Growth of alkali-antimonide films for photocathodes

Photocathodes are widely used nowadays in many practical devices such as image tubes,radiation detectors and photomultipliers. A material often used for the photocathode film isan alkali antimonide. The development of most conventional photocathodes is still more orless empirical. This article describes studies on alkali-antimonide films that may lead to abetter understanding of the growth processes and hence to improvements in the manufactur-ing technology, with better reproducibility and photosensitivity.

Philips tech. Rev. 40,19-28,1982, No. 1

P. Dolizy

General features of photocathodes

A photocathode is a metallic or semiconductingcathode that will give photoemission: it emits elec-trons on illumination if the photon energy is equal toor greater than a certain value, the photoelectricthreshold energy [11. If a photocathode forms part ofan electrical circuit, its photoemission produces ameasurable current. The ratio of this current to theincident luminous flux, the photosénsitivity, is deter-mined by the efficiency of each of the three stages inthe photoemission process: the excitation of photo-electrons, their movement through the material to thevacuum interface, and their escape into the vacuum.

The excitation of electrons by photons of a givenenergy depends on the extinction coefficient, the re-fractive index and the thickness of the photocathodefilm. On their way to the surface the excited electronsmay lose energy through interactions with otherelectrons or with lattice phonons and through recom-binations with holes. This means that only some ofthe excited electrons will reach the surface. The escapeprobability P(W,x) of a photoelectron excited at adistance x from the vacuum interface can be describedto a first approximation by

P(W,x) = P(W,O)exp( -x/L), (1)

where W is the photon energy and L is called theescape depth. This equation is only valid if the thick-ness of the photocathode film is large compared to theescape depth; For photon energies lower than 2.5 eVthe energy losses of the photoelectrons are mainly due

P. Dolizy is with Laboratoires d'Electronique et de PhysiqueAppliquée (LEP), Limeil-Brévannes, Val-de-Marne, France.

19

to the electron-phonon interactions. At these energiesboth P(W,O) and L decrease with decreasing W,'andP(W,O) becomes zero at the photoelectric thresholdenergy.For the conversion of visible light, including the

near infrared and the near ultraviolet, the highestphotosensitivity is obained with the well-known silver-oxygen-caesium cathodes and with photocathodesbased on semiconductors such as various alkali anti-monides or some of the Hl-V-compounds [11. Thealkali antimonides absorb strongly in the visiblerange. The thickness of photocathode films of thesematerials is therefore made small, so that optical i'n-terference takes place in the film. An appropriate sur-face treatment can be applied to ensure that the thresh-old energy is not much higher than the band gap.In conventional photocathodes the bulk material

often has a cubic crystal structure and is a p-typesemiconductor. To lower the threshold energy thematerial is coated with a thin n-type film. The pres-ence of such a film produces band-bending in thevalence and conduction bands, as shown schemat-ically in fig. 1 for the alkali antimonide Na2KSb. Theenergy difference between the vacuum level and thebottom of the bulk conduction level, the 'effectiveelectron affinity' decreases considerably here (0.7 eV),because of the presence of n-type surface states. As aresult the threshold energy decreases from 2.0 eV to1.3 eV.

[1] See for example A. H. Sommer, Photoemissive materials,WiIey, New York 1968, and W. E, Spicer, J. appi. Phys. 31,2077, 1960.

20

For some photoemitters the effective electron affinity may evenbecome negative: the vacuum level is then lower than the bottom ofthe conduction band in the bulk of the material. Electrons in thisband then require no extra kinetic energy to escape into the vac-uum. Materials with a negative effective electron affinity can havea high photosensitivity because of their large escape depth; oneexample is p-GaAs coated with Cs [21. In this article, however, suchmaterials will not be discussed [31.

I---"---Evac

70eV

T03ev

I.OeV

EF'----,----

Ev· ::X/;

Fig. 1. Energy bands at the surface of an Na2KSb film. E; upperedge of the valence band in the bulk of the material. EF Fermi level.E; lower edge of conduction band in the bulk. Evac potential energyof an electron in vacuum. For p-type material (left) the photoelec-tric threshold energy, Evac - Eç, is equal to 2.0 eV. Adsorption ofcaesium at the surface gives rise to n-type surface states. This hasthe effect of bending the valence and conduction bands downwardnear the surface (right), so that the effective electron affinity,Em - Ec, has faUen to 0.3 eV and the threshold energy is only1.3 eV. La loss of electron energy as a result of interaction with avalence electron. An electron excited by a photon with an energy hvof about 2.5 eV can either escape directly to the vacuum, or mayfirst create a new electron-hole pair, and then escape.

Photocathodes have many practical applications,e.g. in image tubes, radiation detectors and photo-multipliers. In most applications the photocathodesconsist of a polycrystalline film deposited on a sub-strate. Various substrates can be used: metals such aspalladium, alloys such as nichrome, or insulators suchas glass. In addition, the substrate surface can be largeor small, flat or curved, without affecting the deposi-tion of the photoemissive film. The light may be in-cident from the vacuum interface (the reflectionmode), or from the substrate interface (the transmis-sion mode). In the reflection mode the substrate doesnot have to be transparent and the film thickness isnot critical. In the transmission mode, however, thesubstrate must be transparent and the film thicknesshas a critical optimum value: on the one hand thethickness should be not much larger than the escapedepth, whereas on the other hand the film should bethick enough to ensure a sufficient absorption of theincident radiation via the optical interference insidethe film. In most applications photocathodes are usedin the transmission mode.

P. DOLlZY Philips tech. Rev. 40, No. I

Tbe scope of the investigations

Most conventional photocathodes are producedmore or less empirically. The growth is usually mon-itored only by measuring the photosensitivity, in themode to be used, at various stages of the growth. Thisdoes not give much information about the variousprocesses that occur during the growth. An under-standing of these processes may be very useful, how-ever, for improving the reproducibility of the growthand the photosensitivity of the photocathodes, or forautomating the growth procedure. In our investiga-tions on conventional alkali-antimonide photocath-odes we therefore carried out a number of additionalmeasurements during the growth.

We applied the alkali-antimonide films in a vacuumchamber by depositing antimony atoms on a substratein the presence of one or more alkali metals. To findout how the film thickness increased during thegrowth we measured the optical reflection and trans-mission. We used atomic absorption speetrometry(AAS) to determine the alkali partial vapour pres-sures, and also measured the photosensitivity in boththe reflection and the transmission modes, to find outhow the escape depth and escape probability variedduring the growth. These measurements were foundto be of great value for examining and improving thephotocathode films during the growth in practicaltubes. Before considering these studies more closely,let us first look at the growth of the alkali-antimonidefilms in more detail.

Growth of alkali-antimonide films

Alkali-antimonide films for photocathodes are usu-ally evaporated from separate sources ('dispensers')for antimony and alkali. The dispensers are locatedinside the evacuated tube being manufactured, andthey are heated electrically under external control.A thin film of antimony evaporated on to the sub-strate reacts exothermally with the alkali-metal va-pour to produce various semiconductor compounds.The chemical reaction between solid antimony and so-dium vapour, for example, can be represented by:

Sb, + 3 Nag - Na3Sb. (2)

This equation provides no information about theexistence of intermediate compounds that could beformed during the reaction.

The actual composition of an alkali-antimonidefilm depends on the partial alkali pressures in thevapour, the vapour temperature and the substratetemperature, as well as on the substrate state, the im-purity content and the growth rate. Since alkali metalsreact very rapidly with oxygen, carbon dioxide and

Philips tech. Rev. 40, No. 1 ALKALI-ANTIMONIDE FILMS FOR PHOTOCATHODES 21

water, it is essential that the growth takes place in avery clean vacuum envelope. The number of corn-pounds formed increases with the atomic number ofthe alkali metal [4] and with the number of differentmetals in the vapour. The control of the partial alkalipressures is therefore very important for obtaining aphotocathode. film of the desired. composition. At apartial pressure p the number of atoms tu incident onthe surface per cm'' per second is given by [5]

ni = p/(2nmkTv)1, (3)

where m is the mass of an alkali-metal atom, k isBoltzmann's constant and Tvis the temperature of thevapour. Some of the adsorbed atoms will leave thesurface again, however. Assuming that this desorp-tion is a first-order proces, the number of atoms nl

leaving the surface per cm2 per second is [5]

where nad is the number of atoms adsorbed percm'' of the surface and.A is the partition function,which is nearly constant to a first approximationand has a value between 1014. and 1016 S-I. Eb isthe binding energy between the metal and the sub-strate; I's is the substrate temperature. When theevaporated film is in equilibrium with the vapour,the value of nl is equal to that of ni, so that nad isgiven by

nad = A exp( -Eb/kI's)

Ifp, Tv, A, Eb and I's are known, this relation can beused to calculate the number of atoms adsorbed onthe substrate. In fig. 2 the calculated monolayer frac-tion f for potassium on glass is plotted against thesubstrate temperature, for three different combina-tions of the partial vapour pressure of potassium andthe binding energy. The value of f decreases stronglywith increasing substrate temperature. The tempera-ture above which f is virtually zero is considerablyhigher for' a higher partial vapour pressure of potas-sium and higher binding energy.

îo:[~\_S=, Io 100 200 300 °C

-TsFig. 2. Calculated monolayer fraction f for potassium evaporatedon to a glass substrate, as a function of the substrate temperature Ts•The three curves correspond to three different combinations of thepartial potassium vapour pressure PK and the binding energy Ebbetween potassium and the glass. a) PK = 1.3 X 10-4 Pa, Eb = 92kj/mol. b) PK = 1.3 X 10-3 Pa, Eb = 92 kj/mol. c) PK =1.3 X 10-4 Pa, Eb = 125 kj/mol.

(4)

In small vacuum chambers the speed of the vacuumpump must be taken into account [6]. If the bindingenergy is low, the alkali partial pressure is also affect-ed by the pumping time. The binding energy dependson the interaction between the alkali metal and thesubstrate. For example, sodium reacts more stronglythan potassium or caesium with a glass substrate.Other important factors for the binding energy are thepurity of the substrate and the presence of other mat-erials on the surface, such as alkali metals and anti-mony, which have been adsorbed in a previous stageof the processing. If the vapour contains differentalkali metals, the adsorption and desorption areassociated with chemical exchanges of the alkaliatoms on the substrate surface.The composition of the compounds formed de-

pends on the adsorption of the various alkali metalsand on the type of reaction occuring at the surface. Insome cases different reaction mechanisms give thesame final compounds. A bialkaline antimonide ofcomposition (A,BhSb, for example, can be obtainedfrom the AsSb compound by a partial substitution ofA by B, or by the addition of solid Sb and gaseous Bvia a diffusion process. The addition mechanism forthe growth of an Na2KSb film can be described as fol-lows:

(5)

2Sbs + ë Na, - 2NasSb2 NasSb + Sb, + 3 Kg - 3 Na2KSb.

The measurements of the partial alkali-vapour pres-sures and the optical properties of the film during thegrowth have shown that such addition processes arealways accompanied by substitutions inside the film.At a typical temperature of about 200 oe and typicalpartial pressures of 10-4 Pa potassium is readily sub-stituted by sodium [7].

An important property of alkali-antimonide films istheir crystalline nature. To obtain a high photosen-sitivity, the microcrystals must be larger than the es-cape depth of the photoelectrons excited inside them.The material then behaves as a single crystal for thephotoelectrons. Since the escape depth must be of thesame order as the film thickness, the growth condi-tions should generally favour the formation of thelargest possible microcrystals.

(6)

[2] J. van Laar and J. J. Scheer, Philips tech. Rev. 29, 54, 196B.[3] More information on photoemitters with a negative 'effective

electron affinity' can be found in the article by J. van Laar andJ. J. Scheertêl , or. in the six articles in Acta Electronica 16,No. 3, 1973 (pp. 215-271).

[4] E. Kansky, Adv. in Electronics and Electron Phys. 33A, 357,1972.

[6] H. Mayer, Vakuum-Technik 4, I, 1955.[6] J. P. Hobson, 1961 Trans. Bth Nat. Vacuum Symp.,

Washington O.C., Vol. 1, page 26.[7] P. Dolizy, O. De Luca and M. Deloron, Acta Electronica 20,

265, 1977.

22

The texture of the films depends on their composi-tion and crystal structure. A strong preferential orien-tation has been observed for hexagonal Na-Sb, withthe (00.1) plane parallel to the surface [7]. On the'other hand, films of hexagonal KsSb often have apoor crystal structure. An Na2KSb film obtained byevaporating K and Sb on to a structured NasSb film isalso strongly oriented. In this case the (111) plane ofthe cubic crystal structure lsl is parallel to the surface.Further evaporations of K and Sb lead to the forma-tion of the hexagonal NaK2Sb phase, coexisting withthe initial Na2KSb phase. This hexagonal phase againhas a texture with the (00.1) plane parallel to the sur-face. It is assumed that during the growth the bulk ofthe films contains only one or two 'neighbour' com-pounds from the series Na-Sb, Na2KSb, NaK2Sb andKsSb, when the antimony atoms are completely sat-urated by the alkali atoms.

Determination of properties during growth

In our investigations we determined the propertiesof interest during the growth of alkali-antimonidefilms on glass by measuring the atomic absorptions ofthe alkali vapours, the optical reflection and transmis-sion of the film, and its photosensitivity in the reflec-tion and transmission modes. We interpreted the re-sults of these measurements in terms of the partialalkali-vapour pressures, the extinction coefficient,refractive index and thickness of the film, and also theescape depth and escape probability of the photoelec-trons. These measurements will now be discussed andthe derivation of the properties during the growth willbe explained.

Optical measurements

Fig. 3 is a diagram of the equipment used for meas-uring the atomic absorptions of sodium and potas-sium in the alkali vapour. The growth takes place in avacuum chamber containing an antimony bead, alkalidispensers, an anode and a glass substrate on whichthe photocathode film is deposited. Two modulatedlight beams, one with the correct resonance line forsodium, and the other with the line for potassium, arepassed through the vacuum chamber. The transmittedlight beam from both beams is detected by a simplearrangement of a focusing lens, an interference filterand a photocell. The electrical signal from this cell issupplied to a synchronous detector, which is only sen-sitive to signals at frequencies close to that of themodulation of the light beam. To measure the atomiccaesium absorption, the transmitted resonance radia-tion was detected with a monochromator on accountof unwanted signals from other elements.

P. DOLIZY PhiJips tech. Rev. 40, No. J

We measured the reflection,. transmission andphotosensitivity. during the growth with the arrange-ment shown in the diagram of fig. 4. Three modulatedlight beams were used in these measurements. Thefirst beam is incident on the photocathode film from

Subc

Fig. 3. Schematic arrangement for determining the partial sodiumand potassium vapour pressures during the growth of an alkali-antimonide photocathode C in a vacuum chamber. Two lightbeams, originating from hollow-cathode lamps LNa and LK,. withthe appropriate sodium and potassium resonance lines are mod-ulated by two different choppers Chl and Ch2 and passed throughthe vacuum chamber. The measured attenuation for the sodium orpotassium resonance radiation is a measure of the partial sodium orpotassium vapour pressure in the chamber. Le lenses. Sub glasssubstrate. Sb 'bead' of antimony. A anode. Di alkali dispensers.P vacuum pump. FNa and FK sodium and potassium interferencefilters. Dl and D2 detectors, each consisting of a photocell and asynchronous detector tuned to the modulation frequency of one ofthe two modulated light beams.

DR

p

Fig. 4. Schematic arrangement for measuring the optical reflectionand transmission and the photosensitivity in the reflection andtransmission modes during the growth. Three light beams are usedin the measurements; they originate from the sources Ll' L2 andLs, and each beam is modulated at a different frequency by thechoppers Chl' Ch2 and Chs. DR is a detector for the light originat-ing from Ll and reflected by the photocathode C. DT is a detectorfor the light originating from L2 and transmitted by C. DeR andDe. are synchronous detectors for the current generated in thevacuum chamber by the light originating from L2 and Ls respec-tively. The photosensitivity of C in the reflection and transmissionmodes can be determined from these currents. M half-silvered mir-rors. Sub glass substrate. Sb 'bead' of antimony. A anode. Di alkalidispensers. P vacuum pump.

Philips tech. Rev. 40, No. 1 ALKALI-ANTIMONIDE FILMS FOR PHOTOCATHODES 23

the substrate side. Measurement of the intensity of thereflected light gives the reflectance of the film. Thesecond beam is incident from the photocathode side,so that synchronous detection of the electric currentgenerated gives the photosensitivity of the film in thereflection mode. The intensity of the light that passesthrough the film is also measured, giving the opticaltransmittance of the film. The third beam is again in-cident from the substrate side; synchronous detectionof the current generated gives the photosensitivity inthe transmission mode.

Derivation of the properties

The partial vapour pressures of sodium, potassiumand caesium can be derived by comparing the ob-served attenuation of the resonant radiation withmeasurements on calibration cells with known alkalivapour pressures. Our experimental arrangementallows the partial vapour pressures to be determinedsimultaneously. An important experimental observa-tion in this respect relates to the exchange of sodiumand potassium. For example, when all the glass sur-faces of a vacuum chamber have been covered withthe two metals, the ratio of the partial pressures PNa

and PK in the non-saturated vapour is only slightlydependent on which dispenser is emitting inside thechamber at that moment. This is illustrated in fig. 5for a glass chamber in which the substrate tempera-

Fig. 5. Sodium partial vapour pressure PNa plotted against thepotassium partial vapour pressure PK in a vacuum chamber in. which all the surfaces are covered with the two metals. Thesubstrate temperature is between 200 and 250°C. If the sodiumdispenser DiN. alone or the potassium dispenser DiK alone isemitting, the partial pressure of the other metal also increasesstrongly, because of substitution reactions at the walls. The ratioPN./PK does not increase greatly when the potassium dispenser isswitched off and the sodium dispenser is switched on.

ture is between 200 and 250 oe. The ratio PNa/PKdoes not increase much when the potassium dispenseris switched off and the sodium dispenser is switchedon. In both situations the ratio does not deviate muchfrom 2. As indicated before, this ratio is strongly de-pendent on the nature of the surfaces.The extinction coeffcient k and the refractive index

n can be derived from the reflection measurements [9] .

Since .the film thickness d is smaller than the wave-length I.. of the incident light, the reflectance R is acomplicated function of k, n, d and 1... If R is plottedagainst d, or against the deposition time t, whichamounts to the same thing, for given values of k, nand 1.., a number of maxima and minima are obtained;see fig. 6. For a given wavelength the value of R at

40%r------------=~~~------------__,R

t 20

°o~----~--~----~----~----~~10 20 30 40 50min

-tFig. 6. Reflectance R of an Na2KSb film at 500 nm, measured as afunction of the deposition time t.

1.4 Rma=0.25

k

t 1.2

1.0

QB 0.055

Q60.065

Q42.0 2.5

0.30 ///

// 0.40/

/

//0.35///

3.5 4.0-n

Fig. 7. Calculated isoreflectance curves in the plane of theextinction coefficient k and the refractive index n, for the firstminimum Rml and the first maximum Rma in the R(t)-curve (see forexample fig. 6). The values of k and n can be derived by comparingthese isoreflectance curves with the measured values of Rml andRma·

[s] J. J. Scheer and P. Zalm, Philips Res. Repts 14, 143, 1959.W. H. McCarroll, Phys. Chem. Solids 16, 30, 1960.

[9] S. E. Webber and S. R. Scharber Jr., App!. Optics 10, 338,1971.V. E. Kondrashov and A. S. Shefov, Eng!. Trans!. Bull. Acad.Sci. USSR, phys. Ser. 28, 1349, 1964.

24

these extreme values can he expressed in terms of kand n. This permits the 'isoreflectance' curves to bedrawn, which give the theoretical reflectances at theseextreme values in a plane of k- and n-coordinates.These curves are shown ui fig, 7 for the first minimumand the first maximum on the R(t)-curve. The k- andn-values can then be deduced 'by comparing the curveswith the extreme values of the reflectance observedduring the growth (fig. 6).If the k- and n-values are known, the optical

reflectance at the wavelength À. can be calculated asa function of d; see fig. 8. A comparison of thecalculated R(d)-curve with the measured R(t)-curve(fig. 6) gives the film thickness d(t) and hence thegrowth rate r(t) as well. A more elaborate way of de-termining d(t) and r(t) depends on the calculation ofR as a function of À. for a large number of d-values.Fig. 9 shows how the theoretical curve for R(À.) varieswith d. Comparison with the experimental curves forR(À.) measured during the growth allows the values ofd(t) and r(t) to be obtained.

In determining the escape depth L and the escapeprobability P(W,O) we have to remember that L canbe of the same order as d. This implies that some ofthe photoelectrons will travel to the interface with thesubstrate, so that eq. (1) is no longer true. If there isno recombination at the interface these electrons can

40%

20

10

50-d

100nm

Fig. 8. Calculated reflectance R at 500 nm for a film with k = 0.8and n = 3.5, as a function of the film thickness d.A comparison ofthe R(d)-curve, calculated from the known k- and n-values of afilm, and the measured R(t)-curve (see for example fig. 6) gives thefilm thickness and the growth rate as a function of time.

P. DOLIZY Philips tech. Rev. 40, No. I

&_~LL_~E_t4::f~

400 600 800nm-11.

400 600 800nm 400 600 800nm-11. -À,

Fig. 9. Calculated optical reflectance R as a function of the wave-length À, for different thicknesses of an Na2KSb film. Comparisonof these curves with the R(À)-curves measured during the growthprovides another method for determining the film thickness and thegrowth rate as a function of time.

still escape on the vacuum side ..The escape probabil-ity for a given photon energy W is then given by

cosh (diL - x/L)P(W;x) = P(W,O) /. (7)

cosh (d L)

However, if the recombination velocity at the sub-strate interface is very high, the electrons arriving atthis interface can no longer escape on the vacuumside. The escape probability is then given by

sinh (d/L - x/L)P(W;x) = P(W,O) . / . (8)

smh(d L)

In practice the recombination velocity will have afinite value, so that P(W,x) is a very complicatedfunction, which will not be discussed here [101.

The values of Land P( W;O) during the growth ofthe film on a substrate with known optical propertiescan be determined from measurements of the photo-sensitivity in the reflection mode ({}R) and in the trans-mission mode ({}T). If AR and AT are the distributionfunctions, dependent on k, nand À., of the light ab-sorbed in the film in the reflection and transmissionmodes, {}R and {}T can be expressed as

d

(}R = J AR(k,n,À.,x) P(W;x) dx

(9)d

(}T = J AT(k,n,À.,x) P(W;x) dx.o

If k, nand d are known, the value of L at a givenwavelength À.can be derived from the ratio {}of {}R to{}T. Infig. 10 the calculated ratio {}is plotted againstL for various combinations of k, n, À. and d. Thevalue of L is obtained by comparing the experimentalvalue of {},determined by relative measurements of {}R

Philips tech. Rev. 40, No. 1 ALKALI-ANTIMONIDE FILMS FOR PHOTOCATHODES 25

Fig. ID. Calculated ratio e of the photosensitivity in the reflection mode to the photosensitivity inthe transmission mode, as a function of the escape depth L of the photoelectrons, for variousvalues of the film thickness d, wavelength À, extinction coefficient k and refractive index n. Theescape depth L can be derived from the e(L)-curve calculated from the known values of À, d, kand n, and the measured value of e, provided that the calculated curve varies sufficiently with L.

and {}T, with the theoretical curves. A necessarycondition for the combination of parameters is that (!

should vary sufficiently withL. In the situations of fig.10 a good determination is possible with blue light ifthe thickness is larger than about 40 nm. If L isknown, the escape probability P(W,O) can be ob-tained from an absolute measurement of l!R or {}T.

5r-----------------~~À= 436nmk= 1.3n= 3.7

4

" \\\\\\\\\\\\\,,

" d=75nm' ......

.....__2 ," " .....

..........................._ 50----1 30-------------10 20 30 40 50nm

-L

Growth monitoring

The evaluation of the properties during the growthof a photocathode provides useful information formonitoring this growth. Let us first consider coating aglass substrate with antimony, and examining the filmby reflection measurements. At a substrate tempera-ture above 140 oe the sticking coefficient of antimonyon a very clean glass substrate is close to zero if theantimony flux is no larger than WIS at/cm's. In thepresence of an alkali vapour, however, this coefficientapproaches· unity for substrate temperatures between140 and 250 oe. When the growth of a photocathodefilm is started at 200 oe, for example, antimony musttherefore be evaporated in the presence of an alkalivapour, to permit the antimony to be satisfactorilydeposited on the glass and the chemical reaction be-tween solid antimony and the alkali vapour to start.

During the chemical reaction the film thickness in-creases. To obtain a thicker film when this reaction hasbeen completed, more antimony must be deposited andsaturated again with alkali vapour. The growth of thefilm can easily be monitored by optical reflection ortransmission measurements,' as described. As an

example, fig. 11 shows the reflectance during thegrowth of an Na-Sb film!on a glass substrate. In thiscase the growth occurs via a number of separate anti-mony evaporations, whereas sodium is continuouslypresent in the vapour. The theoretical curve of the re-flectance as a function of the film thickness (fig. 11a)is similar to that of fig. 8. If the reflectance at the

1.5~--------------Ä-=-5-7t-8-n-m~k = 0.48n=3.1

1.25

et "-- ~75~

0.75

30

~0.5

10 20 30 40 50nm-L

beginning of the growth is measured in detail, theeffects of the various antimony depositions can beseen; see fig. lIb. The change of reflection during onedeposition is shown in fig. lIc. After the antimonydispenser has been switched on, the film thickness firstincreases steeply, since antimony is deposited in thepresence of sodium vapour. When the evaporation istemporarily stopped, the antimony atoms still presenton the film react with the sodium atoms to form acompound close to NajSb. This chemical reactioncauses a further increase in the thickness. When thereaction has been completed, the thickness remainsconstant until the next antimony evaporation begins.Similar results have also been obtained in growingKsSb and essSb layers.The rate of the reaction depends upon the quantity

of antimony and alkali metal at the substrate surface.If there is an excess of antimony, the reaction rate iscontrolled by the local amounts of alkali. The rate ofgrowth of the film during the chemical reaction is thenproportional to the partial alkali vapour pressure.

[la] See C. Piaget, Thèse No. 1892, Université de Paris Sud, 1977.

0 5 10 15minQ -t

R /t50 100nm -t

f

26

However, if more alkali atoms than antimony atomsarrive at the substrate, the growth rate is proportionalto the deposition rate of the antimony., The growth of Na2KSb films can be studied ina similar way to that of the monoalkali-antimonidefilms. To obtain an Na2KSb film with the highestphotosensitivity the addition mechanism of eq. (6) isfound to be the most satisfactory. In this mechanism,the number of antimony evaporations in the presenceof sodium should be twice the number of evapora-tions in the presence of potassium. In practice, how-ever, the situation is more complicated, since potas-sium in the deposited film is partly substituted by so-dium, depending on the substrate temperature and theratio of the sodium and potassium vapour pres-sures [7]. In a tube these vapour pressures are nevernegligible, because of substitution reactions on the

40%r----------...,

Q

P. DOLlZY Philips tech. Rev. 40, No. I

in the presence of sodium are required than the addi-tion mechanism would indicate.

Depending onthe film thickness, the substitution ofpotassium by sodium can have a marked effect on thereflection. Fig. 12 shows how the reflectance at 520 nmvaries with the film thickness for the compoundsNa2KSb and Na-Sb, If the potassium cif Na2KSb issubstituted by sodium, the reflectance at a thicknessof about 30 nm (the first maximum) decreases; where-as at a thickness of 80 nm (the second minimum) thereflectance increases. Thus, to obtain a pure Na2KSbfilm with no Na-Sb, the reflectance at the first max-imum should be as high as possible and the reflectanceat the second minimum should be as low as possible.The homogeneity of a film can be tested by repeat-

ing the determination of k and n at various stages ofthe growth. A film can be said to be homogeneous if

R

t

Fig. 11. Calculated and measured reflectance at 520 nm during the growth of an NasSb film.a) Calculated reflectance R as a function of the thickness d of a film with k = 1.1 and n = 3.3.b) Detail of the reflectance R (in arbitrary units) measured at the start of the growth, as a functionof the evaporation time t. The start and finish of the antimony evaporations in the presence ofsodium are denoted by open and filled circles. The effect of these evaporations on the reflection isclearly demonstrated. c) Detail of (b), showing more clearly the effect of a single antimony evap-oration, starting at B and finishing at E. During the evaporation the reflection increases withtime because of the increase in film thickness as a result of the deposition of antimony in thepresence of sodium vapour. After the antimony evaporation has been stopped the reflectance firstcontinues to increase, because of the formation of a compound of similar composition to NasSb,owing to the reaction of the residual antimony with sodium. As soon as all the antimony hasreacted, the reflection remains virtually constant until the next evaporation.

walls. When both alkali metals are present on the tubewalls, heating one of the two alkali dispensers pro-duces a simultaneous pressure increase of the otheralkali metal (fig. 5). The difference from the rate ofreaction with antimony should therefore also be takeninto account. We have fOttnd that at the same pres-sure sodium reacts 1.3 times as fast as potassium. Asa result of the substitution reactions and the higherreaction rate of sodium, fewer antimony evaporations

the k- and n-values obtained from different pairs ofextreme values in the reflectance-thickness curve arethe same. On the other hand, a change in k or n in-dicates that there have been some changes in the filmduring the growth. We have found that a low and uni-form growth rate and a constant film compositionduring the growth favour the formation of homo-geneous Na2KSb films. The homogeneity is also betteras the growth due to addition reactions increases. In-

Philips tech. Rev. 40, No. 1 ALKALI-ANTIMONIDE FILMS FOR PHOTOCATHODES 27

creasing the homogeneity gives higher photosensitiv-ities because of the reduction in crystal imperfections,grain boundaries and stresses. In addition, the dimen- lIrsions of the microcrystals may become comparable to fthe film thickness, so that the photoelectrons can havea large escape depth.The presence of unwanted deposits on a grown film

can be detected easily, since they usually have a 50marked effect on the reflection. Fig. 13 shows howdrastically the reflectance at 520 nm for differentthicknesses of Na2KSb film can vary because ofadditional deposits of pure antimony and Na-Sb,

To obtain the highest possible photosensitivity thegrowth of a film must be stopped at the optimum film

-- Na2KSb--- Na3Sb

°0~--5~0~----1~00------5~0-n-m-d

Fig. 12. Measured reflectance R at 520 nm for an Na2KSb film andan NasSb film, as a function of the thickness d. The reflectancechanges drastically when Na2KSb is converted into NasSb.

4O%r-------::~~------"C"""O-- Na2KSb--- Na3Sb••••••••. Sb

20

20 40 60 80nm-d

Fig. 13. Measured reflectance R at 520 nm for an Na2KSb film as afunction of the thickness d, and the effects of antimony and NasSbdepositions on the reflection. These depositions were made onNa2KSb films of various thicknesses: 5, 10, 20, 30, 45 and 60 nm.The antimony deposition has a particularly marked effect on thereflection. .

?=800nm

..50 100 150

-dFig. 14. Photosensitivity ër in the transmission mode for differentwavelengths, as a function of the thickness d of an Na2KSb film.The photosensitivities are given in relative units; the maximum ofeach curve has been set equal to 100. The optimum thicknessdepends strongly on the wavelength À of the incident light: about10 nm for À = 430 nm, about 80 nm for À = 546 nm and about120 nm for À = 800 nm.

200nm

thickness. This optimum thickness depends stronglyon the wavelength of the incident radiation. Infig. 14the photosensitivity is shown as a function of thethickness of an Na2KSb film for three different wave-lengths. The optimum thickness increases with thewavelength: 10 nm for À. = 430 nm, 80 nm forÀ. = 546 nm and 120 nm for À. = 800 nm. If the lightto be detected has a low intensity, as in night-visionapplications, the optimum thickness of the photo-cathode film is 120 ± 10 nm.

Surface treatment of Na2KSb films

The photosensitivity of a pure and homogeneousNa2KSb film remains virtually constant for severalhours at 200°C. It increases by a factor of about 3 ifthe temperature falls to 20°C, but nevertheless re-mains fairly low tin A considerably higher photo-sensitivity can be obtained by applying a surface treat-ment in which caesium is evaporated on to theNa2KSb film. The highest efficiencies that we havemeasured for a test cell after the surface treatment areshown in fig. 15 as a function of the wavelength of theincident light. The curves relate to two Na2KSb films,one made mainly by substitution reactions and theother mainly by addition reactions. The additionmechanism gives the highest efficiency; the maximumphotosensitivity is 705 1lA/lm for an Na2KSb film ofabout 120 nm [12].

[u] D. E. Persyk, J. L. Ibaugh, A. F. McDonieand R. D. Faulker,IEEE Trans. NS-23, 186, 1976.

[12] The measurements were carried out with no amplification ofthe signal by an electric field.

28 ALKALI-ANTIMONIDE FILMS FOR PHOTOCATHODES Philips tech. Rev. 40, No. 1

A problem with this caesium treatment is that smallquantities of sodium and potassium are also presentin the evaporation chamber during the evaporation,as we have found from atomic absorption spectro-metry . Most of the sodium and potassium atomscome from the chamber walls as a result of the alkali-displacement reactions mentioned earlier. The caes-ium treatment of an Na2KSb film is therefore ratherdifficult to control and to study.

We have carried out special experiments to measurethe thickness of this top layer and analyse its chemicalcomposition. An Na2KSb film was deposited on halfof a glass substrate, with the other half covered by amask during the evaporation. The mask was thenremoved and the top layer was deposited over the.total substrate area. The evaporation was carried outat a temperature such that both parts were identicallycovered. We can therefore assume that the thicknessand composition of the top layer were the sameforboth halves. The thickness of this top layer was es-timated to be less than 0.8 nm from reflection meas-'urements. Complementary Auger experiments [13]

confirm this and also indicate that the top layer con-tains potassium as well as caesium and antimony. Thephotosensitivity of such a (K,Cs,Sb) top layer is of theorder of 1 !lA/lm, whereas an Na2KSb film of about100 nm coated with this top layer can have a photo-sensitivity of up to 350 !lA/lm.

Investigators have often wondered whether theatoms of the top layer continue to remain on top ofthe Na2KSb film or whether they partly diffuse into it.Although conclusive evidence is not yet available, wehave sufficient indications to lead us to think thatinterdiffusion of the alkali atoms between the toplayer and the Na2KSb film occurs along the grainboundaries. The actual surface composition of a caes-ium-treated Na2KSb film probably does not thereforecorrespond to the top layer originally deposited. Thehighest photosensitivities are obtained if the sodiumand potassium pressures are kept as low as possibleduring the caesium treatment. Slow growth of thephotocathode film is also beneficial: it limits the num-ber of grain boundaries so that the alkali diffusion isless significant.

One of the problems encountered in practice is thatundesirable substances such as sodium andpotassiumappear inside the tube during the caesium evapora-[lsl The equipment used for Auger speetrometry of multi-alkali-

antimonide layers has been described in [7].[14) W. H. McCarroll, R. J. PatI and A. H. Sommer, J. appl.

Phys. 42, 569, 1971.A. A. Dowman, T. H. Jones and A. H. Beck, J. Physics D 8,69, 1975.C. Ghosh and B. P. Varma, J. appl. Phys. 49,4549, 1978.R. Holtom, G. P. Hopkins and P. M. Gundry, J. Physics D12, 1169, 1979.L. G. EstelIa, Thesis No. 8016823, Stanford University, 1980.

100 _mAIW ,--V

.......... _,\\\\\\\I,,I,,,,,\,,

Q1L___ _L ~_L~ __ ~

400 600 800 1000nm

'rJr

t 10~

-À.

Fig. IS. Photoelectric spectral response /'fT for two Na2KSb filmswith a thin top layer containing caesium, measured in the trans-mission mode without electrical amplification. The solid curve wasobtained with an Na2KSb film of thickness about 110 nm, grownmainly by substitution reactions. The maximum photosensitivity ërof this film was about 545 !1A/lm. The.dashed curve was obtainedwith an Na2KSb film of thickness about 120 nm, grown mainly byaddition reactions. This film had a significantly higher spectral re-sponse, with a maximum photosensitivity of 705 )lA/lm.

tion. There have been many investigations of theseeffects [13] [14], but no unambiguous conclusions canas yet be drawn for the best top-layer composition forhigh photosensitivity. We have therefore initiatedspecial experiments using the molecular-beam tech-nique to help us to study and improve the top-layerdeposition. We have for example been able to demon-strate that a top layer with about one monolayer ofcaesium gives an appreciably higher photosensitivitythan a top layer of the (Cs,Sb) or (K,Cs,Sb) type.

In the investigations described here useful discus-sions and technical support were given by J. Beltra-melli, M. Decaesteker, M. A. Deloron, O. De Luca,F. Grolière, J. Houdard, F. Maniguet, G. Marie andC. Piaget, all of LEP.

Summary. The growth of alkali-antimonide films for photocath-odes on an appropriate substrate (such as glass) takes place in a vac-uum chamber by chemical reactions between antimony depositedon a substrate and the vapour of alkali metals. The growth can bestudied by measurements of the atomic absorptions in the alkalivapour, the optical reflection and transmission of the film, and thephotosensitivity in the reflection and transmission modes. Thesemeasurements allow the partial alkali-vapour pressures, the opticalproperties, thickness and growth rate of the film to be evaluatedduring the growth, as well as the escape depth and escape probabil-ity of the photoelectrons. As has been demonstrated for Na2KSbfilms, growth monitoring gives useful information about the prefer-red chemical reaction mechanism, the optimum film compositionand the thickness of the film. It provides a basis for improvement ofthe growth conditions, so as to increase the photosensitivity andreproducibility. To obtain an Na2KSb film of high photosensitivityit must be coated with a very thin top layer of caesium. Molecular-beam experiments indicate that the best results are obtained with amonolayer of pure caesium.

Philips tech. Rev. 40, No. 1 29

Scientific publicationsThese publications are contributed by staff of laboratories and plants that form part ofor cooperate with enterprises of the Philips group of companies, particularly by staff ofthe following research laboratories:

Philips Research Laboratories, Eindhoven, The Netherlands EPhilips Résearch Laboratories, Redhill, Surrey RHI 5HA, England RLaboratoires d'Electronique et de Physique Appliquée, 3 avenue Descartes,

94450Limeil-Brévannes, France LPhilips GmbH Forséhungslaboratorium Aachen, WeiBhausstraBe, 51 Aachen,

Germany APhilips GmbH Forschungslaboratorium Hamburg, Vogt-Kölln-StraBe 30,. 2000 Hamburg 54, Germany HPhilips Research Laboratory Brussels, 2 avenue Van Becelaere, 1170 Brussels

(Boitsfort), Belgium BPhilips Laboratories, N.A.P.C., 345 Scarborough Road, Briarcliff Manor,N.Y. 10510, U.S.A. N

H. A. Algra, L. J. de Jongh (University of Leiden) &.J. Reedijk (Delft University of Technology): Specificheat near the critical concentration for the dilutesimple-cubic magnet COpZnl-p(C5H5NO)6(CI04)2.Phys. Rev. Letters 42,606-609,1979 (No. 9). E

P. M. Asbeck, D. A. Cammack, J. J. Daniele, D. Lou,J. P. J. Heemskerk, W. J. Kleuters & W. H. Ophey:High-density optical recording with (Ga,Al)As DHlasers. .Appl. Phys. Letters 34,835-837,1979 (No. 12). N, E

C. Belouet: Vapour growth in a microgravity environ-ment.Thin Solid Films 58, 1-8, 1~79 (No. I). L

F. Berz, R. W. Cooper & S. Fagg: Recombination inthe end regions of pin diodes.Solid-State Electronics 22,293-301, 1979 (No. 3). R

J. Bloem: Problems in the melt- and vapor growth ofsilicon for integrated circuits and solar cells.J. solid State Chem. 27,19-27,1979 (No. I). E

F. R. de Boer, W. H. Dijkman, W. C. M. Mattens (allwith University of Amsterdam) & A. R. Miedema: Onthe valence state of Yb and Ce in transition metal inter-metallic compounds.J.less-common Met. 64, 241-253,1979 (No. 2). E

M. R. Boudry & J. P. Stagg: The kinetic behavior ofmobile ions in the Al-Si02-Si system.J. appl. Phys. 50, 942-950, 1979 (No. 2). R

M. Boulou, M. Furtado, G. Jacob & D. Bois: Recom-bination mechanisms in GaN:Zn.J. Luminescence 18/19,767-770,1979 (Part Il). L

J. C. Brice & A. M. Coleï Infrared absorption in a- 'quartz.J. Physics D 12, 459-463, 1979 (No. 3). R

J. W. Broer: Professional re-identification in thewriter/scientist.Proc. 26th Int. tech. Comm. Conf., Los Angeles 1979,pp. W 27-31. E

K. H. J. Buschow & N. M. Beekmans: Magnetic andelectrical properties of amorphous alloys of Gd and C,Al, Ga, Ni, Cu, Rh or Pd.Rapid quenched metals Ill, Proc. 3rd Int. Conf.,Brighton 1978,Vol. 2, pp. 133-136. E

K. L. Bye: An X-ray topographic assessment of cad-mium mercury telluride.J. Mat. Sci. 14,619-625, 1979 (No. 3). R

S. R. Chinn (M.I.T., Lexington, Mass.) & W. K.Zwicker: FM mode-locked Ndo.5Lao.5P5014 laser.Appl. Phys. Letters 34, 847-849,1979 (No. 12). N

T. A. C. M. Claasen: Comments on 'The absolutestability of high-order discrete-time systems utilizingthe saturation nonlinearity'.IEEE Trans. CAS-26, 138-140, 1979 (No. 2). E

T. A. C. M. Claasen & W. F. G. Mecklenbräuker:Application of transposition to decimation and inter-polation in digital signal processing systems.1979 IEEE Int. Conf. on Acoustics, speech & signalprocessing, Washington D.C., pp. 832-835. E

J. A. Clarke: Aspheric mirror optical systems for theinfra-red.2nd Int. Conf. on Low light and thermal imaging, Not-tingham 1979 (lEE Conf. Publn No. 173), pp. 18-19.R

D. J. Coe & H. E. Broekman: Corner breakdown inMOS transistors with lightly-doped drains.Solid-State Electronics 22,444-446, 1979 (No. 4). R

N. H. Dekkers: Object wave reconstruction in STEM.Optik 53, 131-142,1979 (No. 2). E

M. DeIfino: A comprehensive optical second harmonicgeneration study of the non-centrosymmetric characterof biological structures.Mol. Cryst.liq. Cryst. 52, 271-284,1979 (No. 1-4). N

P. Delsarte: Bilinear forms over a finite field, withapplications to coding theory.J. combin. Theory A 25,226-241,1978 (No. 3). B

30 SCIENTIFIC PUBLICATIONS Philips tech. Rev. 40, No. 1

P. Delsarte, Y. Genin & Y. Kamp: A simple approachto spectral factorization.IEEE Trans. CAS-25, 943-946, 1978 (No. 11). . B

P. Delsarte, Y. Genin & Y. Kamp: An equivalence rela-tion in planar least squares inverse approximation.Proc. IEEE 66, 1662, 1978 (No. 12). B

P. A. Devijver: Nonparametrie estimation of featureevaluation criteria.Pattern recognition and signal processing, ed. C. H.'Chen, pp. 61-82; Sijthoff & Noordhoff, Alphen aanden Rijn 1978. B

P. A. Devijver: On the amount of information con-veyed by nearest neighbors and its use in pattern re-cognition. 'Théorie de l'information, Coli. Int. CNRS No. 276,Paris 1978, pp. 353-362. B

C. Z. van Doorn & J. J. M. J. de Klerk: Two-frequencylOO-line addressing of a reflective twisted-nematicliquid-crystal matrix display.J. appl. Phys. 50,1066-1070,1979 (No. 2). E

J. W. F. Dorleijn (Philips Lighting Division, Eind-hoven) & A. R. Miedema: The anomalous Hall effectinnickel alloys: contributions from skew scattering andside displacement in the two-current model.J. Magn. magn. Mat. 12,26-30,1979 (No. I). E

G. Engelsma: Effect of daylength on phenol meta-bolism in the leaves of Salvia occidentalis.Plant Physiol. 63, 765-768,1979 (No. 4). E

W. van Erk: A solubility model for rare-earth irongarnets in a PbO/B20s solution.J. Crystal Growth 46, 539-550, 1979 (No. 4). E

R. M. van Essen & K. H. J. Buschow: Hydrogenabsorption in various zirconium- and hafnium-basedintermetallic compounds.J. less-common Met. 64, 277-284,1979 (No. 2). E

L. F. Feiner & R. P. van Stapele: Ontaarde elektronen-banen en structurele faseovergangen.Ned. T. Natuurk. A 44, 111-114,1978 (No. 3). E

W. E. Fischer: PHIDAS - a database managementsystem for CAD/CAM application software.Computer-aided Design 11, 146-150, 1979 (No. 3). H

B. Fitzhenry: Identification of a charging mechanismusing infrared spectroscopy.Appl. Spectrosc. 33,107-110,1979 (No. 2). N

N. Fleurot, M. Nail, R. Verrecchia (all with CEA,Villeneuve-Saint-Georges) & G. Clément: Character-ization of image converter tubes and photodiodes inthe infrared region.Proc. 13th Int. Congress on High speed photographyand photonics, Tokyo 1978,pp. 440-442; 1979. L

J. A. Geurst: Mutual friction in the laminar flow ofsuperfluid helium IJ through capillary tubes.Physics Letters 71A, 78-82, 1979 (No. I). E

J. P. Gex, R. Sauneuf (both with CEA, Villeneuve-Saint-Georges), J. P. Boutot & J. C. Delmotte: Somenew possibilities in direct visible and X-ray measure-ments.Proc. 13th Int. Congress on High speed photographyand photonics, Tokyo 1978,pp. 405-408; 1979. L

R. W. Gibson & R. Wells: Thepotential of SSB forland mobile radio.29th IEEE Vehicular Technology Conf., ArlingtonHeights, Ill., 1979, pp. 90-94. R

A. A. van der Giessen (Philips Electro-acousticsDivision): Audio recording tapes based on iron par-ticles.J. Audio Engng. Soc. 26, 838-842, 1978 (No. 11).

J.-M. Goethals: Combinatorial decoding methods forblock codes.Théorie de I'information, ColI. Int. CNRS No. 276,Paris 1978, pp. 223-231. B

W. J. A. Goossens: Temperature dependence of thepitch in cholesteric liquid crystals: a molecular statisti-cal theory.J. Physique 40, C3/158-163, 1979 (Colloque C3). E

R. G. Gossink & T. P. A. Lommen: Secondary-ionmass speetrometry (SIMS) analysis of electron-bom-barded soda-lime-silica glass.Appl. Phys. Letters 34,444-446, 1979 (No. 7). E

L. H. Guildford: Experiments in real-time image pro-cessing.2nd Int. Conf. on Low light and thermal imaging, Not-tingham 1979 (lEE Conf. Publn No. 173),pp. 55-56. R

P. Guittard, P. Jarry, C. Piaget, J. C. Richard, 'E. Roaux & P. Saget: GaAs photocathodes for lowlight level imaging.2nd Int. Conf. on Low light and thermal imaging, Not-tingham 1979 (lEE Conf. Publn No. 173), pp. 24-25. L

S. Herman: Nonlinear capacitors improve the per-formance of saturable lead ballasts.J. Illum. Engng. Soc. 8,122-125,1979 (No. 3). N

W. J. van den Hoek & J. A. Visser (Philips LightingDivision, Eindhoven): Are oscillations in a horizontalrare-earth metal iodide/cesium iodide/mercury are in-duced by an external magnetic field.Appl. Phys. Letters 34,357-359, 1979 (No. 6).

M. H. H. Höfelt: On the stability of a l-bit-quantizedfeedback system.1979 IEEÈ Int. Conf. on Acoustics, speech & sigrialprocessing, Washington D.C., pp. 844-848. E

K. Holford: Front ends for Doppler with-sense radar.Electronics Letters 15,75-76,1979 (No. 3). R

K. Holford: Microwaves can control traffic lights forfire appliances.Fire, March 1979, pp. 507-508. R

L. Honds & H. Meyer: Auswirkungen von Oberwellenauf die Materialfunktion im Hysteresemotor.etz-Archiv 1, 187-190, 1979 (No. 6). A

Philips tech. Rev. 40, No. 1 SCIENTIFIC PUBLICATIONS 31

IH. Jhrig: Reply to 'Comment on "A systematic exper-imental and theoretical investigation of the grain-boundary resistivities of n-doped BaTi03 ceramics"'.J. appl. Phys. 50,1158-1159,1979 (No. 2). A

G. D. Khoe: Practical machine for electric are splicingof optical fibres in the field.Electronics Letters 15, 152-153, 1979 (No. 5). E

G. D. Khoe, H. G. Koek & L. J. Meuleman: Fiberlesshermetic packaged lens-coupled laser diode for wide-band optieal-fiber transmission.Optical fiber communication, Dig': tech. Papers TopicalMeeting Washington D.C. 1979, pp. 94-97. E

L. C. Kimerling*, P. Blood & W. M. Gibson* (* BellLaboratories, Murray Hill, N.J.): Defect states inproton-bombarded silicon at T< 300 K.Int. Conf. on Defects and radiation effects in semi-conductors, Nice 1978 (Inst. Phys. Conf. Ser. No. 46),pp. 273-280; 1979. R

M. Klinek: Control of the surface-water purificationplant for the Amsterdam Water-Supply Authority.Journal A 20,59-70, 1979 (No. 2). H

G. Kowalski: Multislice reconstruction from twin-conebeam scanning.IEEE Trans. NS-26, 2895.-2903,1979 (No. 2, Part 2). H

G. Kowalski & W. Wagner: Patient dose rate: An ulti-mate limit for spatial and density resolution of scan-ning systems.Biomed. Technik 24, 38-42, 1979 (No. 3). H

H. K. Kuiken: The cooling of low-heat-resistance cyl-inders by radiation.J. Engng Math. 13, 97-106, 1979 (No. 2). E

G. Kurtze, C. KIingshirn (both with Universität Karls-ruhe), B. Hönerlage (Laboratoire de Spectroscopie etd'Optique du Corps Solide, Strasbourg), E. Tomzig(Universität Erlangen) & H. Seholz: Excitation spec-troscopy, Raman scattering and the temperature de-pendence of the luminescence in highly excited redHgh.J. Luminescence 20,151-161,1979 (No. 2). A

J. van Laar: Foto-emissie van vaste stoffen.Ned. T. Natuurk. A 44,91-95, 1978 (No. 3). E

G. Laurenee, F. Slmonder & P. Saget: CombinedRHEED-AES study of the thermal treatment of (001)GaAs surface prior to MBE growth.Appl. Phys, 19,63-70, 1979 (No. I). L

F. H. de Leeuw, W. de Geus & P. Q. J. Nederpel: Highspeed photography, of moving domain walls inmagnetic bubble materials.Proc. 13th Int. Congress on High speed photographyand photonics, Tokyo 1978, pp. 317-320; 1979. E

P. A. Lewis & M. J. Underhill: Quiet tuning andmatching of antennas for radio silence operation.Proc. lEE Conf. on Recent advances in h.f. communi-cation systems and techniques, London 1979 (lEEColloq. Dig. 1979/48), pp. 31-44. R

G. M. Martin, M. L. Verheijke, J. A. J. Jansen & G.Poiblaud (RTC, Caen): Measurement of the chromiumconcentration in semi-insulating GaAs using opticalabsorption.J. appl. Phys. 50, 467-471,1979 (No. I). L, E

R. Memming & F. Sehröppel: Electron transfer reac-tions of excited ruthenium(II) complexes in monolayerassemblies at the SrrOa-water interface.Chem. Phys. Letters 62,207-210, 1979 (No. 2). H

A. D. Mills, J. Mackenzie & R. J. Dolphin': The use ofa microcomputer for flexible automation of a liquidchromatograph.J. autom. Chem. 1, 134-140,1979 (No. 3). R

A. Mireea & D. Bois (INSA, Villeurbanne): A reviewof deep-level defects in Ill-V semiconductors.Int. Conf. on Defects and radiation effects in semicon-ductors, Nice 1978 (Inst. Phys. Conf. Ser. No. 46),pp. 82-99; 1979. L

B. J. Mulder: Unbacked ultra-thin films of berylliumand other metals.J. Physics E 12, 268-269,1979 (No. 4). E

A. van Oostrom, L. Augustus, A. Steinmetz* & G. v.d.Berg* (* Philips Telecommunications Industries, Hil-versum): Analysis of surface resistance and elementalcomposition of a reed contact.Electrical contacts 1978, Proc. 9th Int. Conf. &. 24thAnn. Holm Conf., Chicago, pp. 521-526. : E

R. Orlowski & E. Krätzig: Holographische Speicherungin elektrooptischen Kristallen.Laserspektroskopie, ed. F. Aussenegg (Acta PhysicaAustriaca, Suppl. XX), pp. 241-255; Springer, Wien1979. H

J. A. Pals & J. Wolter: Measurement of the order-parameter relaxation in superconducting Al-strips.Physics Letters 70A, 150-152, 1979 (No. 2). E

J. B. H. Peek & J. M. Sehmidt: Een 'Station Pro-gramma Identificatie' systeem voor F.M.-radio-omroep.T. Ned. Elektronica- en Radiogen. 44, 25-29, 1979(No. I). E

P. van Pelt & E. E. Havinga: Electrochromism oforganic dyes at high fields.Nonlinear behaviour of molecules, atoms and ions inelectric, magnetic or electromagnetic fields, ed. L. Néel,pp. 291-300; Elsevier, Amsterdam 1979. E

D. Polder, M. F. H. Sehuurmans & Q. H. F. Vrehen:Superfluorescence: Quantum-mechanical derivation ofMaxwell-Bloch description with fluctuating fieldsource.Phys. Rev. A 19, 1192-1203, 1979 (No. 3). E

D. Pons, A. Mireea, A. Mitonneau & G. M. Martin:Electron traps in irradiated GaAs: comparison withnative defects.Int. Conf. on Defects and radiation effects in semicon-ductors, Nice 1978 (Inst. Phys. Conf. Ser. No. '46),pp. 352-359; 1979. L

32 SCIENTIFIC PUBLICATIONS Philips tech. Rev. 40, No., 1

G. Prast: A thermal tracking solar collector.Proc. DFVLR Int. Symp. on Solar thermal power sta-tions, Cologne 1978, section 10, pp. 1-8. E

J. L. Robert, B. Pistoulet, A. Raymond, J. M. Dus-seau (all with Centre d'Etudes d'Electronique desSolides, Montpellier) & G. M. Martin: New model ofconduction mechanism in semi-insulating GaAs.J. appl. Phys. 50, 349-351,1979 (No. I). L

P. C. Scholten: Colloid chemistry of magnetic fluids.Thermomechanics ofmagneticfluids, ed. B. Berkovsky,pp. 1-25; Hemisphere, Washington D.C. 1978. E

H. Schomberg: Reconstruction of spatial resistivitydistribution of conducting objects from external re-sistance measurements.Z. angew. Math. Mech. 59, T 41-42,1979 (No. 3). H

H. Schomberg: Monotonically convergent iterativemethods for nonlinear systems of equations.Numer. Math. 32, 97.-104, 1979 (No. I). H

P. J. Severin: Calorimetrie measurement ofthe absorp-tion coefficient of fibre-quality compound glass rods.Óptica hoy y mafiana, Proc. ICO-U, Madrid 1978,pp. 499-502. E .

P. J. Severin: Isotope separation by chemical vapourdeposition and related processes.J. Crystal Growth 46, 630-636,1979 (No. 5). E

M. Sintzoff: Properties, feasibility and usefulness of alanguage for programming programs.ALGOL Bull. No. 43, 84-90, 1978. B

M. J. Sparnaay: Transfert d'électrons et d'ions danset sur les semi-conducteurs.Le Vide 33,236-237,1978 (No. 194). E

W. Spiesberger: Mammogram inspection by computer.IEEE Trans. BME-26, 213-219, 1979 (No. 4). H

B. Steinmüller & R. Bruno: The energy requirementsof buildings.Energy and Buildings 2, 225-235, 1979 (No. 3). A

A. L.N. Stevels: Red Mn2+-luminescencein hexagonalaluminates.J. Luminescence 20, 99-109, 1979 (No·.2). E

F. L. H. M. Stumpers: Some notes on the scientificwork of H. Bremmer .

. Radio Sci. 14, 179-188, 1979 (No. 2). E

J. B. Theeten & F. Hottier: In situ surface analysis ofthe vapor phase epitaxy of GaAs.J. Electrochem. Soc. 126, 450-460,1979 (No. 3). L

J. B. Theeten, F. Hottier & J. Hallais: Ellipsometricassessment of (Ga,Al)As/GaAs epitaxiallayers duringtheir growth in an organometallic VPE system.J. Crystal Growth 46,245-252,1979 (No. 2). L

J. J. P. Valeton (Philips .Blectro-acoustics Division):Electronische signaalbehandeling in televisiecamera's.T. Ned. Elektronica- en Radiogen. 44, 9-16, 1979(No. I).

H. Vantilborgh: Exact aggregation in exponentialqueueing networks.J. Ass. Computing Machinery 25, 620-629, 1978(No.4). B

H. Verweij: Raman study of the structure of alkaligermanosilicate glasses: I. Sodium and potassiummetagermanosilicate glasses, 11. Lithium, sodium andpotassium digermanosilicate glasses.J. non-cryst. Solids 33, 41-53, 55-69, 1979 (No. I). E

H. Verweij, J. H. J. M. Buster & G. F. Remmers(Twente University of Technology, Enschede): Re-fractive index and density of Li-, Na- and K-germano-silicate glasses.J. Mat. Sci. 14, 931-940,1979 (No. 4). E

A. T. Vink, C. J. Werkhoven & C. van Opdorp: Largedefects: observation and influence on minority carrierrecombination.Semiconductor characterization techniques, Electro-chem. Soc. Proc. Vol. 78-3, pp. 259-288, 1978. E

W. Wagner: Reconstructions from restricted regionscan data - new means to reduce the patient dose.IEEE Trans. NS-26, 2866-2869, 1979 (No. 2, Part 2). H

H. W. Werner & A. E. Morgan: Secundaire-ionen-massaspectrometrie.Ned. T. Natuurk. A 44, 122-125, 1978 (No. 3). E

J. S. van Wieringen: Exact and approximate solutionof the regenerator equation for the case of high heatexchange and moderate heat capacity.Appl. sci. Res. 34, 145-158,1978 (No. 2/3). E

J. P. Woerdman: Laser-excited broadband violet emis-sion from sodium vapor.J. Opt. Soc. Amer. 68, 714, 1978 (No. 5). E

M. de Zwart & C. Z. van Doom: The field-inducedsquare grid perturbation in the planar texture ofcholesteric liquid crystals. .J. Physique 40, C3/278-284, 1979 (Colloque C3). E

Published in Gallium arsenide and related compounds,1978 (Proc. 7th· Int. Symp., St. Louis; Inst. Phys.Conf. Ser. No. 45, 1979):

C. E. C. Wood, J. Woodcock & J. J. Harris: Low-compensation n-type and flat-surface p-type Ge-dopedGaAs by molecular beam epitaxy (pp. 28-37). RG. B. Scott & J. S. Roberts: Photoluminescence in111-V compounds grown by MBE (pp. 181-189). RW. J.Bartels &H. Veenvliet: X-ray study of Gal-xAIxAsepitaxial layers grown with the metallorganic VPEtechnique on GaAs substrates (pp. 229-238). ECh. Hurtes, L. Hollan & M. Boulou: Impurity charac-terization of GaAs high-resistivity VPE layers for FETdevices (pp. 342-352). LJ: Hallais, J. P. André, P. Baudet &D. Boccon-Gibod:New MESFET devices based on GaAs-(Ga, AI)Ashetèrostructures grown by metallorganic VPE (pp.361-370). . L

Volume 40, 1982,No. 1 Published 16th March 1982pages 1-32