HELIUM IONMCROSCOP H

Tiie physics of secondary electrongeneration and ion backscatteringallows for the collection of imageinformation not possible with ascanning electron microscope insome applications.First results are discussed here.

L Scipioni"", L A. Stern, j . Nolte,S. SijhrandijCarl Zeiss SMT Inc.Peabody, Massachusetts

B. GriffinUniirrsity of Western Australia

ii, Atislnilia

T he Orion microscope is based on the firstcommercially manufactured helium fieldion gun. Several attempts have been madeto design a microscope based on the gas

field ion source. Now, the Atomic Level Ion Source(ALIS) technology provides a stable helium ionbeam with extremely high brightness and thus asub-nanometer probe size on the sample.

The source is based on field ion emission fromthe apex of a sharp needle held at a high positivevoltage in the presence of a gas. The technology ofthe field ion microscope (FIM) is based on this phe-nomenon, and has itself been used in materialsanalysis. Figure 1 a shows an illustration of the typ-ical hemispherical end-form of a needle of a singlecrystal, and the multiplicity of ion beams emitted.If these are collected on an imaging screen, a pat-*Meinber of ASM International

Iff;

tern such as that in Figure lc results, indicating theposition of the atoms and the crystal facets.

In the ALIS emitter, a tip reforming process causesa small number of atoms to protrude more than theothers on the crystal axis (Fig. Ib). These becometile only emitters; thus, they carry all of the ion cur-rent (Fig. Id). One emitter is selected to travelthrough the ion optics, vielding a single atom emitterwith high brightness (B > 4x10" A-cm- srO- Tlieelectrostatic optics are theoretically capable of pro-ducing a focused probe with a diameter of 0.25 nm(spot size down to 0.5 nm has been measured sofar.)

Material-sensitive imagingIn addition to imaging the topology of a surface or

particle, spatially resolved information is oftenneeded for other properHes, such as chemical com-position, grain structure, density, conducti\'ity, andthe like. Acquiring such information in a chargedparticle microscope requires that the beam interactwith the sample in specific ways.

Following are images that illustrate ways that thehelium ion microscope can provide high-sensitivitymaterial information on surfaces.

Figiire 2a shows surface Information from goldparticles on carbon. The image shows a resolutionof about 2 nm. Notable is the amount of surface de-tail. This is due to the low energy of the secondaryelectrons (SE), mostly below 5 eV, so that the infor-mation is coming from just the top few nanometers.Compare this to the SEM image in Figure 2b, wherethe surface character of the particles is completelyabsent because of the higher energy SE's thatcome from deeper below the surface, as well as tlietype II SE's. These serve to delocalize the probeinformation.

The helium ion microscope image also shows dis-tinct contrast levels corresponding to the grains in

Fig. 1 — (a) Illustration (side viiiv) of the eiid-fonn of a typical field ion tip. (b) Illustration o/ the ALIS entitter. (c) Actual FIM image of thegeometry illustrated in la. (d) Actual FIM imageof ALIS emitter illustrated in Ib.

ADVANCED MATERIALS & PROCESSES/JUNE 2008 27

Fig. 2 — Gold particles U!U!\;CL! :ntli (a) a helium ion beam, (b) an electron beam.Magnification is about 250.000X in each image. Vie arrows in the Orion imageindicate afeiv ofthe many places ivhcre grain kmndaries are seen.

Fig. 3 — NBPT monolayer overagoldfrlm. patterned with an electron beam mask,(a) Orion image, (b) SEM image.

Fig. 4 — Metal whisker formed on a thi-coated copper electrical connector, (n)Orion S£ image, shozrhig cimnges in contrast lei>els along the whisker, (b) OrionRBI image, showing little to no material contrast. Cren: arrows point at identicalsites in each image where a transition is seen in the whisker.

the material. Although grain boundaries are visiblein the SEM image, the resolution is compromised,and the contrast between the various grains is min-imal. Thus, an entirely new view of tliis material be-comes possible with the helium ion microscope.

Another important aspect of imaging is chemicalor material contrast. How well can materials be dif-ferentiated in an image based on the SF signal? Itturns out that the SF yield is a more sensitive fujic-tion of material for an incident ion beam than for anincident electron beam. Thus, in helium ion mi-croscopy, the images of a given sample tend to dis-play a greater range of gray levels than a correspon-ding SFM image. This may be attributable to thefact tliat the energy loss for an ion beam at this en-ergy is dominated by nuclear-nuclear scatteringrather than electron scattering.

The greatly increased surface sensitivity also isbeneficial for the investigation of surface chemicalstates. Tliese factois could benefit the analysis of cat-aiysts, which depend on surface properties for theirbehavior and also for any type of nano-patteming

AcknowledgenientsThe authors thank

Korsten RottondAndre Beyer of the

University ofBielefeld, Germany,

who kindlyprepared andprovided the

patterned NBPTmonolayer somple

discussed herein,and Gary Stupian of

The AerospaceCorporation for thetin whisker sample.

that involves modification of surface states.An example of this is seen in Figure 3. This

sample (courtesy of A. Beyer, Bielefeld University,Germany) is a chemically patterned, self-assembledmonolayer of nitrobiphenyl thiol (NBPT) formedon a gold film. A contact eiectixm beam stencil maskwas placed over this layer. The mask had a 40-nmframe within which were multiple electron trans-parent test windows in the form of circles andellipses.

Upon electron beam exposure, the terminal NO2group at the NIBFT surface beneath these windowswas reduced to NH2 — a change affecting just thetop atomic layer of the sample. Previously, thischange was orily detected in AFM in friction mixJe,for there is neither topology nor detectable SEM con-trast (Fig. 3b).

However, the helium ion microscope makes thepatterned areas highly visible, as demonstrated inFigure 3a, Tbis capability opens tlie door to manymore opportunities to investigate such chemicallymodified surfaces, Note also that it is possible withboth helium ion microscopy and SEM to imagethrough the organic monolayer to the gold filmbelow. Only the ion beam technology allows for de-tection of both.

Rutherford imagesA unique feature of the Orion microscope is the

ability to acquire Rutherford backscattered ion im-ages (RBI). This mode highlights materials in thesample based upon atomic number "Z" contrast,and provides complement<iry information to tlie SEimage. The helium ion microscope is capable of cap-turing both images simultaneously through twoparallel detection channels, so that identical condi-tions are assured for both.

For example, this capability may be utilized in amaterials failure analysis. Tlie images in Fig. 4 arefrom a sample of a tin wliisker that has grown out ofan ordinary tin-coated copper electrical connector.These whiskers grow over time and can eventtiallyextend from one electrical connector to another,causing failures in electronics. Several spacesatellites have even suffered partial or total lossof functionality due to this phenomenon. Althoughcommon and well-studied, the fundamentalsbehind the problem are still not completelyundei-sttxjd.

The images in Fig. 4 a, b show complementary in-formation from this whisker. The SE image from tliehelium ion microscope shows the toptilogy of the ob-ject. The fact tliat tlie different sections ak>ng its lengtliare at different contrast levels indicates that stime dis-crete changes take place as the whiskers grow. Tliesechanges were seen in several places, particularlywhere the whiskers have a kink or bend.

The arrows ui the figuivs point to such areas. TheRBI image (also produced on the helium ion micro-scope), on the other hand, shows that the materialis uniform all along the length. Tliis hints that thephysical or electronic structure of tlie tin is changingin some way, rather tlian a material variation. How-ever, the RBI image does reveal that a different ele-ment (with a higher backscatter yield) is located di-

28 ADVANCED MATERIALS & PROCESSES/JUNE 2008

BrukerAXS Handheld

rectly at the transition points on the whisker. Theseare seen at the two places indicated by arrows,where the contrast is brighter in a small band. Thisinformation could guide analysis of the growth dy-namics, Note also that the many particles seen onthe whisker in the SE image are invisible in the RBIimage, indicating that they are most likely also tin,the same as the whisker.

Elemental analysisTlie collection of backscattered helium ions also

provides the potential for quantitative analyses ofmaterials. Tlie intensity, energy spectrum, and an-gular distribution of the backscattered ions all canyinformation about the sample under the primaryLx'am. We are just beginningto delve into the details ofthese processes to derivechemical element information.

The most easily accessibleinformation is the total inten-sity of the backscatter signalas a function of the Z of thetarget. It Ls generally tRie tJiatthe prt)bability of a back-scatter event increases withZ. We carried out a set of ex-periments to evaluate this. Asample was prepared thatconsisted of a copper Lilockwith 44 holes into which werepressed, with indium as acement, pure elemental ma-terial from every conductingelement from boron to bis-muth.

Relative measurement oftlie RBI yield could be takenby comparing tlie image graylevel from each of the sam-ples. After controlling forbeam current variations andcollector efficiency variationacross the stage (via collectionof a reference copper area ineach image), we determinedtlie RBI yield curve plotted inFig. 5. Although the yieldgenerally increases with Z,pnmiinent peaky are obviovLs,one each in periods 4,5, and6 of the periodic table. Tlieycorrespond to the samples ofcopper, silver, and gold. Thisphenomenon is not predictedin TRIM simulations, whichcia* shown by tlie solid line indie plot. This behavior is as ofyet unexplained. The maximapreclude a unique identifica-tion of an element based onsignal intensity, but the im-ages still provide stime infor-mation about the make-up ofthe sample.

Penod Au6 *

0.0510 20 30 40 50 60 70 80 90 too

Atomic number

Fig. 5 — This graph shows the helium backscatter rate for elemental targets.Experimental (fwhits) and simulated (line) backscatter yields are shown for 25 keVhelium ions in conducting elemental targets.

BRUKER

Positive Material Identification (PMI)

S I TRACER Handheld XRF Analyzer

• Fast, accurate, and dependable analysis• Nondestructive testing with Grade-ID and Chemistry• Largest alloy library on the market• Light Element capability, Mg, AL, Si. P" Prevent material mix-ups CnHfM By I I H Spin Faunlathin

For more information, please visit www.brukerhandhetd.comOr email info@brukerhandheld.com

think forwardJ Handheld

ADVANCED MATERIALS & PROCESSES/JUNE 2008 29

10.000

1000

Oxygen Silicon Copper Gold

Au/Cu

10

10 12 14 18 20 22Helium ton energy

Backscatteredhelium energij

spectra from thingold films onsubstrates of

copper, silicon,and glass (SiOi).

Also marked isthe expected

kinematic edgefor each of the

elements foundin these sauiples.

Primanj beamenergi/ was

b

Quantitative informationFor more quantitati\'e infonnation, we turn to the

measurement of low energy Rutherford backscat-tering spectra. Tlie ion has a probability of under-going a large angle scattering event through whichit is ejected from the material. The maximum en-ergy of a backscattered ion is found theoretically forthe case in which the particles collide at the firstmonolayer. Then the recoil energy is a fraction ofthe incoming energy, as determined by kinematiccalculations: The higher the target atom mass, thelarger the recoil energy.

Ions, which penetrate more deeply into the ma-terial before backscattering, lose energy at a rate de-termined by the stopping power. Thus, collecting asufficient number of these ions with an energy-sensitive detector will reveal a spectioim with anedge at the kinematically defined energy, and a tailextending theoretically down to zero energy. The

shape and position of such a cur\'e can be utilizedto identify materials and measure layer thickness.

An example of our initial explorations into thisteclinique is illustrated in Fig. 6. Here we kxiked atthin films of gold on substrates of copper, silicon, andglass (SiO2). Calculatijig from the primary beam en-ergy of 30 keV and the mass ofthe various materialsbeing probed, one would ideally expect edges in thespectra at the line positions indicated on the plot.

Although tlie instnimentation factors broadenedthe spectral features in this experiment, it is obviousthat we can both detect the substrate and distin-guish the materials one from another. See for ex-ample the evidence of the oxygen in the glasssample, which is absent from the silicon sample.The double humps in these peaks can also enablegold layer thickness determination.

As the technology is utilized by more researchersutilizing more of the analytical options on the ttiol,our understanding of tlie potential of this techniquewill grow even further. € •

For more infonnation: L. Scipioni, Carl Zeiss SM T Inc.,One Corporation Way, Peabody, MA 01960; I.scipioni®smt.zeiss.com; www.smt.zeiss.com.

Bibliography1. V. N. Tondare, 1. Vac. Sci. Technol. A, 73 (6), p.l498 {20)5).2. J. Morgan et. a\.. Microscopy Today, 14 (4), p.24 (2(X16),3. M. K. MiUer and C. D. W. SmitlV, "Atom Probe Micro-analysis: Principles and Applications to Materials Prob-lems", Materials Research StKiety, Pittsburgh, ]9H9.4. J.H. Richardson, and B.R. Lasley, "Tin Whisker Initi-ated Vacuum Metal Arcing in Spacecraft Electronics,"1992 Govemment Microcirctiit Applications Conference,Vol. XVIII, pp. 119 -122, November 10 -12,1992.

Wemanufacture...

patent pendingcomposite alloy

wires madeexclusively forthe Twin Wire

Arc Sprayprocess, anddesigned tosolve your

mostchallenging

coating is

Call Us Toda,.

314-801-6""

47.14 Eiirth City Exp

Bridgelon. MO W044\v\vu .arcmclt.com

04

Bridging the Gap Between Science and Appiication

ADVANCED AAATERiALS & PROCESSES/JUNE 2008