECTION SYSTEM

St. Taut (FZ Rossendorf)

A new computer program has been developed for the control of the GSI Rotating Wheel Multi DetectorApparatus (ROMA) running under the Windows 95 and Windows NT 4.0 operating systems. It can be fittedto a lot of detection tasks in a highly flexible way.

The ROMA detection apparatus [1]has been successfully used in a lot ofchemical investigations of heavyelements. In preparation of ourseaborgium experiment at GSI insummer 1998 [2] a new controlprogram had to be written, becausethe old computer code was not able tofulfill our experimental demands. Thenew program is written entirely in C++using the Borland C++ Builder 1.0programming environment.

The program has to carry out thefollowing tasks:• control of the wheel movements;• start and stop of the nuclear

spectroscopy data acquisition;• sending information about the

actual ROMA status to the nuclearspectroscopy data acquisitionhardware;

• processing control requests of thedata acquisition hardware (e. g.switching to "daughter mode" [2]);

• processing user input before and during anexperiment (e. g. experiment and wheel setupdialog boxes).

The program consists mainly of four software modulesas shown in Fig. 1. Hardware_IO.dll implements aninterface for all interactions with the hardware. The

Slop J

••. i

/ : "••!

I*

1 ^ - 4

{ »'° -,:

.... i

r-. Er- - 2 J

J" -."• n SJ

- . . . * .• ^ § * ' > J

<

i :

1.

•

V!1

it |

, - • * /

' . - • ! - • • ; .•

• •i n! . '<[•:<

S' ' H i '

c ,.. t. -„...' i '1 •

. !e: t T

>! • ; :! . " M

_ 3> "fr'C- K v - r

' . . " ' • " *

-..-i- ... ,

. * •

'.'.'.-Hi

•

Fig. 2: Main Window of the Program

MainWindow.cpp: Main application window,event processing, and user interface

ExpCtrl

AppCtrl

cpp:

xpp:

iExperiment flow control

iLogical hardware control

Hardware_IO.dll: implementation of thehardware control

Operating System: Windows 95 / NT 4.0,including device drivers for the interfaces

other parts, organizing the experiment on a moreabstract level, are using exclusively this interface forcontrolling the hardware.Thus, the program can be adopted to similar detectionsystems with very low programming effort: onlyHardware_IO.dll must be adapted to the otherhardware. This is planned for the new GSI multidetector apparatus which is currently in development.

The program has a very comfortable user interface.Fig. 2 shows the Main window. It is almost selfexplanatory. Control can be accessed in an easy wayeven by not experienced users using dialog boxes.Critical hardware settings are hidden, but all programsettings can be modified in initialization files (Windows3.x style) with a normal editor.It is possible to control the experiment related parts ofthe program solely by the digital I/O interface withoutdirect user input. Thus, the program can be used infully automated systems.

This work was supported by BMBF, 06DR824

REFERENCES[1] K. Summerer et. al., GSI Annual Report 1983,

84-1,246(1984).[2] S. Hubener et. al., this Annual Report.[3] A. Turier et. al., Phys. Rev. B. 57, 1648 (1998).

Fig. 1: Simplified Layer Scheme of the Program.

134

A LINEAR SCANNING DEVICE FOR ON-LINE RADIOACTIVITY MEASUREMENTS

D.T. Jost, D. Piguet, M. Ammann (PSI), H. Muller (Muller Prazisionstechnik GmbH, Klettgau)

A linear scanning device with a usable length of 1m was constructed. The scanning head can be equippedwith a pair of BGO-detectors for coincident measurement of positron annihilation radiation distributionswith a resolution of 20mm or a Ge-detector for general y-activity distributions.

1 INTRODUCTION

In a lot of experiments like thermochromatography andadsorption of reactive species on surfaces thedistribution of the species along a column is of greatinterest. If the species is labelled with a positronemitting isotope a coincidence scanner is a verypowerful tool to measure such distributions. Theannihilation radiation from the positron emission allowsfor an easy determination of the position of the labelledspecies in the column.

2 CONSTRUCTION

Mechanical equipmentA linear motion unit NADELLA MB-BB16 with astepper motor is used to move the detector head alongthe column. The maximum scan length is 1 m. Thehead can be moved with a speed of 200 mm/s. Thestepper motor can be moved in 2.5|im increments.The accuracy for the positioning is better than 1 mm. Itcould be further improved by adding a position sensorfor the feedback controller loop. With different supportseven LN2 cooled columns can be used.

Detectors and electronicsFor the application with positron emitting nuclides acoincidence arrangement with two 1x1" BGO-detectors (Crismatec 25SEA25/BGO) mounted face toface with a gap of 25mm is used. The detectorsoperate with a high voltage of 800 V. The output signalis amplified with an Ortec 474 Timing Filter Amplifierfor maximum time resolution. Discrimination againstnoise and signals from Compton scattering isperformed with a SIN 101-A constant fractiondiscriminator. The use of this fast electroniccomponents leads to a pulse resolution of 50ns,allowing for a single count rate in excess of 1 MHz anda very low random coincidence rate. Pulse from thediscriminator as well as from the coincidence unit arecounted in an Ortec 974 4 channel counter(coincidence, two single and timer).

Fig. 1: Detector and electronics of the coincidencescanner. BGO = bismuthgermanate detector,TFA = timing filter amplifier, CFD = constantfraction discriminator, AND = logical AND unit.

SoftwareA small FORTRAN program was written to control thescanner. Fortran was chosen because thecommunication with the sealer module was alreadyavailable and the program has to run on a DEC VMS-system since the y-spectroscopy software to operatethe Ge-detector head is running on this system andhas a FORTRAN callable interface. A simplecommand line interface for the setting of the scanparameters step-width, counting time per step, startposition, number of repetitions and name of the datafile is provided.

3 PERFORMANCE

The device has been used so far with the BGO-coincidence detector head. For a detector gap of25mm and a point source of 5 mm diameter thecoincidence yield is 12.5%. The background reductionfor this arrangement is 1000 giving a 100 timesimproved signal to noise ratio. The linear resolution isbest demonstrated in Fig. 2. The left peak is the signalfrom a 10 mm wide strip perpendicular to the traveldirection, the middle is a point source of 5 mmdiameter and the right one is a 120 mm long zone.The width of the three peaks at half height is 40mm forthe left, 20mm for the middle, and 120mm for the rightfeature. The width is to a large degree determined bythe step width of 15mm of this experiment. The ratio ofthe applied activity to the three sources was 2:1:2. Theintegration of the three peaks in the plot yields the ratioof 7:5.7:10.5 for the single counts from the detectorswhereas for the coincidence signal the correct ratio isfound.

16000

14000 •

12000 •

10000

aooo

6000

4000

2000-

020

1800

1600

1400

1200 |

1000 O1

BOO 3

600 S

400

200

0SO 60 70

Position [cm]

100

Fig. 2: Example of scanner output with three differentactivity sources. Solid line for coincidencecounts, dashed line for single counts.

135

EFFICIENT TRANSPORT OF BROMINE AND SELENIUM NUCLIDES FROM THER-MAL NEUTRON INDUCED FISSION OF 235U USING THE SINQ GAS-JET FACILITY

M. Wachsmuth, M. Ammann, L Tobler(PSI), H.W. Gaggeler (Univ. Bern & PSI)

With the aim to use short-lived bromine nuciides in tracer studies related to atmospheric chemistry, theSINQ gas-jet facility was used to transport Br and its precursor Se selectively and efficiently as productsfrom nuclear fission of 235U. Selectivity was achieved by adding small amounts of carbon monoxide orpropene to the carrier gas.

Bromine containing compounds are thought to play animportant role in the ozone destruction in variousregions of the atmosphere. Barrie et al. [1] firstreported a relationship between springtime ozonedepletion in the lower Arctic troposphere with a strongnegative correlation between the concentration ofozone and bromine compounds. Short-lived tracers foraerosol experiments have already been usedsuccessfully in the heterogeneous chemistry ofnitrogen oxides [2].

Short-lived bromine nuciides are generated in thenuclear fission of 235U, which had been placed in thethree target chambers of the SINQ gas-jet facility [3].Upon irradiation with thermal neutrons, the fissionproducts are emitted into the adjacent chamber, whichis continuously purged by a gas flow, and stopped tothermal velocity. For the experiments described infollowing, the target with the 6-|i,m Ni degrader foil,which suppresses the heavy mass peak, was used.The classical gas-jet approach uses an aerosol fortransport, in which all nuciides efficiently are attachedto carbon or salt aerosols and transported to the lab.However, the separation in the laboratory may bedifficult, and the transport to the 70 meters distantlaboratories is non-selective. When the aerosol isomitted to avoid all non-volatile fission products, onlythe fission noble gases are swept to the laboratory,and, for instance, in pure helium, no detectableamounts of Br isotopes were observed. Therefore,traces of a reactive gas were added to catch Br andSe selectively. Carbon monoxide (CO) and propene(C3H6) were chosen because no radical initiatedpolymerisation reaction of these two gases are known.

Various trapping systems and geometries at differenttemperatures were tested: Silver is known as an idealtrapping substance for halides and halide-likeelements and compounds. Therefore, silver wool anda silver filter, likewise at room temperature, were firsttried. Since mainly fission noble gases and their decayproducts were visible, the trapping system waschanged to silver coated glass globules, which wereheated. By using this system, we reached the bestresults by increasing the reactive gas concentrationsup to 600 ppm in helium.

It was observed that CO and C3H6 react moreefficiently with selenium than with bromine atoms. Themost likely transport form is carbonyl selenide [4]

(COSe) and propane selenal [5] (C3H6Se). Therefore itwas decided that the "generator approach" is thepreferred technique to get a high output of Br nuciides:the transport products were cracked in a 800°C ovento elemental selenium, which was hold on quartz woolin a distance of 13 cm from the oven outlet. Theremaining nuciides, i.e. the fission noble gases andtheir decay products, were trapped on activatedcarbon. Only approximately 7 % of all Se nuciidespassed the quartz wool. Most of the bromine nuciideswere retained on the quartz wool as well. By varyingits temperature and perhaps the gas composition, theseparation of Br from Se will be addressed in the nextsteps.

700

500 1000

Energy [keV]

1500 2000

Fig. 1: y-spectrum of fission products transportedusing 600 ppm CO in the gas-jet selective CO-transport, collected on quartz wool afterpassing a 800 °C oven. 1:90Kr, 2: "Se, 3:101Tc,

0 M ^4: 5: 104Tc, 6: '"Se, 7: MSe, 8: 9:annihilation (511 keV) gammas, 10: '"Br, 11:90Kr, 12:86Br

ACKNOWLEDGEMENTThis project is supported by the Swiss NationalScience Foundation.

REFERENCES[1] LA. Barrie et al., Nature 334, 138 (1988).[2] M. Ammann et al., Nature 395, 157 (1998).[3] D.T. Jost et al., PSI Annual Report 1997, Annex

MIA, p. 20.[4] R.S. Baldwin et al., Radiochim. Acta 16, 94

(1971)[5] G.A. Krafft, J. Amer. Chem. Soc. 108/6, 1314

(1986).

136

SYNTHESIS OF AN ARCTIC HALOGEN RESERVOIR SPECIES:HYPOBROMOUS ACID (HOBr)

M. Wachsmuth, M. Ammann (PSI), H.W. Gaggeler (Univ. Bern & PSI)

Hypobromous Acid (HOBr), an important arctic halogen reservoir species and known as an atmospherictrace compound with high ozone depletion potential, was synthesized according to three different methods.

HOBr plays an important role in the arctic springtimeozone depletion [1] The main route of the formation ofHOBr in the atmosphere is the reaction of HO2 withBrO:HO2 + BrO->HOBr + O2 (1)The hydrolysis of BrONOa,BrONO2 + H2O -> HOBr + HNO3 (2)on or in aerosol particles also leads to HOBr formation.Destruction of HOBr in the lower atmosphere is domi-nated by photolysis:HOBr + hv -> OH + Br \,reshold = 619.8 nm (3)The Br atoms enter the catalytic ozone depletion cycle.

In view of future experiments with radioactivelylabelled HOBr (e.g. HO^Br), we tested three differentmethods of HOBr synthesis.

1 SODIUM BROMATE METHOD [2]

NaBrO3 + H2SO4 -> HOBr + O2 + NaHSO4

which can further decompose to bromine:(4)

(5)4 HOBr -» 2 Br2 + 2 H2O + O2

Experimental procedure:

2 cm3 of a 1 M aqueous NaBrO3 solution were addeddropwise to a continuously stirred mixture of 3 cm3 ofH2O and 5 cm3 of H2SO4 (96% w/w) at -10 °C. Themixture was slowly warmed to 40 °C. When the vigor-ous oxygen evolution ceased, the reaction mixture wasdiluted to 50 cm3 with deionisated water. The solutionwas freed from Br2 by four successive extractions withCCI4.

2 SILVER NITRATE METHOD [3]

AgNO3

BrONO2

Br2 -> BrONO2

H2O HOBrAgBrlHNO3

(6)(7)

Experimental procedure.

In a typical run, 9 g AgNO3 were dissolved in 50 cm3

H2O and 2.5 cm3 Br2 were added dropwise at roomtemperature. The resulting HOBr (in aqueous solution)was vacuum distilled (0.1 mbar) into a separate vesselcontaining some stabilising H2SO4 (20% w/w), where itwas stored at -40 °C. This sample was then purified byrepeated washing with CCI4. An UV/VIS spectrum wastaken immediately after the distillation (Fig. 1).

3 MERCURY OXIDE METHOD [4]

HgO -> Br2O + HgBr2

Hg°2 Br2HgBr H2O

Br2OHOBr + HBr

(8)(9)

Experimental procedure:

A continuous slow flow of Br2(g) and H2O(g) through apyrex glass reaction chamber, equipped with an inletand outlet as well a separate water vapour inlet and aloose package of yellow HgO, was maintained at apressure of 0.11-0.13 mbar. Separation of the reactionproducts was realised by three cooling traps in series,where the resulting compounds were collected at dif-ferent temperatures.

4.8866

IAbs)

-0.888

A.r M

*

J\\

V

x. k*K—"

196.B Have length (niu 809.8

Fig. 1: UV/VIS-spectrum of HOBr: absorption at 268nm, 286 nm and saddle point at 320 nm.

Table 1: Example of a typical separation run, whichwas repeated several times with different tempera-tures with the intention to optimise the separation effi-ciency.

Cooling trapnumber

12

3

temperature[°C]-35-65

-198

condensedproduct

H2OHOBr, traces

ofBr2

Br2O, HBr

ACKNOWLEDGEMENTThis project is supported by the Swiss NationalScience Foundation.

REFERENCES

[1] J.J. Orlando, J.B. Burkholder, J. Phys. Chem. 99,1143(1995).

[2] Z. Noszticzius, E. Noszticzius, Z.A. Schelly, J.Am. Chem. Soc. 104, 6194 (1982).

[3] Crowley, J.N., personal communication.[4] G.A. Me Rae, E.A. Cohen, J. Mol. Spectrosc. 138,

467(1989).

137

HONO FORMATION ON SOOT AEROSOL PARTICLES - INFLUENCE OF HUMIDITYAND OZONE

M. Kalberer, H. W. Gaggeler (Univ. Bern & PSI), M. Ammann, F. Arens, U. Baltensperger (PSI)

The reaction of NO2 and soot aerosol particles to HONO was investigated as a function of relative humidityand ozone.

Nitrous acid (HONO) is an important compound inatmospheric chemistry. On daytime, the photolysis ofHONO produces OH radicals, the most importantoxidant in tropospheric chemistry. Thus HONO is animportant factor initiating photochemical smog.

In a recent study we measured the kinetics of theheterogeneous reaction of NO2 to HONO on laboratorygenerated soot aerosol particles by use of the 13Ntracer technique [1,2]. These experiments showed avery fast initial reaction of NO2 to HONO on the sootsurface according to reaction (1)

at higher rh making the reactive surface inaccessibleor less reactive for NO2.

{C}re HONO + {C}O reaction (1)

In this study we investigated the influence of possiblecompeting gas molecules, i.e., H2O and O3, on thisreaction to assess the importance of HONO formationon soot under more realistic ambient atmosphericconditions [3].

1e+15

10% 20% 30% 40% 50% 60% 70%

relative humidity

Fig. 1: HONO production as a function of the relativehumidity from 2% to 70% at 20 ppb NO2 and20 s reaction time.

The HONO production shows a non-linear behavior asa function of the relative humidity (rh) between 2 -70% rh (Fig. 1). Starting at low rh the amount ofHONO formed increases with increasing rh up to 30%rh suggesting that H2O is either directly involved in thechemical reaction of NO2 to HONO or builds up aprecursor (e.g. by hydrolysis) which is the reactionpartner for NO2. Above 40 % rh the amount of HONOformed decreases again. This decreasing HONOproduction with higher rh may be due to the fact thatthe particle surface becomes covered with liquid H2O

0 20 40 60 80 100

ozone dose ^concentration • time) [ppb-min]

Fig. 2: Relative reduction in HONO production as afunction of the O3 dose applied to the aerosol.

Aging of the soot aerosol by O3 lowers its HONOformation potential dramatically (Fig. 2(b)), so thatafter an O3 dose of 105 ppb-min less than 10% ofHONO is formed compared to a fresh aerosol. Incontrast, simultaneous mixing of O3, NO2 and theaerosol affects the HONO formation only slightly (Fig.2(a)). These experiments show that O3 is able tooxidize the same surface sites on the soot particles asNO2. However, it seems that the kinetics is faster forNO2 than for O3.

Summarizing, these results show that other gas phasemolecules may significantly influence the reaction ofNO2 to HONO on soot aerosol particles. Thus, forambient conditions the influence of O3 on HONOproduction is rather small.

ACKNOWLEDGMENT

We appreciate the stable proton beam of the Philipscyclotron provided by the staff of the PSI acceleratorfacilities. This work was supported by the Kommissionfiir Technologie und Innovation (KTI).

REFERENCES

[1] U. Baltensperger et al., J. Phys. Chem. 97,12325-12330(1993).

[2] M. Ammann et al., Nature, 395,157-160 (1998).

[3] M. Kalberer et al., J. Geophys. Res., submitted.

138

FORMATION OF HONO FROM THE REACTION OF NO2 WITH DIESEL SOOT

F. Arens, M. Ammann, U. Baltensperger (PSI), M. Kalberer, H. W. Gaggeler (Univ. Bern & PSI)

Soot aerosol particles prepared from a graphite spark discharge generator have been shown to be an effi-cient substrate for the conversion of NO2 to HONO, an important trace constituent of the atmosphere. Inthis study, this reaction is measured with samples of soot from the spark discharge generator and withsamples of soot particles extracted under various conditions from diesel exhaust.

Nitrogen oxides are important constituents of theatmosphere and play a crucial role in ozone formation.Thereby, nitrous acid (HONO) may be able to startphotochemistry in the morning due to its photolysis toNO and OH radicals by the rising sun. The currentunderstanding of the HONO formation mechanisms isstill uncomplete. It is assumed that it forms heteroge-neously from NO2 on ground or airborne surfaces. Inour previous work we have shown that NO2 can reactvery efficiently to HONO on the surface of soot aerosolparticles and therefore have provided a possibleexplanation of HONO formation in polluted air masses[1]-

Most previously reported results on the NO2 soot reac-tion have been obtained using a commercial sootsource, namely a spark discharge generator (PALASGfG100). Although the resulting particles chemicallyand physically resemble real soot particles with re-spect to content of organic carbon, carbon and oxygenvalency, fractal dimension, diameter and other pa-rameters [2], similar experiments with real combustionsoot are of course highly desirable. The main difficul-ties arrising from extracting soot particles directly fromthe exhaust of a diesel engine are the very high con-centrations of NOX and volatile organic compounds(VOC) which both would heavily interfere with ourdetection systems. Therefore, a novel diluting device(Matter MD19-1E) was installed which is able to pro-vide highly stable diluting ratios of 10 to 2000 andoperates at temperatures above the dew point of theexhaust gas (typically 75°C). After dilution with syn-thetic air, the aerosol passed a sequence of silicagel,activated carbon and cobalt oxide denuders to removewater, VOC, and No, respectively, from the gas-phase. The downstream NOX levels were always below1ppb.

In a first series of measurements, samples of dieselexhaust soot particles were collected on glass fiberfilters. The size distribution of the particles was moni-tored on-line using a scanning mobility particle sizingsystem, which provided the basis for the estimates ofthe total soot surface area collected on the filter. Sam-ples were taken with and without the activated carbondenuders to test for effects of removing VOC's fromthe gas phase, and the engine was operated underidling and full-load (4kW) conditions. Immediately aftersampling, the filter were exposed to 13N labelled NO2

at 30% relative humidity. The output of HONO down-stream of the filter was measured using our standarddenuder detection system [1] (Fig. 1).

COXIn

CM

oCDamJH3.2o

O

1.0E+11 -

1.0E+10 -

1.0E+09 -

Hi• • m

1 1 1 1 1 h—•)—I—I—I—I—0 5 10 15 20 25 30 35 40 45 50 55 60

time [min]• Palas-soot (31 ppb NO2)• diesel-soot (full load, sampled with activated carbon denuder, 22 ppb NO2)& diesel-soot (full load, sampled without activated carbon denuder, 22 ppb NO2)<;• diesel-soot (idling, sampled without activated carbon denuder, 22 ppb NO2)

Fig 1: Comparison of filter samples of spark dis-charge and diesel soot in off-line bulk experi-ments, influence of sampling conditions onamount of HONO formed.

The results depicted in Fig. 1 clearly show that HONOforms considerably more slowly on diesel soot than onsimilar samples from the spark discharge generator.Eventually, also the total amount of HONO per sootmass is lower. It is also evident that HONO formationstrongly depends on the combustion and samplingconditions. The fact that the HONO output depends onwhether the activated carbon denuder is in line or notindicates that semivolatile organic compounds areinvolved in the reaction.

ACKNOWLEDGEMENT

The technical support by U. Matter from Matter Engi-neering AG (Wohlen, Switzerland) is highly appreci-ated. This work was supported by the Kommission furTechnologie und Innovation (KTI).

REFERENCES

[1] M. Ammann et al., Nature 357,157-160 (1998).

[2] M. Kalberer et al., Atmos. Environ., in press(1999).

139

REACTION OF NO2 TO HONO ON SULPHURIC ACID COATED SOOT AEROSOLPARTICLES

F. Arens, M. Ammann (PSI), J. Kleffmann (Univ. Wuppertal), U. Baltensperger (PSI), M. Kalberer,H.W. Gaggeler (Univ. Bern & PSI)

In the polluted troposphere gaseous nitrous acid (HONO) maybe formed via a reaction between nitrogendioxide (NOJ and soot particles. The long term significance of this process could be influenced by adsorp-tion of strong oxidants like sulphuric acid. We have investigated the formation of HONO on soot aerosolparticles coated with sulphuric acid. Such particles are found in remote areas as well in jet contrails.

Gaseous nitrous acid (HONO) is an important traceconstituent of the polluted troposphere which is formedby heterogeneous processes involving nitrogendioxide (NO2). It has been shown that NO2 may bereduced to HONO on the surface of soot particleswhereby organic soot components are oxidised [1].

The long-term significance of this reaction depends onpossible recycling processes making the soot surfacereactive again, and on competing ageing processes.Besides the reaction with O3, adsorption of sulphuricacid is an important ageing process throughout theatmosphere. Using the 13N tracer technique [1], wehave investigated the formation of HONO on sootaerosol particles coated with sulphuric acid.

After passing the soot particles over liquid sulphuricacid at a controlled temperature, they were humidifiedto a relative humidity of 30 to 80%. In a flow reactorthe resulting aerosol was mixed with NO2. Thepossible reaction products (gaseous HNO3, HONO,NO2, NO and particle-adsorbed NO2) were measuredon-line using denuders and a filter [1].

Figure 1 shows the growth of sulphuric acid coatedsoot particles with relative humidity measured with ascanning mobility particle sizing system. Because ofthe hygroscopicity of the sulphuric acid, the particlediameter increases with increasing water vapourpressure in the gas phase, whereas uncoated sootparticles do not increase in their size accordingly [2].

As sulphuric acid can be a strong oxidant, it seemslikely that a part of the decrease in HONO formation,shown in Fig. 2, is caused by the oxidation of reducedsurface sites by sulphuric acid. Additionally, theparticle growth by condensation reduces the ability ofNO2 to access the reaction sites on the solid surface ofthe incorporated soot particle. Although for sulphate tosoot mass ratios greater than one, a completeoxidation of any reduced surface site on the sootwould be likely, the HONO output did not completelydrop to a value expected for sulphuric acid alone [3],and soot still seemed to contribute to HONOformation. Eventually, the soot particle was notcompletely incorporated in sulphuric acid, but waslocated on the surface of the sulphuric acid dropletsstill exposing some unreacted sites to the gas phase.

6.0E+10

"Z 5.5E+10•O

"S S.0E+10

5 4.5E+10

2 , 4.0E+10

0 3.5E+10

5 3.0E+10

U 2.5E+10

§ 2.0E+10

1 1.5E+10

•g 1.0E+10

M S.OE+09J BUM

A

1

i?MA

!

&1j

to

1

\\\

1 M1 1 1-•-without wat

vapour-*»r.hV = 30%

-*-r.h. =40%

- * • • r .h. = SO %

— r.h. = 60 %

-•-r.h. =70%

-<~ r.h. =80%

> \m mtit

er-

.-

. .

""

•Jj10 100particle-diameter [nm]

Fig 1: Growth of sulphuric acid coated soot particleswith relative humidity.

1E+14

S 1E+13

ozoX

1E+12

0 1 2 3 4sulphate to soot mass ratio

Fig 2: Amount of HONO molecules per surface areaas a function of the sulphate to soot mass ratioof the particles formed after a reaction time of25 seconds.

ACKNOWLEDGEMENT

This work was supported by the Kommission furTechnologie und Innovation (KTI).

REFERENCES

[1] M. Ammann et al., Nature 395, 157-160 (1998).

[2] E. Weingartner et al., Atmos. Environ. 31, 2311-2327(1997).

[3] J. Kleffmann et al., Atmos. Environ. 32, 2721-2729(1998).

140

MODELING THE NO2+SOOT REACTION IN THE LOWER TROPOSPHERE

M. Ammann, M. Kalberer, U. Baltensperger (PSI), B. Aumont, F. Brocheton (Univ. Paris XII),S. Madronich, D. Hauglustaine (NCAR, Colorado)

The contribution of the NOs+soot reaction to the chemistry of the lower troposphere is investigated using amodeling approach. The NO2+soot reaction was shown to affect only to a very minor extent theOx/NO^/HO^ chemistry both under typical urban and rural environment.

INTRODUCTION

Recent modeling studies have suggested that soot is akey component of tropospheric chemistry in remoteregions, acting to reduce HNO3 to NO2 and possiblyNO2 to NO. Soot may thus also affect the chemistry ofrural and urban areas, where soot concentrations aretypically several orders of magnitude higher than in theremote troposphere. To test this hypothesis, theNO2/soot chemistry was introduced in a detailedhydrocarbon-NOx mechanism and simulations wereconducted for 2 scenarios representing typical urbanand rural areas.

MODEL

General conditions:Simulations were performed using a 2-layers boxmodel, a mixed layer in which species are subjected toemission and deposition and an upper residual layer.Parameterizations include a diurnal variation for T,mixing height, emission rates, deposition, photolysisfrequencies, ... Values are set to be representative oftypical urban and rural areas. Details are given in [1].

General soot parameters:Soot concentrations were set to 10 |xg m'3 and 1 jj.g m'3

for the urban and rural scenarios, respectively. Weused a soot specific surface area of 140 m2 g"1.

RESULTS AND CONCLUSION

We first set the reactive uptake coefficient j to thevalue proposed in the previous modeling studies(y=2.8x10'2). It was found that the NO/soot reactionslead to some unrealistic results, e.g. negligible O3

formation in the urban case. This strongly suggeststhat the contribution of the NO2 reduction process onsoot was considerably overestimated in previousmodeling studies.

Based on [2,3], we consider the following alternativescheme for the NO2/soot interactions:

(1a) NO2 + soot ->{S.NO2}{S.NOJ -»NO{S.NOJ -»{S.N}NO, + soot -»HONO

(1b)(1c)(2)

k=103s1

y=1.6x104

where {S.NO2} denotes NO2 chemisorbed on the sootsurface and {S.N} nitrogen permanently bound to soot.Whether soot is reactivated after reaction by othergaseous species or perhaps via thermal orphotochemical processes remains an open question.Therefore we consider 2 cases, with and withoutdeactivation.

Deactivation case:For the urban scenario, it was found that thedifferences with and without the NO/soot reactionnever exceed 1% for all species in the OX-NOX-HOX

pool. Differences are totally negligible for the ruralscenario.No deactivation case:For the rural case, differences in O3 and NOX

concentration are typically of a few percent (see Figure1). Similar results were obtained for HOX. For theurban case, differences in O3 concentration are alsofairly small (a few %) but reach up to a factor of 2 forNOX during nighttime. The early morning HOX

concentration was also found to be affected to a majorextent, due to the photolysis at sunrise of any HONOproduced on soot during the night. Finally, our resultssuggest a potentially important contribution of theNO2/soot reaction to the production of HONO.

O3 - urban scenario

ft•S 2003=

Aa so

I-i-

O3

-

rural scenario

/ \

f\4A

\l -time (day)

NO, NO2 - urban scenario

time (day)

NO, NO2 - rural scenario

time (day)

HONO - urban scenario

time (dny)

HONO - rural scenario

1 I Utime (day) time (day)

Fig. 1: O3, NOX and HONO concentration profiles forthe urban and rural scenarios. Solid curve:without NO2/soot chemistry; dotted curve: withthe NO2/soot chemistry and for the nodeactivation case

REFERENCES

[1] B. Aumont et al., J. Geophys. Res., in press .[2] M. Kalberer et al., J. Phys. Chem. 100, 15487

(1996).[3] M. Kalberer et al., PSI Proc. 97-02, 7-9 (1997).

141

THE INFLUENCE OF HETEROGENEOUS PROCESSES ON WATER SOLUBLEANIONS OF CARBONACEOUS AEROSOL PARTICLES

M. Ammann, E. Rossler, S. Brutsch, M. Schwikowski, U. Baltensperger (PSI)

Carbonaceous aerosol particles were sampled on filters and exposed to NO2 and humidity. The reactionproducts NO and HONO were recorded downstream of the filter. Before and after exposure, the filters wereextracted in water. The data clearly show that carboxylic acids disappear from the particle surface duringexposure. These acids are important in determining the hygroscopic properties of the particles in theatmosphere.

INTRODUCTION

Heterogeneous reactions between gas phasemolecules and the surface of carbonaceous aerosolparticles affect the hygroscopicity of these particles,which is an important parameter when their influenceon climate is addressed. On the other hand, thesereactions also produce new reactive compounds whichmay influence the oxidizing capacity of thetroposphere. For instance, the reaction of NO2 withsoot aerosol particles can lead to nitrous acid (HONO)which produces the strongly oxidising OH radical uponphotolysis [1]. In view of the importance ofcarbonaceous particles in the atmosphere it istherefore important to investigate especially the watersoluble surface species and their behaviour duringheterogeneous processing in the atmosphere.

EXPERIMENTAL

Soot aerosol particles from a laboratory soot source(PALAS GfG-100) were sampled on glass fiber filters.A detailed description of the aerosol techniques usedis described elsewhere [2]. Immediately after asampling period of 3h, the filters were exposed to 0and 50ppb NO2 at 50% relative humidity during 16h.The concentrations of NO, NO2 and HONO wererecorded downstream of the filter using achemiluminescence detector. Then, the filters weretransferred into ultrapure water and extracted in anultrasonic bath during 45min at 45°C. The aqueoussamples were stored at -20°C until analysis by ionchromatography. All sampling and exposure wasperformed in four replicates, and four controls ofempty filters and for each extraction step wereanalysed in parallel.

RESULTS AND CONCLUSIONS

During exposure of the soot particles to NO2, HONOand small amounts of NO were observed (data notshown) which is consistent with the aerosol flowreactor experiments described elsewhere [1]. Inaqueous extracts of fresh soot particles, substantialamounts of carboxylic acids, nitrite, nitrate andsulphate were observed, which are all common sootcomponents. Exposure of the samples to humidifiedair did not significantly affect the concentration oftheses species. However, when they are exposed to a

mixture of humidified air and NO2 (Fig. 1), carboxylicacids are strongly depleted, whereas the nitrateconcentration increases. NO2 reacts slowly with soot toform organic nitrates which can then be hydrolysed inwater. When shorter exposure periods are applied,nitrite is also enriched on the particle surface, but itdisappears again for longer exposures presumablydue to desorption in the form of HONO.

40x101 2

Acetate Formate Nitrite Nitrate

Species

Fig. 1: Major extractable anions on soot aerosolparticles after exposure to NO2 (shaded bars)and to synthetic air only (empty bars).

The results show that under the influence of humidityand NO2 carboxylic acids are released from thesurface or are no more extractable in water. It will beinvestigated in further studies whether they play a rolein the chemical mechanism of the reaction of NO2 toHONO or whether the observed effect is theconsequence of pH changes on the soot surface.However, because carboxylic acids are major speciesdetermining the hygroscopicity of soot, heterogeneousprocessing of soot particles in the atmosphere directlyaffects their ability to act as cloud condensation nuclei.

REFERENCES

[1] M. Ammann etal., Nature 357,157-160 (1998).

[2] M. Kalberer et al., Atmos Environ, in press (1999).

142

THE INFLUENCE OF ADDITIVES ON THE SIZE DISTRIBUTION AND COMPOSITIONOF PARTICLES PRODUCED BY DIESEL ENGINES

G. Skillas (ETHZ& PSI), K. Siegmann, U. Natter (ETHZ), H. Burtscher (Fachhochschule Aargau),U. Baltensperger (PSI)

Measurements with different diesel engines and fuel additives revealed a correlation between the minimumadditive concentration needed to induce exhaust particles consisting of additive products, and the elemen-tal carbon emission factor of the engines.

Airborne particles are potentially harmful to humanhealth [1,2]. The most important source of anthropo-genic ultrafine particles are diesel combustion engines.To reduce emissions, particle traps which also reducethe engine efficiency, and fuel additives are underconsideration. Many studies have been performed toinvestigate the efficiency and mechanism of a varietyof additives [3,4,5]. The main result of these studies isthat the benefit of additives is quite limited as an insuf-ficient emission reduction is obtained.

For the measurements two different diesel engines areused. A 6.11 litre TDI heavy-duty engine with a nomi-nal load of 105 kW is run at an engine test bench. Thesecond engine is a small YAMAHA diesel driving anelectrical power generator, (Model EDA 4700T). Loadsup to 3.0 kW can be attained with it. The engines areoperated either with regular diesel, or with diesel con-taining an additive (Ce or Fe).

In both cases a similar setup is used, the aerosol be-ing sampled from the exhaust pipe, 3 meters after theengine (Fig. 1).

Diesel engine DMA

Exhaust

Dataacquisition

Aethalometer PA sensor

Fig. 1: Experimental setup for the L1 engine.

By coulometric analysis of filter A and of aethalometerdata the engine emission factor (particulate mass perburnt fuel mass) is computed. Selecting the particleswith the DMA, counting them with the CPC and ana-lysing filter B, the composition of a specific particle sizefraction is found. An overdose of additive leads to theappearance of large amounts of particles. These parti-cles have a very small diameter of about 10 nm. Thechemical analysis shows that the particles with di-ameters Dp < 50 nm consist mainly of metal oxidesstemming from the additive. The reduction of particleswith Dp > 50 nm is about 50%. The additive concentra-tion required to form new particles, consisting mainlyof the additive metal, increases when more soot is

emitted. In Fig. 2 this concentration is plotted versusthe emission factor both for the L1 and the YAMAHAdiesel engine.

100

70

a 40

10

L1 diesel eng.

YAMAHA diesel gen.

1400 rpm/eOONm

2000 rpm/250Nm

I +—\

11400 rpm/300Nm

2000 rpm/500Nm

ir.1-10 2 -10 ' 4-10 1-10"*

EF[]

Fig. 2: Plot of the minimum cerium concentrationneeded for additive particles to be formed,against the BC emission factor, (YAMAHA en-gine) and the PM emission factor (L1 engine).

The onset of particle formation correlates well with theemission factor. The two different types of engines, onthe other hand, do no seem to have much influence.These results permit an optimal dosing of the additive,in order to apply as little excess additive as possible. Itis hoped that a combination of self-regenerating filtersand soot suppression with additives will drasticallyreduce the diesel emissions.

REFERENCES[1] J.B. Howard, W.J.Jr. Kausch, Soot control by

Fuel Additives, Prog. Energy Combust. Sci., 6,263-276(1980).

[2] J. Lemaire, W. Mustel, Fuel additive supportedparticulate trap regeneration possibilities by en-gine management system measures, 942069 inSAE technical paper series, Society ofAutomotive Engineers (1994).

[3] G. Lepperhoff et al., Quasi-continuous particletrap regeneration by cerium additives, SAEtechnical paper series, Society of AutomotiveEngineers, 950369 (1995).

[4] G. Oberdorster et al., Increased PulmonatoryToxicity of Inhaled Ultrafine Particles: Due toLung Overload Alone?, Annals of OccupationalHygiene, 38 Suppl. 1, S295--S302 (1994).

[5] A. Peters et al., Respiratory Effects are associ-ated with Number of Ultrafine Particles, Am. J.Respir. Crit. Care Med., 155,1376—1388 (1997).

143

AEROSOL MEASUREMENTS DURING SUMMER SMOG EVENTS IN THE MILANAREA

N. Streit, E. Weingartner, S. Nyeki, U. Baltensperger (PSI), ft. van Dingenen, J.-P. Putaud (JRC, Ispra),A. Even, H. ten Brink (ECN,Petten), A. Blatter, A. Neftel (IUL, Liebefeld), H.W. Gaggeier (Univ. Bern & PSI)

Aerosol measurements were performed in the Milan suburbs in early summer of 1998 as a part of an inter-nationally coordinated field campaign run under the auspices of EUROTRAC. The main aims of the fieldcampaign were the study of VOC or NOx limitation in ozone production and secondary aerosol formation.

During early summer this year, we measured variousaerosol parameters in the Milan metropolitan area.The measurements were performed during aninternationally coordinated field campaigninvestigating the subject of NOx or VOC limitation ofozone production in urban, suburban and ruralsettings. The field campaign PIPAPO (PianuraPadana pjoduzione di ozono) was part of the LOOP(Limitation of ozone p/oduction) program, a Eurotrac-2 subproject.

Meteorological and gas parameters were measuredat several ground stations stretching from south ofMilan up to Lake Maggiore. During two intensiveobservation periods (IOP), two research aircraftsmeasured the spatial distribution of the parameters ofinterest. The lOPs lasted from May 12-14 and fromJune 1-6. In addition to the traditional homogeneousgas phase chemistry, several selected aerosolparameters were also measured in order to gaininsight into heterogeneous processes. The aerosolparameters were measured at two sites north ofMilan, one situated at six km from the city center atBresso local airport, the other at a distance of about35 km (Verzago).

The following aerosol parameters were measured:

• the number concentration of condensation nucleiin the 0.003-1 |j,m range (TSI condensation nucleicounter 3025)

• the chemical composition of the aerosol in two sizeclasses (filter measurements)

• the black carbon concentration (Aethalometer AE-10)

• the size distribution of the aerosol from 0.01 to 0.8\m\ (TSI differential mobility particle sizer DMPS)

• PM25, the mass of particulate matter with a diame-ter smaller than 2.5 jxm (Eberline FH 62 I-Rbetameter)

• the concentrations of ammonium, nitrite, nitrate,sulfate and HONO, HNO3, SO2 (wet effluent de-nuder/aerosol collectors)

• the volatility of the aerosol (thermal denuder sys-tem coupled to a TSI scanning mobility particlesizer SMPS)

• the hygroscopicity of the aerosol (hygroscopicitytandem differential mobility analyzer).

In the following, a few preliminary results are pre-sented. First, we compare the performance of our wet

effluent denuder/aerosol collector system with tradi-tional filter measurements. As an example, the nitratevalues measured during the second IOP are shown,where the more numerous values of our semi-continuous system have been averaged according tothe longer filter sampling times. Both methods showvery good agreement of the nitrate concentration.

24 -i

1 6 H

8 -

01.06.98 04.06.98 07.06.98 10.06.98

Fig. 1: Comparison of filter and aerosol collectormeasurements.

Second, we show measurements of the volatility ofthe aerosol in the submicrometer range. The graphpresents the aerosol number size distributionsurviving thermal treatment in a thermal denuder atfour alternating temperatures. The volume at 280 °Cshows a good correlation with black carbon. Thedifference in volume between 30 and 280 °C can beexplained by the evaporation of organic carbon,nitrate, and sulfate.

1200 -,

100Diameter [nm]

1000

Fig. 2: Aerosol volatility size distribution.

These are just two impressions from the vast PIPAPOdata sets. Further investigations planned with thedata include secondary aerosol formation and thesearch for HONO sources, both in collaboration withother research laboratories.

ACKNOWLEDGEMENTThis work was supported by the Kommission fiirTechnologie und Innovation (KTI).

144

SUBMICROMETER AEROSOL SIZE DISTRIBUTIONS AT A HIGH-ALPINE SITE(JUNGFRAUJOCH, 3580 m a.s.l.)

E. Weingartner (PSI), S. Nyeki (PSI & Univ. Essex) N. Streit, M. Lugauer, U. Baltensperger (PSI)

During a long-term field campaign at a high-alpine site (Jungfraujoch; 3580 m asl, Switzerland) from March1997 to May 1998, the particle number size distribution (diameter D =18- 750 nm) and number concentra-tion N (D > 10 nm) were measured on a continuous basis (every 30 minutes). The data offer insight into thedifferent mechanisms responsible for shaping the background number size distribution.

On the Jungfraujoch, the seasonal variation of aerosolparameters related to the surface or volume concen-tration, are characterized by a distinct seasonality withconcentrations in summer approximately 10 timeshigher than in winter [1]. During summer, thermal con-vection of air from the planetary boundary layer to thesampling site is responsible for higher aerosol con-centrations [2], Figure 1 shows that this phenomenonis also seen in the number concentration /Vof particlesin the accumulation mode (with diameter D = 100 -1000 nm). In this size range, N exhibits a summer towinter ratio (S/W) of about 6.6. The seasonality of N inthe Aitken mode size range (20 - 100 nm) is muchsmaller (S/W = 1.6). Nucleation mode particles (D< 20nm) even show an inverse seasonality with higherconcentrations during winter (S/W = 0.8) [3]. This de-crease of S/W values with decreasing particle size isexplained by the enhanced formation of new particlesvia homogeneous nucleation during the colder sea-sons. During wintertime, this process is most probablyfavored by lower temperatures and lower surface areaconcentrations.

400O SpringA SummerV AutumnO Winter

10 1000Diameter D [nm]

Fig. 1: Seasonal variation of the number size distribu-tion (3-month averages) for dry aerosols.Symbols are measured data and solid linesrepresent fitted bimodal curves.

Average number size spectra are characterized by adistinct bimodal shape. The spectra were well-fittedwith the sum of two lognormal distributions. Fittedmodal diameters were surprisingly constant throughoutthe year (D = 39 ± 3 nm; 126 ± 6 nm).

Figure 2 illustrates how this bimodality is generatedand maintained by cloud processes. In clouds, solubleparticles with D > 80 nm are activated to droplets

which absorb trace gases. Aqueous phase reactionsthen produce additional aerosol material. Upon evapo-ration of the droplet, the residue is larger than theoriginal particle on which the droplet was formed. Inthe Aitken mode, a narrowing of the distribution isclearly observed during in-cloud conditions. This ismost probably due to preferential scavenging of thesmallest particles by cloud droplets. Besides precipita-tion scavenging, these two cloud processes are be-lieved to be mainly responsible for the observed dou-ble-peak shape of the measured spectra. The detailedmechanisms leading to the higher in cloud aerosolconcentrations are at present ill-defined and will be thesubject of further work.

350-

5-300-

jo 250-

Q200-c- 1 5 0 -

S 100-• o

50-

0-

trtL/ /

4rscavenging

— i — t i i i 1

•

^ \ \growth T ^ \

due t ° % ^aqueousTsJjl

phase T^reactions

dense cloud •clear sky

-

-

-

10 100Diameter D [nm]

1000

Fig. 2: For the period February to May 1998 inclusive,one-hour values were classified into differentliquid water content (LWC) ranges, and aver-age number size distributions of the driedaerosol for clear-sky (LWC < 10 mg m"3) anddense cloud conditions (LWC > 200 mg m3)were calculated. Symbols and solid lines rep-resent measurements and lognormal distribu-tion fits, respectively.

ACKNOWLEDGEMENT

This work was supported by SMI, KTI and BUWAL.

REFERENCES

[1] S. Nyeki et al., J. Geophys. Res., 103, 6097-6107,(1998).

[2] M. Lugauer et al. Tellus, 50B, 76-92, (1998).[3] E. Weingartner et al., J. Geophys. Res., submit-

ted, (1998).

145

CONDENSATION NUCLEI (CN) AND ULTRAFINE CN IN THE FREE TROPOSPHERETO 12 KM: A CASE STUDY OVER THE JUNGFRAUJOCH

S. Nyeki (PSI & Univ. Essex), M. Kalberer, U. Baltensperger (PSI), I. Colbeck (Essex Uni, UK),A. Petzold, F. Schroder (DLR)

An airborne campaign was conducted over the Jungfraujoch (JFJ) station to determine vertical profiles ofcondensation nuclei (CN) concentration to just below the tropopause.

A theme of current interest in the study of aerosoleffects on climate concerns the origin and formationmechanisms of natural aerosols in the remoteatmosphere. Modeling [1] and measurements ofaerosol vertical profiles [2, 3] suggest that the FT maybe a source of CN to the planetary and marineboundary layers.To address these aspects, an airborne campaign wasrecently conducted over the JFJ high-alpine researchstation (3454 m; Switzerland) to investigate: 1) thevertical distribution of CN and 2) the CN concentrationat 8.0 km a.s.l. during horizontal transects. CN resultsfrom a case study on July 30, 1997 are analyzed frommorning and afternoon profiles, comprising a databaseof about 5 hours. A stacked vertical flight pattern wasflown from 4.5 to 12 km a.s.l., with intermittenthorizontal flight legs 25-35 km long at specificaltitudes.Vertical profiles of CN concentration for d> 5 nm (AQ,d> 15 nm (W,15) and ultrafine CN (5 > d> 15 nm; A/5.15)exhibited a distinct increase in particle production withaltitude, as indicated by the ratio NJNiiS. Figure 1illustrates these results for the morning flight (0730 to1000 hrs LST; LST=UTC+1), as well as the aerosolsurface area concentration for d > 100 nm, S2100. Acomparison of CN concentration at the lowest altitudeand simultaneously at the JFJ compare well [4]. Near-constant values of d(CN)/dz with altitude were found indiscrete layers (-1.4-2.8 km in depth), and coincidedwith different airmasses as delineated by changes inthe potential temperature profile d9/dz at 5.6, 6.4, 9.0and 10.5 km. These were corroborated by upwindradiosondages and back-trajectory analysis, using the48-hour mesoscale Swiss Model, which indicated thatairmasses had had no recent contact with land andorginated from over the North Atlantic.In addition, the CN concentration was mapped over a6.1 x 31 km horizontal grid at 8.0 km altitude centredover the JFJ. A pronounced variation in Ni5, W,15, A4300

(FSSP-300X) and specific humidity with respect to theunderlying alpine topography was found and suggeststhat mountain waves may have been responsible. Thecase study provides further evidence suggesting thatthe upper FT appears to be a source of new particles.

1200010°

11000 -

10O00 -

9000 -

3 8000-

7000 -

6000-

5000-

4000

102

x a

: ^"^iymv^wmt^'

W 10'

j and Nil5 Concentration (scm3)

Fig. 1: Aerosol parameters (STP conditions) over theJFJ during the morning profile on July 30,1997. Symbols indicate median values duringhorizontal flight legs.

ACKNOWLEDGMENTSOur thanks are extended to the DLR, PSI, SMI andJFJ staff. This work was kindly funded through the EUSTAAARTE program, the NERC, the Swiss GAWprogram and BUWAL.

REFERENCES[1 ] F. Raes, J. Geophys. Res. 100,2893-2903 (1995).[2] A. D. Clarke, J. Geophys. Res. 98, 20633-20647

(1993).[3] F. Schroder and J. Strom, Atmos. Res. 44, 333-

356(1997).[4] S. Nyeki, F. Li, E. Weingartner, N. Streit, I.

Colbeck, H. W. Gaggeler and U. Baltensperger, J.Geophys. Res. 103, 31749-31762 (1998).

146

SEASONAL VARIATION OF 212Pb-ACTIVITY ON AEROSOLS AT THEJUNGFRAUJOCH

M. Lugauer, U. Baltensperger, M. Furger, D.T. Jost, M. Schwikowski (PSI), H. W. Gaggeler(Univ. Bern & PSI)

The density of 212Pb-activity on aerosol particles at the Jungfraujoch exhibits a pronounced seasonalvariation, which is probably caused by the annual variation of the alpine snow-cover. The high and densesnow-cover in spring inhibits the production of 212Pb from high-alpine slopes. The transport of 212Pb fromlow-elevation valley sites is to slow to compensate the decrease of the alpine212Pb source.

The aerosol surface area (AS) and the 212Pb activityconcentrations are measured at the high-alpine siteJungfraujoch (JFJ, 3454 m a.s.l.) within a project thatis dedicated to investigate the role of aerosol particleswithin the climate system [1]. It was shown by [2] thatthermally-driven convection (TDC), developed withinthe Alps, causes a pronounced annual cycle in ASconcentration at the JFJ. Thus, it is crucial tounderstand, how TDC is determined and influenced bythe complex alpine topography.

Radioactive 212Pb (T1/2=10.64 h) is a good parameter toinvestigate TDC, because 212Pb is only produced closeto the ground surface by the decay of 220Rn (T1/2=55.6s), which emanates from the soil. 212Pb is attached toaerosol particles and is measured at the JFJ by asimilar technique as the AS [3].

The annual cycle of 212Pb density on the AS at the JFJis shown in Fig. 1. The 212Pb density is lowest in April,increases nearly linearly until October and exhibitsanother peak in December. The 212Pb density dependsproportional to the time period At, between attachmentat time t and detection at the JFJ at t+At by

10

where

At = :

= ln2/T1 / 2(2 1 2Pb),

(1)

A(t): 212Pb-activity density at time t.

Since A(t) is unknown, the transport time of individualair parcels cannot be calculated. However, theaverage difference in At between different seasonscan be calculated. The 212Pb-density increases by afactor of 3.25 from April to October on average (Fig.1). This change yields a time difference of 18 hours(Eq. 1).

This time difference may be caused either by anincrease in transport velocity or by a shift of the 212Pbsource closer to the JFJ. The first possibility is unlikely.The second possibility could be explained by thealpine snow-cover, which undergoes large seasonalvariations. The snow-cover is highest in April, retreatsduring summer until October and starts to grow inNovember. The snow-cover in December is oftenlower than in November. In addition, the snow volumein early winter has a much lower water equivalent thanthat in spring.

iCD

Eo b

daily median- Q - runavg(30 days)

'•§ 2 5 -

.a a)CM CD

5 ]=CO

i i i i i i i i i i i i r

J F M A M J J A S O N D

Month

Fig. 1: Seasonal variation of 212Pb-activity density onaerosol surface area at the Jungfraujoch.Database 1991-1997.

Above snow-covered surfaces 212Pb production is belimited, because of limited ^Rn emanation. Snow-cover also prevents the development of TDC, by which212Pb could be transported to the JFJ.

If the alpine snow-cover controls the 212Pb density atthe JFJ, it can be concluded that the overall verticaltransport time from the boundary layer, where themain 212Pb source is located, up to the JFJ is longerthan the half-life of 212Pb. Otherwise, 212Pb would becompensated in spring and summer, when the airexchange between the boundary layer and the JFJ ishigher than in autumn. Thus, the vertical transport byTDC between the boundary layer and the JFJ is not amotion in a single step within one day, but a multi-stepprocess, accomplished within a few days.

the Bundesamt fur

ACKNOWLEDGEMENT

This work was supported byUmwelt, Wald und Landschaft.

REFERENCES[1] U. Baltensperger et al., J. Geophys. Res. 102,

19707(1997).[2] M. Lugauer et al., Tellus 50B, 76 (1998).[3] H.W. Gaggeler et al., Atmos. Environ. 29, 607-

616(1995).

147

MEASUREMENT OF REACTIVE NITROGEN (NOy) AT THE JUNGFRAUJOCH

C. Zellweger (PSI & EMPA), E. Weingartner, M. Lugauer, U. Baltensperger, N. Streit, M. Ammann (PSI),P. Hofer (EMPA)

Speciated and total NOy was measured at the Jungfraujoch during the summer 1997. Under specialmeteorological conditions the Jungfraujoch seemed to be influenced by planetary boundary layer airthroughout the whole day.

During a field campaign in summer 1997 at theJungfraujoch, speciated reactive nitrogen compoundswere measured in conjunction with total NOy. Suchmeasurements have proven to be valuable inassessing the photochemical processing that hasoccurred in an air parcel. The measured individualnitrogen species included nitric oxide (NO), nitrogendioxide (NO2), nitric acid (HNO3), paniculate nitrate,and peroxyacetyl nitrate (PAN). During themeasurement campaign, PAN was the most abundantNOy species with an average fraction of 36%. Secondlargest fraction with a total contribution of 24% to totalNOy was inorganic nitrate (HNO3 and particulatenitrate). On average, nitric acid amounted to 32% oftotal inorganic nitrate, indicating that most of the nitricacid was neutralized, most probably by ammonia. NOX

was with 22% of total NOy the third largest fraction oftotal NOy during the measurement campaign. Exceptfor a few occasional events, NO2 was the dominatingNOX species. This clearly indicates the absence of apollution source in the immediate vicinity of themeasurement station. The NOj/NOy ratio averaged0.25 during the campaign. The NOx/NOy ratio reflectsthe transformation of NOX to reservoir species andaccounts for dilution effects, but wet and drydeposition and decomposition processes can makethe ratio complicated to interpret in regions remotefrom NOX emissions. At the Jungfraujoch, lowNOx/NOy ratios are often accompanied by high NOy

concentrations. Figure 1 shows an overview of dailyaverage concentrations of speciated and total NOy

between 18 July and 23 August.

Nitrate

HNO3

Fig. 1: Daily average concentrations of speciated andtotal NOy at the Jungfraujoch during thesummer field campaign.

The day to day variability was mainly caused bydifferent meteorological conditions. For example the

sharp concentration drop in all measured speciesexcept NOX on 1 August was caused by a cyclone.

The data were further analyzed with respect to aweather type classification based on the AlpineWeather Statistics [1]. During the measurement periodthe convective weather type dominated. A very distinctdiurnal pattern was observed for convective days withwind direction from N to NW at 500 Pa. This thermallyinduced diurnal behavior is illustrated in Fig. 2 for NOy

and PAN.

Fig. 2: Time series with a very distinct diurnal patternof NOy, and PAN between 27 and 30 July.

In contast, for convective days with low wind speed at500 hPa only little diurnal change was observed.Figure 3 shows the time series for NOy and PAN for aperiod with the above meteorological conditions.

3000

2500-

Fig. 3: Time series of NOy, and PAN between 16 and22 August.

Concentration changes during this period wereepisodic rather than governed by thermally inducedtransport processes, and the Jungfraujoch seemed tobe influenced by planetary boundary layer airthroughout the day.

ACKNOWLEDGEMENT

The support of the staff of the Jungfraujoch ResearchStation is gratefully appreciated.

REFERENCE

[1] M. Lugauer et al., Tellus50B, 76-92, (1998).

148

CHEMICAL COMPOSITION STUDY OF THE FINE ATMOSPHERIC AEROSOL AT AHIGH-ALPINE SITE (JUNGFRAUJOCH, 3580 m a.s.l.)

Z Krivacsy (Univ. Veszprem, Hungary), E. Weingartner, S. Nyeki, U. Baltensperger (PSI)

Quantities of inorganic ions, organic substances and black carbon were determined in fine atmosphericaerosol collected at the Jungfraujoch. Organic substances comprised the major portion of the characterizedaerosol mass.

Knowledge of the fine atmospheric aerosol chemicalcomposition is essential to study different phenomena(e.g. light scattering and absorption, cloud formation)in the atmosphere. More recently, evidence concern-ing the importance of organic substances in atmos-pheric processes has been accumulating [1, 2]. In thiswork, the chemical composition of PM2.5 atmosphericaerosol collected at the Jungfraujoch (JFJ) researchstation during July and August 1998 was studied todetermine: i) the contribution of organic substances toaerosol loading and ii) bulk characterization of organicmatter [3, 4].

"heavier") compounds. The average fractions of thetwo types of carbon are 36 and 64%, respectively.

• other inorganics• ammonium

B organic substances• black carbon

E3)

3

2

1

0 ,

II

CO CO GO

I ll

• UL• U , - P VCO CO CO CO if.

samples

Fig.1: Chemical composition of JFJ fine aerosol.

In Fig.1 concentrations of inorganic ions, black carbon(measured by aethalometer) and organic substances(calculated from the mass of organic carbon deter-mined by TOC analysis using a conversion factor of1.4) of the samples are shown. It is clearly seen thatorganic compounds dominate the aerosol mass, wheretheir contribution is between 40 and 76% (averagevalue 58.5%) of the identified aerosol means.

The thermal behavior of carbonaceous compoundsduring catalytic burning in oxygen at 680°C is pre-sented in Fig. 2. In this curve, two peaks can be distin-guished. The first peak presumably corresponds to themore volatile (and/or more oxidisable; "lighter") andthe second to the less volatile (and/or less oxidisable;

0,25 -r

100 200 300

Integration time (s)

400

Fig. 2: Curve of evolving CO2 during catalytic burningof the aerosol.

The water solubility of organic compounds in JFJsamples was also investigated. On average 75% ofthe organic carbon proved to be water-soluble. It isinteresting to note that the water solubility of "heavier"compounds was much higher (over 80%) than that of"lighter" compounds (35%). One of the classes of"lighter" water-soluble organics are the low-molecularweight carboxylic acids (oxalic, malonic, succinic andformic acid). Their mean concentration was about 50ng m3, and may account for 10-20% of water soluble"lighter" organic substances. The "heavier" speciesmay be similar polar macromolecular compoundsfound recently in fine aerosol at other places.

While these are preliminary results, it is tentativelyconcluded that organic substances may play an im-portant role in atmospheric processes at the Jung-fraujoch.

REFERENCES

[1] T. Novakov and J. Penner, Nature 365, 823(1993).

[2] M.O. Andreae and P.J. Crutzen, Science 276,1052(1997).

[3] N. Havers et al., J. Atmos. Chem. 29, 45 (1998).

[4] S. Zappoli et al., Atmos.Environ., in press.

149

HISTORICAL RECORD OF CARBONACEOUS PARTICLE CONCENTRATIONSOBTAINED IN A EUROPEAN HIGH-ALPINE GLACIER (COLLE GNIFETTI) IN

COMPARISON WITH ESTIMATED EMISSIONS

V.M.H. Lavanchy, T. Boring, H.W. Gaggeler (Univ. Bern & PSI), U. Schotterer, M. Schwikowski,L. Tobler, U. Baltensperger (PSI)

Historical records of black carbon (BC), elemental carbon (EC), organic carbon (OC) and total carbon (TC)concentrations were obtained for the time period 1757-1975. Comparison of these records estimated withparticulate carbon emissions indicates that this ice core reflects emissions of Western Europe.

The reconstruction of the historic evolution of BC, ECand OC was performed by analyses of an ice coredrilled at Colle Gnifetti (4450 m a.s.l.), Monte Rosamassif, Switzerland. Ice samples were melted in aglass container and filtered through quartz-fibre filters.With an optical technique the attenuation of a lightbeam through the filters was measured. This valuewas then converted into BC concentrations using theinstrumental absorption efficiency CCAPI- For this workwe used an absorption efficiency of 9.3 m2g'1,corresponding to an annual mean value measured forhigh-alpine aerosols [1]. Afterwards, the filters weretreated with 0.1 M HCI to remove HCO3', and the OCand EC contents of the filters were measured with ourthermal technique [1].

1750 1950

Fig. 1: 10-year average (solid line) and ice coresection (dashed line) concentrations of OC,EC and BC.

Figure 1 shows the historic evolution for the OC, ECand BC concentrations in the Colle Gnifetti ice core.During the period from 1750 to 1890 theconcentrations of all three species are fairly constantat a relatively high level. After 1890 an increase inconcentrations is found due to anthropogenicactivities. Between the time period (1750-1890) and(1950-1975) the concentrations increase by a factor of

3.7, 3.0, and 2.5, for OC, EC, and BC, respectively.These factors are comparable to those found for SO4

(5.8) and NO3" (2.3) between two similar time periodsin the same ice core [2]. A high correlation (R2 = 0.57)between sulphate and BC indicates the same origin (inaddition to similar vertical transport processes).

Consumption data of hard and brown coal and ofpetroleum products for the last 150 years are availableusing production and trade data of each country [3].The data from Germany, France, Italy andSwitzerland, i.e., the countries surrounding the drillingsite, were used for the comparison with the totalcarbon concentrations. Figure 2 shows that theincrease of coal consumption and of the TC record arehighly correlated after 1890. Between 1890 and 1950,coal was the most important source of carbonaceousparticles. After 1950, consumption of petroleumproducts e.g. diesel fuel, increased exponentially.

5e+12

a 4e+12

en.20)

o

36+12

2e+12

1e+12

—o— Hard coal—©— Brown coal

Petroleum productsSum of all sources

••-&•••• T o t a l carbon

concentrations

1800 1850 1900Year

1950 2000

Fig. 2: 10-year averages of the TC concentrations inthe ice core. In addition, the sum of theestimated emission of the countries Germany,France, Italy, and Switzerland is shown forhard coal, brown coal, and petroleum alongwith the sum of these sources.

REFERENCES

[1] V.M.H. Lavanchy, H.W. Gaggeler, S. Nyeki, andU. Baltensperger (1998), Atmos. Environ., inpress.

[2] A. Doscher, H.W. Gaggeler, U. Schotterer and M.Schwikowski (1995), Water, Air, Soil Poll. 83,603-609.

[3] B.R. Mitchell, Stockton Press, New York (1992).

150

ENVIRONMENTAL RECORD PRESERVED IN AN ICE CORE FROM CERROTAPADO, CHILE

P. Ginot (Univ. Bern), M. Funk (ETHZurich), H.W. Gaggeler, U. Schotterer (Univ. Bern & PSI), S. Hubener(Univ. Bern), M. Schwikowski (PSI), W. Stichler (GSF Neuherberg), F. Stampfli (SF-lnventor)

Ice cores drilled on the summit glacier of Cerro Tapado in Chile, affected by precipitation of the Westerlies,permit to reconstruct 15 years of past climate and El Nino events.

The cold ice caps of the highest peaks of the Andesrepresent natural archives of past environmentalconditions. During the past 15 years, Thompson et al.have recovered several ice cores from sites within thetropical circulation regime (Quelccaya, Huascaran, andSajama). Their analyses allow to reconstruct theclimatic history of the tropics since the last glaciation[1], In order to find a glacier preserving the chemicaland isotopic signal from a precipitation regime affectedby the Westerlies, we performed an exploratory drillingon top of Cerro Tapado (5550m, 30°08'S, 69°55'W) inNorthern Chile in February 1998.

Cerro Tapado is one of the northernmost glacierssouth of the "South American Arid Diagonal" within theWesterlies circulation regime. At the drilling site, theglacier has a maximum thickness of 42 m. Two icecores, 10 m each and 50 m apart could be recoveredwith a portable solar powered electromechanical 2-inch drill. From ice core stratigraphy, chemicalanalyses, isotope measurements as well as theevidence of various regional volcanic eruptions weconclude that these cores cover the period between1980 and 1998 (Fig. 1).

Precipitation events in the area of Cerro Tapado occurprimarily between May and September (i.e. winter).Because of their predominant Pacific moisture sourceand their regular annual deposition period, we assumethe chemical concentrations of the measured speciesto be rather uniform in fresh precipitation. Hence, theobserved strong variability of the chemical signal maybe related to post-depositional effects such assublimation (estimated to be between 3 and 6 mmwater per day [2]) or dry deposition. We therefore usethe structure of the chemical signal and particleconcentration to distinguish between dry (highconcentration) and humid conditions (lowconcentration) during the time period covered by theice cores.

Dating was supported by volcanic eruptions (Pinatubo- June 1991, and El Chichon - April 1982) identifiedby high concentrations of sulphate, and in addition, byseveral distinct 818O depletions, which may be causedby extreme precipitation events in connection with thephenomenon El Nino. By applying this time scale, it ispossible to reconstruct 15 years of precipitation historyof the Cerro Tapado ice cap. The results are inagreement with climatic data (precipitation andtemperature) recorded at a meteorological station in

Elqui Valley [3] with El Nino events established withSea Surface Temperature Anomalies.

Summer 82/83— El Chichon

April 1«B2

Kg2* c i 2

Fig. 1: Chemical concentration (p.g/1) and isotopic ratioalong Core 3, Cerro Tapado (Chile).

During summer 1998, we developed a 3-inch lightsolar powered electromechanical 50-m-depth drill. InFebruary 1999, we intend to drill to bedrock on the topof Cerro Tapado, install a meteorological station andperform various experiments to study post-depositionaleffects. This 40 m long ice core should recover the20th century and more and will allow to evaluateanthropogenic influences and to reconstruct theintensity of past El Nino-Southern Oscillation events.

ACKNOWLEDGMENT

This work was supported by the Swiss NationalScience Foundation, Project #. 2100-050854.9.

REFERENCES

[1] L.G. Thompson et al., A 25,000-Year TropicalClimate History from Bolivian Ice Cores, Science,282, 1858-1864(1998).

[2] M. Vuille (1996): Zur raumzeitlichen Dynamik vonSchneefall und Ausaperung im Bereich dessiidlichen Altiplano, Sudamerika. Dissertation,University of Berne, Switzerland, Institute ofGeography.

[3] Ing. T. Cuevas, Direccion General de Aguas,Santiago, Chile.

151

ICE THICKNESS AND FIRN TEMPERATURE MEASUREMENTS ON THE CERROTAPADO GLACIER (CHILE)

P. Ginot, S. Hubener, U. Schotterer (Univ. Bern), M. Funk (ETH Zurich), M. Schwikowski (PSI),H. W. Gaggeler (Univ. Bern & PSI)

A radar survey on the Cerro Tapado (Chile) summit glacier indicates that the maximal ice thickness is44 m. Firn temperatures of -9°C and -10.5°C at 10 and 9 m below the surface, respectively, suggest thatthis is a cold glacier which is frozen to bedrock.

Cerro Tapado (5550m, 30°08'S, 69°55'W) is one ofthe first mountains south of the "Dry Diagonal" inSouth America where glacier studies and ice coredrilling are possible.

In addition to exploratory drillings of shallow firn cores[1], the glacier thickness was determined by a radarsurvey. On 21 February 1998, the first day on the top,radar profiles were obtained from a weak depressionin the south-east of the principal summit (depression 1,profiles L, Q1, and Q2 in Fig. 1). Measurements werefocused mainly on this depression, the outflow of the"Col Ginot", since topography and the snow coverwere more favourable than in depression 2 (Fig. 1).Depression 2, nearer to the summit, was profiled on27 February (profile Q3).

The radar instrumentation is composed of a transmit-ter and a receiver connected to dipolar antennas. Thetransmitter consisted of a generator of mono-fre-quency pulsation whereas the receiver was a digitaloscilloscope coupled to a computer [2].

Cerro Tapado

Verlauf dot

( h 2 < h i j

100 m

Fig. 1: Summit region on Cerro Tapado with radar pro-files (L, Q1, Q2 and Q3) and drilling sites (BL1and BL2).

The transmitting and receiving antennas were laid outin parallel on the surface of the glacier with a distanceof 30 m. They were oriented perpendicular or parallelto the prolongation of the depressions. The emitted

electrical signal is propagated towards the receiverwhich measures the differences in time between thedirect beam (propagated into the air) and the beamreflected by the bed-rock of the glacier. Knowing thepropagation velocities of the signal in snow and ice,this difference in time allows to determine thethickness of the glacier.

Fig. 2: Radar profile L, glacier thickness.

With the 13 radar shots carried out, the thickness ofthe glacier could be precisely determined (2 to 5%error). Depression 1 has a thickness of 42 m possiblyincreasing in the direction of the flow (Fig. 2), whereasdepression 2 has a thickness of 44 m.

The temperature measurements in the two boreholeswere performed with thermistors (YSI 44006). Thetemperatures at the bottoms of the boreholes 1 (10 m)and 2 (9 m) were -9°C and -10.5°C, respectively.These values suggest that the glacier maintains atemperature below 0 °C down to bedrock and, thus,preserves the chemical stratigraphy, which is aprerequisite for paleo-envi ran mental studies.

REFERENCES

[1] P. Ginot et al., this Annual Report.

[2] M. Funk, G.H. Gudmundsson, F. Hermann,(1994). Geometry of the glacier bed of the Un-teraarglacier, Bernese Alps, Switzerland. Zeit-schrift fur Gletscherkunde und Glazialgeologie,30, 1-8.

152

QUANTIFICATION OF SULPHATE SOURCES FROM AN ALPINE ICE CORE

M. Schwikowski (PSI), A. Doscher, H.W. Gaggeler, U. Schotterer (Univ. Bern & PSI)

Opposite to greenhouse gases, sulphate aerosol particles are expected to cause climate cooling, but un-certainties exist about source variability and strength. We analysed an ice core from a European glacier toquantify source strengths of aerosol-borne sulphate over a 200-year period. Anthropogenic emissions ofS02 were responsible for about 80% of total sulphate in the industrial period. In the pre-industrial periodmineral dust was the dominant contributor.

In this study, we analysed an ice core from a high-elevated glacier in the Swiss Alps (Colle Gnifetti,Monte Rosa massif, Swiss Alps, 4450 m asl,7°51'52"E, 45°53'33"N). Being located in the Euro-pean Alps, which are surrounded by highly populatedand industrialised countries, the Colle Gnifetti glacier iswell suited as an archive for the history of air pollutionin Europe. The site has been glaciologically charac-terised and meets the requirements of an environ-mental and climate archive. We aim to identify differentsources of aerosol SO/' and to estimate their contribu-tions to the observed concentration trend by usingSO4

2" concentrations in the ice extending back overtime through the industrialised period to the pre-indus-trialised period. Concentration records consist of 1451analysed samples for the top 70 m ice core, coveringthe time period 1756-1982. In order to investigate theindustrial and pre-industrial impurity inventory of snowprecipitation, two time periods were selected: 1963-1981 (industrial, 265 samples) and 1756-1870 (pre-industrial, 572 samples). The industrial period repre-sents the maximum in the SO/ concentration record,whereas the definition of the pre-industrial period wasbased on the finding that the concentrations of SO4

2'began to increase significantly at about 1880 [1].

For quantifying the contribution of sea-spray to thesulphate concentrations, Na+ was selected as sea-salttracer, taking a SO/" to Na+ ratio of 0.12. For the in-dustrial period a contribution of 0.23 u.eq I'1 (2%) ofsea-salt SO4

2' to the total SO/" was obtained, whereasfor the pre-industrial time the contribution amounted to0.20 |i.eq I'1 (7%). Thus, no change in the absolutecontribution of sea-salt SO/' was observed from pre-industrial to industrial times, reflecting the concentra-tion record of Na+ which does not show a trend overthe studied time period.

Mineral dust aerosols, particularly those originatingfrom the Sahara, play an important role for the chemi-cal inventories of Alpine glaciers, since Saharan dustdepositions occur almost annually, but with greatvariation in intensity. Although Ca2+ and Mg2+ are onlyminor constituents of mineral dust aerosols, which arein the case of Saharan dust dominated by quartz orcalcite, both are highly correlated, corroborating acommon source. Thus, they can both be used in orderto quantify the terrestrial source of SO/". However, incontrast to sea-salt, no universal SO/7Ca2+ or SO/'/Mg2+ ratio exist for mineral dust. Nevertheless, a good

correlation between non sea-salt SO/" (nssSO/) andnssCa2+ was observed for the pre-industrial period([nssSO/] = 0.175 • [nssCa2+], r2 = 0.68, concentra-tions in u.eq I"1). The observed slope of 0.175 is inter-preted as the pre-industrial nssSO/YnssCa2* ratio ofmineral dust in snow at the Colle Gnifetti glacier. Theterrestrial SO/" contribution can now be estimated forthe entire time period by using the pre-industrialnssSO/7nssCa2+ ratio. For the industrial period mineraldust SO/" amounted to 13%, and in the pre-industrialperiod to 57% of the total SO/', thus being the domi-nating contributor in this period.

We assume that the exSO/, which remains after cor-rection for the sea-salt and the terrestrial contribution,originated exclusively from oxidation of SO2 in theatmosphere. This SO2 was emitted either from anthro-pogenic or natural sources. Temporal changes in SO2

emissions can thus be deduced from the exSO/' rec-ord. During the industrial period exSO/" constituted themajor portion, i.e. 85% of the total SO/", whereas inpre-industrial time it accounted only for 36%. In Fig. 1the historical development of exSO/" concentrations isillustrated, revealing a strong increase from 1870 untilabout 1970, which we interpret directly as a record ofanthropogenic SO2 emissions. ExSO/ increasedfrom 0.047 mg I"1 in the pre-industrial to 0.602 mg I"1 inthe industrial time period, equivalent to a factor of 13.After 1970 a pronounced downward trend is observed,likely reflecting air pollution control measures and pos-sibly also the economic downturn at that time.

.9•5

1760 1780 1B0O 1820 1840 1860 1880 1900 1920 1940 1960 19;Year[A.D.]

Fig. 1: Records of 5-year averages of exSO/, and ofsea-salt as well as mineral dust derived SO/".

REFERENCE

[1] A. Doscher et al., Water, Air and Soil Pollution 85,603(1995).

153

CALIBRATION OF A HISTORICAL AEROSOL CONCENTRATION RECORD FROMAN ALPINE ICE CORE BY LONG-TERM AEROSOL MONITORING

M. Schwikowski, S. Brtitsch (PSI), H.W. Gaggeler, U. Schotterer (Univ. Bern & PSI)

Chemical analyses of environmentally relevant species were performed on an ice core from Fiescherhornglacier, Swiss Alps (3890 m asl). This glacier site is unique since it is located nearby the high-alpineresearch station Jungfraujoch (3580 m asl). Good agreement between the sulphate palaeo record anddirect atmospheric measurements was found (f=0.41). Thus, a "calibration" of the palaeo record over asignificant period of time could be conducted, revealing an average scavenging ratio of 180 for sulphate.

Chemical analyses of aerosol related, environmentallyrelevant species (Ca2+, Cl\ K+, Mg2+, NH4

+, NO3, Na+,and SO4

2") were performed on an ice core fromFiescherhorn glacier, located in the north of the mainAlpine chain, in the Bernese Alps, Switzerland (3890m asl). The 77 m long ice core was dated by annuallayer counting using the seasonally varying signals ofthe tritium concentration and of the isotopic ratio 818Oresulting in a time scale with an uncertainty of only + 1year. The core covers the time period 1945-1989, andthe observed high net accumulation of 1 to 2 m waterequivalent per year allows for high resolution to beobtained in the concentration record.

Since 1973, aerosol SO/' concentrations have beendetermined with a time resolution of 24 h at the high-alpine research station Jungfraujoch [1]. The Jung-fraujoch is located about 6 km west of the Fiescher-horn glacier at an elevation of 3580 m asl, and weassume that atmospheric concentrations are compa-rable at both sites due to the small horizontal and ver-tical distances. In order to calibrate the FiescherhornSO/' ice record with the atmospheric data, annualaverages were calculated for both over the time periodwith complete annual data sets (1974-1988). A highertime resolution could not be deduced from the palaeorecord, since precipitation occurs rather episodicallymaking an exact timing impossible. Both records cor-relate well as illustrated in Fig. 1 ( [ S O / ^ J = 180[SO4

2'(air)], 1^=0.41, concentrations in |ig kg'1). The con-centrations of SO4

2' in the atmosphere are, thus,indeed preserved in the ice core. Although no long-term atmospheric measurements exist for the otherspecies investigated in this study, it can be assumedthat their concentrations in the air are equally wellrecorded in the glacier ice, since as aerosol-borneconstituents they are expected to show a similarbehaviour. From the linear regression analysis, a massscavenging ratio W of 180 for SO4

2' (range for individ-ual years: 89-455; average: 186±111) was obtained(W = pa C,/Ca, with Cs = concentration in snow in \ig I'1,Ca = concentration in air in u.g m'3, and pa = density ofair with a value of 0.875 g I"1 for the conditions at theJungfraujoch). This value is in good agreement withscavenging ratios for SO4

2' deposited on the Green-land ice sheet at Dye 3 (180+120, three summersnowfalls, [2]), at Dye 3 calculated using atmosphericconcentrations measured at coastal sites (100-200 inwinter, 200-400 in summer [3]), and at Summit,

Greenland (220±129, eight summer snowfalls [4]). Ahigher scavenging ratio of 350 was reported fromsimultaneous aerosol and snow sampling at the Alpinesite Weissfluhjoch Davos, Switzerland, for a threemonth sampling period [5], whereas values from singleevent studies at the Jungfraujoch and the high-alpinesite Sonnblick, Austria are significantly higher (710 [6],880 and 1500 [7]). The variety of reported data possi-bly illustrates the complexity of the scavenging proc-ess. However, it seems that event based studiesreveal higher values, probably because they reflectonly the ambient conditions during synoptic situationswith precipitation. Hence, it is particularly important forthe interpretation of palaeo records to establish therelationship between atmospheric and palaeo data forthe respective conditions at the glacier site.

80 82Year

Fig. 1: Annual SO42' averages in the Fiescherhorn ice

core (solid line) and in atmospheric samplesfrom Jungfraujoch (dashed line).

REFERENCES

[1] R. Gehrig, Annual EMEP Reports (1996).[2] C.I. Davidson et al., Atmos. Environ. 19, 2065

(1985).[3] C.I. Davidson et al., Atmos. Environ 21, 871

(1987).[4] C.I. Davidson et al., NATO ASI Series, Springer,

Berlin, pp. 275-306(1996).[5] U. Baltensperger et al., Atmos. Environ 27A,

1881 (1993).[6] M. Schwikowski et al., Atmos. Environ 29, 1829

(1995).[7] D. Wagenbach, in: Cloud Multi-phase Processes

and High Alpine Air and Snow Chemistry,Springer, Berlin, pp. 177-199 (1997).

154

COMPARISON OF DATING METHODS FOR AN ICE CORE FROM AN ALPINE SITE

A. Eichler (Univ. Bern & PSI), V. Furrer, M. Schwikowski (PSI), H.W. Gaggeler (Univ. Bern & PSI)

Three different methods were used for dating of a high-alpine ice core: annual layer counting usingammonium summer maximum concentrations, dating by the exponential decay of the radioactive isotope210Pb, and application of a kinematic ice flow model.

Ice cores of high-alpine glaciers are important archivesto study the regional air pollution and climate history.One crucial point for interpreting ice core data is theexact dating of the core. The methods used for thedating of alpine ice cores are:

• Annual layer counting of seasonally varyingparameters (e.g. concentrations of NH4

+ and 81BO)• Dating by radioactive decay of isotopes• Use of stratigraphic markers like well known

Saharan dust events or volcanic eruptions• Application of ice flow models.

In October 1994, a 125 m ice core was drilled at upperGrenzgletscher (Monte Rosa massif, 4200 m a.s.l.)[1].So far, the complete core (2320 samples,104 m w.eq.)was analysed for major and minor ions [2]. 210Pb analy-ses were performed only for the upper 90 m w.eq.

25

20

£ 102

;1993

: 1 90

Mflrl

85

• ! • •

81

il§V

79

I

II

7.7k

74

|

ill1

66

A 61

m55

46

2000 4000 6000 8000

Depth (cm water)

10000

Fig. 1: Record of NH4+ (means of 50 cm core pieces).

First, dating was carried out using the large seasonalfluctuations of NH4

+ (Fig. 1) with significantly highersummer than winter concentrations, which are due tothe emission and transport characteristics described in[2]. This layer counting seems successful down to theyear 1977 (45 m w. eq.), which is marked by largeamounts of calcium from the well known Saharan dustevent in this year. Because of the thinning of annuallayers with depth, dating is more uncertain in thedeeper parts.

Because of a half-life of 22.3 years dating with 210Pbcould be possible for a time scale of at most 200years. For dating, a constant input has to be assumed.In the Grenzgletscher ice core the decay of the isotopeis superimposed by transport phenomena with higherconcentrations in summer, and lower concentrations inwinter snow [3]. In order to take into account thetransport term we divided the 210Pb activities by theNH4

+ concentrations, and this ratio shows the expectedstrong exponential decrease with depth (Fig. 2), which

allows a dating of the core. In comparison to the NH4+

dating we got a good agreement with an uncertainty of±4 years for the period 1952-1994.

y = 0.916.e"c

R = 0.79

2000 4000 6000Depth (cm water)

8000

Fig. 2: 210Pb/NH4+ratio of the upper 90m w.eq. (points)

along with an exponential fit (solid line).

In addition, dating was performed by modelling of theglacier with a simple kinematic ice flow model [1,4].This model assumes a constant ice velocity v from theglacier surface down to a depth h and then a linear de-crease of v to zero between h and the bedrock. Byusing constants (e.g. glacier thickness: 170 m w.eq.)the age-depth relation was fitted with the given fitfunction [1,4] through 14 reference points receivedfrom dating with NH4

+ (Fig. 3).

2000h=136±25m w.eq.

•+•

2000 4000 6000 8000Depth (cm water)

10000

Fig. 3: Reference points and fitted curve (solid line).

This simple model seems applicable to describe theflow pattern at the Grenzgletscher, and yields a meanaccumulation rate of 1.9 m w. eq.

From our three dating methods we determined the ageof the deepest part in the ice core to be 1940±5. Otherparameters (e.g. tritium activity) will be investigated inorder to corroborate the dating.

REFERENCES

[1] A. Doscher, Dissertation, Universitat Bern 1996.[2] A. Grund, PSI Ann. Rep. 1997, Annex MIA, 43.[3] A. Grund, PSI Ann. Rep. 1997, Annex MIA, 44.[4] S. J. Johnson et al., Nature 235, 429 (1972).

155

SAMPLING AND ANALYSIS OF ICE CRYSTALS AS A FUNCTION OF SIZE

S.H. Ehrman, U. Baltensperger, M. Schwikowski, S. Brutsch (PSI), H.W.Gaggeler (Univ. Bern & PSI)

A new method was applied to collect size-classified snow samples. Coarse and accumulation modeaerosol constituents showed a different dependence of concentration on crystal size.

In addition to the so-called 'greenhouse gases',aerosols are believed to affect global climate. In thiscontext, two sub-populations of aerosol are ofparticular importance, cloud condensation nuclei(CCN) and ice nuclei (IN). The magnitude of thisaerosol climate forcing is uncertain, particularly themagnitude of the indirect effects associated with cloudformation. To reduce this uncertainty, an improvedunderstanding of the relationship between clouds andaerosols is needed. This information is also useful forinterpreting paleoclimatology data from glacier icecores since incorporation into cloud droplets and snowcrystals during precipitation events, either as CCN andIN or via riming (accretion of supercooled clouddroplets), is an important mechanism for deposition ofaerosols onto glaciers.

Size-classified snow samples were collected duringSpring, 1998, at the Jungfraujoch High-AlpineResearch Station (3454 m). A procedure modifiedfrom the Guttalgor method, originally developed forsize-selective sampling of raindrops by Bachmann etal. [1] was used. Ice crystals were collected into adewar containing liquid nitrogen and a cascade ofstainless steel sieves, 10 centimeters in diameter, withmesh diameters of 1.0 mm, 0.20 mm, and 0.025 mm.Because the density of ice is greater than that of liquidnitrogen, the crystals sink upon impact. In addition tothe size-classified samples, formvar replicas weretaken every 20 to 30 minutes for determination of thedegree of riming and to record the crystal habit of theice crystals.

The samples were stored and transported frozen, andanalysed within a week of collection using ionchromatography for Na+' NH4 \ K+, Ca2+, Mg2+, CI",NO3" and SO4

2~. Although there is some dispute as tothe precise lower cutoff, it has been observed in fieldstudies that growth by riming is negligible below acertain size, 0.3 mm in diameter for hexagonal snowcrystals and 0.05 mm for columnar crystals [2]. Themotivation for size-classified sampling of ice crystalswas the hypothesis that the contributions of thenucleation and riming processes to the bulk snowcomposition could be estimated by comparing thecomposition of the smallest ice crystals to thecomposition of larger crystals.

Size dependent concentrations of the ions Caz+ andNO3" are shown in Figure 1. Ca2+, Na+, \C, Mg2+, andCI' are representative of coarse mode aerosol, and theresults were similar for each ion; the concentrationwas greatest in the smallest size fraction. For theremaining ions, NH4

+, NO3", and SO42", no significant

variation in concentration with ice crystal size wasobserved. The results for NO3", shown in Figure 1b,are representative. These preliminary results suggest

that the species associated with the coarse aerosolmay preferentially nucleate ice crystals from the vaporphase or act as crystallisation nuclei in cloud droplets.In this case, the chemical species associated with icenucleating aerosols should be concentrated in thesmallest sized ice crystals. Since these species arepreferentially removed from the aerosol populationearly on, they are not available to participate in theriming process, and hence should show a decrease inconcentration with increasing size, resulting fromdilution as the ice crystal grows. The relativelyconstant concentration of NO3" with increasing crystalsize suggests crystal growth by riming or gas-phasediffusion.

a)

fo

oo

1200

1000

800

600

400

200

01. Apr 98

02. Apr 98

19. Apr 98

- - - 11-Jun-98#1

11-Jun-98#2

^ 4 6 8Sieve diameter, mm

10

b)

4000

3 3000c

2000

S 1000co0 0

0 2 4 6 8 10Sieve diameter, mm

Fig. 1: Concentration of Ca2+ (a) and NO3" (b) in sizeclassified snow samples.

ACKNOWLEDGEMENT

S. Ehrman acknowledges support from the U.S.National Science Foundation, grant INT-970428.

REFERENCES[1] K. Bachmann, A. Roder, I. Haag, Atmos.

Environ., 26A, 1795-1797 (1992).[2] T. Harimaya, J. Meteorol. Soc. Jpn., 53, 384

(1975).

156

- • • • -

T. »« *

f

Fig. 1: SEM image of a small hexagonal crystal sam-pled directly using liquid nitrogen. Letters aand b indicate areas analyzed using EDS.

IMAGING AND ANALYSIS OF INDIVIDUAL ICE CRYSTALS

S.H. Ehrman, O. Gebhardt, U. Baltensperger, M. Schwikowski (PSI), B. Frey (WSL)

Single ice crystals were preserved by the formvar technique as well as directly onto precooled substratesand chemically analyzed by SEM/EOS and SIMS.

Scavenging of aerosol particles by ice crystals is an - • • • •important mechanism for removal of pollutants from ", •", ' . _ , - •the atmosphere. Because of the multiple scavengingprocesses which can occur, including nucleation,inertial impaction, Brownian motion, and accretion ofsupercooled cloud droplets (riming), the relationshipbetween the concentrations of pollutants in theatmosphere and in bulk snow is quite complex [1]. Thelocation of the particulate matter in the crystalsuggests the scavenging mechanism, i.e. nucleationscavenging for particles at the center of the crystal,and riming (growth by accretion of cloud droplets) ordirect impaction for particles located at or near theedges.To estimate the relative importance of the nucleationand riming mechanisms for aerosol scavenging,attempts were made to determine chemical composi-tion as a function of position in the individual icecrystals using cryo scanning electron micros-copy/energy dispersive X-ray spectrometry(SEM/EDS) and secondary ion mass spectrometry(SIMS). The objective was to differentiate betweenthe chemical species contained in the ice nuclei, pre-sumably located at the center of the crystals, and thecomposition of the aerosol contained in the rimedroplets visible on the surface of the crystals. Crystalreplicas were collected onto nickel foil substratesusing the formvar method described by Mosimann etal. [2]. This technique preserves the shape of thecrystal and any non-volatile non-aqueous material inthe crystal is retained in the formvar resin. Ice crystalswere also collected directly onto precooled methylcellulose-coated SEM substrates as described byRango et al. [3], stored in liquid nitrogen, and analysedusing a liquid nitrogen-cooled sample stage to preventevaporation of the crystal under the electron beam. ASEM image of a small hexagonal crystal is shown inFig. 1. A SEM image of a formvar ice crystal replica ofis shown in Fig. 2.

Chemical analysis using EDS of both the ice crystalsand formvar replicas proved difficult. It was not pos-sible to resolve statistically significant peaks corre-sponding to the aerosol material contained in thecrystals; background radiation generated from theinteraction of the electron beam with the ice crystal orformvar dominated the spectra. However, with sec-ondary ion mass spectrometry (SIMS), a much moresensitive technique than EDS, at magnifications of2000x or greater, enriched areas a few microns indiameter were observed in the formvar. These areasare believed to correspond to particulate materialcontained in the ice crystal. Future work will be con-cerned with making a qualitative comparison betweenthe chemical composition, measured using SIMS, ofthe rimed and unrimed areas of the formvar crystalreplicas.

500 im

Fig. 2: SEM image of a formvar snow crystal replica.Squares indicate areas analysed using SIMS.

ACKNOWLEDGEMENTS. Ehrman acknowledges support from the U.S. Na-tional Science Foundation, grant INT-970428.

REFERENCES[1] H.R. Pruppacher, J.D. Klett, Microphysics of cloud

and precipitation, D. Reidel Publishing Company,Holland, p. 497(1980).

[2] L. Mosimann, E. Weingartner, A. Waldvogel, J.Atmos. Sciences., 58, 1548-1558 (1994).

[3] A. Rango, W.P. Wergin, E.F. Erbe J. HydrologicalSci. 41, 219-233 (1996).

157

INVESTIGATING SCAVENGING PROCESSES OF ICE CRYSTALS

S. Tenberken, M. Schwikowski (PSI)

First laboratory experiments for separate sampling of ice crystals and supercooled cloud drops wereperformed.

In order to relate concentrations of trace species inglacier ice cores to past atmospheric conditions it isnecessary to investigate transfer mechanismsbetween the atmosphere and glacier ice. Since glacierice is formed by snow, one important point is the in-corporation of trace species in snow which consists ofaggregated ice crystals. Consequently, more informa-tion about transfer mechanisms of pollutants to glacierice will be obtained by investigating scavengingprocesses of ice crystals.Concentrations of pollutants in ice crystals are mainlydetermined by following processes:

• nucleation scavenging

• condensation of water vapour

• riming of ice crystals by supercooled cloud drops

• aerosol particle scavenging

• gas scavenging

Especially the small (d < 50 u.m) ice crystals which arejust generated are of interest because the process ofice crystal formation is not completely understood. Toinvestigate the riming of ice crystals both, supercooledcloud drops as well as ice crystals have to be sampledand analysed. Therefore, it is necessary to separateice crystals from cloud drops which normally coexist inclouds. One approach is the application of an inertialimpaction system. Supercooled cloud drops shouldfreeze when contacting the first impactor stagewhereas ice crystals should bounce off because theywill not stick to the surface of the impactor stage. At asecond stage ice crystals can then be sampled as bulksample. In order to examine this assumption, first labo-ratory experiments were performed. A scheme of theexperimental setup is shown in Fig.1.

compressed air-ultra-filter mass flow

controller(0 - 30 Lmin'')

Fig. 1: Scheme of experimental setup.

Drops are generated by nebulizing a solution using apressure-nebulizer. The fog is cooled by liquid nitrogenin a cooler in order to freeze and afterwards suckedthrough an impactor, which is a 5-stage Small DepositArea Impactor (SDI) [1]. Table 1 shows characteristiccut-offs of the impactor. The gas flow is controlledusing a mass-flow controller. Vacuum is obtained byapplication of two pumps.

Table

20 °C,

0°C,

1 : Charactenstical cut-offs of

1010 mbar

1010 mbar

8.45 4.07 2.59

8.14 3.92 2.49

the SDI

1.66

1.6

1.06

1.02

0.35

0.34

The fog freezes in the cooler and is sampled by theimpactor. Sampled drops will evaporate in the impac-tor and leave the solutes originally dissolved in thenebulizer solution. After sampling, the residues aredissolved again and analysed. If the frozen drops infact will not stick on the fist stage and will bounce offno residue should be found on the first stage. Two ex-periments were performed. The first experiment wasperformed at ambient conditions at 20 °C. A solution of100 u.M Cs+ was nebulized and sampled for 2 h with amass-flow of 11 Lmin'1. After sampling, the impactorstages were extracted by 400 \iL of Milli-Q water.Analysis was performed by capillary electrophoresis.Fig. 2a shows the result of this experiment. Cs+ isfound on the first stage as well as on the last stages.

& 8

I S

S 2

I 0diameter / urn diameter/jim

Fig. 2: Concentration of Cs+ residue: a) no cooling, b)with cooling.

In a second experiment the fog was cooled so thatfrozen drops were formed. Sampling and analysis wasperformed at the same conditions as in the firstexperiment. In contrast to the first experiment a resi-due of Cs+ was only found on the last stage of theimpactor (see Fig. 2b). This shows that frozen dropsindeed bounce off and do not stick on the first stage sothat they are sampled on a later stage of the impactor.The same experiment was repeated with an ice layeron the first impactor stage, and again no residue ofCs+ was detected on the first stage.

To conclude, these results show that it is possible toseparate ice crystals from supercooled drops. There-fore, this technique will be used in future field experi-ments at the Jungfraujoch to sample ice crystals andsupercooled cloud drops separately.

REFERENCE

[1] W. Maenhaut et al., Nucl. Inst. Meth. Phys. Res.B 109, 482-487(1996).

158

CONTINUOUS MELTING AND ANALYSIS OF ICE CORES

T. Huber (Univ. Bern & PSI), M. Schwikowski (PSI), H. W. Gaggeler (Univ. Bern & PSI)

A heatable head for continuous melting of ice cores has been developed which will in combination with twoion chromatographs allow for simultaneous on-line analysis of the most common anions andcations.

The common way of analysing ice cores is to cut thecores in pieces, melt them and analyse each pieceseparately. This procedure is time consuming and therisk of contamination is relatively high. By continuouslymelting and analysing the ice cores, these two effectscan be reduced. This so called continuous flowanalysis (CFA) was successfully applied for recordingconcentration profiles of different tracers in Greenlandice [1, 2]. Most of the tracers are determined byfluorescence or absorption spectrometry, thus, foreach species an individual detector is needed. Instead,our approach is to couple the melting device with amulticomponent analytical method such as ionchromatography or ICP-mass spectrometry.

For this purpose, an aluminium melting head was builtaccording to Sigg et al. [1]. The surface of the head iscoated with PTFE. The head can be heated by 3 heatsources (100 W each), which are controlled by athermocouple combined with a transistor. A crosssection of maximum 12.5 cm2 of an ice core is placedover the melting head, stabilized by a sample holder.The amount of melt-water can be regulated bychanging the heating temperature and the diameter ofthe inner circle of the melting head. This is necessarybecause of the different densities of firn and ice. Themelt-water is separated in an inner and outer fraction.For the following analyses, only the inner fraction isused, while the outer section is rejected. This results ina decontamination of the samples to be analysed. Theinner sample is filtered and pumped over a valve ontwo different concentrator columns, one for anions andone for cations. By switching the valve after a definitetime, the collected ions are eluted from theconcentrator columns to the corresponding ionchromatographs, where the ions are separated anddetected. During this separation, the newly melted ice-water is concentrated on two other concentratorcolumns. These samples are analysed afterwards onthe same ion chromatographs. The analysed ions arefluoride, chloride, nitrate and sulphate for the anionsand sodium, ammonium, potassium, magnesium andcalcium for the cations. A scheme of the apparatus isshown in Fig. 1.

There are two main problems of using concentratorcolumns: the first is column overload, which results ina non-quantitative stripping of the ions of interest. Thesecond problem is, that with high concentrations ofhigh affinity ions such as calcium and sulphate, lowaffinity ions such as sodium and chloride can bedisplaced on the ion exchange resin and can thereforeget lost.

Both problems have to be considered in our case, asthere can be high concentrations of calcium andsulphate in ice cores [3]. Concentrations of inorganicions in an ice core from a high-alpine glacier [3]showed that with 1 ml of sample over 99.5 % of allanalysed samples could be determined without columnoverload or losses. Only samples containing Saharandust would show loss in low affinity ions. This can beavoided by reducing the amount of concentratedsample, which means reducing the meltingtemperature or the inner diameter of the melting head.

ice core

IC anions

waste

Fig. 1: Scheme of the melting head combined withtwo ion chromatographs.

Future work includes determination of the melting ratein dependence of temperature and the density of theice. Another point is the calibration of the method. Forthis, standard solutions have to be frozen andanalysed by melting and concentrating. Furthermore,memory effects have to be considered. At last, acomparison between conventional sample preparationand analysis and the melting technique has to bedone.

REFERENCES[1] A. Sigg et al., Environ. Sci. Technol. 28, 204

(1994).[2] K. Fuhrer et al., J. Geophys. Res. 101, 4147

(1996).[3] A. Doscher, Dissertation, Bern 1996.

159

THE INFLUENCE OF STRATOSPHERIC INTRUSIONS ON ALPINE OZONECONCENTRATIONS

AStohl, N. Spichtinger-Rakowsky (Univ. Munchen), P. Bonasoni (ISFFC Bologna), H. Feldmann,M. Memmesheimer (Univ. Koln), H. E. Scheel, T. Trickl (IFU Garmisch-Partenkirchen), S. Hubener

(Univ. Bern), L Tobler(PSI), H.W. Gaggeler (Univ. Bern & PSI)

Two independent climatologies of direct stratospheric intrusions for several sites in the Alps were estab-lished and used to evaluate the influence of stratospheric intrusions on Alpine ozone concentrations.

During pre-industrial times, surface O3 concentrationswere much lower than today [1], and stratospheric in-trusions were the most important source for tro-pospheric O3. The magnitude of their contribution tothe present-day tropospheric O3 budget is still underdiscussion.

In this work two climatologies of ozone brought downinto the troposphere with intruding stratospheric airwere established. The first one is based on measure-ments of 7Be, relative humidity and ozone at Zug-spitze, Sonnblick, and Jungfraujoch, the other one re-sults from a three-year transport model simulationconsidering stratospheric ozone as a passive tracer.The O3 and relative humidity data used to establish thefirst climatology are 1-h averages. The 7Be concentra-tions are 24-h averages except Jungfraujoch whereonly 48-h data are available. In the screening algo-rithm applied to the measurement data in order to de-tect stratospheric intrusions a day is considered to bean "intrusion day" if (i) the 7Be concentration is higherthan 8 mBqm"3, (ii) the daily minimum of relative hu-midity is lower than 30 %, and (iii) the daily maximumO3 concentration is at least 10 % above the monthlymean concentration. The frequency of stratosphericintrusion days as found with this detection algorithm isshown in Fig. 1.

ZugspitzeSonnblick

Fig 1: Frequency of stratospheric intrusion days atZugspitze, Sonnblick, and Jungfraujoch.

The average frequency of intrusions is 5.5 % at Zug-spitze, 10 % at Sonnblick, and 17 % at Jungfraujoch.

The second climatology is based on a simulation ofstratospheric intrusions using the Lagrangian particlemodel FLEXPART [2] and gives average concentra-tions of both an O3 and an air tracer originating in thestratosphere.

A significant agreement between the two climatologiesis found for Zugspitze and Sonnblick. The deep sum-mer minimum of the frequency of stratospheric intru-sions is confirmed by the model results. The agree-ment for Jungfraujoch is less good. The observationsshow no clear seasonal variation, whereas the mod-eled O3 tracer concentrations show a late winter/earlyspring maximum and a deep summer minimum alsofor Jungfraujoch. This disagreement is reduced whenthe different periods of the climatologies are consid-ered. The reason for the high frequency of observedintrusions at Jungfraujoch can hardly be explainedonly by the high altitude. Possibly the detection algo-rithm, which was set up by studying intrusions mostlyfor Zugspitze, is less suited for Jungfraujoch. In anycase, more measurements are needed to establish amore realistic climatology.

The model results indicate that stratospheric intrusionscontribute only a few ppb to the ozone found in theboundary layer but they are an important term in theozone budget of the free troposphere, especially inlate winter/early spring. During that season strato-spheric intrusions occuring within the limited modeldomain contribute between some 15 and 25 % to theozone concentrations measured at high Alpine peaks.The total contribution, including intrusions from outsidethe model domain, will be even higher. Further studieswith an extended model domain are planned to quan-tify this contribution.

ACKNOWLEDGMENT

This study was part of the EU research projectVOTALP and was funded by the European Commis-sion and the governments of Switzerland and Austria.

REFERENCES

[1] A. Volz, D. Kley, Nature 332, 240 (1988).

[2] A. Stohl, Atmos. Environ. 32, 947 (1998).

160

CHARACTERIZATION AND USE OF THE NEW IRRADIATION FACILITY AT SINQFOR NEUTRON ACTIVATION ANALYSIS

L Tobler, D.T. Jost (PSI)

The new pneumatic tube transfer system for neutron activation analysis (NAA) installed at SINQ has beencharacterised for operation and irradiation parameters, i.e., absolute neutron flux, flux variations within thetransport container, variations with time, and interfering reactions.

Instrumental neutron activation analysis (INAA) is asensitive, non-destructive analytical method for thedetermination of trace elements in preferably solidsamples. The new pneumatic tube transfer system [1]installed at the spallation neutron source SINQ allowsthe irradiation and fast transfer of samples for INAA.The operation and irradiation characteristics of thisfacility are important for accurate and reliable results inINAA. Different parameters have been evaluated thisyear during the operational period of SINQ.The absolute flux for thermal and epithermal neutronswere determined by a 30 min irradiation of a Zr foil andmeasurement of the y-lines of the (n,y)-reactions94Zr(n,y)95Zr and 96Zr(n,y)97Zr -> 97mNb -* 97Nb. The fluxesfound a r e :<&th = 2.8 • 1012 (n/(cm2-s)) / mA, and<Depi = 7.3 • 109 (n/(cm2-s)) / mA.

The second, redundant tube system has fluxes whichare only 5 % lower.The neutron flux distribution in axial and radial direc-tion of the transport container was measured by plac-ing copper wires along its centre axis and inside itsbottom circumference. After irradiation the wires werecut into 5 mm long pieces for counting. Neutron fluxvariations in the radial direction of the transport con-tainer were below measurement uncertainty (2%),while along its central axis (length 4 cm) a decrease inneutron flux of 14 % was found.Reproducibility of successive irradiations (n = 9) wasdetermined by irradiating samples of 100 ^.g Al for 15seconds. Count rates of the measured 28AI - activitiesvaried by 5% among different irradiations after correc-tion for fluctuations of the neutron flux during the irra-diations. This variation is larger than the counting errorand might be due to uncertainties in the irradiationtime and the correction of the neutron flux variations.Beside the thermal and epithermal neutrons, high en-ergy neutrons from the spallation process are presentat the irradiation position. These fast neutrons induce(n,xn), (n,p) and (n,oc) - reactions, which can causeinterferences by producing the same activity as a (n,y)- reaction from a different element. A few reactionswith relatively high levels of interference were investi-gated. The level of interference is given by the activityratio (activity of interfering reaction / activity of thermalreaction) for equal masses of the target elements.Results are shown in Tab le i .

As long as the concentrations of the element to beanalysed and of the interfering element are of thesame order of magnitude, the interference is negligi-

ble. However, in cases, where the concentration ratio(interfering element / analysed element) is greaterthan 10, the interference is in the percent range andhas to be corrected for.

Table 1 : Interferences by fast neutron reactions.

Intendedthermal reaction

23Na(n,y)24Na

26Mg(n,Y)27Mg

27AI(n,Y)2aAI

Interferingreaction

27AI(n,a)24Na

27AI(n,p)27Mg

30Si(n,a)27Mg

28Si(n,p)28AI

level ofinterference

0.4- 10"'

6.9- 1 0 '

5.0- 10"'

2.4- 10"'

Unfortunately, the fast neutrons induce also a high "C- activity in the polyethylene (-[C2HJn-) transport andsample containers by the reaction 12C (n,2n) 11C, whichhas a threshold energy of 20 MeV. 11C is a pure fT -emitter with the characteristic 511 keV y-line frompositron annihilation and with a half-life of 20.38 min-utes. This 11C - activity hampers the measurement ofisotopes with short decay times. Unpacking of thesamples from the irradiated sample vials is sometimesa prerequisite in order to improve the measurementconditions (i.e. better counting geometry, lower deadtime, lower background) [2]. Furthermore, this activityis also disturbing in view of radiation protection.As a first application of the irradiation facility traceelements in size fractionated aerosols were analysed

[2].During the first full operation period of the transportsystem about 160 irradiations with irradiation timesbetween 5 seconds and 40 minutes were performed.Transportation time of samples between the irradiationposition in SINQ and the laboratory is 50 seconds.Minor software and hardware problems occurred withthe pneumatic transfer system. These will be solvedduring the next shutdown period of SINQ, in order toincrease the performance and reliability of the systemfor routine operation.

ACKNOWLEDGEMENTS

Special thanks are due to H.U. Aebersold and A. El-lenberger (ASQ Dept.) for providing the irradiation fa-cilities. The help of O. Morath (radiation safety) andthe work of the staff of SINQ are highly appreciated.

REFERENCES[1] L Tobler et al., PSI Annual Report Annex IIIA,1997.[2] D. Horcajada et al., this Annual Report.

161

GADOLINIUM COMPLEXES FOR TARGETED MOLECULAR RADIOBIOLOGY ANDRADIOTHERAPY

C. Salt, M. Argentini, R. Weinreich (PSI)

In order to evaluate DNA-seeking compounds for targeted molecular radiobiology, several benzimidazolederivatives were coupled with Gd-chelating agents (DTPA, DOTA). The complete conjugates will beactivated in cell cultures by neutrons generated at SINQ.

With respect to make the high-LET short-rangedAuger electrons accessible for targeted radiobiologyand radiotherapy, several DNA ligands were labelledwith radioactive Auger emitters or with inactive butactivatable isotopes. The compound classes to belabelled were DNA components (pyrimidinederivatives) and DNA fluorescence stains (Hoechst33258).

In particular, the gadolinium neutron capture therapytakes advantage of the extraordinarily high neutroncapture cross section of 157Gd (254'000 barns)accompanied by the release of therapeuticallyeffective high-LET Auger electron showers. The limitedrange (approx. 22 nm) of the energy depositionproduced by the Auger and Coster-Kronig transitionsrequires essentially that the gadolinium has to beplaced in the immediate vicinity of the DNA helix [1].In order to assess the reactivity of the specificfunctionalities involved in the key reaction stepsleading to the modified Hoechst 33258 conjugates,several corresponding model ligands and modelchelates were synthesized and characterized.

The fluorophores thus designed represent a simplifiedversion of the Hoechst DNA linkers where instead ofthe bisbenzimidazole one single benzimidazole ringsystem carries the modified phenolic moiety. Thisterminal aminophenyl residue is the functional groupthat will form the amide covalent attachment to the Gd-chelating agent (DTPA, DOTA) via the latters'carboxylic side-chain as will be the case with the morecomplex Hoechst derivatives [2]. Also the N-metnylpiperazinyl residue at the other end of theHoechst molecule is omitted.

The synthesis and characterization of the followingfour DNA ligands (1)-(4) with open-chainedcoordination sites based on the bifunctional chelatingagent DTPA has been achieved (Fig. 1). These DTPA-based model ligands all display the same pronouncedfluorescence as their corresponding Hoechstanalogues. Therefore, the DNA minor-groove bindingproperties of these smaller fluorophores will also be

tested on cell culture samples at the SINQ neutronbeam facility with regard to the enhancement of thefluorescence yield upon groove-binding events in thecells.

The chelation of the model ligands (1) and (2) withGdCl3 has lead to the corresponding gadoliniumchelates which are in the process of characterization.

Fig. 1: Model Compound Open Chain DTPA Ligands.

REFERENCES

[1] A.D. Whittaker et al., in Advances in NeutronCapture Therapy (A.H. Soloway et al., eds),Plenum Press, New York 1993, 383-387.

[2] K. Wiederholt et al., J. Am. Chem. Soc. 118(1996)7055-7062.

162

MEASUREMENT OF THE RADIONUCLIDE INVENTORY IN ACCELERATOR WASTE

R. Weinreich, M. Argentini (PSI)

As an essential part of waste management the nuclide inventories of radioactive waste which has beengenerated in the PSI facilities must be determined. Preliminary results show an agreement within a factorof 5 with model calculations.

Paul Scherrer Institut is Switzerland's largest producerof so-called middle-active radioactive waste, whichgenerally must be prepared for final storage. Whereasuseful routine methods for the measurement of radio-activity in waste from nuclear power stations havebeen available for a couple of years, the correspond-ing experience in the field of accelerator waste islacking. As a general indication, the main waste frompower stations is located in the fuel elements andconsists of long-lived fission products and a-emittingtransuranes. The accelerator waste, however, is lo-cated in machine and beam-lining elements and con-sists of long-lived activation products, originating bynuclear interactions of accelerated protons and sec-ondary neutrons onto (mostly metallic) targets.Cross section data for nuclear reactions of protonswith metals are well-known mostly for proton energiesbelow 50 and above 500 MeV. However, for our rele-vant energy region between 50 and 500 MeV, the dataare relatively scarce. Many groups have startedmeasuring programs, in order to eliminate this gap [1-3], and in the meantime, it seems that even in thisenergy region data are available which are sufficientfor the development of mathematical models in orderto estimate the activation of accelerator parts duringthe years.

A further uncertainty factor is the inexact knowledge ofthe history of the irradiated material. While irradiationdates, durations and beam intensities generally havebeen recorded, the beam intensity is normally distrib-uted inhomogeneously onto the target material andcauses errors which cannot be recorded quantitativelyby the operator team. Thus, model calculations arehandicapped by inherent uncertainties, and in the firstphase of determining the radionuclide inventory, themodel calculations must be validated by measure-ment.Unfortunately, most radionuclides which can bemeasured easily, for instance by y-spectroscopy, are ofno significance for problems of final disposal: gen-erally their half-life is too short. The really significantnuclides have half-lives of more than 103 years but aremostly p-emitters und must consequently be sepa-rated carefully from the irradiated matrix by chemicalprocesses before measuring. In many cases, themeasurement must even be carried out by acceleratormass spectroscopy.

Table 1 shows the dismantled parts of the PSI accel-erator equipment from which samples were taken, inorder to measure the radionuclide inventories. Theeasily measurable radionuclides Co and Na, arenot nuclides significant for final disposal but might beof some relevance for model validation.

Table 1 : Parts of the PSI accelerator equipment readyfor determination of the radionuclide inventories.

Dismantledpart

"Beamdump" ofTarget E

Target BX2

VariousAluminiumparts

Material

Copper

CopperNormal SteelStainless St.Cast iron

Copper

CopperStainless St.ConcreteLeadCeramics

AluminiumAluminium

activated by

Protons (590MeV)NeutronsNeutronsNeutronsNeutrons

Protons (72MeV)NeutronsNeutronsNeutronsNeutronsNeutrons

ProtonsNeutrons

Main nuclide

60Co

60Co60Co60Co60Co

60Co

60Co6 0 ^ _

Co152Eu,cc-emit.

fission prod.?

22Na22Na

PSI and NAGRA (National Society for Storage of Ra-dioactive Waste) agreed in carrying measurements ofthe following radionuclides in the beam dump: 12.3-a3H, 1.6x106-a 10Be, 5.7x103-a 14C, 7.3x105-a Al, 132-a 32Si, 3.0x105-a 36CI, 1.3x109-a 40K, 1.0x105-a 41Ca,57-a ^Tl , 2.73-a 55Fe, 7.6x104-a 59Ni, 1.5x106-a 60Fe,5.27-a 60Co, 99.6-a 63Ni, and 127-a 10BmAg. For thevalidation of models also developed at PSI [4], how-ever, the measurement of more radionuclides is re-quested.

Preliminary measurements of some y-emitters in ma-terials of the target E beam dump showed an agree-ment within a factor of 5 between the measured andthe calculated data. This error limit seems to be toler-able for the specific requests of the final disposal ofradioactive waste. On the other hand, it seems to beunsatisfactory for a correct scientific validation state-ment.

REFERENCES[1] R. Michel et al., Nucl. Instr. Meth. Phys. Res.

B129 (1997) 153-193.

[2] R. Bodemann et al., Nucl. Instr. Meth. Phys. Res.B 82 (1993) 9-31.

[3] M. Fassbender et al., Appl. Radiat. Isot. 48 (1997)1221-1230.

[4] F. Atchison, PSI Berichte No. 98-09, 98-10 and98-11 (Sept. 1998).

163

CEMENT CHEMISTRY: QUALITY CONTROL AND DEVELOPMENTS 1998

H.P. Zimmermann, M. Patorski, L Dohring, M. Egloff, R. Ivankovic (PSI)

The main fields of our work are mentioned and some numbers relating here to. The distribution our workbetween our customers is shown.

This year, the majority of work was approximatelyequally divided between quality control of radioactivewaste solidifications, conducted for the Swiss nuclearpower plants, including improvements and develop-ment and validation of testing methods.

Since January 1998, 26 new sets of samples con-taining radioactive waste and 17 inactive sets havebeen tested with respect to quality control and for de-velopment purposes (1 set consists of about 20 singlesamples).

This resulted in more than 520 measurements ofcompressive strength, more than 890 y-spectrometricmeasurements, about 600 I of radioactive leachingliquids, and about 10 I radioactive used cement sam-ples.

The distribution of working hours from January toNovember 1998 is shown in Fig. 1. This year thehighest share of working capacity was dedicated tothe development and validation of a new testing pro-cedure for bituminous samples as a replacement forneedle penetration and ring-ball softening pointmeasurement [1]. The second in size was KKM(Muhleberg) due to the amount of delivered samplesets and data evaluation.

For ZWILAG we are setting up some quality controlprocedures for glassy and metallic products of theplasma oven. As a preliminary study some measure-ments of corrosion of inactive metallic samples havebeen done by the LWR group (Light Water ReactorChemistry group from the Laboratory for MaterialsBehaviour). As part of our work it was planned to acti-vate some samples with neutron activation at SINQand to leach it similarly to the cement samples. A testactivation of a metallic sample has been done andfurther activations are planned. An estimate on thebasis of the y-spectrometric measurements of theactivated sample, the results of the LWR corrosiontests and the estimated contents of activity in the realZWILAG metallic waste packages showed that thisprocedure, analogue to the cement tests, shouldwork. A verification by a leach test with small acti-vated samples is planned for 1999, also for glassspecimens.

In connection with the dismantling of the formerresearch reactor DIORIT, we plan some recipe testswith milled graphite from the reactor at the beginningof 1999.

At the begin of 1998 we also produced some series ofspecial cement samples with a high water to cementratio for research done by the Water Constituentsgroup of the Laboratory for Waste Management.

zwi KKM

KKL

KKW

7 Infrastructure

Adm.

NAGRA

Fig. 1: Distribution of working capacity between ourcustomers and several special fields.(ZWI: ZWILAG, KKW: all Swiss nuclearpower plants).

REFERENCE

[1] H.P. Zimmermann et al., this Annual Report.

164

APPLICATION OF THE PARALLEL-PLATE SQUEEZING-FLOW TECHNIQUE FORQUALITY CONTROL OF BITUMINOUS SAMPLES

H.P. Zimmermann, M. Egloff (PSI), R. Gubler (EMPA), R. Dressier, M. Patorski, L Dohring,R. Ivankovic (PSI)

A 'Texture Analyser" has been evaluated as a "parallel-plate squeeze-flow viscometer" as replacement forneedle penetration and ring-ball-softening point measurements.

Measurement of needle penetration and the ring-ball-softening point are widespread for quality control in thebitumen industry and road construction business. Theyhave been and are still applied for quality control ofradioactive waste embedded in bitumen but theycause some radiation protection problems due tocontamination, exposure times, and secondary (liquidorganic) waste.

Looking for a simpler procedure to measure somerheological properties of bituminous samples forquality control, we have chosen the parallel-platesqueezing-flow technique, as reported in Huschek etal.[1].

By squeezing a small cylindrical bituminous specimen(h0 = 4 mm, 0 20 mm) between two steel discs (0 20mm) with a constant force and a very low shear rate ofthe order of 102 s'1 or less we measure a "dynamicviscosity", in the range 103 to 108 Pa-s.

For that purpose we use a so-called "TextureAnalyser" from Stable Micro Systems, GB. This is acomputer controlled device which is able to control andregister the force applied to the sample, the distancetravelled by an inter-changeable probe, the time, andthe temperature, by cooling with liquid/gaseousnitrogen or electrical heating of the surrounding air in a"closed" temperature cabinet.

A specimen, consisting of a small disc approx. 4 mm inheight, is cut from a cylinder of bitumen with a motordriven knife, which makes a slow but steady cutpossible.

The specimen is placed between the steel discs,thermostatted over 12 h, and pre-squeezed to astarting height h0 of 4 mm for some minutes. Then thesample is usually squeezed for around 2 h.

Distance, time and applied force data are recorded ata rate of at least 1 Hz, the temperature at 10 Hz.

From these data the dynamic viscosity ri may becalculated under assumptions of a Newtonian liquid,lubrication approximation, no inertial effects, with thefollowing equation (see also e.g. [2]).

w(t) = h0 - 14(7

3 r 2

w(t) = h0 - h(t)a = pressure; load per disc area rcr2

r| = viscosityr = disc radius

We tested reproducibility, variation of applied force,variation of pre-squeezing time and the time necessaryfor reaching thermal equilibrium. The results weresatisfying according to our requirements.

It should be kept in mind that bitumen, especially thosetypes with high viscosities, are not pure Newtonianliquids but at least visco-elastic fluids. However, thedescription of the system with these Newtonianviscosities, as calculated above, meets our require-ments in most cases. The full data sets may be storedand the viscosity recalculated with other models ifnecessary.

Measurements by EMPA with a DRS (dynamic shearrheometer) and a supply of several bituminoussamples with known properties (B20/30, B40/50,B80/100 [B15 and B15/KKG from own stock]) enabledus to calibrate our system absolutely and to establish aviscosity - softening point - relationship as a link toformer quality control parameters (see Fig.1).

10'50 60 70 80

Ring and ball softening point [°C]90

h0, h(t) = height at time t0 = 0 resp. at time t

Fig. 1: Dynamic viscosity at 40 °C vs. ring-ball-softening point. From left to right: B80/100,B40/50, B20/30, B15 and B15/KKG (the lattertwo from own stock).

REFERENCES

[1] S. Huschek et al., ISETH, Report 44, 1980.

[2] Bird et al., Dynamics of Polymeric Liquids, Vol. 1.,2nd ed. Wiley & Sons, 1987, p.20.

165

Ion Beam P h y s i c s

166

THE PSI/ETH TANDEM ACCELERATOR FACILITY

H.-A. Synal, M. Döbeli, H. Fuhrmann, P.W. Kubik(PSI), G. Bonani, A. Duhr, P. Gartenmann, I. Hajdas,R. Huber, S. Ivy-Ochs, S. Jacob, C. Maden, R. Mühle, M. Suterand S. Tschudi (ETHZ)

The 1998 operation of the PSI/ETH tandem accelerator at ETH Hönggerberg is summarised in a detailedcompilation of beam time allocation and a statistics of AMS samples for the radioisotopes measured, andfor the major fields of research.

AMS Be-10

C-14AI-26CI-36

1-129Heavy Elements

SubtotalMaterials SciencesAccelerator SIMS

Tests

ConditioninqTotal

Hours199655172164

219184168

1907412178

201

912789

1997653

71871

227121106

189629510782

1892569

1998475

881103161

11181

1812

416171119

2572775

%199619.8

25.92.37.9

6.66.068.4

14.86.4

7.2

3.3100

199725.4

27.92.88.84.74.173.811.54.23.2

7.4100

199817.1

31.73.7

5.8

4.02.965.315.06.24.3

9.3100

199610841387

125458

20310

3257

Samples1997

1193

1429129534

3105

3600

199810981755

231401291

2

3778

Beam time statistics 1996-1998

Research FieldOceanographyPolar iceLimnologyAtmosphereEnvironmental MonitoringExposure Ape Datinq, OthersEarth Sciences (Total)Meteorite/Cross SectionsArchaeologyOthers, TestsSubtotalStandardsBlanksTotal

Be-10874654579

13681215

2338142106

1098

C-14217

4631

111405

673376

104922477

1755

AI-26

787826

861122813

231

CI-36

258

43

17318

14144524401

1-129

2381112

21610

9193719

291

Total3047231142341123421829516735081232476239

3776

8%19%3%6%3%9%48%1%18%13%33%13%6%

100%

Compilation of measured AMS samples at the PSi/ETH AMS facility in 1998.

In 1998, the PSI/ETH tandem accelerator ran for 2775hours, approximately 8% more than in 1997. In addi-tion, the total number of analysed samples could beincreased by 5%. Of the total beam time 65% was de-voted to accelerator mass spectrometry, 15% to mate-rial sciences, and 6% was used to measure stabletrace elements in connection with the focused ionbeam sputter source. Less than 10% of the total beamtime was needed for conditioning the accelerator.The number of measured radiocarbon samples hassignificantly increased. The increase in oceanographyand archaeology samples has required a highermeasurement effort because of the high precisionneeded in these research fields. The measurements ofthe 10Be and 35CI concentrations in polar ice cores have

been continued and the analysis of 129I samples hasbeen established as a tool for environmental monito-ing. Surface exposure dating using 10Be and 26AIicreased again its share in the Earth Sciences applic-tions. The high current ion source was routinely usedfor 10Be and 26AI measurements. Due to the higherbeam currents the measurement time per sampleanalysed could be reduced. The installation of thesmall AMS system for radiocarbon dating has beencompleted. First tests were successfully conductedwhich clearly demonstrate that the molecular comp-nent can be sufficiently destroyed. Blank samplesgave background values similar to those observed onthe large AMS system. Repeated measurements ofstandard samples gave highly reproducibly results.

167

A NEW SMALL ACCELERATOR FOR RADIOCARBON DATING

M. Suter, ft. Huber, S.A.W. Jacob (ETHZ), H.-A. Synal (PSI), J.B. Schroeder(NEC, Middleton Wl, USA)

A new small and compact radiocarbon dating facility based on a 500 kV Pelletron accelerator has beenbuilt. The novel feature is that it operates with 1+ ions. The interfering molecules are destroyed by collisionsin the gas stripper. The results of first test measurements demonstrate that stability, background andtransmission are almost equal to those of conventional AMS systems based on larger accelerators.

Experiments recently performed at our existing AMSfacility [1] and studies at other facilities [2, 3] havedemonstrated the feasibility of 14C analysis at energiesbelow 1 MeV. To remove interfering molecules, a spe-cial method can be used to destroy them in multiplecollisions inside the stripper of the accelerator [2]. Thetechnique opens the possibility of developing a newgeneration of AMS facilities which are much smallerthan the existing ones. The main concern when oper-ating an AMS facility at these low energies is thebackground [4]. First of all, interfering molecules ofmass 14 have to be destroyed and molecular frag-ments have to be removed efficiently from the 14Cbeam. It is known that the molecular components atmass 14 have intensities in the range of 10'4 relative tothe extracted 12C current.

A new instrument has been built to demonstrate thepotential of this principle and to investigate the rele-vant processes of molecular destruction and chargestate changing. The system is operational since July1998 and first test measurements were performed tostudy the molecular suppression capability at 970 keVbeam energy. They clearly demonstrate that the mo-lecular component can be destroyed. A gas density ofmore than 2 ug/cm2 of argon is necessary to achievethe required 10 to 11 orders of magnitude in suppres-sion. Processed blanks having molecular 13CH beamsof less than 10"4, gave background values of 2-8 times10"15. Fig. 1 shows the spectra observed from a proc-essed blank sample and an Oxalic I standard. Themeasured background values agreed with the valuesobtained for the same material at our large facility. It isassumed that these events originate mainly from real14C introduced during sample preparation. To investi-gate optimum operation conditions, charge state frac-tions (1+, 2* and 3+) have been measured for 12C in theenergy range from 400 - 500 keV for various stripperdensities (0.2 -3 ug/cm2 argon). This study clearlyshowed that charge state 1+ is the most abundant inthe operating energy interval (more than 50%). At thepressure needed to destroy all the molecules thetransmission is presently about 42%.

We found very good stability in the measured 13C/12Cratios. Within measurements of one sample, variationsin observed 13C/12C ratios are typically less than 0.05%.Also repeated measurements on the same samplegive highly consistent results. For different targets,

which deliver different currents, a small current de-pendent mass fractionation was observed. The pre-liminary tests have demonstrated the potential of thenew system. More measurements, and in particular,experience in routine operation, are needed to char-acterize its ultimate performance. But we clearly dem-onstrated that interfering molecules can be eliminatedat sub MeV energies. This opens new possibilities forradiocarbon dating and for biomedical applications. Itseems possible to reach performance equivalent tolarger facilities, but at significantly lower cost.

10000

1000

100

ICxdiciBBL

0 10 20 30 40 50 60 70 80 90 100 110 120

Chcnnels

Fig. 1: Two spectra measured with the 14C silicon de-tector during test measurements. Data acqui-sition time for the Oxalic I standard (grey) andfor the processed blank sample BL124 (black)were 200 seconds each. By integrating thecounts in the 14C gate, a 14C/12C ratio of 2.4 10"15 was obtained for the blank material. Thiscorresponds to a radiocarbon age of 48 ky.

REFERENCES

[1] M. Suter, S. Jacob, and H.-A. Synal, Nucl. Instr.and Meth. B123, 148(1997).

[2] A. E. Litherland, Nucl. Instr. and Meth. B5, 100(1984).

[3] B. J. Hughey, R. E. Shefer, R. E. Klinkowstein, X.L. Zhao, W. E. Kieser, and A. E. Litherland, Nucl.Instr. and Meth. B123, 186 (1997).

[4] M. Suter, Nucl. Instr. and Meth. B139, 150(1998).

168

EXPOSURE AGE DATING OF GLACIAL DEPOSITS IN THE LITANG AREA,EASTERN TIBET

S. Tschudi), C. Schliichter (Univ. Berne), S. Ivy-Ochs (Univ. Berne & ETHZ); P.W. Kubik (PSI); J. Schafer(ETHZ); Z. Zhao (Inst of Geomechanics, CAGS, Beijing), X. Wu (Xi'an Laboratory of Loess and Quater-

nary Geology, CAS, Xi'an)

In 1997, various erratic boulder surfaces of well defined glacial deposits in the Litang area, Eastern Tibet,were sampled. The 10Be and26AI data show a clustering within an exposure age interval between 12ka and18ka BP (minimal ages). All samples are older than Younger Dryas except one, which has a very youngexposure age, probably due to a collapsing event of the erratic.

The glaciation of the Tibetan Plateau during the Qua-ternary is a subject of controversial discussion [1,2].Almost no information on the timing of glacial events isavailable. Exposure age dating using stable and radio-active cosmogenic isotopes (21Ne, 10Be and 26AI) canprovide useful information on the ages of glacial de-posits. The Litang sampling area was chosen in col-laboration with Chinese geologists of the Institute ofGeomechanics, Beijing and the Xi'an Laboratory ofLoess and Quaternary Geology, Xi'an.

SAMPLE DESCRIPTION

Fig. 1 : Typical erratic boulder on a lateral moraine ofunknown age in the Litang area.

In the Litang area (N 30°17', E 99°30'), 7 similar-shaped erratic boulders with granitic lithology (Fig. 1)from different geological units [3] were sampled. Thealtitude above sea level ranges between 4'500 m and4'900 m. Thus the whole area lies above the tree lineand vegetation is limited to bushes. All boulders showerosion phenomena (e.g. spalling).

METHOD

The samples were prepared using the method de-scribed in [4]. 10Be and 26AI were measured at the AMSfacility of the PSI and ETH Zurich, neon meas-urements were performed in the noble gas laboratoryat the Institute of Isotope Geology, ETH Zurich.

RESULTS

Minimum exposure ages of 3 erratics (Litang 4, 5and 6) cluster around 15 ka, whereas one sample(Litang 1) shows a very young age. This young age isprobably due to a collapsing event of the erratic.Litang 7 is younger than the main age cluster, whichagrees with the geological observation. All samplesare older than Younger Dryas.

Table 1 : Results of the

Sample

Litang 1al0Be

*AI

Litang 1b10Be

*AI

Litang 4a 10Be

aAI

Litang 4bc'°Be

*AI

Litang 5a1°Be

*AI

Litang 5b10Be

MAI

Litang 6 "Be

^Al

Litang 7 "Be

^Al

Altitude/m

4'900

4'900

4'575

4'575

4'605

4'605

4'585

4'515

radioisotope analysis.

Atoms/gr

82.1 x10"

965 x10"

115.8x10"

550.3x10"

143.3x10"

809.9x10"

119.8x10"

781.8x10"

125.3x10"

739.1 x10"

124.7x10"

737.9x10"

124.8x10"

783.8x10"

102.2x10"

712.5x10"

Minimum Age/y

8'500 ± 600

16700 ±1'100

12'400±900

9'800 ± 600

17'500±1'200

16'500±1'200

14'500±1'000

15'800±1'100

15'100±1'000

14'800 ±1'000

16'100±1'000

15 900 ±1 '800

15'200±r000

15'900±1'300

12700 ±900

14700 ±1700

REFERENCES

[1] M. Kuhle, Quaternary International 45/46, 71-108(1998).

[2] Z. Benxing and N. Rutter, Quaternary Interna-tional, 45/46, 109-122 (1998).

[3] X. Wu and Y. Li, Quaternary Sciences 2, 157-158(1990).

[4] C. P. Kohl and K. Nishiizumi, Geochimica etCosmochimica Acta 56, 3583-3587 (1992).

169

PRELIMINARY RESULTS ON THE USE OF PYROXENE FOR 10Be SURFACEEXPOSURE DATING

S. Ivy-Ochs (ETHZ& Univ. Bern), P.W. Kubik (PSI), J. Masarik (Komensky University, Slovakia),R. Wieler, L Bruno (ETHZ), C. Schluchter (Univ. Bern)

Test were performed to see whether pyroxene is suitable for exposure dating with 10Be or whether meteoricwBe cannot be quantitatively removed from this mineral.

We performed two sequential dissolution series of py-roxenes separated from Antarctic dolerites with expo-sure ages around 2 Ma [1]. The goal of these experi-ments was to test whether pyroxene is suitable for ex-posure dating with 10Be or whether meteoric 10Be can-not be quantitatively removed from this mineral. Forcomparison, we show leaching results from threequartz samples, two from sandstones and one fromgranite. Because pyroxene retains 3He and 21Ne quan-titatively, it is an attractive mineral for exposure dating.The measurement of 10Be in the same mineral wouldcomplement the noble gas results by revealing possi-ble episodes of prior exposure (when there has beenan intervening period of burial).

The samples investigated are part of a project to de-termine the age of the Sirius Group sediments andrelated landscape elements in the Dry Valleys regionof Antarctica, using in situ produced cosmogenic 3He,10Be, 21Ne and 26AI [2, 3, 4]. Quartz-bearing clasts un-ambiguously associated with the Sirius Group are rareso that the possibility to analyze pyroxene for 10Be, inaddition to 3He and 21Ne is crucial.

In the first of two dissolution series (sample 4 in figure1), the final 10Be concentration was 26% greater thanthe concentration expected based on the 3He and 21Neexposure age of 2.4 Ma. Given production rate uncer-tainties of perhaps up to 20%, it is possible that mete-oric 10Be was completely removed from this sampleafter two dissolution steps. However, in the second se-ries (Sample 5a) the final 10Be concentration was al-most four times greater than that expected from thenoble gas exposure age of 1.8 Ma. This indicates that,using the standard hydrofluoric leaching technique,meteoric 10Be is in some cases not removed com-pletely from pyroxene grains. This is in contrast toquartz, where after only one hydrofluoric leaching stepmeteoric 10Be is completely removed. One possibleexplanation is that crystallographically controlled dis-solution during weathering of the pyroxene accompa-nied by clay formation allows meteoric 10Be to pene-trate and to be locked in the interior of the grains. Thisis in contrast to quartz, where all of the meteoric 10Be isonly adsorbed on the surface. It may be that, similar toolivine, 10Be exposure ages can be measured in py-roxenes from young rocks where individual grains areunaltered. The multi-nuclide approach is important forchecking for gain or loss of one of the isotopes. Fi-nally, we would like to point out that the open structure

of pyroxene could just as easily lead to prematureleaching of in situ produced 10Be during the chemicalcleaning which would also lead to spurious results.

3x107

2x10c

20 40 60

% Dissolved

80 100

Fig. 1: 10Be concentrations normalized to sea level.Filled symbols are for pyroxene, open symbolsare for quartz. The shaded band indicates thesea level secular equilibrium concentration of10Be for pyroxene.

REFERENCES

[1] S. Ivy-Ochs, P.W. Kubik, J. Masarik, R. Wieler,L.A. Bruno, C. Schluchter, Schweiz. Mineral. Pe-trogr. Mitt. 78, 375-382 (1998).

[2] S. Ivy-Ochs, C. Schluchter, P.W. Kubik, B. Dit-trich-Hannen, J. Beer, Geology 23, 1007-1010(1995).

[3] S. Ivy-Ochs, C. Schluchter, M. Prentice, P.W. Ku-bik, J. Beer, In: C.A. Ricci, (ed.): The AntarcticRegion: Geological Evolution and Processes,Terra Antarctica Publication, Siena, 1153-1158(1997).

[4] L.A. Bruno, H. Baur, T. Graf, C. Schluchter, P.Signer, R. Wieler, Earth Planet. Sci. Lett. 147, 37-54(1997).

170

CATCHMENT-WIDE DENUDATION RATES FROM COSMOGENIC NUCLIDES INRIVER SEDIMENT

F. von Bianckenburg, M. Schaller, J.D. Kramers (Univ. Bern), P.W. Kubik (PSI)

Erosion rates from large geographical areas can be measured using cosmogenic nuclides in river-bornequartz. The validity of the method is investigated in a field-based study on European rivers of which thecatchment denudation rates are known from conventional gauging data.

The concentration of in situ-produced cosmogenic nu-clides [1] in river sediment-borne quartz reflects spa-tially averaged rates of denudation in a drainage basin.This technique is applicable for estimation of time-integrated mean denudation rates in small catchments[2, 3]. In this study the method was applied to medium-sized catchments to evaluate the validation of themodel beyond the local scale.

Mechanical |12|4O| Chemicalmm/ka

(BfG, LfU Ba-WQ)

50 km

Fig. 1: Neckar denudation rates from river load data.

Denudation rates inferred from 10Be in quartz in riversediment from middle-European rivers were comparedwith conventional rates calculated from the present-day flux of daily monitored suspended and dissolvedload in these streams. Denudation rates calculatedfrom in stfu-produced cosmogenic nuclides [1] are 18-24 mm/ka for the river Regen (SE Germany), whichcompares to 8 mm/ka as estimated from conventionaldata. Cosmogenic rates are 50-90 mm/ka for the riverNeckar (SW Germany) (Fig. 2), whereas conventionalrates are 35-55 mm/ka (Fig.1). Denudation rates fromthe two methods agree remarkably well, given thatconventional data represent sampling intervals of afew years at most, whereas cosmogenic data integrateover several 1000 years. No regional patterns are visi-ble for samples that have been collected from

upstream down to the river mouth. The fact that cos-mogenic rates in both rivers are sometimes higher bya factor of two to three may be explained by variationsin erosion rates, the difficulties in accurately measuringsuspended loads in streams, much of which is as-sociated with low-frequency flood events, or by the un-certainty associated with the amount of pollution in thesolute load.

However, the concentration of in-situ cosmogenic nu-clides in river sediments is a valuable and accuratetechnique to estimate time-integrated denudationrates.The effects of climate, tectonics, lithology andtopography on denudation rates may be studied byapplying this method to different river basins.

Heidelberg

Total CosmogenicDenudation Rate

[mm/ka]

50 km

Fig. 2: Neckar total denudation rates from cos-mogenic nuclides.

REFERENCES[1] D. Lai, Earth Planet. Sci. Lett. 104, 424-439

(1991).

[2] E.T. Brown, R.F. Stallard, M.C. Larsen, G.M.Raisbeck, F. Yiou, Earth Planet. Sci. Lett. 129,193-202(1995).

[3] D.E. Granger, J.W. Kirchner, R. Finkel, J. Geo-logy 104, 249-257 (1996).

171

THE DETERMINATION OF SURFACE EXPOSURE AGES IN QUEEN-MAUD-LAND(ANTARCTICA) VIA "IN-SITU" PRODUCED COSMOGENIC RADIONUCLIDES

M. Altmaier, U. Herpers (Uniyersitat zu Koln, Germany), G. Delisle (Bundesanstalt fur Geowissenschaftenund Rohstoffe, Hannover, Germany)

We have determined surface exposure ages in Queen-Maud-Land (Antarctica) by measuring "in-situ" pro-duced 10Be and 26AI in purified quartz minerals. The results call for a rather complex giaciation and exposurescenario for the investigated region of Queen-Maud-Land.

The interaction of cosmic radiation with terrestrialmatter leads to the production of the long-lived cos-mogenic radionuclides 10Be and 26AI. Accelerator massspectrometry (AMS) enables us to quantitativelymeasure these "in-situ" produced nuclides in quartz-rich minerals. This ultimately allows the determinationof surface exposure ages, erosion rates and otherprocesses of landscape evolution in a glacier domi-nated scenario.Our work concentrates on Queen-Maud-Land (Antarc-tica) - a region for which scientific data are still veryscarce.

EXPERIMENTAL PROCEDUREAs the samples taken in the field usually do not consistof pure quartz, a quartz separation and purificationprocedure has to be employed. Moreover, the AMS-technique itself also necessitates careful samplepreparation.First, we use a chemical mineral separation procedure(repeated digestion of the sample with HF) to isolatethe quartz and decontaminate the material from at-mospheric 10Be. The dissolved quartz is then treatedby radiochemical separation techniques (anion, cationexchange chromatography and precipitation steps) toyield BeO and AI2O3 as required by AMS.

RESULTS• Some samples from the "Petermann-Kette" (from

locations of up to 3000m above sea level) show nu-clide concentrations in saturation corresponding tovery high exposure ages of at least 6 Ma. Othersamples, with lower nuclide concentrations, haveexposure ages of 0.3-2 Ma. For the "Petermann-Kette" we find no simple correlation between expo-sure ages and the altitude of the analysed sample.One explanation could be that crystalline surfacematerial had been exfoliated under the influence offrost shattering leading to "too short" exposureages.

• The samples from the "Schussel" region ("Eck-hb'rner" and "Seitental 1, Seitental 2") were col-lected at 1640-2110 m, with most of them locatedlower then the "Petermann" samples. The exposureages also tend to be lower ranging from 0.04 to 2.5Ma. Exposure age and altitude appear to be corre-lated.

• The "Morane" in the "Schussel" region is a verylarge glacier field located at 1650 m above sealevel. Exposure ages of the large boulders in the

moraine are surprisingly low - mostly around0.01-0.1 Ma. One rock with an exposure age of 1.0Ma, however, suggests a long residence time of thiscomplex feature.

• For the locations "Untersee-Ostgrad" and "Unter-see-Westgrad" (1040-1430 m) we have determinedexposure ages ranging from 0.3-0.9 Ma.

• Most samples from the "Dallmann-Berge"(1720-1940 m) also seem to have been exposedrather recently, about 0.1-0.3 Ma ago. Erratic rocksamples have larger exposure ages of 0.4-0.5 Ma.

• Samples from the location "Schirmacher-Oase"(100-120 m) showed extremely low radionuclideconcentrations leading to very young exposure agesof 0.02-0.04 Ma.

• The nuclide concentrations measured correspond tosmall erosion rates as is expected for samples fromAntarctica. Erosion is predominantly controlled byfrost shattering and tafoni formation.

Table 1 lists the locations analysed in this work to-gether with the numbers of 10Be and 28AI measure-ments and the resulting exposure ages (erosionrate=10'6cm/a).

Table 1 : Survey of the analysed locations / results.

Location

SchirmacherOase

Dallmann-Berge

"Schussel"Eckhorner

"Schussel"Morane

"Schiissel"Seitenlaler

"Untersee"Westgrad

"Untersee"Ostgrad

Petermann-Kette

Number of samplesanalysed (Be / Al)

3 / 1

9 / 6

4 / 4

13/3

8 / 7

3 / 3

3 / 3

9 / 7

Exposure ages(Ma) / mean value

0,02-0,04 / 0,03

0,1-0,6/0,25

0,04-2,4 / 0,9

<0,005-1,0/0,15

0,2-1,9/0,7

0,2-0,9 / 0,6

0,3-0,8/0,5

0,3->6 / 2,4

ACKNOWLEDGEMENT

This work was supported by the German ResearchFoundation (DFG) and the Paul Scherrer Institut.

172

MAGNETIC, SEDIMENTOLOGICAL AND GEOCHEMICAL PROPERTIES OF2 LOESS PROFILES FROM THE LOESS PLATEAU OF CHINA

M. Sartori, F.Heller, P. Sroubek (ETHZ), J. Beer (EAWAG), P.W. Kubik (PSI), A. Tsatskin (University ofHaifa, Israel), T. Evans (University of Alberta, Canada); E. Vincent (C.E.A., France)

Susceptibility (%), grain size and wBe concentration measurements performed at two sections of the Chi-nese Loess Plateau (Xiagaoyuan and Houzhuang) show that wBe concentration is inversely proportional tothe grain size rather than proportional to c.

Beer et al. [1] tried to determine the paleoprecipitationduring the last interglacial. 10Be measurements per-formed at the loess section at Louchuang (CentralLoess Plateau, as the section of Houzhuang), wereused to discriminate detrital to pedogenic susceptibil-ity. A good correlation between magnetic susceptibil-ity and 10Be was found. To verify this result, 10Be con-centration was measured at two new section in Houz-huang and Xiagaoyuan. A good correlation was foundfor the soil S, in Houzhuang, whereas in Xiagaoyuana bad correlation between 10Be and the susceptibilityin the pedocomplex S, was observed. On the otherhand, a good relationship between 10Be and mediangrain size (Md) was observed. In the pedocomplex S,of Xiagaoyuan, a general smaller Md

HOUZHUANGMedian Grain Size (urn)

XIAGAOYUANMedian Grain Size (run)

35 30 35 20 IS 10 S 35 30 2S 20 IS 10 5

" B e 10+8 [a/g] 10+e [a/g]

Fig. 1: 10Be concentration, grain size and magneticsusceptibility % in Xiagaoyuan and Houz-huang as a function of depth. The grain sizewas measured from 7 to 9.80 m depth inHouzhuang and from 6 to 23 m inXiagaoyuan. 10Be concentration was meas-ured on bulk samples from 7.50 to 10 m inHouzhuang (every 10 cm) and from 14 to 23m in Xiagaoyuan (every 10 to 20 cm). "S" in-dicated paleosols (inter-glacial), "L" indicatesloess (glacial). S, correspond to theRiss/Wurm interglacial (from ca. 70-120 ka).

and a higher 10Be concentration was observed,whereas in the loess layer L, Md increase abruptlyand 10Be concentration diminishes (Fig.1).To verify the relationship between 10Be concentrationand grain size distribution, nine fractions for four dif-ferent samples (36 samples) were measured (a loessand a paleosol sample for both section of Xiagaoyuanand Houzhuang (Fig. 2). The results of the fractions0.39-0.228, 0.228-0.091 and 0.091-0.030 were sum-marized in a unique fraction labeled "<0.39"). An in-crease of 10Be concentration is observed toward finergrain size fraction. In the fraction < 0.39 mm a de-crease of 10Be concentration was observed in theloess and paleosol sample of Houzhuang. This couldbe due to 10Be loss from the finest fraction during thecentrifugation procedure used to obtain the smallestgrain size fraction.

10HOUZHUANG XIAGAOYUAN

• PALEOSOL Sin LOESS L2

60

i 40

HOUZHUANG

grain size ftlm)

XIAGAOYUAN• PALEOSOL Sio LOESS L2

ill Hi ill A

grain size (|xm)

Fig. 2: 10Be concentration variation as a function ofgrain size for 4 different samples: Houzhuang(1 from paleosol S1 and 1 from loess L1) andXioagaoyuan (1 from paleosol S1S3 and 1from Loess L2). A general concentration in-crease is observed towards smaller grainsizes. The decrease observed in the fraction<0.39u.m is thought to be caused by labora-tory effect during grain size fractionating.

REFERENCE[1] J. Beer, CD. Shen, F. Heller, T.S. Liu, G. Bo-

nani, B. Dittrich, M. Suter, P.W. Kubik, Geophys.Res. Lett. 20, 57-60 (1993).

173

RADIOCARBON TIME SCALE OF THE MEDITERRANEAN CORE CT85-5: THELAST 40'000 YEARS

/. Hajdas, G. Bonani (ETHZ), J. Beer(EAWAG), G. Bonino, G. Cini Castagnoli, C. Taricco (Istituto diCosmogeofisica del CNR, Turin, Italy).

Although there is an increasing interest in absolute dating of ciimatic changes using absolute chronologiessuch as tree rings, 14C provides reasonable time scales and more importantly allows correlation betweenrecords around the world. One of the important aspects is correlation between terrestrial and deep-searecords. The goal of this study was a detailed radiocarbon dating of the Mediterranean deep-sea coreCT85-5.

The sediment core, which was recovered from theTyrrhenian Sea (40°19'02"N, 11°15'42"E) at 2833 mwater depth, has a length of 616 cm. A high sedimen-tation rate of ca. 10 cm/kyr has been estimated usingthe presence of historically dated tephra layers of thePollena eruption (472 AD) (17 cm) and Pompei (79AD) at 20 cm. Older tephra layers at the depth of 290-334 cm and 400-426 cm have been correlated with theterrestrial deposits of the Campanian Ignimbrite andCitara, respectively. A relatively constant sedimenta-tion rate was suggested by the fact that the spectralanalysis of the carbonate content for the upper (lateHolocene) and lower part (Late Glacial) of the coreshow the same periodic components.

Because only 10 to 20 mg of carbonate is required forAMS 14C dating, hand picked shells of planktonic fo-raminifera can be used. Since planktonic foraminiferalives in the mixed layer of the ocean, a correction forreservoir age of this layer (ca. 400 years) must beapplied to correct for a different 14C content of the at-mosphere and the mixed layer of the ocean. Anotherdifficulty in establishing a radiocarbon chronology ofdeep sea sediments is a possible effect of bioturba-tion. However for the cores with higher sedimentationrate (8 cm/kyr) , such us the core CT85-5, the effect ofbioturbation on the radiocarbon chronology should besmall [1].

Samples, which contained around 2 g of sediment,were soaked in 10% H2O2 and then placed in an ultra-sonic bath for a few seconds before washing themthrough the sieve (180 jj.m). Shells of Globigerina Bul-loides or mixed planktonic were picked out for radio-carbon dating. Most of the samples contained morethan 2 mg C (more than 20 mg of CaCO3). Carbonateof the sample was acidified using phosphoric acid(80%) and the released CO2 was then reduced to

graphite in a reaction with hydrogen using cobalt pow-der as a catalyst. The graphite-cobalt mixture fromeach sample was pressed onto copper discs.

The graphite targets were measured together withstandards and with blank-samples made from opticalcalcite. 14C/12C and 13C/12C isotopic ratios for each tar-get were measured quasi-simultaneously at theETH/PSI accelerator facility.Samples selected from the sediment between 95 and395 cm were dated. The resulting radiocarbon chro-nology (Fig. 1) provides time scale for sediments de-posited between 40 and 10 kyr BP.

S^ 50

oo

6000 12000 18000 24000 3000014Cage BP

Fig. 1: The carbonate content of sediment betweenthe top and 280 cm is plotted versus radiocar-bon time scale.

REFERENCE

[1] T.H. Peng, W.S. Broecker, Quaternary Research11, 141 (1979).

17411C AGES OF TERRESTRIAL MACROFOSSILS FROM LAGO GRANDE Dl

MONTICCHIO (ITALY)

/. Hajdas, G. Bonani (ETHZ), B. Zolitschka (GFZ, Potsdam, Germany)

Lago Grande di Monticchio (Italy) contains long sedimentary records of more than 75,000 years. In a joinedEuropean project, vegetation history and climatic changes for this part of Mediterranean has been recon-structed and Radiocarbon dating was applied to construct the chronology of the last 40,000 years. Previousstudies have shown that the age of bulk sediment from Lago Grande di Monticchio, which is a maar lake,had been influenced by 'dead carbon' of volcanic origin, which is dissolved in lake water. Thus radiocarbondating of the record is problematic and was limited to dating macrofossils of terrestrial origin.

Lago Grande di Monticchio is located in southern Italy(45N, 15E). Two of four piston sediment cores LGM-C(40 m) and LGM-D (52 m) were combined to one rec-ord with a composite sediment depth of 52 m. Thisprofile has been studied for pollen, geochemistry,sedimentology, physical sediment properties, paieo-magnetic, tephras and varve chronology. A set of 19sediment samples was submitted to the 14C lab. Inorder to gain on material, subsamples from three cor-related cores were combined.

Sediment samples were treated with 10% HCI (6 hrs)and 10% KOH (24 hrs) in order to disperse sedimentparticles. Then sediment was sieved through a 500micron sieve. Organic matter which remained on thesieve was examined using a microscope. A mixture ofbark, leaf fragments, seeds of trees was picked outwhen only identification was possible. In one case asample made up of a mixture of unidentified fragmentsof plants was dated after terrestrial macrofossils werepicked out (LGM5.2) (Fig 1). From the total of 19 sam-ples which were washed only seven contained suffi-cient amount of organic matter/terrestrial macrofossilsfor AMS dating.

Standard acid-base-acid treatment was applied inorder to remove carbonates and humus contaminants.The solutions of 0.5 M HCI and 0.1 M NaOH were usedat 60°C. Each step lasted ca. 1 hour and was followedby rinsing with destilled water to the normal pH. Sam-ples were combusted in Vycor tubes and CO2 wasconverted to graphite in a reaction with H2 over cobaltas a catalyst.

Our chronology adds to the existing dating of LagoGrande di Monticchio and improves its resolution. TheHolocene sediments are now dated by five samples.The sediments corresponding to the Younger Dryasperiod (10 to 11 kyr BP) are now dated by threepoints. They all fall into to the range of the YoungerDryas radiocarbon ages i.e. 10 to 11 kyr BP. Only onepoint dates the Lateglacial part of the pollen record.There were attempts to date the Lateglacial sedimentsbut all the ages have been rejected as they turned outto be too old. In fact, the age of 23,000±190 yr BPobtained on the rest material, presumably of aquatic

origin, of the sample LGM5.2 AQ is close to the agesreported previously [1].

The older sediments contain little macrofossils. Pollenfound in sediments down to 1650 cm depth suggestmainly herbaceous vegetation [2] but Juniperus andPinus were also present. Therefore, improvement ofthe chronology between 13 kyr and 23 kyr BP seemspossible.

Improving present chronology as well as an extensionof the radiocarbon chronology beyond the age of theoldest sample 24400±460 yr BP is essential for inter-pretation of the Lago Grande di Monticchio recordwhich contains important paleoclimatic information.

25000

Q. 20000

m

<0 15000

o

10000

5000

LGM5.2 AQ

LGM5.2

600 800 1000 1200

Depth cm1400 1600

Fig. 1: Radiocarbon ages plotted versus compositedepth in sediment core from Lago Grande diMonticchio. The arrows point to the ages ob-tained on mixed organic matter, possibly ofaquatic origin, and terrestrial macrofossils.This illustrates a discrepancy between agesobtained on bulk samples and terrestrialmacrofossils, which is typical for Monticchiorecord.

REFERENCES

[1] C. Robinson, Chemical Geology 118, 235 (1994)..

[2] W.A. Watts et al, Quaternary Science Reviews15, 133(1996).

175

GEOMAGNETIC MODULATION OF THE 36CI FLUX IN THE SUMMIT GRIP ICE CORE

G. Wagner, J. Beer, St. Baumgartner, J. Masarik, R. Muscheler (EAWAG), P.W. Kubik, H.-A. Synal (PSI),M. Suter (ETHZ)

A comparison of measured XCI GRIP data with production rate calculations based on a paleomagneticreconstruction for the last 100 ka gives a good agreement which holds true not only for the long term trend,but also on a shorter time scale. In particular the 36C/ peak at 38 ka BP previously found in the '"Be recordfrom Vostok can be explained by a period of low geomagnetic field intensity.

The 36CI concentration in deep ice cores from Green-land is predominately affected by three processes:cosmogenic production, atmospheric transport, anddeposition. By using accumulation information basedon the 818O data, a 36CI flux can be calculated. Thissignal should reflect changes of the production ratecaused by geomagnetic modulation. However, highfrequency components such as the changing activity ofthe and the short-term atmospheric transport as wellas deposition phenomena are still visible. Therefore,the 36CI flux data from the GRIP ice core (CentralGreenland) have been low-pass filtered (Fig. 1a) [1].

Traditionally, the paleomagnetic intensity is recon-structed from remanence measurements on cores ofdeep-sea sediments. We used such a record from theSomali Basin located east of Africa. This record waschosen from all available paleomagnetic data setscovering the last 100 ka because of its by far highesttemporal resolution. The data are plotted here in Fig.1d) on their original time scale. From this paleomag-netic intensity the expected mean global 36CI produc-tion rate has been calculated using a new productionmodel [2]. This is shown in Fig. 1c).

The comparison of the 36CI flux (Fig. 1a) with the cal-culated production rate (Fig. 1c) shows several strikingsimilarities. The most prominent feature, the peak (p,)at about 38 ka B.P. is present in both records, but alsosmaller structures exist, for example the peak (p2) atabout 60 ka B.P. and the lows at about 70 ka B.P. (v,)and 85 ka B.P. (v2). Fig. 1b) was made by slightlyshifting the time axis of Fig. 1c) within the uncertaintiesof the time scales to match the major wiggles in Fig.1a). Originally, the peak (p1) at 38 ka B.P. has beendiscovered by Raisbeck et al. [3] in the 10Be record ofthe Vostok core.

The good agreement between measured 36CI flux andcalculated 36CI production from a paleomagnetic deep-sea record confirms, that the geomagnetic field in-deed modulates the 36CI flux to Central Greenland, andthat during the last 100'000 years the field intensitychanged considerably. The peak at 38 ka B.P. can beexplained by a period of very low field intensity, usuallycalled the Laschamp event.

In conclusion, the cosmogenic radionuclide 3eCI pro-vides a new and independent tool to reconstruct thehistory of the geomagnetic dipole field. The additionalanalysis of 10Be data from the same ice core will

consolidate the information on the geomagnetic dipolefield intensity for the past 100 ka.

So 20 30 40 50 60 70 80 90 100age [ka B.P.]

Fig. 1: a) 36CI flux [104 atoms/cm2 year] as calculatedfrom the measured 36CI concentrations in theGRIP ice core from Summit, Greenland: rawdata (thin curve) and low-pass filtered data(cut-off frequency 1/2000 years; thick curve);b) is the same as Fig. 1c), but shifted in time tomatch the major wiggles in Fig. 1a); c)geomagnetically controlled 36CI production rateon its own time scale, calculated from the datashown in Fig. 1d); d) geomagnetic field inten-sity as reconstructed from three sedimentcores from the Somali basin.

REFERENCES

[1] S. Baumgartner et al., Science 279, 1330-1332(1998).

[2] J. Masarik and J. Beer, J. Geophys. Res., (1999)in press.

[3] G. M. Raisbeck et al., Nature 326, 273-277(1987).

176

THE ROLE OF GALLIUM AND INDIUM AT THE ALUMINUM-OXIDE INTERFACE

N. Margadant, M. Textor, N.D. Spencer (Surface Science, ETHZ), M. Dobeli (PSI)

The enrichment of gallium and indium in the oxide-metal interface in aluminum has been quantified bymeans of Rutherford Backscattering Spectrometry. Under anodization Ga and In are nearly completely ex-pulsed from the oxide bulk to the oxide front. Alkaline etching causes a continuos interface enrichment lim-ited by a saturation value.

Gallium and indium are present as impurities in baux-ite and end up in the aluminum metal in a typical con-centration range of 50 to 150 ppm. This work aims atelucidating the influence of Ga and In on the surfaceproperties of high purity Al (99.99) modified by chemi-cal and electrochemical surface treatments of indus-trial relevance [1]. Anodization, chemical and electro-chemical brightening, lithographic etching and alkalineetching have been examined. A continuous interfaceenrichment of Ga and In in a layer of a few nm thick-ness occurs only in the alkaline and in the neutralelectrolytes. Crucial for the product properties is thefact that an element specific change of the surfaceproperties (oxide growth, topography, reactivity, etc.)after the treatment is always correlated with anenrichment of Ga or In in the oxide-metal interface.This enrichment has been quantified by RutherfordBackscattering Spectroscopy (RBS) (Fig. 1).

0

100 150 200 250 300 350 400

Channels

Fig. 1: 2 MeV He RBS Spectrum of an anodized AI2O3

layer. The inset shows the Ga profile whichreveals the interface enrichment, and the con-centration in the oxide and the Al bulk.

Ga and In are enriched in the metal-oxide interfaceduring the barrier layer formation at anodizing voltagesup to 400V. The enriched amount is proportional to themolecular units of the formed aluminium oxide indi-cating a nearly complete expulsion of Ga and In fromthe oxide bulk to the oxide front (Fig. 2). The alkalineetching causes a continuous interface enrichment. TheRBS measurements reveal that for both elements asaturation value of the interface concentration isreached with time (Fig. 3) and a critical bulk concen-tration exists above which no further enrichment can

be achieved. The saturation concentration in thesegregation zone is close to the solid solubility of Ga inAl and the eutectic composition of Al-ln, respectively.

0.25

200 400

Oxide Thickness / nm

600

Fig. 2: Indium enrichment at the oxide-metal interfaceas a function of anodized oxide thickness. Thedashed line is a linear fit to the data points.The solid line is the enrichment expected fortotal repulsion of In from the oxide.

n(3

500 1000Etching Time/s

1500

Fig. 3: Ga interface enrichment after alkaline etchingas a function of etching time.

REFERENCE

[1] M. Textor, M. Amstutz, Analytica Chimica Acta297,15-26(1994).

177

IMAGING OF FLUORINE ON GEOLOGICAL SAMPLES

M. Dobeli (PSI), J. Schafer, R. Wieler, A. Halliday (ETHZ)

Raster scan images of the fluorine concentration on the surface of rock samples have been determined byProton Induced Gamma Emission (PIGE). The fluorine content in mineral grains is one of the criteria for thefeasibility of21Ne exposure dating of geological surfaces.

Exposure dating of a rock surface (and hence a land-scape element) can be obtained by measuring theconcentrations of in situ produced cosmogenic iso-topes. 21Ne is one of the isotopes of interest. Cos-mogenic neon is produced in quartz grains e.g. by aspallation reaction with 28Si. The 21Ne/22Ne concentra-tion in a surface exposed to cosmic rays increaseslinearly with time as long as erosion can be neglected.

4000

3000 -

2000 -

1000 -

Energy / MeV

Fig. 1: Total y-energy spectrum obtained from thesample shown in Fig. 2.

However, it has to be made sure that no other neonproducing reactions take place within the mineral. Apossible interfering reaction is 19F(a,n)22Na -> p' ->^Ne which can yield substantial amounts of Ne espe-cially in U/Th containing materials.

1 mm

Fig. 2: Left: photograph of a polished surface of agranite sample. Right: fluorine concentrationmap (measured by PIGE) of the same sample.White corresponds to 30 ppm, black to a con-centration of approx. 1 % (weight).

Proton Induced Gamma Emission (PIGE) is acovenient method to determine fluorine concentrationswith a detection limit of approx. 10 ppm in most mate-rials. It has been implemented at the Laboratory forIon Beam Physics to determine fluorine profiles inmeteorites and ancient bones [1,2].

Sections of stone samples have been raster scannedby a 2.7 MeV focused proton beam. The produced y-rays have been measured with a Nal detector. Eachevent falling into the y-energy range of the 19F(p,ocy)160reaction (see Fig. 1) is filled into a 2-dimensional map.

An apatite sample with known F concentration is usedfor calibration. Fig. 2 shows the fluorine map of a pieceof granite from the region of Litang, Eastern Tibet,which exhibited a neon dating anomaly. This map ledto the hypothesis that the fluorine is mainly concen-trated in the biotite grains of the stone. Measurementson extracted pieces of biotite confirmed this assump-tion (Fig. 3).

" Ik- ' ' '

Fig. 3: F concentration map of extracted biotitegrains.

In order to explain the excess 22Ne concentration inthis kind of sample, two questions still remain to beanswered. On the one hand, it has to be proven thatthere is enough U/Th in this material to account for theF to Ne transmutation. On the other hand, it has to beexplained how the neon moved from the biotite to thequartz grains which are analysed for dating.

REFERENCES

[1] U. Krahenbuhl et al., Meteoritics and PlanetaryScience 33, 665 (1998).

[2] K. Noll, M. Dobeli, PSI Annual Report, Annex III,1997.

178

MEV ION TRACKS IN PMMA: FORMATION OF NANOWIRES

C. Musil, M. Dobeli (PSI), A. Duhr (ETHZ)

PMMA has been irradiated by a variety of MeV single and cluster ions, including Cm. The relationshipbetween the deposited energy and the developed track volume points towards a diffusive mechanism ofenergy dissipation. Nanowires have been formed by the galvanoforming of the cylindrical tracks.

Ionized molecules and clusters at MeV energies candeposit extremely high energy densities along theirtracks in solids. Since the total energy loss per tracklength is approximately equal to the sum of the energylosses of the cluster constituents, stopping powers canbe obtained which are much higher than that of anymonoatomic particle [1]. We investigated the effect ofisolated impacts of a wide variety of MeV ions inPMMA (PolyMethylMethAcrylate), which is widely usedas resist in X-ray and particle beam lithography.

*-

PMMA surface. The topography of the track openingswas analysed by Scanning Force Microscopy (SFM).Two examples are shown in Fig. 1. Nonlinear effects,e.g. due to saturation of the density of broken bondsand a dependence of the track diameter on the rangeof secondary particles (electrons, recoil atoms), havebeen expected. However, a very simple behavior wasfound. The area of the developed tracks increaseslinearly with the energy per track length deposited intothe electronic system of the material (Fig. 2). Thismeans that the deposited energy per developed trackvolume is a constant. A linear fit to the data yields avalue of 7.3 eV/nm3 which is very close to the 6.7eV/nm3 found in focused ion beam lithography withions in the 100keV range. This points towards a'diffusive' (not ballistic) energy transport mechanism.The deposited energy fills the track volume like anincompressible liquid. So far, a detailed understandingof this phenomenon has not been attained.

Fig. 1: SFM image of developed isolated MeV iontracks in PMMA.

PMMA layers of 85 nm thickness were spun onto thegold covered surface of oxidized silicon wafers. Afterirradiation, the PMMA was developed for 60 sec in 1:3MiBK:IPA and heated for 10 minutes at 100 °C in orderto remove solvents and absorbed water from the

4-25 nm

3000

1000

1000

n

-

6.7 eV/nm3 (from keV ions)

'If

o ^"Ge3

c« I-sp^T,,

sT, . . i . . . . i . . . .

_,.

1 . . . . 1 . -i

- 60

- 50

- 40 g-

-; 30 I

20

10

0 5 10 15 20

Energy Loss into lonization in keV/nm

Fig. 2: Measured area (and diameter) of developedholes in PMMA as a function of electronicenergy loss per track length (calculated byTRIM). The solid straight line is a fit to thedata, the dashed line is an extrapolation fromresults with keV focused ion beams.

Fig. 3: Scanning electron micrograph of a nanowireobtained by galvanoforming of an iodine iontrack.

In order to make a larger part of the ion track visibleand to explore possible applications of single ionlithography, the developed holes have been galvani-cally filled with nickel. After dissolving the PMMA layer,free-standing nanowires were obtained. Fig. 3 showsan example of a wire produced by a 77 MeV iodine ionin 375 nm thick PMMA. Features with an aspect ratioof approx. 15:1 can be achieved.

REFERENCE

[1] M. Dobeli et al., Nucl. Instr. and Meth. B143, 503(1998).

179

INVESTIGATIONS ON THE MECHANISM OF A NOVEL FOCUSED ION BEAM BASEDLITHOGRAPHY TECHNIQUE

H. Fuhrmann, M. Dobeli, R. Muhle (PS1)

We have developed a novel focused ion beam lithography technique for the structuring of silicon surfaceson the sub-ym scale. The technique is based on the desorption of adsorbed hydrogen from the surface of asilicon sample, induced by irradiation with focused ion beams for which any ion species can be used.

Commercially available focused ion beam (FIB) sys-tems have reached the same resolution as electronbeam systems (about 5 nm), so that in some casesFIB based lithography processes may be an interest-ing alternative to electron beam techniques, especiallyif an adsorbed monolayer can be used as a resist [1].We have developed a new process of this kind.

CONCEPT OF THE PROCESS

In our process, a Si sample's surface is first saturatedwith hydrogen by dipping it into a HF bath (Fig.1a).The sample is then introduced into the UHV chamberof the FIB system and irradiated locally (Fig.1b) In-duced by the energy which is deposited by nuclearcollisions between the incident ions and the sample,this leads to desorption of the adsorbed hydrogen. Thesample is oxidised by exposing it to air (Fig.1c) Thenatural oxide can then be used as an etch mask forKOH solutions (Fig.id). Fig. 2 shows an example of asilicon surface structured by this technique.

adsorbed H adsorbed H

Si (100)n

Ga+,Si+5 keV - 30 keV

(b)

adsorbed Hnatural oxide

Anatural oxide

Si (100)

Fig. 1: Concept of the lithography process (schema-tically).

40nm 40

2010

30

i.m

1 00

0

Fig. 2: Atomic force microscopy picture of a siliconsurface after FIB lithography and KOH etching.

CRITICAL FLUENCES

In a first series of experiments, we measured the criti-cal ion fluence for the desorption process, i.e. the flu-ence for which a masking effect in KOH is observed,as a function of ion energy and species. This wasdone by analysing the profiles of the samples afteretching in KOH using a profiler and SEM. For theetching, we used a solution of 10wt% KOH,60wt% H2O and 30wt% isopropyl alcohol at 40°C. Thecritical fluences are given in Table 1.We found that the nuclear deposited energy at thiscritical fluence is essentially the same for Ga+ and Si*ions with energies between 6 keV and 30 keV, a typi-cal value being 0.3 eV/A3.

Additionally, we used 2.5 keV electrons to irradiate oursamples. In this case, a masking effect in KOH wasonly observed for fluences as high as 1017cm"2, whichis probably due to sticking of residual gases inducedby the irradiation.

Table 1 :Critical fluences for various ions and energies.

Ion

Ga+

Ga+

Ga+

srsr

Ion Energy

in keV

5.7

15

30

15

30

Critical Fluence

in 10l3cm2

3.3 + 1.8

2.5 + 1.3

1.4 + 0.6

11 +2

7.4 ±2.2

CONCLUSION

Our experiments have shown that hydrogen that isadsorbed at a silicon surface can be desorbed by irra-diation with focused ion beams, whereas electronbeams are not suitable for this task. The natural oxidecan be used as an etch mask for KOH solutions.

The desorption is a function of the nuclear depositedenergy. This means that any ion species can be used.In particular, we have demonstrated that Si+ ions aresuited for the process, avoiding unwanted doping ofthe samples.

REFERENCE

[1] E.T. Ada et al., J. Vac. Sci. Technol. B13, 2189(1995).

180

APPLICATION OF A NOVEL LITHOGRAPHY PROCESS TO SILICON AND CoSi^SiHETEROSTRUCTURES

H. Fuhrmann, M. Dobeli, R. Miihle (PSI)

We have tested the applicability of the novel focused ion beam lithography technique presented in thisannual report [1] to silicon and CoSi/Si heterostructures. We could demonstrate that step heights of up to430nm can be produced. Induced damage was investigated as a function of the substrate temperatureduring irradiation. The lithography technique is also suitable for the structuring of CoSi/Si heterostructures.

The new focused ion beam (FIB) lithography processwe have reported [1] is of interest for various applica-tions. For instance, it is compatible to AFM lithography[2] and, if Si+ ions are used, undesired chemical altera-tions of the sample are avoided. We have investigatedseveral aspects of the process regarding its applica-tion.

MAXIMUM HEIGHT OF STRUCTURES

The maximum height of the structures is limited by theetch resistance of the natural oxide. We have meas-ured this after KOH etching as a function of the etchingtime. The profile height increases linearly with etchingtime, up to 430nm and with an etch rate of almost50nm/min. After nine minutes, the natural oxide hasbeen removed by KOH.

MINIMIZATION OF INDUCED DAMAGE

In addition to desorbing adsorbed hydrogen from thesurface of the silicon sample, the irradiation causesunwanted damage of the substrate. However, heatingthe substrate during implantation can reduce damage.We have measured the substrate damage in terms ofthe surface recombination velocity s, which is propor-tional to the number of point defects, using the TWAmethod [3,4]. Some results are shown in Fig. 1. Athigh fluences, s (and hence the number of point de-fects) is reduced by a factor of six.

I 40000

120000

100000

80000

cm/s 60000

40000

200GO

0

10

•iiiiifT1

ooooo-ooc

— • - T=23C

—O-T=200C

1 0 " 1 0 "

fluence in c m *2

1 0 ' • 1 0 "

Fig. 1: Surface recombination velocity s after irradia-tion with 30 keV Si+ ions at two different sub-strate temperatures.

HETEROSTRUCTURES

Finally, we demonstrated that the process is applicableto Si/CoSL/Si heterostructures. The structures used inthe experiment consisted of a 4nm CoSi2 layer grownon a Si(100) substrate and were covered by 4nm ofamorphous silicon (a-Si). We used the FIB lithographymethod to locally oxidise the a-Si (Fig. 2a,b). The sur-face was then etched down to the CoSi2 layer by KOH(Fig. 2c). After this, we used the remaining parts of thea-Si as mask for the final etching in HF which removestheCoSi2(Fig. 2d).

4nm a-Si4nm C0SI2SK100)

4nm a-Si4nm CoSi2Si-(100)

adsorbed H

Ga+,Si*5 kev - 30 kev

(a)

(c)

4nm a-Si4nm CoSlj

Si-(100)

adsorbed H

^ ^ ^ ^ - ^

4nm a-Si4nm CoSi2

SI-(100)

natural oxide

(b)

(d)

Fig. 2: Application of the FIB lithography process toCoSi/Si heterostructures (schematically).

The resulting structureswere analysed by atomicforce microscopy (AFM).They showed widths downto 160 nm and the ex-pected profile height of8 nm, corresponding to thesum of thicknesses of theCoSi2 layer and the a-Silayer. An example isshown in Fig. 3.

Fig. 3: AFM picture.

REFERENCES

[1] H. Fuhrmann et al., this Annual Report.

[2] E. S. Snow et al., Appl. Phys. Lett. 64, 1932-1934(1994).

[3] M. Wagner et al., Laser and Optoelektronik 26(1), 63, (1994).

[4] P.W. Nebiker, Diss. ETH Nr. 12046 (1997).

181

A CoY LIQUID ALLOY ION SOURCE FOR FOCUSED ION BEAM IMPLANTATION

ft Muhie, H. Fuhrmann, M. Dobeli (PSI)

A CoY liquid alloy ion source has been developed for focused ion beam implantation of cobalt into silicon.The alloy with the rare earth element Y was selected to achieve a low melting point and a low vapour pres-sure. Using a time-of-flight spectrometer, we have measured the mass spectrum as a function of the emis-sion current. The relative amount and the energy width of the main ions in the spectrum have been deter-mined in the current range between 0.5 and 30 j i A The ion source has been operated successfully over along time in a focused ion beam system.

The formation of buried or surface cobalt silicide layersby focused ion implantation of cobalt into silicon [1]necessitates liquid metal ion sources (LMIS) emitting ahigh percentage of cobalt ions. Due to its high meltingpoint (1495°C), cobalt cannot be used in a pure LMIS,but only in a liquid alloy ion source (LAIS) as constitu-ent of an alloy with considerably lower melting point.Earlier on [2] we reported on our investigations on aAuCo LAIS. This ion source shows a stable ion emis-sion but the lifetime is limited to several days due tothe relatively high working temperature (see Table 1).We continued this work with the investigation of a CoYLAIS promising a high percentage of cobalt ions and alow reactivity with air. Table 1 summarizes some pos-sible binary alloys for Co LAIS together with values formelting temperature and vapour pressure.

The CoY alloy was prepared by electron beam meltingunder high vacuum conditions in a cooled Cu boat.The ion source consists of a tungsten needle emitter,spot welded to a tungsten hairpin, and a molybdenumextraction electrode with a hole diameter of 1 mm,positioned 0.6 mm in front of the emitter. For a goodperformance of the ion source the tungsten wires hadto be prewetted with Au before loading the alloy.

Table 1: Some possible binary alloys CoxM1 x for Co

LAIS. Te: eutectic temperature, Tm(M): melting tem-

perature of the second component M, Pm(M): vapour

pressure of M at its melting point, P1495=C(M): vapour

pressure of M at the melting point of Co.

Alloy

atomic-%

Co«Au715

CoMCeRR

Co,7Ge7,

Co,RNdM

CoMPrRR

Co^Snv

Co,7YRa

Te

°C

997

435

817

566

533

575

738

Tm(M)

°C

1064

798

937

1010

931

1072

1523

Pm(M)

mbar

2.3*10"°5

6.7*10"12

1.3 1 0 ^

3.9*10"°5

2.4*10"°7

4.3*10+°°

2.7*10"°3

mbar

4.7*10"°2

4.2*10"°4

9.3*10"°2

9.3*10"°2

9.7*10"°3

2.7*1 O^2

1.8*10"°3

The mass spectrum shown in Fig. 1 was obtained witha time-of-flight spectrometer, described in detail in [3],for a source current of 5 jiA and a source temperatureof 780°C. Peak #3 (Co+ ions), can be used for cobaltimplantation and peak #2 (Y** ions), for sputteringapplications. The relative amount and the energy widthof the emitted ions have been determined in the cur-rent range between 0.5 and 30 jxA. The relativeamount of Co+ ions in the spectrum increases with thesource current and reaches a maximum of about 10 %at 5 jiA. The relative amount of Y++ ions has a maxi-mum value of about 70 % at low source currents, de-creases with increasing current and approaches aconstant value of about 40 % above 5 uA. The mini-mum energy spread of Co+ and Y~ has been deter-mined to 13 and 18 eV, respectively.

1000 2000channel

3000 4000

Fig. 1: Mass spectrum of the CoY LAIS for a sourcecurrent of 5 jiA. The numbers above the peaksindicate the following ions: 1: C o ^ + Y * 2: Y",3: Co+, 4: CoY", 5: Co3"+Y+, 6: Co2Y", 7: Co/,8: CoY+, 9: Au+, 10: AuCo+, 11: AuY+.

REFERENCES

[1] L Bischoff, E. Hesse, D. Panknin, W. Skorupa,J. Teichert, Microelectronic Engineering 23, 115(1994).

[2] R. Muhle, M. Dobeli, P. Zimmermann, PSI AnnualReport 1996 Annex MIA.

[3] R. Muhle, M. Dobeli, C. Maden, J. Phys. D: Appl.Phys. 32, 187(1999).

183

LIST OF PUBLICATIONS

PARTICLE PHYSICS

R-81-02P. Hauser, K. Kirch, F. Kottmann, L.M. SimonsAbsolute X-ray yields of light muonic atomsNucl. Instr. and Meth. A 411, 389-395 (1998).

R-87-01.2K. Assamagan, Ch. Brönnimann, M. Daum, R. Frosch, P.-R. Kettle, C. WiggerSearch for a heavy neutrino in the decay TC+-> JJ++ V ^

Phys. Lett. B 434, 158(1998).

R-87-03J. Kaulard, C. Dohmen, H. Haan, W. Honecker, D. Junker, G. Otter, M. Starlinger, P. Wintz, J. Hofmann,W. Berti, J. Egger, B. Krause, S. Eggli, R. Engfer, Ch. Findeisen, E.A. Hermes, T. Kozlowski, C.B. Niebuhr,M. Rutsche, H.S. Pruys, A. van der SchaafImproved limit on the branching ratio of | i --> e* conversion on titaniumPhys. Lett. B 422, 334-338 (1998).

E.A. Hermes, F. Rosenbaum, H.P. WirtzThe large, low mass AL-CFC-ARAMID compound vacuum chamber for the SINDRUM-llii-e experimentat PSI (design, testing and operation)Nucl. Instr. and Meth. A 413, 185-190 (1998).

R-87-12A. Ahmidouch, J. Arnold, B. van den Brandt, M. Daum, Ph. Demierre, R. Drevenak, M. Finger, M. Finger Jr.,J. Franz, N. Goujon-Naef, P. Hautle, R. Hess, A. Janata, J.A. Konter, H. Lacker, C. Lechanoine-LeLuc,F. Lehar, S. Mango, C. Mascarini, D. Rapin, E. Rössle, P.A. Schmelzbach, H. Schmitt, P. Sereni,M. Slunecka, A. Teglia, B. VuaridelSpin observables in neutron-proton elastic scatteringEuropean Physical Journal C2, 627-641 (1998).

J. Arnold, B. van den Brandt, M. Daum, R. Drevenak, M. Finger, M. Finger Jr., J. Franz, N. Goujon-Naef,P. Hautle, R. Hess, A. Janata, J.A. Konter, H. Lacker, C. Lechanoine-LeLuc, F. Lehar, S. Mango, D. Rapin,E. Rössle, R. Schirmaier, P.A. Schmelzbach, M. Schmidt, H. Schmitt, P. Sereni, M. Slunecka, A. Teglia,B. VuaridelKinetic energy spectrum and polarization of neutrons from the reaction 12C(p, n )X at 590 MeV

European Physical Journal A2, 411-415 (1998).

R-87-13B. Kotlinski, D. Androic, G. Backenstoss, D. Bosnar, H. Breuer, H. Döbbeling, T. Dooling, M. Furie,P.A.M. Gram, N.K. Gregory, A. Hoffart, C.H.Q. Ingram, A. Klein, K. Koch, J. Köhler, M. Kroedel, G. Kyle,A. Lehmann, A.O. Mateos, K. Michaelian, T. Petkovic, M. Planinic, R.P. Redwine, D. Rowntree,U. Sennhauser, N. Simicevic, R. Trezeciak, H. Ullrich, H.J. Weyer, M. Wildi, K.E. WilsonInitial state interaction in the (j&,3p) reaction on N, Ar andXeEuropean Physics Journal A1, 435-445 (1998).

A.O. Mateos, D. Androic, G. Backenstoss, D. Bosnar, H. Breuer, H. Döbbeling, T. Dooling, M. Furie,P.A.M. Gram, N.K. Gregory, A. Hoffart, C.H.Q. Ingram, A. Klein, K. Koch, J. Köhler, B. Kotlinski, M. Kroedel,G. Kyle, A. Lehmann, K. Michaelian, T. Petkovic, M. Planinic, R.P. Redwine, D. Rowntree, U. Sennhauser,N. Simicevic, R. Trezeciak, H. Ullrich, M. Wang, M.H. Wang, H.J. Weyer, M. Wildi, K.E. WilsonTotal and partial pion absorption cross sections on"He in the A resonance regionPhysical Review C58, 942-952 (1998).

184

N.K. Gregory, D. Androic, G. Backenstoss, D. Bosnar, H. Breuer, H. Döbbeling, T. Dooling, M. Furie,P.A.M. Gram, A. Hoffart, C.H.Q. Ingram, A. Klein, K. Koch, J. Köhler, B. Kotlinski, M. Kroedel, G. Kyle,A. Lehmann, A.O. Mateos, K. Michaelian, T. Petkovic, M. Planinic, R.P. Redwine, D. Rowntree, U. Sennhauser,N. Simicevic, R. Trezeciak, H. Ullrich, M. Wang, M.H. Wang, H.J. Weyer, M. Wildi, K.E. WilsonTwo step processes in pion single charge exchange on2H at Tn+ = 239 MeVPhysical Review C58, 3469-3479 (1998).

R-89-03K. Fohl, R. Bilger, H. Clement, J. Gräter, R. Meier, J. Pätzold, D. Schapler, G.J. Wagner, O. Wilhelm,W. Kluge, R. Wieser, M. Schepkin, R. Abela, F. Foroughi, D. RenkerPionic double charge exchange on N = Z doubly closed shell nucleiPhys. Rev. Lett. 79, 3849-3852 (1997).

J. Pätzold, R. Bilger, H. Clement, A. Denig, F. Fohl, J. Gräter, W. Kluge, R. Meier, M. Schepkin, G.J. Wagner,O. WilhelmPionic double charge exchange on33Nb at low energiesPhys. Lett. B 428, 18-22 (1998).

J. Pätzold, R. Bilger, H. Clement, F. Fohl, J. Gräter, R. Meier, D. Schapler, G.J. Wagner, A. Denig, W. Kluge,M. SchepkinThe reaction 7Li (n+,nr) 7B and its implications for7BPhys. Lett. B 443, 77-81 (1998).

H. Clement, R. Bilger, A. Denig, F. Fohl, J. Gräter, W. Kluge, R. Meier, J. Pätzold, M.G. Schepkin, G.J. WagnerPionic double charge exchangeActa Physica PolonicaB29, 2999-3007 (1998).

R-89-03R. Bilger, H. Clement, A. Denig, K. Fohl, J. Gräter, W. Kluge, J. Pätzold, R. Meier, M. Schepkin, G.J. Wagner,P.A. Amaudruz, A. Ambadar, F. Bonuti, P. Camerini, J. Clark, L. Felawka, E. Friedman, N. Grion, G.J. Hofman,P. Hong, M. Kermani, E.L Mathie, D. Ottewell, O. Patarakin, R. Rui, M.E. Sevior, G.R. Smith, R. Tacik,G. Tagliente, W. Brodowski, H. Calén, C. Ekström, K. Fransson, L. Gustafsson, S. Häggström, B. Höistad,A. Johansson, T. Johansson, K. Kilian, S. Kullander, A. Kupsc, G. Kurz, P. Marciniewski, B. Morosov, J. Moehn,A. Mörtsell, W. Oelert, V. Renken, R. Ruber, U. Siodlaczek, J. Stepaniak, A. Sukhanov, A. Turowiecki,Z. Wilhelmi, J. Zabierowski, A. Zernov, J. ZlomanczukHunting the dibaryon d' (2065)Acta Physica PolonicaB29, 2415-2423 (1998).

R-91-13H. Anklin, L.J. deBever, K.I. Blomqvist, W.U. Boeglin, R. Böhm, M. Distler, R. Edelhoff, J. Friedrich, D. Fritschi,R. Geiges, J. Götz, A. Honegger, P. Jennewein, J. Jourdan, M. Kahrau, M. Korn, H. Kramer, K.W. Krygier,G. Kubon, V. Kunde, A. Liesenfeld, G. Masson, K. Merle, R. Neuhausen, E.A.J.M. Offermann, Th. Petitjean,Th. Pospischil, M. Potokar, L.M. Qin, A.W. Richter, R. Rokavec, G. Rosner, P. Sauer, S. Schardt, I. Sick,S. Sirca, Ph. Trueb, M. Tuccillo, B. Vodenik, A. Wagner, Th. Walcher, G. Warren, S. Wolf, J. Zhao, M. Zeier,B. ZihlmannPrecise measurements of the neutron magnetic form factorPhys. Lett. B 428, 248 (1998).

R-93-02P. Ackerbauer, D.V. Balin, V.N. Baturin, G.A. Beer, W.H. Breunlich, T. Case, K.M. Crowe, H. Daniel, J.Deutsch,J. Govaerts, Yu.S. Grigoriev, F.J. Hartmann, P. Kammel, R. King, B. Lauss, E.M. Maev, V.E. Markushin,J. Marton, M. Mühlbauer, C. Petitjean, Th. Petitjean, G.E. Petrov, R. Prieels, W. Prymas, W. Schott,G.G. Semenchuk, Yu.V. Smirenin, A.A. Vorobyov, N.I. Voropaev, P. WojciechowskyA precision measurement of nuclear muon capture on3HePhys. Lett. B 417, 224-232 (1998).

185

R-93-06M. Bregant, D. Chatellard, J.P. Egger, P. Häuser, E. Jeannet, K. Kirch, F. Kottmann, E. Milotti, C. Petitjean,L.M. Simons, D. Taqqu, E. ZavattiniMeasurement of K X-ray intensity ratios in muonic hydrogen at low gas densitiesPhys. Lett. A 241, 344-350 (1998).

R-94-01D. Anagnostopoulos, M. Augsburger, D. Belmiloud, G. Borchert, D. Chatellard, M. Daum, J.-P. Egger,P. El-Khoury, H. Gorke, D. Gotta, P. Häuser, P. Indelicate, K. Kirch, S. Lenz, Th. Siems, L.M. SimonsA new determination of the mass of the charged pionPhys. Lett. B 416, 50-55 (1998).

D. Anagnostopoulos, M. Augsburger, G. Borchert, D. Chatellard, J.-P. Egger, P. El-Khoury, D. Gotta,P. Häuser, P. Indelicato, K. Kirch, Th. Siems, L.M. SimonsNew precision measurement of the pionic deuterium S-wave strong interaction parametersPhys. Rev. C 58, 1869-1872 (1998).

R-94-03B. Lauss, P. Ackerbauer, W.H. Breunlich, B. Gartner, M. Jeitler, P. Kammel, J. Marton, W. Prymas, J. Zmeskal,D. Chatellard, J.P. Egger, E. Jeannet, H. Daniel, F.J. Hartmann, A. Kosak, C. PetitjeanX-ray emission during the muonic cascade in hydrogenPhys. Rev. Lett. 80, 3041-3044 (1998).

S. Tresch, R. Jacot-Guillarmod, F. Mulhauser, C. Piller, L.A. Schaller, L. Schellenberg, H. Schneuwly,Y.A. Thalmann, A. Werthmüller, P. Ackerbauer, W.H. Breunlich, C. Cargnelli, B. Gartner, R. King, B.Lauss,J. Marton, W. Prymas, J. Zmeskal, C. Petitjean, D. Chatellard, J.P. Egger, E. Jeannet, F.J. Hartmann,M. MühlbauerMuon transfer rates from hydrogen to 3He and 4He measured at low temperaturePhys. Rev. A 57, 2496-2501 (1998).

S. Tresch, R. Jacot-Guillarmod, F. Mulhauser, L.A. Schaller, L. Schellenberg, H. Schneuwiy,Y.A. Thalmann, A. WerthmüllerCharge transfer from the ground state of muonic hydrogen to "He at room temperatureEur. Phys. J.D 2, 93(1998) .

S. Tresch, F. Mulhauser, C. Piller, L.A. Schaller, L. Schellenberg, H. Schneuwly, Y.A. Thalmann, A. Werth-müller, P. Ackerbauer, W.H. Breunlich, C. Cargnelli, B. Gartner, R. King, B.Lauss, J. Marton, W. Prymas,J. Zmeskal, C. Petitjean, D. Augsburger, D. Chatellard, J.P. Egger, E. Jeannet, T. von Egidy, F.J. Hartmann,M. Mühlbauer, W. SchottMeasurement of the formation rate and the radiative decay of the muonic molecules (p\i3He)* and (p \i4He)*Phys. Rev. A 58, 3528-3536 (1998).

R-94-04R. Schmidt, H. Daniel, F.J. Hartmann, P. Hauser, F. Kottmann, M. Mühlbauer, C. Petitjean, W. Schott,D. Taqqu, P. WojciechowskiMeasurement of the Barkas effect in hydrogenEur. Phys. J. D 3,119-122 (1998).

R-95-01Y.-A. Thalmann, R. Jacot-Guillarmod, F. Mulhauser, L.A. Schaller, L. Schellenberg, H. Schneuwly, S. Tresch,A. WerthmüllerMuon transfer from excited states of hydrogen and deuterium to nitrogen, neon, and argonPhys. Rev. A 57, 1713(1998).

A. Werthmüller, A. Adamczak, R. Jacot-Guillarmod, F. Mulhauser, L.A. Schaller, L. Schellenberg,H. Schneuwly, Y.-A. Thalmann, S. TreschEnergy dependence of the charge exchange reaction from muonic hydrogen to oxygenHyperfine Interactions 116,1 (1998).

186

R-98-02V.F. Boreiko, V.M. Bystritsky, V.l. Datskov, A.N. Fedorov, V.N. Pavlov, V.A. Stolupin, A. Del Rosso,R. Jacot-Guillarmod, F. Mulhauser, L.A. RivkisNew target cryostat for experiments with negative muonsNucl. Instr. and Meth. A 416, 221 (1998).

Z-90-09Ch. Herren, B. Böschung, J.-CI. Dousse, B. Galley, J. Hoszowska, J. Kern, Ch. Rhême, M. Polasik,T. Ludziejewski, P. Rymuza, Z. SujkowskiProbabilities for M-shell ionization in intermediate velocity collisions of medium-mass atoms with 4He2* ionsPhys. Rev. A 57, 235-245 (1998).

Z-91-05H. Anklin, J. Jourdan, L.M. Qin, I. Sick, P.A. Schmelzbach, P. Trueb, D. Vetterli, B. Zihlmann4He(D,D) scattering as a polarization standardNucl. Instr. and Meth. A 404, 394 (1998).

Z-93-01.1N. Warr, S. Drissi, P.E. Garrett, J. Jolie, J. Kern, H. Lehmann, S.J. Mannanal, J.P. VorletStudy of 124Te by the 122Sn(<x,2n,y) reaction and by the decay of 124INucl. Phys. A 636, 379-418 (1998).

Z-93-08H. Anklin, L.J. de Bever, S. Buttazzoni, W. Glöckle, J. Golak, A. Honegger, J. Jourdan, H. Kamada, G. Kubon,T. Petitjean, L.M. Qin, I. Sick, Ph. Steiner, H. Witala, M. Zeier, J. Zhao, B. ZihlmannTensor analyzing power A^ of d - p radiative captureNuclear Physics A 636, 189 (1998).

E-88-01R. Luescher, J. Farine, F. Boehm, J. Busto, K. Gabathuler, G. Gervasio, H.E. Henrikson, V. Jorgens, K. Lou,A. Paie, D. Schenker, A. Tadsen, M. Treichel, J.L. Vuilleumier, J.M. Vuilleumier, H. WongSearch for fifi-decay in 136Xe: New results from the Gotthard experimentPhys. Lett. B 434, 407-414 (1998).

E-88-05D.H. Wright, S. Ahmad, D.S. Armstrong, G. Azuelos, W. Berti, M. Blecher, C.Q. Chen, P. Depommier,B.C. Doyle, T. von Egidy, T.P. Gorringe, P. Gumplinger, M.D. Hasinoff, D. Healey, G. Jonkmans, A.J. Larabee,J.A. Macdonald, S.C. McDonald, M. Munro, J.-M. Poutissou, R. Poutissou, B.C. Robertson, D.G. Sample,E. Saettler, CM. Sigler, G.N. Taylor, N.S. ZhangMeasurement of the induced pseudoscalar coupling using radiative muon capture on hydrogenPhys. Rev. C 57, 373-390 (1998).

CERN, CPLEAR COLLABORATIONA. Angelopoulos et al.,P. Bargassa, F. Blanc, P.-R. Kettle, T. Nakada, 0. WiggerMeasurement of the CP-violation parameter r\m using tagged anti-K ° and K °Phys. Lett. B 420, 191 (1998).

An EPR experiment testing the non-separability of the K° anti-K ° wave functionPhys. Lett. B 422, 339 (1998).

Search for CP violation in the decay of tagged anti-K0 and K° to j r WPhys. Lett. B 425, 391 (1998).

The neutral kaons decays to n*nn°: A detailed analysis of the CPLEAR dataEur. Phys. J. C5, 389 (1998).

Measurement of the KL Ks mass difference using semileptonic decays of tagged neutral kaonsPhys. Lett. B 444, 38(1998).

187

First direct observation of time-reversal non-invariance in the neutral kaon systemPhys. Lett. B 444, 43 (1998).

A determination of the CPT violation parameter Re(8) from the semileptonic decay of strangeness-taggedneutral kaonsPhys. Lett. B 444, 52 (1998).

CERN, L3 COLLABORATIONM. Acciari et al.,A. Barczyk, V. Brigljevic, K. Deiters, M. FabreMeasurement of the anomalous magnetic and electric dipole moments of the tau leptonPhys. Lett. B 434, 169-179 (1998).

Measurement of the B°d-B°d oscillation frequency

Eur. Phys. J. C5, 195-203 (1998).

Search for the standard model HIGGS boson in e+e~ interactions at square root (s)=183 GeVPhys. Lett. B 431, 437-450 (1998).

Measurement of W-pair cross sections in e+e~ interactions at^ls =183 GeV and W-decay branching fractionsPhys. Lett. B 436, 437 (1998).

Production of single W bosons in e+e~ interactions at 130 < Vs < 183 GeV and limits on anomalouscouplingsPhys. Lett. B 436, 417-27 (1998).

Study of the hadronic photon structure function F2r at LEP

Phys. Lett. B 436, 403 (1998).

Measurement of the inclusive charmless semileptonic branching fraction of beauty hadrons and adetermination ofNJ at LEPPhys. Lett. B 436, 174 (1998).

Study of anomalous ZZy and Zyy couplings at LEPPhys. Lett. B 436, 187 (1998).

Test of CP invariance in Z-> p.+^~y decayPhys. Lett. B 436, 428-436 (1998).

Search for neutral HIGGS bosons of the minimal supersymmetric standard model in e+e- interactions atVs= 130-183 GeVPhys. Lett. B 436, 389-402 (1998).

Search for new physics phenomena in fermion-pair production at LEPPhys. Lett. B 433, 163 (1998).

Measurement of% polarisation at LEPPhys. Lett. B 429, 387-398 (1998).

Determination of the number of light neutrino species from single photon production at LEPPhys. Lett. B 431, 199 (1998).

Angular multiplicity fluctuations in hadronic Z decays and comparison to QCD models and analyticalcalculationsPhys. Lett. B 428, 186-196 (1998).

Measurement of the weak dipole moments of the x leptonPhys. Lett. B 426, 207-216 (1998).

188

Local multiplicity fluctuations in hadronic Z decayPhys. Lett. B 429, 375-386 (1998).

Search for scalar leptons, charginos and neutralinos in e+e- collisions at^ls= 161-172 GeVEur. Phys. J. C4, 207-19 (1998).

Measurement of the average lifetime of b-hadrons in Z decaysPhys. Lett. B 416, 220-232 (1998).

Missing mass spectra in hadronic events from e+e- collisions at^ls = 161-172 GeV and limits on invisibleHIGGS decaysPhys. Lett. B 418, 389 (1998).

DESY, H1 COLLABORATIONC. Adloff et al.,R. Eichler, K. Gabathuler, J. Gassner, R. HorisbergerPhotoproduction of *F2S mesons at HERAPhys.Lett. B421, 385-394 (1998).

Thrust jet analysis of deep inelastic large rapidity-gap-eventsEur.Phys.J. C1, 495-507 (1998).

Measurement of the inclusive dijet cross-section in photoproduction and determination of an effective partondistribution in the photonEur.Phys.J. C1, 97-107 (1998).

Differential (2+1) jet event rates and determination ofas in deep inelastic scattering at HERAEur.Phys.J. C5, 625-639 (1998).

Observation of events with an isolated high-energy lepton and missing transverse momentum at HERAEur.Phys.J. C5, 575-584 (1998).

Multiplicity structure of the hadronic final state in diffractive deep inelastic scattering at HERAEur.Phys.J. C5, 439-452 (1998).

Hadron production in diffractive deep inelastic scatteringPhys.Lett. B 428, 206-220 (1998).

CERN, CMS COLLABORATIONE. Auffray et al.,K. Deiters, Q. Ingram, D. RenkerBeam tests of lead tungstate crystal matrices and a silicon strip preshower detector for the CMSelectromagnetic calorimeterNucl. Instrum. Methods Phys. Res. A 412, 223 (1998).

H.F. Chen, K. Deiters, H. Hofer, P. Lecomte, F. Nessi-TedaldiRadiation damage measurements of undoped lead tungstate crystals for the CMS electromagneticcalorimeter at LHCNucl. Instrum. Methods Phys. Res. A 414, 149 (1998).

THEORY GROUPE. Accomando et al. (ECFA/DESY LC Physics Working Group), F. Cuypers, A. DennerPhysics with e+e- linear collidersPhys. Rept. 299, 1-78 (1998).

P. Ackerbauer et al. and V.E. MarkushinA precision measurement of nuclear muon capture on3HePhys. Lett. B 417, 224-232 (1998).

189

C. Alexandrou, R. Rosenfelder, A. W. SchreiberVariational field theoretic approach to relativistic meson-nucleon scatteringNucl. Phys. A 628, 427-457 (1998).

C. Alexandrou, R. Rosenfelder, A.W. SchreiberVariational treatment of quenched QED using the worldline techniqueNucl. Phys. A 631, 635c-639c (1998).A. Angelopoulos et al. (CPLEAR Collaboration), M.P. Locher, V.E. Markushin

Direct determination of two-pion correlations for ppto2n+ 2x~ annihilation at rest

Eur. Phys. J. C1, 139-148 (1998).

U. Baur, S. Keller, D. WackerothElectroweak radiative corrections to W boson production in hadronic collisionsPhys. Rev. D 59, 013002 (1999).

A. Boriçi, R. RosenfelderScaling in Su(3) theory with a MCRG improved lattice actionNucl. Phys. Proc. Suppl. 63, 925-927 (1998).

F. Cuypers, S. DavidsonBileptons: Present limits and future prospectsEur. Phys. J. C 2, 503-528 (1998).

A. Denner, S. Dittmaier, M. RothNonfactorizable photonic corrections to e+e- -» W+W- -> 4 fermionsNucl.Phys. B 519, 39-84 (1998).

A. Denner, T. HahnRadiative corrections to M/W-> WW in the electroweak standard modelNucl.Phys. B 525, 27-50 (1998).

A. Denner, S. Dittmaier, M. RothFurther numerical results on non-factorizable corrections to e+e' —> 4 fermionsPhys. Lett. B 429, 145-150 (1998).

A. Denner, W. BeenakkerRadiative corrections to off-shell gauge-boson pair productionActa Phys. Polon. B 29, 2821-2838 (1998).

D. Graudenz (ed.)Proceedings of the Zuoz summer school on hidden symmetries and HIGGS phenomenaPSI Proceedings 98-02, ISSN 1019-6447 (1998).

N. Fettes, R. RosenfelderInclusive and deep inelastic scattering from a dressed structureless nucléonFew-Body Syst. 24, 1-25 (1998).

P. Hasenfratz, V. Laliena, F. NiedermayerThe index theorem in QCD with a finite cutoffPhys. Lett. B 427, 125-131 (1998).

T. Herrmann, R. RosenfelderA consistent calculation of dispersion corrections in elastic electron-deuteron scatteringEur. Phys. J. A 2, 29-40 (1998).

H. Kamada, M.P. Locher, T.-S.H. Lee, J. Golak, V.E. Markushin, W. Glöckle, H. WitalaPion absorption cross section for 2H and 3He in the A-isobar regionNucl. Phys. A 631, 519c-523c (1998).

190

M.P. Locher, V.E. Markushin, H.Q. ZhengStructure off0(980) from a coupled channel analysis of S-wavenn scatteringEur. Phys. J. C 4, 317-326 (1998).

M.P. Locher, V.E. MarkushinReply to the comment: Central production off0(980) in pp -•> ppXPhys. Rev. D 58, 38504 (1998).

V.E. Markushin, M.P. LocherOZI rule violation in pp annihilation into tynn by two step processes

Eur. Phys. J. A 1, 91-98 (1998).

F. Cuypers, P. H. Frampton, R. RucklProduction of two non-conjugate leptoquarks in e+e~ collisionsPhys. Lett. B 390, 221-226 (1997).

ASTROPHYSICS

A. O. Benz, J. Conway, M. GCidelFirst VLBI images of a main-sequence starAstron. & Astrophys. 331, 596 (1998).

F. De Paolis, G. Ingrosso, Ph. Jetzer and M. RoncadelliMachos and molecular clouds in galactic halosin Proceedings of "Dark Matter in Astro- and Particle Physics" (Heidelberg, September 1996) (World Scientific,Singapore 1998), 214-221.

F. De Paolis, G. Ingrosso, Ph. Jetzer and M. RoncadelliX-ray emission from dark clusters of machosAstron. & Astrophys. 329, 74 (1998).

F. De Paolis, G. Ingrosso, Ph. Jetzer and M. RoncadelliBinary brown dwarfs in the galactic halo?Mont. Not. R. Astr. Soc. 294, 283 (1998).

F. De Paolis, G. Ingrosso, Ph. Jetzer and M. RoncadelliHalo dark clusters of brown dwarfs and molecular cloudsAstrophys. J. 500, 59 (1998).

Ph. Jetzer, M. Strassle and U. WandelerGravitational microlensing by globular clustersAstron. & Astrophys. 336, 411 (1998).

Ph. JetzerGravitational microlensingNapoli Series in Physics and Astrophysics, Vol. 1, p. 75 -100 (1998).

Ph. JetzerKalte Wasserstoffwolken als dunkle Materie?Beitrag zu Forschung und Technik, Neue Zurcher Zeitung, 18. 11.1998.

Ph. Jetzer, M. Strassle and N. StraumannOn the variation of pulsar periods in close binary systemsNew Astronomy 3, 619 (1998).

191

R. Mewe, S. A. Drake, J. S. Kaastra, C. J. Schrijver, J. J. Drake, M. Güdel, J. H. M. M. Schmitt,K. P. Singh, N. E. WhiteAlpha Centauri: Coronal temperature & abundances from ASCA observationsAstron. & Astrophys. 339, 545 (1998).

R. Mewe, M. Güdel, F. Favata, J. S. KaastraA Beppo SAX/LECS X-ray observation of alpha CentauriAstron. & Astrophs. 340, 216(1998).

K.A. Moldowanov, M.A. Samsonov, L.S. Kim, R. Henneck, O.H.W. Siegmund, J. Warren, S. Cully, D. MarshHighly absorptive coating for the vacuum ultraviolet rangeAppl. Optics 37, 93-97 (1998).

D. Puy, M. SignorePrimordial LiH: the chemistry in a collapsing protocloudNew Astronomy 3, 27 (1998).

D. Puy, M. SignoreCooling balance and Thermal Instability of a collapsing molecular protocloudNew Astronomy 3, 247 (1998).

D. Puy, P. PeterPrimordial magnetic field and spectral distortion of CBRInt. J. Mod. Phys. D 7, 489 (1998).

MUON SPIN SPECTROSCOPY

RA-86-07E. Roduner, M. Stolmâr, H. Dilger, I. D. ReidReorientational dynamics of cyclohexadienyl radicals in high-silica ZSM-5J. Phys. Chem. A102, 7591 (1998).

M. Stolmâr, E. RodunerComplexation of copper by cyclohexadienyl in Cu/ZSM-5 zeoliteJ. Amer. Chem. Soc. 120, 583 (1998).

RA-90-07C. M. Aegerter, S.H. Lloyd, C. Ager, S.L. Lee, S. Romer, H. Keller, E.M. ForganEvidence for a square vortex lattice in Sr2RuO4 by muon-spin rotation measurementsJ. Phys. Cond. Mat. 10, 7445 (1998).

C. Ager, S.L. Lee, CM. Aegerter, F.L Pratt, S.J. Blundell, S.H. Lloyd, T. Sasaki, E.M. Forgan, H. KellerAngular dependent muon-spin rotation on the mixed state of the organic superconductorK-(BEDT-TTF)2CU(SCN),

J. Mag. Mag. Mat. 177-181, 561 (1998).

RA-90-16A. Schenck, M. Pinkpank, F.N. Gygax, K.U. Neumann, K.R.A. Ziebeck, A. AmatoMagnetic properties of U14Aus1: An exemplary study by muon spin rotation faSR) spectroscopyJ. Phys.: Condens. Matter 10, 8059 (1998).

RA-90-17H. Dilger, M. Stolmâr, U. Himmer, E. RodunerKinetics of the gas-phase addition of the ethyl radical and the tert-butyl radical to NOJ. Phys.: Chem. A102, 6772 (1998).

192

RA-93-05W.Wagener, H.-H. Klauss, M. Hillberg, M.A.C. de Melo, W. Kopmann, M. Birke, F.J. Litterst, B. Büchner,H. Micklitz^i+SR on La2.x.yRExSryCuO4

J. Magn. Magn. Mat. 177-181, 545-546 (1998).

W. Wagener(iSR-Untersuchungen an Selten-Erd-dotiertem La2.x.ySryCuO4

PhD thesis, TU Braunschweig, ISBN 3-89720-210-7 (1998).

RA-93-06Th. Jestädt, R.I. Bewley, S.J. Blundell, W. Hayes, B.W. Lovett, F.L. Pratt, R.C.C. WardSpin dynamics in the spin-gap system CaV4Og studied using muon-spin relaxationJ. Phys.: Condens. Matter 10, 259 (1998).

C. Ager, S. L. Lee, C. M. Aegerter, F. Pratt, S. J. Blundell, S. H. Lloyd, T. Sasaki, E. M. Forgan, H. KellerAngular-dependent muon-spin rotation on the mixed state of the organic superconductorK-BEDT-TTF)2CU(SCN)2

J. Mag. Mag. Mat. 177, 561 (1998).

R. M. Valladares, A.J. Fisher, S.J. Blundell, W. HayesStudies of implanted muons in organic radicalsJ. Phys.: Condens. Matter 10,10701 (1998).

RA-93-10M. Schefzik, J. Schmidl, R. Scheuermann, L. Schimmele, A. Seeger, D. Herlach, O. Kormann, J. Major, A. RockEvidence for a novel muon species in crystalline siliconSolid State Comm. 107, 395 (1998).

RA-94-03G.J. Nieuwenhuys, S. Suellow, A.A. Menovsky, J.A. Mydosh, R.H. Heffner, L.P. Le, D.E. MacLaughlin,O.O. Bernai, A. Schenckji+SR in the random bond spin glass URh2Ge2

J. Mag. Mag. Mat. 177-181, 803 (1998).

RA-94-04S.L. Lee, CM. Aegerter, S.H. Lloyd, E.M. Forgan, C. Ager, M.B. Hunt, H. Keller, I.M.Savic, R. Cubitt, G. Wirth,K. Kadowaki, N. KoshizukaObservations of suppression of static and dynamic disorder in Bi,1s Sr1B5CaCu2OM crystals by columnar defectsPhys. Rev. Lett. 81, 5209 (1998).

CM. Aegerter, J. Hofer, I.M. Savic, H. Keller, S.L. Lee, C. Ager, S.H. Lloyd, E.M. ForganAngular dependence of the disorder crossover in Bi215 SrIBSCaCu2OM from muon-spin rotation and torquemagnetometryPhys. Rev. B 57, 1253 (1998).

S.L. Lee, CM. Aegerter, C. Ager, E.M. Forgan, S.H. Lloyd, H. KellerMuon-spin rotation measurements on Bi2Sr2CaCu2OM in the presence of columnar defectsJ. Mag. Mag. Mat. 177-181, 523 (1998).

RA-94-14V.S. Egorov, E.P. Krasnoperov, F.V. Likov, G. Solt, C Baines, D. Herlach, U. ZimmermannObservation of diamagnetic domains in berylliumPhysics of the Solid State (Fiz.Tverd.Tela)40, 524 (1998).

RA-95-10V.Yu. Pomjakushin, A.M. Balagurov, A.I. Beskrovny, V.N. Duginov, Yu.V. Obukov, A.A. Zakharov,A.N. Ponomarev, F.N. Gygax, A. Schenck, A. Amato, D. Herlach, S.N. BariloMicroscopic phase separation in LafiuO^ induced by the superconducting transitionPhys. Rev. B 58, 12350 (1998).

193

RA-95-14A.M. Mulders, P.CM. Gubbens, U. Gasser, C. Baines, K.H.J. BuschowTwo types of magnetism in the magnetic superconductor TmNisB2C related to the degree of carbon-siteoccupancyPhys. Rev. B 57, 10323 (1998).

A.M. Mulders, P.CM. Gubbens, U. Gasser, C. Baines, K.H.J. BuschowTwo magnetic moments observed at the Tm site in the magnetic superconductor TmN^B2CJ. Mag. Mag. Mat. 177-181, 555-556 (1998).

RA-95-17D. Herlach, C. Buhrer, D.M. Herlach, K. Maier, C. Notthoff, D. Platzek, J. ReskeMagnetic ordering in a supercooled Co-Pd melt studied by muon-spin rotationEurophys. Lett. 44, 98-102 (1998).

RA-97-19A.I. Mihut, L.E. Spring, R.I. Bewley, S.J. Blundell, W. Hayes, T. Jestadt, B.W. Lovett, R. McDonald, F.L. Pratt,J. Singleton, P.D. Battle, J. Lago, M.J. Rosseinsky, J.F. VentePhysical properties of the n =3 Ruddlesden - Popper compound Ca4Mn3O10

J. Phys.: Condens. Matter 10, 727 (1998).

RA-97-21T.N. Mamedov, I.L. Chaplygin, V.N. Duginov, V.N. Gorelkin, D. Herlach, J. Major, A.V. Stoykov, M. Schefzik,U.ZimmermannInvestigation of acceptor centers in semiconductors with the diamond crystal structure by the \iSR-methodJETP Lett. 68, 64-70 (1998). (Pis'ma Zh. Eksp. Teor. Fiz. 68, 61-66 (1998).

RA-98-06A. Schenck, D. Andreica, F.N. Gygax, M. Pinkpank, Y. Onuki, P. Ahmet, M. Abliz, R. Settai, A. Amato,N. KaplanHigh-temperature magnetic order in the singlet-ground state compound PrCu2

Phys. Rev. B 58, 5205 (1998).

RA-98-10W. Wagener\iSR-Untersuchungen an Selten-Erd-dotiertem La2.ySryCuO4

PhD thesis, TU Braunschweig, ISBN 3-89720-210-7 (1998).

RA-98-14A. de Visser, R.J. Keizer, R. van Harrevelt, A.A. Menovsky, J.J.M. Franse, A. Amato, F.N. Gygax, M. Pinkpank,A. Schenck\x.SR study of magnetism in heavy-fermion U(Pt,Pd)3

J. Mag. Mag. Mat. 177, 435 (1998).

Low Energy MuonsT. Prokscha, E. Morenzoni, M. Meyberg, Th. Wutzke, B.E. Matthias, A. Fachat, K. Jungmann, G. zu PutlitzMuonium formation by collisions of muons with solid rare-gas and solid nitrogen layersPhys. Rev. A 58, 3739 (1998).

MICRO AND NANO TECHNOLOGYF. Coppinger, J. Genoe, D.K. Maude, X. Kleber, L.B. Rigal, U. Gennser, J.C. Portal, K.E. Singer, P. Rutter,T. Taskin, A..R. Peaker, A.C. WrightMagnetic characterization of self-organized ErAs clusters using telegraph noise spectroscopyPhysical Review B 57, 7182 (1998).

194

C. DavidNanometer structuring at the Paul Scherrer InstitutreSOLUTIONI (1998).

C. David, D. R. Kayser, H. U. Muller, B. Volkel, M. GrunzeA new method for the manufacture of large-area condenser zone plates with small outermost zone width,In: X-ray microscopy and spectromicroscopy, Eds.: J. Thieme, G. Schmahl, D. Rudolph, E. Umbach,Springer Verlag (1998).

P. Gassot, U. Gennser, D.M. Symons, A. Zaslavsky, D.A. Grutzmacher, J.C. PortalStress and pressure effects on a Si/SiGe double-barrier structure studied by magnetotunneling spectroscopyPhysicaE 2, 758(1998).

D. Grutzmacher, R. Hartmann, P. Schnappauf, U. Gennser, H. Sigg, S. Brosig, D. BachleMagneto-transport studies of Si/SiGe and Si/SiGeC quantum well structures grown by molecular beamepitaxy at low temperaturesJ. Vac. Sci. Technol. B16, 1670 (1998).

D. Griitzmacher, R. Hartmann, P. Schnappauf, U. Gennser, E. Muller, D. Bachle, A. Dommann,Low temperature Molecularbeam epitaxy of SiGeC/Si quantum well structures: electrical and optical propertiesThin Solid Films 321, 26 (1998).

G.M. Gusev, U. Gennser, X. Kleber, D.K. Maude, J.C. Portal, D.I. Lubyshev, P. Basmaji, M. de P.A. Silva,J.C. Rossi, Yu.V. NastaushevPercolation network in a smooth artificial potentialPhysical Review B 58, 4636 (1998).

G.M. Gusev, P. Basmaji, U. Gennser, X. Kleber, D.K. Maude, J.C. Portal, D.I. Lubyshev, M. de P.A.Silva,J.C. Rossi, Yu.V. NastaushevUniversal diagonal and Hall conductance fluctuations in a random magnetic fieldBrazilian Journal of Physics 27A, 219 (1997).

R. Hartmann, D. Grutzmacher; E. Mueller; U. Gennser; A. Dommann; P. Schroeter; P. WarrenGrowth of Sii-yCy/Si- and Sii-x-yGexCy/Si-multiple quantum wells using molecular beam epitaxyThin-Solid-Films 318, 158 (1998).

R. Hartmann, U. Gennser, H. Sigg, and D. Grutzmacher, K. EnsslinBand gap and band alignment of strain reduced Si/S^^y Ge^Cy multiple quantum well structuresobtained by photoluminescence measurementsApplied Physics Letters 73,1257 (1998).

R. Hild, C. David, H. U. Muller, B. Volkel, D. R. Kayser, M. GrunzeFormation and characterization of self-assembled monolayers of octadecyltrimethoxysilane on chromium:Applications in low-energy electron lithographyLangmuir 14, No. 2, 342-346 (1998).

F.J. Himpsel, T.A. Jung, P.F. SeidlerElectronic states in magnetic nanostructuresIBM Journal of Research and Development 42/1, 33-42 (1998).

R.W. Jaszewski, H. Schift, J. Gobrecht, P. SmithHot embossing in polymers as a direct way to pattern resistMicroelectronic Engineering 41/42, 575-578 (1998).

T.A. Jung, F.J. Himpsel, R.R. Schlittler, J.K. GimzewskiChemical information from scanning probe microscopy and spectroscopyBook Chapter in: Scanning Probe Microscopy Analytical Methods. Ed. by R. Wiesendanger, Springer Seriesin Nanoscience and Technology, Series Editors. R. Wiesendanger, K. v. Klitzing; Springer, Berlin, Heidelberg,New York (1998).

195

X. Kleber, G.M. Gusev, M.V. Budantsev, M. Cassé, U. Gennser, D.K. Maude, J.C. Portal, Z.D. Kvon,A.E. Plotnikov, A.I. Toropov, and N.T. MoshegovMagnetoconductance of two independently tunable parallel point contacts using an alliptical antidotPhysics of Low-Dimensional Structures 11/12, 75 (1997).

S. Kossek, C. Padeste, H. Siegenthaler, L. TiefenauerLocalization of individual biomolecules on sensor surfacesBiosensors & Bioelectronics 13, 31-43 (1998).

H. Kühne, A. Fischer, T. Morgenstern, D. GrützmacherHow boron incorporated during SiGe chemical vapor deposition degrades the Ge profileSolid State Electrochem. 2, 273 (1998).

B. Larsson, N. Crompton, J. Gobrecht, C. Higgs, G. Kühne, E. Lehmann, H. Reist, J. Stepanek, T. SudaA silicon based compact cell culture diskJ. Ass. Laboratory Automation 3, 58 - 62 (1998).

O. Leifeld, B. Müller, D.A. Grützmacher, K. KernAn UHV-STM forin-situ characterization of MBE/CVD growth on 4-inch wafersAppl. Phys. A 66, 993(1998)

C. Padeste, A. Grubelnik, L. TiefenauerAmperometric immunosensing using microperoxidase MP-11 antibody conjugatesAnal. Chim. Acta 374, 167 -176 (1998).

D.J. Petrovykh, F.J. Himpsel, T.A. JungWidth distribution of nanowires grown by step decorationSurface Science 407, 189-199 (1998).

E. Ribeiro, E. Mueller, T. Heinzel, H. Auderset, K. EnsslinInAs self-assembled quantum dots as controllable scattering centers near a two-dimensional electron gasPhys. Rev. B 58, 1506 (1998).

R. Ros, F. Schwesinger, D. Anselmetti, M. Kubon, R. Schäfer, A. Plückthun, L. TiefenauerAntigen binding forces of individually addressed single-chain Fv' antibody moleculesProc. Natl. Acad. Sei. USA 95, 7402-7405 (1998).

C. Rosenblad, H.R.Deller, T. Graf, E. Mueller, H. von KaenelLow temperature epitaxial growth by LEPECVDJ. Cryst. Growth 188, 125-130 (1998).

C. Rosenblad, H.R. Délier, M. Doebeli, E. Mueller, H. von KaenelLow temperature epitaxial growth by LEPECVDThin Solid Films 318, 11 (1998).

H. Schift, J. SöchtigLIGA-technology for the fabrication of positioned planar structuresMicrosystem Technologies 4, 132-134 (1998).

H. Schift, J. Söchtig, M. Rossi, S. WestenhöferReplicated high precison micro-optical insert elements for optical bench arrangementsSPIE Proc. 3513, 122-134 (1998).

M. Schildenberger, Y. Bonetti, M. Aeschlimann, L. Scandella, J. Gobrecht, R. PrinsPreparation of model catalysts by laser interference nanolithography followed by metal cluster depositionCatalysis Letters 56, 1-6 (1998).

196

RADIO- AND ENVIRONMENTAL CHEMISTRY

HEAVY ELEMENTS

W. Brüchle, E. Jäger, V. Pershina, M. Schädel, B. Schausten, R. Günther, J.V. Kratz, W. Paulus,A. Seibert, P. Thörle, S. Zauner,, D. Schumann, B. Eichler, H.W. Gäggeler, D.T. Jost, A. TürlerChromatographie studies of Rf (element 104) with tributylphosphate (TBP)J. Alloys Comp. 271-273, 300 (1998).

B. Eichler, S. Hübener, N. Erdmann, K. Eberhardt, H. Funk, G. Herrmann, S. Köhler, N. Trautmann,G. Passler, F.-J. UrbanAn atomic beam source for actiniae elements: Concept and realizationRadiochim. Acta79, 221-223 (1997).

H.W. GäggelerChemistry gains a new element: Z= 106J. Alloys Comp. 271-273, 277-282 (1998).

R. Günther, W. Paulus, J.V. Kratz, A. Seibert, P. Thörle, S. Zauner, W. Brüchle, E. Jäger, V. Pershina,M. Schädel, B. Schausten, D. Schumann, B. Eichler, H.W. Gäggeler, D.T. Jost, A. TürlerChromatographie study of rutherfordium (element 104) in the system HCI/tributylphosphate (TBP)Radiochim. Acta 80,121 (1998).

D. Schumann, H. Nitsche, St. Taut, D.T. Jost, H.W. Gäggeler, A.B. Yakushev, G.V. Buklanov, V.P. Domanov,Din Thi Lien, B. Kubica, R. Misiak, Z. SzeglowskiSorption behaviour of rutherfordium and thorium from HCI/Hf containing aqueous solutionJ. Alloys Comp. 271-273, 307-311 (1998).

S. Taut, S. Hübener, B. Eichler, A. Türler, H.W. Gäggeler, S.N. Timokhin, I. ZvaraThermochromatography of heavy actinides adsorption of No-259 on Ti, V, Nb, Ta and MoJ. Alloys Comp. 271-273, 316 (1998).

S. Taut, S. Hübener, B. Eichler, H.W. Gäggeler, M. Schädel, I. ZvaraThermochromatography of heavy actinides - determination of the sublimation enthalpy of EsRadiochim. Acta 78, 33-38 (1998).

A. Türler, R. Dressler, B. Eichler, H.W. Gäggeler, D.T. Jost, M. Schädel, W. Brüchle, K.E. Gregorich,N. Trautmann, S. TautDecay properties of265Sg (Z=106) and!66Sg (Z=106)Phys. Rev. C57, 1648(1998).

A. Türler, G.V. Buklanov, B. Eichler, H.W. Gäggeler, M. Grantz, S. Hübener, D.T. Jost, V.Ya. Lebedev,D. Piguet, S.N. Timokhin, A.B. Yakushev, I. ZvaraEvidence for relativistic effects in the chemistry of element 104J. Alloys Comp. 271-273, 287 (1998).

A. Weidenkaff, A. Steinfeld, A. Wokaun, P.O. Auer, B. Eichler, R. RelierDirect solar thermal dissociation of zinc oxide: Condensation and crystallisation of zinc in the presence ofoxygenSolar Energy 65, 1, 69-69 (1999).

SURFACE CHEMISTRY, AEROSOL CHEMISTRY, ANALYTICAL CHEMISTRY

M. Ammann, M. Kalberer, D.T. Jost, L. Tobler, E. Rössler, D. Piguet, H.W. Gäggeler, U. BaltenspergerHeterogeneous production of nitrous acid on soot in polluted air massesNature 395, 157-160 (1998).

U. Baltensperger, S. NyekiAtmospheric aerosolsin: Physical and Chemical Properties of Aerosols, I. Colbeck (ed.), Blackie Academic & Professional, London,1998, pp. 280-330.

197

U. Baltensperger, M. Schwikowski, D.T. Jost, H.W. Gaggeler, O. PoulidaScavenging of atmospheric constituents in mixed phase clouds at the high-alpine site Jungfraujoch; Part I:Basic concept and cloud scavengingAtmos. Environ. 32, 3975-3983 (1998).

A.M. Hering, J. Staehelin, U. Baltensperger, A.S.H. Prevot, G.L Kok, R.D. Schillawski, A. WaldvogelAirborne measurements of atmospheric aerosol particles and trace gases during photosmog episodes overthe Swiss Plateau and the Southern Pre-Alpine RegionAtmos. Environ. 32, 3381-3392 (1998).

M. Lugauer, U. Baltensperger, M. Furger, H.W. Gaggeler, D.T. Jost, M. Schwikowski, H. WannerAerosol transport to the high alpine sites Jungfraujoch (3454 m) and Colle Gnifetti (4453 m)Tellus50B, 76-92(1998).

S. Nyeki, U. Baltensperger, I. Colbeck, D.T. Jost, E. Weingartner, H.W. GaggelerThe Jungfraujoch high-alpine research station (3454 m) as a background continental site for the measurementof aerosol parametersJ. Geophys. Res. 103, 6097-6107 (1998).

S. Nyeki, F. Li, E. Weingartner, N. Streit, I. Colbeck, H.W. Gaggeler, U. BaltenspergerThe background aerosol size distribution in the free troposphere: an analysis of the seasonal cycle at a high-alpine siteJ. Geophys. Res. 103, 31749-31762 (1998).

O. Poulida, M. Schwikowski, U. Baltensperger, J. Staehelin, H.W. GaggelerScavenging of atmospheric constituents in mixed phase clouds at the high-alpine site Jungfraujoch; Part II:Influence of riming on the scavenging of particulate and gaseous chemical speciesAtmos. Environ. 32, 3985-4000 (1998).

U. Schotterer, P. Schwarz, V. RajnerFrom the prebomb levels to industrial times. A complete tritium record from an alpine ice core and itsrelevance for environmental studiesIn Isotope Techniques in the Study of Environmental Change, Proc. of an International Symposium on IsotopeTechniques in the Study of Past and Current Environmental Changes in the Hydrosphere and theAtmosphere, Vienna, 14-18 April 1997, IAEA, 1998.

M. Schwikowski, U. Baltensperger, H.W. Gaggeler, O. PoulidaScavenging of atmospheric constituents in mixed phase clouds at the high-alpine site Jungfraujoch; Part III:Quantification of the removal of chemical speciesAtmos. Environ. 32, 4001-1010 (1998).

P. Seibert, H. Kromp-Kolb, A. Kasper, M. Kalina, H. Puxbaum, D.T. Jost, M. Schwikowski, U. BaltenspergerTransport of polluted boundary layer air from the Po valley to high-alpine sitesAtmos. Environ. 32, 3953-3965 (1998).

G. Skillas, S. Kiinzel, H. Burtscher, U. Baltensperger, K. SiegmannHigh fractal-like dimension of diesel soot agglomeratesJ. Aerosol Sci.29, 411-419 (1998).

W. Tinner, M. Conedera, B. Ammann, H.W. Gaggeler, S. Gedye, R. Jones, B. SagesserPollen and charcoal in lake sediments compared with historically documented forest fires in southernSwitzerland since AD 1920The Holocene 8, 1, 31-42 (1998).

H.R. von GuntenRadium-226 and alkaline earth elements in the water column, interstitial sediment water, and sediments ofZurichsee (Switzerland)Croatica Chemica Acta 71 (2) 391 -404 (1998).

198

A. Wyttenbach, L. ToblerEffect of surface contamination on results of plant analysisCommun. Soil Sci. Plant Anal., 29, 809-823 (1998).

A. Wyttenbach, V. Furrer, P. Schleppi, L. ToblerRare earth elements in soil and in soil-grown plantsPlant and Soil, 199, 267-273, 1998.

A. Wyttenbach, L. Tobier, P. Schleppi, V. FurrerVariation of the rare earth element concentrations in the soil, soil extract and in individual plants from thesame site.J. Radioanal. Nucl. Chem.,231, 101-106, 1998.

PROJECT RADWASTE

I. Guenther, L Wyer, E.J. Knust, R.D. Finn, J. Koziorowski, R. WeinreichRadiosynthesis and quality assurance of 5-(z"l]lodo-2'-deoxyuridine for unctional PET imaging ofcell proliferationNucl. Med. Biol. 25, 359-365 (1998).

J. Koziorowski, C. Henssen, R. WeinreichA new convenient route to radioiodinated N-succinimidyl 3- and 4-iodobenzoate, zwo reagents forradioiodination of proteinsAppl. Radiat. Isot. 49, 955-959 (1998).

R. Weinreich, L. Wyer, N. Crompton, M.C. Nievergelt-Egido, I. Guenther, U. Roelcke, K.L. Leenders,E.J. Knust, R.G. Blasberg1-124 and its applications in nuclear medicine and biologyReport IAEA-TECDOC-1029, Int. Atomic Energy Agency, Wien 1998, 399-418.

M. Argentini, D.F. Dos Santos, R. Weinreich, H.-J. HansenSynthesis of an o-carboranyl derivative of4-[5-(4-Methyl-1-piperazinyl)-2,5'-bi-1H-benzimidazol-2'-yl]phenolInorganic Chemistry 37, 6018-6022 (1998).

ION BEAM PHYSICS

R. F. Anderson, N. Kumar, R. A. Mortlock, P. N. Froelich, P. Kubik, B. Dittrich-Hannen, M. SuterLate-Quaternary changes in productivity of the Southern OceanJ. Marine Systems 17 no. 1-4, 497-514 (1998).

S. Baumgartner, J. Beer, J. Masarik, G. Wagner, L. Meynadier, H.-A. SynalGeomagnetic modulation of the 36CI flux in the GRIP ice core, GreenlandScience 279, 1330-1332 (1998).

W. Beck, D. J. Donahue, A. T. J. Jull, G. S. Burr, W. S. Broecker, G. Bonani, I. Hajdas, E. MalotkiAmbiguities in direct dating of rock surfaces using radiocarbon measurementsScience 280, 2132-2135 (1998).

A. Bischoff, D. Weber, R. N. Clayton, T. Faestermann, I. A. Franchi, U. Herpers, K. Knie, G. Korschinek,P. W. Kubik, T. K. Mayeda, S. Merchel, R. Michel, S. Neumann, H. Palme, C. T. Pillinger, L Schultz,A. S. Sexton, B. Spettel, A. B. Verchovsky, H. W. Weber, G. Weckwerth, D. WolfPetrology, chemistry, and isotopic compositions of the lunar highland regolith breccia Dar al Gani 262Meteoritics & Planet. Sci. 33,1243-1257 (1998).

W. S. Broecker, D. Peteet, I. Hajdas, J. Lin, E. ClarkAntiphasing between rainfall in Africa's Rift Valley and North America's Great BasinQuaternary Research 50,12-20 (1998).

199

P. Brims, W.-C. Dullo, W. W. Hay, M. Frank, P. KubikHiatuses on VBring Plateau: sedimentary gaps or preservation artifacts?Marine Geology 145, 61-84 (1998).

M. Dobeli, F. Ames, C. R. Musil, L. Scandella, M. Suter, H.-A. SynalSurface tracks by MeV Cm impacts on mica and PMMANucl. Instr. and Meth. B 143, 503-512 (1998).

M. Dobeli, R. M. Ender, V. Liechtenstein and D. VetterliTime-of-flight spectrometry applied to 2 MeV He RBSNucl. Instr. and Meth. B 142, 417-424 (1998).

D. Fink, R. Klett, W. H. Chung, R. Grunwald, M. Dobeli, F. Ames, L. T. ChaddertonDoping of Cn * (n=1,3,5,8) cluster ion tracks in polyimideRadiat. Eff. Def. Solids 140, 3-20 (1998).

D. Fink, M. Muller, P. Szimkowiak, M. Dobeli, A. Spieler, M. Grether, A. Richter, L. T. Chadderton, F. Zawislak,P. F. P. Grande, M. Behar, V. Hnatowicz, J. VacikHigh fluence ion irradiation of thin fullerite filmsFullerene Sci. and Techn. 911, 6 (1998).

I. Hajdas, G. Bonani, P. Boden, D. M. Peteet, D. H. MannCold reversal on Kodiak Island, Alaska, correlated with the European Younger Dryas by using variations ofatmospheric 14C contentGeology 26 no 11, 1047-1050 (1998).

I. Hajdas, G. Bonani, B. Zolitschka, J. NegendankUC ages of terrestrial macrofossils from Lago Grande di Monticchio (Italy)Radiocarbon 40, no. 2, 803-807 (1998).

S. R. Hemming, P. E. Biscaye, W. S. Broecker, N. G. Hemming, M. Klas, I. HajdasProvenance change coupled with increased clay flux during deglacial times in the western equatorial AtlanticPalaeogeography, Palaeoclimatology, Palaeoecology142, 217-230 (1998).

S. Ivy-Ochs, H. Heuberger, P. W. Kubik, H. Kerschner, G. Bonani, M. Frank, C. SchluchterThe age of the Kofels event. Relative, UC and cosmogenic isotope dating of an Early-Holocene landslide inthe central Alps (Tyrol, Austria)Zeitschrift f. Gletscherkunde und Glazialgeologie34,1, 57-68 (1998).

S. Ivy-Ochs, P. W. Kubik, J. Mazarik, R. Wieler, L. Bruno, C. SchluchterPreliminary results on the use of pyroxene for10Be surface exposure datingSchweiz. Mineral. Petrogr. Mitt. 78, 375-382 (1998).

U. Krahenbiihl, K. Noll, M. Dobeli, D. Grambole, F. Herrmann, L. ToblerExposure of Allan Hills 84001 and other achondrites on the Antarctic iceMeteoritics and Planet. Sci. 33, 665-670 (1998).

M. Krejci, A. N. Tiwari, H. Zogg, M. Dobeli, V. Zelezny, J. Oswald, V. VorlicekEpitaxial CulnxSey layers of improved crystal quality and determination of the crystal orientation withRaman SpectroscopyInst. Phys. Conf. Ser. No. 152, 377-380 (1998).

P. W. Kubik, S. Ivy-Ochs, J. Masarik, M. Frank, C. Schluchter10Be and 26AI production rates deduced from an instantaneous event within the dendro-calibration curve,the landslide ofKofels, Otz Valley, AustriaEarth and Planet. Sci. Lett. 161, 231-241 (1998).

200

J. C. Lin, W. S. Broecker, S. R. Hemming, I. Hajdas, R. F. Anderson, G. I. Smith, M. Kelley, G. BonaniA reassessment of U-Th and C-14 ages for late-glacial high-frequency hydrological events at Searles Lake,CaliforniaQuaternary Research 49, 11-23 (1998).

F. Machalett, W. Wesch, R. Muhle, S. BarthInvestigations on liquid alloy ion sources for rare earth elementsReview of Scientific Instruments 69, no. 3, 1336-1339 (1998).

A. Mangini, M. Lomitschka, R. Eichstaedter, N. Frank, S. Vogler, G. Bonani, I. Hajdas, J. PetzoldCoral provides way to age deep waterNature 392, 347-348 (1998).

R. Menzel, T. Bachmann, F. Machaiett, W. Wesch, U. Lang, M. Wendt, C. Musil, R. MuhleSurface smoothing and patterning ofSiC by focused ion beamsAppl. Surf. Sci. 136, no. 1 and 2, 1-7 (1998).

K. Noll, M. Dobeli, U. KrahenbCihlFluorine profiles in Antarctic meteorites by Nuclear Reaction AnalysisFresenius J. Anal. Chem. 361, 713-715 (1998).

C. Rosenblad, H. R. Deller, M. Dobeli, E. Muller, H. v. KanelLow-temperature heteroepitaxy by LEPECVDThin Solid Films318, 11-14(1998).

A. Schmidt, C. Schnabel, J. Handl, D. Jakob, R. Michel, H.-A. Synal, J. M. Lopez, M. SuterOn the analysis of iodine-129 and iodine-127 in environmental materials by accelerator mass spectrometryand ion chromatographyThe Science of the Total Environment 223, 131-156 (1998).

M. SuterA new generation of small facilities for accelerator mass spectrometryNucl. Instr. and Meth. B 139, 150-157 (1998).

201

CONTRIBUTIONS TO CONFERENCES AND WORKSHOPS

ASTROPHYSICS

M. Audard, M. Gudel, E.F. GuinanThe EUV corona of the young solar analogs 47 Cas and EK Dra: An EUVE spectroscopic and timing studyof coronal heatingProc. of the First XMM Workshop, Ed. M. Dahlem (1998), Noordwijk, The Netherlands, Sept. 30 - Oct. 2,1998.

A. O. Benz, M. Gudel, J. ConwayFirst image of the corona of a main-sequence starTenth Workshop on Cool Stars, Stellar Systems, and the Sun. Eds. R. Donahue and J. A. Bookbinder(San Francisco: Astronomical Society of the Pacific), 1942 (1998).

A. Brinkman, H. Aarts, A. den Boggende, T. Bootsma, L. Dubbeldam, J. W. den Herder, J. Kaastra,P. de Korte, P. van Leeuwen, R. Mewe, F. Paerels, C. de Vries, J. Cottam, T. Decker, S. Kahn,A. Rasmussen, J. Spodek, G. Branduardi-Raymont, P. Cuttridge, K. Thomsen, A. Zehnder, M. GudelThe reflection grating spectrometer on board XMMProc. of the First XMM Workshop, Ed. M. Dahlem (1998), Noordwijk, The Netherlands, Sept. 30 - Oct. 2, 1998.

M. Gagne, J. Valenti, C. Johns-Krull, J.L Linsky, A. Brown, M. GudelCoordinated VLA, optical, EUVE, and RXTE monitoring of flares on EQ Pegasi and AU MicroscopiiTenth Workshop on Cool Stars, Stellar Systems, and the Sun. Eds. R. Donahue and J. A. Bookbinder(San Francisco: Astronomical Society of the Pacific), 1484 (1998).

M. GudelAstrophysics at PSIHESSI Meeting ETH Zurich, 22 March, 1998.

M. Gudel, E. F. Guinan, S.L SkinnerThe X-ray and radio Sun in time: Coronal evolution of solar-type stars with different agesTenth Workshop on Cool Stars, Stellar Systems, and the Sun. Eds. R. Donahue and J. A. Bookbinder(San Francisco: Astronomical Society of the Pacific), 1041 (1998).

M. Gudel, E. F. Guinan, P.B. Etzel, R. Mewe, J.S. Kaastra, S.L. SkinnerAssembling the pieces of the puzzle: A nearby, rapidly rotating young Sun in 47 Cas?Tenth Workshop on Cool Stars, Stellar Systems, and the Sun. Eds. R. Donahue and J.A. Bookbinder(San Francisco: Astronomical Society of the Pacific), 1247 (1998).

M. Gudel, R. MeweX-ray spectroscopy diagnostics with XMM andAXAF: Prospects and challengesTenth Workshop on Cool Stars, Stellar Systems, and the Sun. Eds. R. Donahue and J. A. Bookbinder(San Francisco: Astronomical Society of the Pacific), 1051 (1998).

M. GudelProspects for cool star astronomy with the Square Kilometer arrayInvited Review, International Square Kilometre Array Science Meeting, Calgary, Canada, 19-22 July, 1998.

M. Gudel, R. Mewe, A.C. Brinkman, J.S. Kaastra, F. Paerels, S. Kahn, A. Rasmussen, L Harra-MurnionX-ray spectroscopy diagnostics with XMM: New prospects for the physics of thermal, optically thin plasmasProc. of the First XMM Workshop, Ed. M. Dahlem (1998), Noordwijk, The Netherlands, Sept. 30 - Oct. 2, 1998.

Ph. JetzerMicrolensing and the macho massSeminar, 4th Workshop on microlensing in Paris, 16.1.1998.

202

Ph. JetzerLenti gravitazionali e materia oscuraKolloquium, Universitat "LaSapienza" in Rom, 10.2.1998.

Ph. JetzerGround observations from the solar telescope in LocarnoHESSI meeting ETH Zurich, 22.3.1998.

Ph. JetzerGravitational microlensing by globular clustersSeminar Universitat Neapel, 12.5.1998.

Ph. JetzerGamma rays from the galactic haloSeminar CAPP-98 Meeting CERN, 8.6.1998.

Ph. JetzerMikrolinsenereignisse und Dunkle MaterieKolloquium Universitat Potsdam, 7.7.1998.

Ph. JetzerMateria oscura e radiazione gammaSeminar Universitat Neapel, 10.9.1998.

Ph. JetzerGravitational microlensing and dark matter in the galactic haloSeminar Euroconference 3K Cosmology Rom, 7.10.1998.

Ph. JetzerCold molecular clouds and machos as halo dark matterTexas Symposium on Relativistic Astrophysics, Paris, France, 14-18 December, 1998.

Ph. LerchApplication de la supraconductivite a la detection de rayonnement XInstitut de Microsysteme, Ecole Polytechnique Federale de Lausanne, 20 January, 1998.

Ph. LerchApplication de la supraconductivite a la detection de rayonnement XDepartement de Physique de la Matiere Condensee, Universite de Geneve, 19 May,1998.

R. Mewe, G. Rauw, K.A. van der Hucht, J.-M. Vreux, E. Gosset, M. Gudel, W. SchmutzX-ray spectroscopy with XMM: A new powerful tool to determine fundamental parameters of early-type starsProc. of the First XMM Workshop, Ed. M. Dahlem (1998), Noordwijk, The Netherlands, Sept. 30 - Oct. 2,1998.

F. Paerels, A. Brinkman, J. W. den Herder, C.de Vries, A. den Boggende, R. Mewe, J.S. Kaastra, S. Kahn,A. Rasmussen, T. Decker, M. Stern, J. Cottam, J. Spodek, G. Branduardi-Raymont, M. Gudel, D.A. Liedahl,C. ErdHigh-resolution spectroscopy with XMMProc. of the First XMM Workshop, Ed. M. Dahlem (1998), Noordwijk, The Netherlands, Sept. 30 - Oct. 2,1998.

D. PuyPrimordial chemistry in early UniverseSeminar, Institut fur Astronomie der ETH, 20 January 1998

D. PuyMolecules aux grands redshiftsSeminar, Observatoire de Geneve, 17 March 1998.

203

D. PuyChimie Primordiale, Nuages moleculaires primordiauxSeminar, Observatoire de Marseille (France), 7 May 1998.

D. PuyMolecules, Nuages moleculaires primordiauxSeminar, Observatoire de Toulouse (France), 4 June 1998.

D. Puy, M. SignorePrimordial molecular cloudsInvited talk, 3K Cosmology Euroconference, Roma (Italy), 5-10 October 1998.

D. Puy, M. SignoreMolecules in a gravitational collapse3rd Cologne-Zermatt Symposium, Zermatt, 22-25 September 1998.

D. Puy, M.SignoreMolecules at high redshiftsESA Conference, UNESCO-Paris (France), 20-23 October 1998.

D. Puy, M. SignoreCold molecular clouds and dark matterParis (France), 14-18 December 1998.

G. Rauw, K.A. van der Hucht, R. Mewe, M. Gudel, J.-M. Vreux, E. Gosset, W. Schmutz, I.R. StevensOn the perspectives of using XMM to study fundamental parameters of early-type starsWolf-Rayet Phenomena in Massive Stars and Starburst Galaxies. IAU Symposium 193, Puerto Vallarta, Mexico,3-7 November, 1998.

MICRO AND NANO TECHNOLOGY

Ch. DavidE-beam exposure of optical components using a defocused beamEOS Topical Meeting on Optical Spectral Analysis Microsystems, Engelberg, 6-8 April, 1998.

Ch. DavidLine width control in low voltage e-beam lithographyInternational Workshop on Nanostructures at Surfaces and Interfaces, Ascona, 19-24 April, 1998.

Ch. DavidLine width control in low voltage e-beam lithography using a defocused beamGordon Research Conference on the Chemistry and Physics of Nanostructure Fabrication, Tilton School,NH, USA, 21-26 June, 1998.

Ch. DavidPre-structuring of silicon substrates to investigate MBE-growth ofSiGe layersInternational Conference on Micro- and Nano-Engineering, Leuven, Belgien, 22-24 September, 1998.

Ch. DavidLine width control using a defocused low voltage electron beamInternational Conference on Micro- and Nano-Engineering, Leuven, Belgien, 22-24 September, 1998.

J. H. Elgamel, J. GobrechtHigh efficiency solar cells on sheet ribbon grown silicon materialNationale Photovoltaik-Tagung, Bern, 5 May, 1998.

204

H. Elgamel, J. GobrechtImproving the quality of polycrystalline silicon string ribbon for fabricating high efficiency solar cellsPOLYSE 98, Schwabisch Gmund, September 1998.

S. Gehrsitz, H. Sigg, N. Herres, K. Bachem, K. Kohler, F.K. ReinhartThe elastic constants, the lattice parameter, and the composition ofAIGaAs24th International Conference Physics of Semiconductor (ICPS 24), Jerusalem, 2-7 August 1998.

J. GobrechtMikro- und Nanotechnik: Von der Wissenschaft zur praktischen AnwendungGemeinsame Veranstaltung des PSI und der Aarg. Kantonalbank, PSI, 5 May, 1998.

J. GobrechtWas ist Nanotechnologie?Vortragsreihe der Swiss Engineering, FH Aargau, Windisch, 22 October, 1998.

J. GobrechtMikrosysteme fur die Nanotechnik - ausgewahlte BeispieleSeminar Inst. F. Mikrotechnik, Albert-Ludwigs-Universitat, Freiburg, 16 December, 1998.

S. Graf, H. C. Sigg, W. BachtoldFar infrared optical rectification in asymmetrically doped GaAs/AIGaAs quantum wellsSpring Meeting of the Swiss Physical Society, Bern, 26-27 February, 1998.

S. Graf, H. Sigg, K. Kohler, W. BachtoldHigh Frequency Optical Rectification in Bulk GaAs and Asymmetric AIGaAs/GaAs Quantum Wells24th International Conference Physics of Semiconductor (ICPS 24), Jerusalem, 2-7 August, 1998.

A. Grubelnik, C. Padeste, L. TiefenauerAmperometric immunosensing of penicillin in milkBiosensors '98, Berlin 1998.

D. GrutzmacherNanostrukturen im Materialsystem Si-Ge-CSeminarvortrag, Universitat Wiirzburg, Wilrzburg, Germany, 12 February, 1998.

D. GrutzmacherPhotolumineszenz an Quantenstrukturen im Materialsystem Si-Ge-CSeminarvortrag, Max Planck Institut fur Festkorperphysik, Stuttgart, Germany, 17 March, 1998.

D. Grutzmacher, C. David, R. Hartmann, E. Muller, U. GennserMesa design for dislocation density reduction of relaxed SiGe buffer40th Electronic Materials Conference, Charlottesville, Virginia, USA, 22-25 June, 1998.

D. Grutzrnacher, J.A.D. Anna Selvan, R. Morf, J. GobrechtActivities of the PSI within the project ,,Crystal"Mid-Term Assessment Report, Paris, France, 23 September, 1998.

D. GrutzmacherSiGeC Nanostrukturen: Jenseits des Si HorizontesSeminarvortrag, Technische Universitat Dresden, Dresden, Germany, 10 Dezember, 1998.

U. GennserExperiences sur des structures de bandes dans des heterostructures IV-IV: Une approche "hors piste"par rapport au Si?Seminar Institut d'Electronique Fondamentale, Universite Paris-Sud, Orsay, France, 6 November, 1998.

R. Hartmann, U. Gennser, H. Sigg, D. Grutzmacher, G. DehlingerSi/SiGeC heterostructures: A path towards high-mobility channelsAdvanced Research Workshop, Future Trends in Microelectronics: 'Off the Beaten Path', 1998.

205

R. Hartmann,D. Grützmacher, U. GennserPhotolumineszenz an verspannungsreduzierten Si/SiGeC-Heterostrukturen62. Frühjahrstagung der Deutschen Physikalischen Gesellschaft, Regensburg, Germany, 23-27 March, 1998.

R. Hartmann, U. Gennser, D. Grützmacher, H. Sigg, E. Müller, K. EnsslinPhotoluminescence in strain compensated Si/SiGe Multiple Quantum WellsProc. Materials Research Society Spring Meeting, San Francisco, CA, USA, 12-17 April, 1998, ed. E. Fitzgerald,D. Houghton, P. Mooney, Vol 533, pp. 251-256.

R. Hartmann, O.Leifeld, U. Gennser, C. David, E. Müller, H. Sigg, H. Siegwart, T. Mezzacasa, D. Bächle,D. GrützmacherFabrication and characterization of nanostructures in the Si-Ge-C material systemSeminars at: University of California Los Angeles, Los Angeles, CA, USA, 29 June, 1998; Hughes ResearchLaboratories, Malibu, Ca, USA, 30 June, 1998; IBM Hopewell Junction Science Park, East Fishkill, NY, USA,2 July, 1998.

R. Hartmann, H. Sigg, C. David, E. Müller, C. Panitz, U. Gennser, D. GrützmacherMesa design for dislocation density reduction of relaxed SiGeMBE Workshop '98, Hamburg, Germany, 21-22 September, 1998.

R.W. Jaszewski, H. SchiftHot embossing lithography as a tool to replicate nanostructuresSeminar Institute of Inorganic Chemistry (IOC), Univ. of Bern, Switzerland, 24 June, 1998.

R.W. Jaszewski, H. Schift, C. David, J. GobrechtNanostructuring and fabrication by hot embossing lithography3rd Workshop on Nanoscience, Hasliberg, Switzerland, 12-16 October, 1998.

T.A. Jung, R.R. Schüttler, J.K. GimzewskiSingle molecule mechanics and electronicsSeminar at the University of California, Santa Barbara, USA, 7 January, 1998.

T.A. Jung, F.J. HimpselMetallic nanowires: Model systems for chemical identification and low dimensional electronic systemsSeminar at Dartmouth College, Hanover NH, USA, 15 January, 1998.

T.A. Jung, J.K. Gimzewski, R.R. SchüttlerKonzepte für molekulare Schalter: Adressierbare Bewegung und Konformationsänderungen individuellerMoleküle an OberfllächenSeminar Max Planck Institute für Polymerforschung, Mainz, Germany, 10 February, 1998.

T.A. Jung, J.K. Gimzewski, R.R. SchüttlerMolecular nanotechnology: A brief introductionJoint Seminar of the Physical Chemistry Dept, ETH Zürich and the Max Planck Institut für Polymerforschung,Mainz; Davos, Switzerland, 23-27 March, 1998.

T.A. Jung, R.R. Schüttler, J.K. GimzewskiMolecular Assembly at InterfacesNanostructures at Surfaces and Interfaces Workshop, organized by the Swiss Federal Institutes of Technology;Monte Verita, Ascona, Switzerland, 19-24 April, 1998.

O. Leifeld, B. Müller, D. Grützmacher, K. KernGe-Quantum dot formation driven by sub-monolayer C-coverages on Si(001)Discussions Meeting 'Nanostructures at Surfaces and Interfaces', Ascona, CH, 19-24 April, 1998.

T.A. Jung, J.K. Gimzewski, R.R. SchüttlerLots of nanoscience with increasing relevance for technology: Molecular nanotechnology?Gordon Research Conference on the Chemistry and Physics of Nanostructure Fabrication, Tilton College, NH,USA; 21-26 June, 1998.

206

T.A. Jung, H. Rauscher, J.K. Gimzewski, F.J. HimpselMolecular layers, wires and dots built by self assembly and step decoration from 2 dimensional and1 dimensional interfacesUniversity of Madison, Madison Wisconsin, 21 August, 1998.

T.A. Jung, H. Rauscher, J.K. Gimzewski, F.J. HimpselAligning molecules at 2 dimensional and 1 dimensional interfaces: How to build molecular wires and dots,and what they could be used forNational Institute of Standards and Technology (NIST), 28 August, 1998.

T.A.JungKonzepte fur molekulare Schalter: Adressierbare Bewegung und Konformations-Aenderungen einzelnerMolekule an GrenzflachenUniversitaet Zurich, Institut fur Physikalische Chemie, 9 November, 1998.

T. A. Jung H.Rauscher, J.Gimzewski, F.J. HimpselMolecular positioning and molecular bi-stability at interfaces: What we already know about futurenanoscale devicesUniversity of Basel, Condensed Matter, 16 November, 1998.

T.A. Jung, H. Rauscher, J. Gimzewski, F.J. HimpselAdressing molecular motion and conformational changes: Real space experiments with individual moleculesadsorbed at surfaces, steps and nanoscale wiresETH-Zuerich, 19 November, 1998.

T.A. Jung, H.Rauscher, J. Gimzewski, F. J. HimpselMolecular nanoscience: Building molecular structures from individual units and understanding theirproperties and functions.Hans Sigrist Symposium, University of Bern, 4 December, 1998.

O. Leifeld, D. Griitzmacher, B. Muller, K. KernMassgeschneiderte Germanium-Quantenpunkte durch Nukleation aufC/Si(001)62. Fruhjahrstagung der Deutschen Physikalischen Gesellschaft, Regensburg, Germany, 23-27 March, 1998.

O. Leifeld, D. Grutzmacher, B. Muller, K. KernGe-Quantum dots on Si(001) tailored by carbon predepositionProc. Materials Research Society Spring Meeting, San Francisco, CA, USA, 12-17 April, 1998, ed. E. Fitzgerald,D. Houghton, P. Mooney, Vol 533, pp. 183-189.

0. Leifeld, R. Hartmann, E. Muller, B. Muller, K. Kern, and D. GrutzmacherSelf organized growth of Ge quantum dots on Si(001) substrates induced by sub-monolayerC coveragesIEEE Nanoelectronics workshop1998, Honolulu, HI, USA, 7-8 June, 1998.

O. Leifeld, E. Kaxiras, D. Grutzmacher, K. KernSTM study of Ge dot formation on the C-alloyed Si(001) surface3rd Conference on Developement and Technological Application of Scanning Probe Methods, SXM3,Basel, CH, 14-17 September, 1998.

O. Leifeld, E. Kaxiras, D. Grutzmacher, K. KernSTM study of Ge dot formation on the C-alloyed Si(001) surface3rd Hasliberg Workshop on Nanoscience, Hasliberg, CH, 12-19 October, 1998.

E. Mueller, E. Ribeiro, T. Heinzel, K. EnsslinTEM Study of the formation of InAs self-assembled quantum dots in GaAsMaterials Research Society Spring Meeting (E-MRS 98), Symposium D, Strasbourg, France, 16-19 June, 1998.

207

E. Muller, R. Hartmann, C. David, D. GrutzmacherStructural characterization of SiGe step graded buffer layers grown on pre-structured Si[001] substratesby molecular beam epitaxyMaterials Research Society Spring Meeting (E-MRS 98), Symposium D.Strasbourg, France, 16-19 June, 1998.

R. Morf, J. GobrechtOptimized structures for light trapping10lh Workshop on Quantum Solar Energy Conversion, Bad Hofgastein, Austria, 8-14 March, 1998.

S. Nowak, Ch. Kurtsiefer, C. David, T. Pfau, J. MlynekHigh-order Talbot fringes for atomic matter wavesProc. of the International Quantum Optics Conference '98, 4-7 May, San Francisco (1998).

R. Ros, F. Schwesinger, C. Padeste, A. Pluckthun, H.-J. Giintherodt, L. TiefenauerForce spectroscopy on individual single-chain Fv antibody moleculesSXM 1998 & Workshop on Nanoscience, ,,Hasliberg3", 12-16 October, 1998.

H. Schift, J. SochtigLIGA-technology for the fabrication of high precision micro-optical elementsEOS Topical meeting (16th EOS/TM) on optical spectral analysis microsystems, Engelberg, Switzerland,6-8 April, 1998.

H. SchiftHerstellung von Mikro- und Nanostrukturen durch ReplikationSeminar on "Mikrostrukturen und Mikroprazisionsspritzguss als Chance fur die Zukunft" of the SchweizerTechnischer Verband (STV), Aarau, Switzerland, 16 June, 1998.

H. Schift, J. Sochtig, M. Rossi, S. WestenhoferReplicated high precison micro-optical insert elements for optical bench arrangementsSPIE Conf. on Micromachining and Microfabrication (MF'98), Santa Clara, CA, USA, 21-22 September, 1998.

H. Schift, R.W. Jaszewski, C. David, J. GobrechtNanostructuring of polymers and fabrication of interdigitated electrodes by hot embossing lithographyMicro- and Nanoengineering (MNE'98), Bruxelles, Belgium, 22-24 September, 1998.

H. SchiftNanoreplication in polymers3rd Workshop on Nanoscience, Hasliberg, Switzerland, 12-16 October, 1998.

M. Sillasen, H. SiggAngular Momentum Photon Drag in 2 dimensional electron systemsWinterschool on "New Frontiers in Low Dimensional Physics" Mauterndorf, Austria, 23-27 February, 1998.

H. Sorribas, C. Padeste, U. Gennser, P. Sonderegger, L. TiefenauerMicrostructured biofunctionalized glass for neuron cell guidance and recordingBiosensors '98, Berlin 1998.

L. Tiefenauer, R. Ros Seigel, F. Schwesinger, A. Pluckthun, H.-J. GiintherodtSensing forces of individual ligand-receptor interactions by sanning probe microscopyBiosensors '98, Berlin 1998.

L. TiefenauerMolecular Nanotechnology: Vision, tools and projectsWorkshop on Nanoscience ,,Hasliberg 3", 12-16 October, 1998.

R. Zingg, R. Derrer, H.J. Grossmann, T.A. Jung, J.K. GimzewskiMolecular motion detection and tracerNanostructures at Surfaces and Interfaces Workshop, organized by the Swiss Federal Institutes of Technology;Monte Verita, Ascona, Switzerland, 19-24 April, 1998.

208

RADIO- AND ENVIRONMENTAL CHEMISTRY

HEAVY ELEMENTSB. Eichler, A. Jakushev, J. Adams, A. Türler, H.W. Gäggeler, J. PetersonStabilität und Flüchtigkeit der Chloride schwerer Aktinoide - Erster experimentelle Nachweis einesflüchtigen BkCI4Vortragstagung der GDCh-Fachgruppe Nuklearchemie, Dresden, 7-9 September, 1998

R. Dressler, B. Eichler, H.R. von Gunten, D.T. Jost, A. Türler, H.W. Gäggeler, M. Gärtner, W. Brüchle, M.Schädel, P. Thörle, N. Trautmann, S. ZaunerSynthese schwerster Elemente in (Hl, a-xn) RadktionenVortragstagung der GDCh-Fachgruppe Nuklearchemie, Dresden, 7-9 September, 1998.

R. DresslerChemie mit exotischen Elementen: Experimente zu gaschemischen Charakterisierung von SeaborgiumFachtagung Aachen, Aussenstelle Jülich, 25 November, 1998.

H.W. GäggelerEinweihung Neutronenaktivierungsanlage und Gas-Jet SINQPSI, 26 January, 1998.

H.W. GäggelerLaudationes auf Prof. J. V. Kratz und Dr. N. TrautmannOtto-Hahn Preisverleihung 1998, Kaisersaal des Römers, Frankfurt, 5 July, 1998.

H.W. GäggelerAnalysenmethoden mit NeutronenIndustrietag: Neutronen im Dienst der Industrie, PSI, 29 September, 1998.

H.W. Gäggeler, B. Eichler, A. TürlerAdsorption phenomena and transport of volatile radionuclidesOak Ridge National Laboratory, 2 November, 1998.

H.W. Gäggeler, B. Eichler, A. TürlerAdsorption phenomena and transport of volatile radionuclides15th Int. Conf. On the Application of Accelerators in Research and Industry, Denton, Texas, USA,4-7 November, 1998.

H.W. GäggelerSchwerelementforschung der PSI/Uni Bern Gruppe der letzten 3 JahreWorkshop Schwerelementkollaboration, Hinterzarten, Deutschland, 19-20 November 1998.

H.W. GäggelerChemische Untersuchungen exotischer Elemente mit einzelnen AtomenTU-Freiberg, Germany, 10 December, 1998.

D.T. JostReal-Time SymposiumMini-Symposium 19-20 March 1998, organizer.

D.T. JostZerfallseigenschaften von!6SSg(Z=106) und26eSg(Z=106)Vortragstagung der GDCh-Fachgruppe Nuklearchemie, Dresden, 7-9 September, 1998

A. TürlerDie Chemie exotischer Elemente: von Berkelium zu Seaborgium (invited talk)Zentrumsseminar "Aktuelle Fragen der Radiochemie", Forschungszentrum Rossendorf, 27.11.98.

209

A. TiirlerGas Phase Chemistry of the Transactinides13th Radiochemical Conference, Marianske Lazne, Czech Republic, April 1998.

A. Turler, R. Dressier, B. Eichler, H.W. Gaggeler, D.T. Jost, M. Schadel, W. Bruchle, K.E. Gregorich,N. Trautmann, S. TautDecay Properties of2B5Sg (Z=106) and266Sg (2=106)International Nuclear Physics Conference 98, Paris,. France, August 1998.

A. TurlerChemie des Seaborgium (Element 106), eine experimentelle Bestimmung seiner Fiuchtigkeit alsOxychloridVortragstagung der GDCh-Fachgruppe Nuklearchemie, Dresden, Germany, 7-9 September, 1998,

A. Turler, R. Dressier, Ch. Dullmann, B. Eichler, R. Eichler, H.W. Gaggeler, M. Gartner, D.T. Jost,D. Piguet, P. Thorle, N. Trautmann, A. Yakushev, S. ZaunerSynthese von Dubnium (Element 105) in der Reaktion 24sCm(9F, 5n) zszDb; eine Alternative zur249Bk(aO, SnfDb Reaktion?

Vortragstagung der GDCh-Fachgruppe Nukiearchemie, Dresden, Germany, 7-9 September, 1998.

SURFACE CHEMISTRY, AEROSOL CHEMISTRY, ANALYTICAL CHEMISTRYM. Ammann, M, Kalberer, F. Arens, E. Rossler, L. Tobler, D. Piguet, D.T. Jost, H.W. Gaggeler,U. BaltenspergerKinetics of HONO formation on sootEUROTRAC Symposium '98 Garmisch-Partenkirchen, Germany, 23-27 March 1998.

M. Ammann, M. Wachsmuth, U. Baltensperger, D.T. Jost, H.W. GaggelerA new laboratory approach to heterogeneous halogen chemistryEuropean Geophysical Society General Assembly, Nice, France, 20-24 April 1998.

F. Arens, M. Ammann, M. Kalberer, J. Kleffmann, U. Baltensperger, H.W. GaggelerHONO formation on sulphuric acid coated soot aerosol particlesSecond Workshop of EUROTRAC-2 Subproject CMD, Karlsruhe, Germany, 23-25 September 1998.

M. Ammann, M. Kalberer, F. Arens, V. Lavanchy, H.W. Gaggeler, U. BaltenspergerNitrous acid formation on soot particles: surface chemistry and the effect of humidityInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sci. 29, Suppl. 1,S1031-S1032 (1998).

M. AmmannHONO formation on the aerosol of the polluted troposphere (1998), Review lecture, Second Workshop of theEUROTRAC II Subproject CMD (Chemical Mechanisms Development), September 23-25, Karlsruhe,Germany.

M. AmmannNitrous acid formation on soot aerosol particles, plenary lecture, 6th FECS Conference on Chemistry andEnvironment (Atmospheric Chemistry and Air Pollution), August 26-28, 1998, Kopenhagen, Denmark.

B. Aumont, S. Madronich, M. Ammann, M. Kalberer, U. Baltensperger, D. Hauglustaine, F. BrochetonContribution of the NOs+soot reaction to the chemistry of the lower troposphere: a modeling studySecond Workshop of EUROTRAC-2 Subproject CMD, Karlsruhe, 23-25 September 1998.

U. BaltenspergerDie Auswirkungen von Aerosolpartikeln auf die AtmospharenchemieBereichskonferenz F3A, Paul Scherrer Institut, 12 January 1998.

U. BaltenspergerHeterogene Wechselwirkung von Stickoxid-Molekulen mit AerosolpartikelnBergische Universitat - Gesamthochschule Wuppertal, 29 January, 1998.

210

U. BaltenspergerInteraction of NO2 with soot particles, heterogeneous formation ofHONOUniversity of British Columbia, Vancouver, 14 August, 1998.

U. BaltenspergerHeterogeneous reaction ofNO2 on carbon aerosol particlesUniversity of Washington, Seattle, 18 August, 1998.

U. Baltensperger, E. Weingartner, S. NyekiAerosol measurements: providing the data for the quantification of aerosol impact on climateProc. Global Atmosphere Watch (GAW) Conference, 14-15 October 1998, Zurich, pp. 90-93.

U. Baltensperger, E. Weingartner, M. Lugauer, S. Nyeki, N. Streit, C. Zellweger, I. Colbeck, H.W. GaggelerSeasonal variation of aerosol parameters at the high alpine site Jungfraujoch, Switzerland: vertical transportand new particle formationJoint International Symposium on Global Atmospheric Chemistry, Seattle, 19-25 August, 1998.

U. Baltensperger, L. JalkanenAerosol studies in the WMO's Global Atmosphere Watch programmeInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sci. 29, Suppl. 1, S165(1998).

U. Baltensperger, M. Ammann, M. KalbererGas molecule interaction with aerosol surfaces: sticking coefficients and reaction rates; basic propertiesInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sci. 29, Suppl. 1,S637-S638(1998).

S. Ehrman, M. Schwikowski, U. Baltensperger, H.W. GaggelerChemical characterization of individual ice crystals and size classified snow samplesInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sci. 29, Suppl. 1,S323-S324(1998).

S.H. Ehrman, B. Frey, M. Schwikowski, H.W. GaggelerImaging and analysis of individual snow crystalsIn P.M. Borrell, P. Borrell, T. Cvitas, W. Seiler (eds.), Proc. EUROTRAC-2 Symposium '98, Garmisch-Partenkirchen, 23-27 March, 1998.

S.H. EhrmanCombustion synthesis of multicomponent nanoparticlesSeminar Radio- und Umweltchemie, Universitat Bern, 23 January 1998.

H.W. GaggelerRadionuclide detection at Jungfraujoch as a probe for the study of atmospheric processesVOTALP WORKSHOP, Vienna, Austria, 19-20 January, 1998.

H.W. GaggelerGletscherbohrkampagne auf dem Cerro Tapado, ChileSchweizerischer Alpenclub, Sektion Grenchen, 3 April, 1998.

H.W. GaggelerContents and behaviour of radionuclides in the atmosphere13th Radiochemical Conference, Marianske Lazne, Czech Republic, 19-24 April, 1998.

H.W. GaggelerWorkshop on ice core drilling in Chile and BoliviaJungfraujoch, 9-11 August, 1998.

H.W. Gaggeler, S. Hubener, P. Zanis, M. LugauerBe-7 und Radonzerfallsproduktmessungen auf dem Jungfraujoch zum Studium atmospharischer ProzesseVortragstagung, Ges. Deutscher Chemiker, Fachgruppe Nuklearchemie, Dresden, 7-9 September, 1998.

211

H.W. GäggelerQuellen und Senken atmosphärischer Radionuklide40 Jahre Überwachung der Radioaktivität in der Schweiz, Hôpital des Bourgeois, Fribourg,24 November, 1998.

H.W. GäggelerVom Hobby zum Beruf: Wie Radiochemiker hochalpine Gletscher missbrauchenFestkolloquium aus Anlass des 70. Geburtstages von Prof. H.R. vonf Gunten, PSI, 11 December, 1998.

P. GinotBericht von den Probebohrungen in ChileSeminar Radio- und Umweltchemie, Universität Bern, 17 April 1998.

P. Ginot, M. Funk, H.W. Gäggeler, S. Hübener, B. Pouyaud, U. Schotterer, M. Schwikowski, W. StichlerEnvironmental record preserved in an ice core from Cerro Tapado, Chile16th Latin American Conference on Geosciences, Bayreuth, Germany, 14-16 October 1998.

P. GinotEisbohrkerne vom Cerro Tapado, ChileSeminar VAW-ETH Zürich, 26 November 1998.

A. Grund, M. Schwikowski, H.W. GäggelerHigh resolution determination of trace components in an ice core from Grenzgletscher (Monte Rosa,4200 m a.s.l.)28lh Annual International Symposium on Environmental Analytical Chemistry, Geneva, 1-5 March 1998.

T. HuberKontinuierliches Schmelzen und Analysieren von Eisbohrkernen: Erste VersucheSeminar Radio- und Umweltchemie, Universität Bern, 18 December 1998.

S. HübenerKosmogene Radionuklide am JungfraujochSeminar Radio- und Umweltchemie, Universität Bern, 27 November 1998.

C. Hüglin, R. Gehrig, C. Moor, U. Baltensperger, C. MonnCharacterization of PM10 and PM25 in SwitzerlandInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sei. 29, Suppl. 1,S119-S120(1998).

V.M.H. Lavanchy, U. Baltensperger, H. W. Gäggele, M. SchwikowskiReconstruction of the history of black carbon and elemental carbon concentration from a high-alpineglacier (Colle Gnifetti)28lh Annual International Symposium on Environmental Analytical Chemistry, Geneva, 1-5 March 1998.

M. Lugauer, U. Baltensperger, M. Furger, H.W. Gäggeler, D.T. Jost, M. SchwikoskiVertical transport processes in the lower troposphere above the alpine regionProc. Eighth Conference on Mountain Meteorology, Flagstaff, USA, 3-7 August, 1998, AmericanMeteorological Society, Boston, pp. 159-162.

S. Nyeki, M. Kalberer, F. Schröder, A. Petzold, E. Weingatner, M. Lugauer, I. Colbeck, U. BaltenspergerCondensation nuclei (CN) and ultrafine CN in the free troposphere to 12 km above the Jungfraujoch stationInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sei. 29, Suppl. 1,S711-S712(1998).

M. SchwikowskiPaläoatmosphärenstudien anhand von Eisbohrkernen aus den AlpenSeminar Klima- und Umweltphysik, Universität Bern, 6 April 1998.

212

M. SchwikowskiRekonstruktion der Aerosolkonzentrationen anhand eines Eisbohrkernes vom FiescherhornSeminar Radio- und Umweltchemie, Universität Bern, 17 April 1998.

M. SchwikowskiAnalyse anorganischer Spurenstoffe in alpinen Schnee- und Eisproben zur Rekonstruktion der Geschichteder LuftverschmutzungSeminar Anorganische und Angewandte Chemie, Universität Hamburg, 29 June 1998.

M. Schwikowski, H.W. Gäggeler, U. SchottererAnnually resolved air chemistry record from an Alpine ice coreJoint International Symposium on Global Atmospheric Chemistry, Seattle, USA, 19-25 August 1998.

M. Schwikowski, H.W. Gäggeler, U. SchottererCalibration of a historical aerosol concentration record from an Alpine ice core by long-term aerosolmonitoringInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sei. 29, Suppl. 1,S517-S518(1998).

M. Schwikowski, T. Döring, L. ToblerBestimmung von Schwermetall-Konzentrationen in Gletschereis mittels doppelfokussierender ICP-MSICP-MS Anwendertreffen Schweiz, Villigen, 10 November 1998.

M. Schwikowski, S. Brütsch, T. Döring, A. Döscher, A. Eichler, H.W. Gäggeler, V. Lavanchy, U. SchottererPaläoatmosphärenstudien anhand von Gletschern aus den Mittleren BreitenSeminar Klimaphysik der EAWAG, 20 November 1998.

G. Skillas, K. Siegmann, U. BaltenspergerThe density and fractal-like dimension of particles sampled from a laminar diffusion flameInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sei. 29, Suppl. 1,S1277-S1278 (1998).

N. Streit, E. Weingartner, M. Schwikowski, H.W. Gäggeler, U. BaltenspergerThe relative contribution of natural and man-made sources to the chemical and physical properties of aerosolfrom a high alpine siteEuropean Research Course on Atmospheres, Grenoble, 12 January-11 February 1998.

N. Streit, E. Weingartner, M. Schwikowski, F. Hegedüs, H.W. Gäggeler, U. BaltenspergerCharacterization of size fractionated aerosol from the Jungfraujoch (3571 m asl) using total reflection X-rayfluorescence (TXRF)28th Annual International Symposium on Environmental Analytical Chemistry (ISEAC 28), Geneva,1-5 March, 1998.

N. StreitGrössenfraktionierte chemische AerosolanalytikSeminar Radio- und Umweltchemie, PSI, 3 April 1998.

N. Streit, E. Weingartner, M. Schwikowski, H.W. Gäggeler, U. BaltenspergerCharacterization of size fractionated aerosol from the Jungfraujoch (3580 m asl) using total reflection X-rayfluorescenceInternational Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sei. 29, Suppl. 1,S743-S744(1998).

B. TenberkenGrundlagen der Kapillarelektrophorese und mögliche Anwendungen zur Untersuchung des Scavenging-Prozesses auf dem JungfraujochSeminar Radio- und Umweltchemie, PSI, 30 October 1998.

213

L. ToblerDie Neutronenaktivierungsanlage an der SINQEinweihungskolloquium Neutronenaktivierungsanlage und Gas-Jet SINQ, PSI, 26 Januar 1998.

L Tobler, E. Lehmann, D. Jost, H.W. GaggelerDie Bestrahlungseinrichtungen an der neuen Spallationsneutronenquelle SINQ des Paul Scherrer Instituts17. Seminar Aktivierungsanalyse, Geesthacht, 6-8 May 1998.

L. Tobler, A. WyttenbachWie zerstorungsfrei ist die Neutronenaktivierungsanalyse ?17. Seminar Aktivierungsanalyse, Geesthacht, 6-8 May 1998.

L. Tobler, T. Doring, M. Schwikowski, H.W. GaggelerBestimmung von Schwermetallkonzentrationen und Pb-lsotopenverhaltnissen in Gletschereis mitdoppelfokussierendem ICP-MS4. Symposium MS-Verfahren der Elementspurenanalyse & 14. ICP-MS Anwendertreffen, Mainz, Germany,28 September-1 October 1998.

H.R. von GuntenApplication of natural radionuclides in environmental- andgeosciences13lh Radiochemical Conference, Marienbad, Czech Republic, 19-24 April, 1998.

E. WeingartnerAerosol size distributions at a high-alpine site (Jungfraujoch 3580 masl)Workshop on Particle Measurement, ETH Zurich, 7 August, 1998.

E. Weingartner, S. Nyeki, N. Streit, U. BaltenspergerSeasonal variation of size distributions and fine particle formation of continental aerosol at a high alpine site(Jungfraujoch, 3454 m a.s.l.)International Aerosol Conference '98, Edinburgh, 14-17 September 1998, J. Aerosol Sci. 29, Suppl. 1,S509-S510(1998).

A. WyttenbachNeutronenaktivierungsanalysen von EdelmetallmunzenSeminar MPI, Tubingen, 30 March, 1998.

PROJECT RADWASTER. Weinreich, L. Wyer, N. Crompton, M.C. Nievergelt-Egido, I. Guenther, U. Roelcke, K.L. Leenders,E.J. Knust, R.G. Blasberg1-124 and its applications in nuclear medicine and biologyIAEA-Symposium "Modern Trends in Radiopharmaceuticals for Diagnosis and Therapy, Lissabon, Portugal,30 March - 3 April, 1998.

D.F. Dos Santos, M. Argentini, R. Weinreich, H.-J. HansenCurrent status of selenium coupling to Hoechst-Carborane8th International Symposium on Neutron Therapy for Cancer, La Jolla, CA, USA, 13-17 September 1998.

ION BEAM PHYSICS

M. DobeliBeschleuniger-SIMS: Spurenelementanalyse unterhalb des ppb-BereichsFruhjahrstagung d. Deutschen Physikalischen Gesellschaft, Konstanz, Germany, March 1998.

M. DobeliCm tracks in PMMAInvited Talk., 15th Int. Conf. on the Appl. of Accelerators in Research and Technol., Denton, Texas, USA,November 1998.

214

M. DöbeliMeV Cluster-lonenstrahlenSeminar über Kernphysik, ETH-Zürich, Switzerland, December 1998.

H. FuhrmannH- Desorption von Si-Oberflächen mit dem fokussierten lonenstrahlSeminar, Forschungszentrum Rossendorf, Dresden, Germany, 26 March 1998.

H. Fuhrmann, M. Döbeli, R. Mühle, M. SuterFocused Ion Beam Structuring of Si and Si/CoSi2 heterostrucures using adsorbed hydrogen as a resistE-MRS Spring Meeting, Strasbourg, France, 16-19 June 1998.

P. Gartenmann, M. Suter, H.A. SynalAnwendung von Beschleunigermassenspektrometrie mit^FeFrühjahrstagung d. Deutschen Physikalischen Gesellschaft, Konstanz, Germany, March 1998.

R. Huber, S. Jacob, M. Suter, H.A. SynalMessung des Phasenraumes an einer AMS InjektionsstufeFrühjahrstagung d. Deutschen Physikalischen Gesellschaft, Konstanz, Germany, March 1998.

S. Jacob, R. Huber, M. Suter, H.A. SynalDetektion und Identifikation von Ionen in derBeschleunigermassenspektrometrie im Energiebereich von 1 MeVFrühjahrstagung Deutsche Physikalische Gesellschaft, Konstanz, Germany, March 1998.

P.W. KubikNeuigkeiten von der Beschleuniger Massenspektrometrie Anlage in ZürichSeminar, Abt. Nuklearchemie, Universität zu Köln, Germany, 11 December, 1998.

M. Pouchon, M. Döbeli, C. DegueldreRetention of iodine in yYttria stabilized zirconiaInt. Conf. on Ion Beam Modification of Materials, Amsterdam, 1998.

H.A. SynalBeschleunigermassenspektrometrie : Neue Entwicklungen und ZukunftsperspektivenPlenarvortrag, Frühjahrstagung d. Deutschen Physikalischen Gesellschaft, Konstanz, Germany, March 1998.

H.A. SynalDetection of heavy radionuclides at the PSI/ETH AMS systemInvited Talk, 15th Int. Conf. on the Appl. of Accelerators in Research and Technol., Denton Texas, USA,November 1998.

215

LECTURES AND COURSES

Prof. Dr. R. EichlerETH Zurich, SS98:• Erganzendes physikalisches Praktikum fur UNW• Physik II (with others)• Physikpraktikum fur Vorgeriickte (with others)• Physik der Elementarteilchen (with others)• Teilchenphysik am PSI (with others).

ETH Zurich, WS98/99:• Physik II (with others)• Physikpraktikum fur Vorgeruckte (with others)• Physik der Elementarteilchen (with others)• Teilchenphysik am PSI (with others).

ASTROPHYSICS

Dr. M. GudelETH Zurich und Universitat Zurich, WS98/99:• Hochenergie-Astrophysik (with Prof. A. Benz)

PD Dr. Ph. JetzerUniversitat Zurich, WS98/99:• Weisse Zwerge, Neutronensterne und Schwarze Locher

Universitat Salerno, May 1998:• Nachdiplom-Vorlesungen: Lenti gravitazionali

Universitat Zurich:• Mitwirkung an der Organisation des Theorie-Seminars (gemeinsam mit der ETH)

RADIO- AND ENVIRONMENTAL CHEMISTRY

PD Dr. U. BaltenspergerETH Zurich, WS 98/99:• Aerosole I (with PD Dr. H. Burtscher)

ETH-Zurich, WS 98/99:• Analytische Chemie V (with others)

Prof. Dr. H.W. GaggelerUniversitat Bern, SS98:• Anwendung von Radioisotopen• Seminar in anorganischer und physikalischer Chemie (with others)• Seminar in Radio- und Umweltchemie (with Dr. M. Ammann), in collaboration with Paul Scherrer Institut

Universitat Bern, WS 98/99:• Radiochemie• Eigenschaften Schwerster Elemente• Radiochemiepraktikum (with Dr. A. Turler)• Interdisziplinares Seminar "Klimadynamik - gestern, heute, morgen" (with others)• Seminar in anorganischer und physikalischer Chemie (with others)• Seminar in Radio- und Umweltchemie (with Dr. M. Ammann), in collaboration with Paul Scherrer Institut

ION BEAM PHYSICS

Dr. M. SuterETH Zurich, WS 98/99:• Kernphysik II, Wahlfachvorlesung.

216

HIGHER DEGREES AWARDED

P. Auer: Title of thesis:Verdampfung von Schwermetallen in Temperaturgradientrohren als Methode furdie Verwertung von Ruckstanden aus der MullverbrennungPSI

Thesis advisors:Prof. Dr. H.W. Gaggeler (Universitat Bern & PSI)Dr. S. Stucki (PSI)Prof. Dr. B. Eichler (PSI).

St. Gehrsitz: Title of thesis:Investigation of the Elastic Properties of Crystals by Near infrared Brillouin ScatteringPSI

Thesis advisors:

Prof. F.K. Reinhart (EPFL)Dr. H. Sigg (PSI)

R. Har tmann: Title of thesis:Growth and Characterization of Semiconductor Nanostructures in the SystemSi-Ge-CPSI

Thesis advisors:

Prof. K. Ensslin (ETH Zurich)Dr. D. Grutzmacher (PSI)

A. Hofer: Title of thesis:Niederenergetische Myonen: Eigenschaften und AnwendungenUniversitat Konstanz / PSI

Thesis advisors:

Prof. G. Schatz (Universitat Konstanz)Dr. E. Morenzoni (PSI)

M. Kalberer: Title of thesis:Heterogeneous Reactions of NO2 with Soot Aerosol Particles at AmbientAtmospheric ConditionsUniversitat Bern / PSI

Thesis advisors:

Prof. Dr. H.W. Gaggeler (Universitat Bern & PSI)PD Dr. U. Baltensperger (PSI)Dr. M. Ammann (PSI).

M. Lugauer: Title of thesis:Vertical Transport of Atmospheric Trace Species in the AlpsPSI

Thesis advisors:

Prof. Dr. H.W. Gaggeler (Universitat Bern & PSI)Prof. Dr. H. Wanner (Universitat Bern)PD Dr. U. Baltensperger (PSI).

217

K. Noll: Title of thesis:Instrumentelle Bestimmung der Fluoranreicherung auf antarktischen Meteoritendurch NRA (Nuclear Reaction Analysis)Universitat Bern / PSI

Thesis advisors:

Prof. Dr. U. Krahenbuhl (Universitat Bern)Prof. Dr. H.W. Gaggeler (Universitat Bern & PSI).

Th. Nussbaumer: Title of thesis:Supraleitende Rontgen-DetektorenPSI

Thesis advisors:

Prof. Dr. H.R. Ott (ETH Zurich)Dr. A. Zehnder (PSI)Dr. Ph. Lerch (PSI).

G. Skil las: Title of thesis:Carbon Nanostructures from Combustion: Morphology, Density, and ApplicationsETHZ / PSI

Thesis advisors:Prof. Dr. H.-C. Siegmann (ETH Zurich)PD Dr. U. Baltensperger (PSI).

P A U L S C H E R R E R I N S T I T U T

Paul Scherrer InstitutCH-5232 Villigen PSI

Internet

Phone +41 56 310 21 11Fax +4156 310 2199http://www.psi.ch

7346