, ,

Note: This is a preprint of a paper being submitted for publication. Contents of this papershould not be quoted nor referred to without permission of the author(s).

5ti International Conference on Advanced Physical FieldsMarch 6-9,2000Tsukuba-Shi, JapanInvited Talk

Surface Intensive Materials Processing for Multi-Functional Purposes

D. Ila, E. K. Williams, C. I. Muntele, and MAlabama A&M University

Normal, AL

A. George w

D. B. Poker and D. K. HensleyOak Ridge National Laboratory

OakRidge, TN

D. J. LarkinNASA Glenn Research Center

Cleveland, OH

‘The submitted manuscript has beenauthored by a contractor of the U.S.Government under contract No. DE-AC05-960R22464. Accordingly, theUs. Government retains anonexclusive, royalty-free Iicense topubIish or reproduce the pubtkhedform of this contribution, or allowothers to do so, for U.S. Governmentpurposes.”

Prepared by theOak Ridge National LaboratoryOakRidge, Tennessee 37831

managed bvLOCKHEED MARTIN EN~RGk RESEARCH CORP.

for theU.S. DEPARTMENT OF ENERGY

under contract DE-AC05-960R22464

January 2000

SURFACE INTENSIVE MATERIALS PROCESSING FORMULTI-FUNCTIONAL PURPOSES

D. ILA8, E. K. Williamsa, C. I. Muntele8, M. A. Georgeb, D. B. Pokerc, D. K. Hensl#, David J. Larkm*

a Center for Irradiation of Materials, Alabama A&M University,Normal, AL 35762, USA

b Dept. of Chemistry, University of Alabama in Huntsville,Huntsville, AL 35802, USA

CSolid State Division, Oak Ridge National Laboratory,Oak Ridge, TN 37831, USA

dNASA Glenn Research Center,Clevelan~ 0~ 44135, USA

v

We have chosen silicon carbide (Sic) as a multi-tkmtionrd material to demonstratethe applicationofsurfhceintensive processing for device t2t&kation. We will highl@ two devieeswhich are producedin houseat the Center for Irradiation of materials of Alabama A&M universi~ A) High temperature electronic gassensor, B) High temperature optical propertieskensor.

INTRODUCTION

As a high temperature wide bandgap semiconductor (3.23 ev) with high index of retraction (2.7), high flacturetoughness (3.1 MPa-m]’2), high thermal conductivity (-5.0 W/cm), high saturated electron drifi velocity (-2.7x 107cm/s)and high breakdown electric field strength (-3 M’V/cm), SiC is a material of choice for high temperature, high voltage,high frequency and high power applications. The large S-C bonding energy makes SiC resistant to chemical attack andradiation which make it a suitable candidate as a material for multi-fi.umtional purposes, e.g. fm device fabrication inharsh environments [1-3].

Recently, it has been shown that SiC can be employed as both an oxygen and a hydrogen sensor that operates ina temperature regime considerably higher than conventional sensors such as tin oxide (SnQ or silicon. Because of itsoutstanding thermal stability, silicon carbide can be employed as a hydrogen and hydrocarbon sensor that can potentiallyoperate at temperatures up to 10OO°C [4-9]. Potential uses of elevated temperature SiC sensors include automotiveapplications, process gas monitoring aeronautics, remote optical sensors, and aerospace applications.

A: SiC GAS SENSOR

The deposition of a catalytic metal such as Palladium (Pal) onto silicon carbide (SiC) results in Schottky diodebehavior. The adsorbing gas changes the space charge region under the metal clusters which in turn affect theconductivity of the crystal. This change in conductivity is measured and can be correlated to surfiwe concentrations andto the levels of the sampled gas in the ambient. The high sensitivity fm hydrogen containing combustible gases isenhanced by the presence of catalytic metals. Hydrogen containing species d-iatively adsorb to the metal andhydrogen atoms migrate to the Pd/SiC interfhce where they affkct the current-voltage (I-V) properties of the SiC [8].Rather than applying a palladium film [9], we have implanted palladium ions into the silicon fice of 6fi n-type SiCsamples.

The implantation was performed at 500”C, in order to minimize the induced implantation damage, to fluencesbetween 3x1014 atkxnz and 3.2x1016 atkmz, and implantation energies of 70 keV and 130 keV. The implantation energieswere chosen using the SRIM code [10] in order to get high near surface concentrations of Pd ions. Also, from the samecode, the approximate implantation range was established as being 40 nm and 65 mn (see figure 6).

The current measurements were done for temperatures near 23, 70, 145, 215°C in a closed gas environment.Voltages of *IV were applied on the backside of the sample, and the current was measured using a probe on severalplaces on the implanted side. The temperature was also monitored in order to correct for current fluctuations due totemperature changes. For this experiment, air was cycled with an Hz-Ar gas mixture with 4% hydrogen, and the cycleswere chosen to be “2 minutes in length (a compromise between the signal rise time and the length for a completemeasurement of 7 cycles). A drawing of the experimental setup is shown in Fig. 1. The metdic contacts were made of

1

,“

.

g’ilded copper, and the insulator was thick Teflon, so that the capacitance with the bottom side of the aluminum enclosure(connected to the common ground as a Faraday cage) was negligible.

.

T monitor

I read

insulator

Figure 1. Schematic view of the setup.

RESULTSFigures 2 through 5 show that at room temperature (23”C) the current increases in the presence of the 4% Hq in

the Hq-Ar mixture, for all of the SiC implanted samples.

air

6:W

d--u——

gas air air

a

Qas

230”C

0 120 240 360 480 600 720 8402.6X1 04-

Figure 2. Current response

2.4x104- 150”C for 6H SiC sample implantedwith 3 x 1014/cm2, 130 keV

2.2X10+PA positive voltage applied.

o 120 240 360 480 600 720 840

l.exloa -

— 80”C1.4X10=-

I

o 120 240 3io 480 6ti0 ?io 840

1.6x10”’D 23°C

1.4X1O’O

o 120 240 360 480 600 720 840Current (A)

Time (s)

2

,

,

.

8.0X10-7- air gas air gas sir gas air

7.oxlcr7 -

6.OXI0-7- 230°C

5.OX10-’ Io 120 240 360 460 600 720 840

5.4X10-0- ~

5.2x1& - 150°t3

!3.0X70-=-

0I

120 240 360 480 600 720 840

2.5X10-’0

2.5x10-’0 80”C

2.5x10-’0

0 120 2401.8x10-” -

360 460 600 720 840

1.7X1O” -23°C

1.6x10-” -

f I 1

Figure 3. Current responsefor the same sample as in Fig.2, negative voltage applied.

o 120 240 360 480 600 720 840Current (A)

Time (s)

3.6X1O-8- air gas air air 207°C3.4X10-8-

3.2x?06 -

3.OX10-8-

2.8X1060 120 240 360 480 600 720 840 Figure 4. Current responseJ

1.2X10-7 - 140”C for a sam le implanted with!

i.lxlo”’- 3.2 x 101 /cmz, 130 keV Pd;

1.OXI0“7- ——————— positive voltage applied.

9.OX1Oe -

0 120 240 360 480 600 720 840

4.8x1O_’

4.6X10-9 74°C

4.4X1d

4.2X10-’

0 120 240 360 480 600 720 840.3.OXIO-’” -

2.5XI0’O -

2.0xlol° -24°C

1.5xlo’0 - ~

oI

120 240 360 460 600Current (A)

720 840

Time (s)

3

,

..

*air gas air gas air gas air

3.6X10-7- 207°C

3.4X107 -

3.2x10-7-,

0 120 240 360 480 660 720 $itl

3.2x10-0

3,OX1O-” %4Ll”c

2.8x1O-B

o 120 240 360 480 600 720 840

l.oxlo’”

9.5X1o-” 74°C

9.OX10’ ‘

o 120 240 360 480 600 720 840

Figure 5. Current response forthe same sample as Fig. 4,negative voltage applied.

1.6X1 0-’‘

1.4XI0-” 24°C

1.2X1O-”

Current120 240 360 480 600 720 840

Time (s)

When the Hz–Ar environment is replaced by air, the current reverts to its initial value. At higher substratetemperatures, the differential current for Hz–Ar ambient and air ambient decreases as the substrate temperatureapproaches 60°C to 84°C and then it is reduced to zero. As the temperature increases, as shown in these figures, thesubstrate current in the presence of Hz–Ar gas is smaller than when air is present. Measurements were, also, performedon pristine SiC samples which revealed no current differential with the gas change for temperatures from 23°C to 240”C.The current measurement of the sensors produced at high implantation temperatures shows behavior while exposed tohydrogen, opposite to that for sensors reported in the literature [4,6,9] for SiC with palladium deposited on the suri%ce.There, the measured current in the presence of hydrogen is more for every temperature level, and reverts to the initialvalue when air is supplied instead of hydrogen.

A similar behavior was observed for both positive and negative voltages applied (the currents for negativevoltage are shown in the Figures 3 and 5 in absolute values). For absolute voltages above 1.2 V, the p-n junction breaksdown, and no more sensing behavior can be observed.

The current measurement for sensors produced at low implantation fluence is smaller than the current measuredfor sensors produced at high implantation fluence, at the same implantation energy. The sensors produced at lowerimplantation energies showed higher current sensitivity than the sensors produced at higher implantation energies (deeperimplanted layer, Figure 6). Also, the fict that afler the p-n junction breakdown no more sensing behavior can be observedmakes us believe that the sensing properties are related to the rectifying properties of the device rather than strictly due tothe modification in the conductivity of palladium, or palladium hydride. If we accept this observation, then the similarbehavior for both positive and negative voltages applied might be due to a structure more like n-p-n rather than simple p-n junction. This is expected, since the erystallinity of the SiC layer remains unchanged during high temperatureimplantation.

The relatively long rise time that can be observed on all the graphs presented in figures 2 to 5 is due to the factthat the hydrogen has to diffise through the surfhce layer to reach the deeper palladium ions as the ones close to thesurfhce become saturated. Similarly, when air is introduced, the ditlhsion process is reversed, from the depth to thesurface, resulting in the long recovery time. The recovery process is based on the capability of removal of the hydrogenthat was adsorbed in the device. Two processes occur: the chemical affinity for hydrogen by the oxygen in the air, andthe outgassing due to the device’s operation at elevated temperatures. At room temperature the first process is dominant,while the outgassing begins to contribute at temperatures above 100”C.

4

.

* 0.350

0.300

0.250

0.200

0.150

0.100

0.050

0.000

-0.050

A Xx&

1I , 1 , I I # I 1 I , I , I , I

“o 200 400 600 “Boo 1000 1“200 1400 1600 “

Depth (A)

Figure 6. SRIM simulation for the rangeofPdionsinSiC(10000 events for each energy).

B: NANO-CLUSTERS IN SiC

It has been shown that introducing metal colloids such as gold into an optical material, such as glass matrix,will change the color of the material. These materials have long been used for decoration and more recently have beenused to t%bricate optical devices. In recent years, more attention has been given to both linear and nonlinear properties ofthe material caused by optical absorption due to the surface plasmon resonance flequency, which depends on the index ofrefraction of the host substrate and the electronic properties of the colloids formed in the host material. The nonlinearproperties allow the manufacturing of a wide range of optic@ devices. Ion implantation followed by thermal annealinghas been used to introduce similar effects near the surfke as well as to change the nonlinear optical properties [10-14].An attractive property of ion implantation is that the ions can be focused to introduce the linear and nonlinear propertiesin a well defined space in an optical device.

We present the results of our investigation of producing nanoclusters of gold, silver, copper and tin in 6H and4H-SiC. In this work we study the changes in the linear optical properties as a function of ion fluence and post-implantation heat treatment. We implanted 1.0 MeV A% 2 MeV Ag and Cu and 120 and 160 keV Sn into the Si face ofSiC at 30 or 500°C followed by annealing at various temperatures.

It has long been known that small metallic particles or colloids embedded in dielectrics produce colorsassociated with optical absorption at the surlice plasmon resonance frequency [15,16]. For clusters with diameters muchsmaller than the wavelength of light (k), the theories of Mie [17] can be used to calculate the absorption coefficient of thecomposite:

18C@3a= N

& (~+2n&+rrg(1)

where Q is the volume fraction occupied by the metallic particles, n. is the refractive index of the host medium, and

&land cz are the real and imaginary parts of the frequency-dependent dielectric constant of the bulk metal. Equation 1 isa Lorentzian function with a maximum value when the light frequency equals the surface plasmon resonance frequency

(o) = o@, where

5

4

.. ()S1 Op + 2n~ =0 (2).

.

Values of El for the metals as a fhnction of wavelength are tabulated in [18] and the published index ofretraction for SiC is 2.655. The index of refraction measured by prism coupling for the SiC used in this work was 2.26.From Equation (2) one can predict the photon wavelengths for the surtice plasmon resonance ik.quencies for metalliccolloids in the photorefiactive host materials, as shown in figure 7 (81vs. Wavelength). This figure shows that taking no =2.7, the theoretical value of@ for SiC, the absorption bands should be near 691 nm, 578 nm, 668 nm, and 446 nm forAu, Ag, Cu, and Sn colloids, respectively. Similarly for the in-house measured index, no = 2.26, the absorption bands for ,SiC with Au, Ag, Cu, and Sn colloids shift to 625 nm, 509 nm, 617 nm, and 382 mn, respectively.

-20 -, \v, , ~, ,1

\\ ~

1 I300 400 500 600 700 800

Wavelength (m)

6

5

g’

j4nO= 226

1!“&$a3o

mo=2.7

2

1

I 1 1 1 I= 4E16 RT implant

L ● 3E17500°C imphmtA 1E17500°C impklt

■

A. -m *

A ‘* m

100 500 660 700 800Wavelength (rim)

Figure 7. Permittivity vs. wavelength for Ag, C% Au Figure 8. opticaldensity vs. wavelength for 1.5and Sn. The surfhce plasmon resonance occurs when Eq. MeV Si implanted into 4fi p-tyke SiC at 28°C2 is satisfied. Measured and expected refractive indices and 500”C.of SiC are shown at the horizontal dashed lines.

We have implanted SiC samples from CREE Research Inc., (6H-N type, 3.5° off axis, doped with nitrogen at9.2 x 10*7/cm3, Si i%ce)with ions such as 1.5 MeV Si, 1.0 MeV Au, 2.0 MeV Ag, 2.0 MeV Cu, and 160 keV Sn atfluences between 8 x 1015/cm2to 5 x 1017/cm2both at room temperature and at 500”C.

In our previous work on various other optical materials, we used either an energetic inert ion such as He or Ar oran ion which is one of the constituents of the host crystal, such as Si or O, to observe the change in the optical propertiesof the bombarded material as a fi.mction of the implantation parameters. Figure 8 shows the optical absorption spectra forSiC bombarded by 1.5 MeV Si ions at various fluences, at room temperature and at 500”C,. The room temperaturebombardment caused severe darkening, even at fluences as low as 4 x 101b/cm2,red-shifting the absorption edge from350 nm to 650 run and causing an overall increase in the absorption baseline. Implantation at elevated temperature,500”C, displayed a textbook behavior. That is, red-shifting the absorption edge at higher fluences, while allowing thedamage to recover, even at such low temperature. The reason we selected Si as a test ion is because Si is one of theconstituents of the SiC crystal and the absorption band for Si nanocrystals is in the far IR region and will not interferewith our measurements. These spectra iiom the Si hmbardment were then used as baselines to subtract the opticalabsorption due to ion bombardment damage.

Figure 9 shows optical absorption spectrum for a 6~ N doped SiC sample implanted with 3 x 10’7 Au/cm2 at1.0 MeV and at a temperature of 500”C then annealed at I()()O”Cin an argon ambient for one hour. Also shown is theabsorption spectrum of the sample prior to implantation and the difference between the two. At Au fluences less than 3 x

6

. 1“0‘7/cm2 no peaks due to plasmon resonance were discernible. It is clear from the difference spectrum that significantimplantation damage remains after the heat treatment. Similar effects were obsemed in 6H.j N-doped SiC, as shown in

. Fig. 10. The peak was at 535 nm after heat treatment at 1100°C in air. The peak was expected to appear in the range of630 to 680 nm, depending upon the value of the refractive index afler implantation. Using Eq. 2 and Fig. 7 the indexwould need to change tlom 2.7 to 1.5 to account for the location of the plasmon resonance peak. One possibleexplanation is that SiOz is formed during the annealing process. If this is so it indicates that the Ar atmosphere used inannealing the sample is Fig. 9 fiiled to purge the O completely. Rutherford backscattering ~alysis has indicated thepresence of oxygen in the both implanted and unimplanted mess of the sample, but the oxide layer was 20 percent thicker(170 nm vs. 140 nm) on the implanted side than on the unimplanted side. Approximately 10 to 20 percent of the goldatoms were in the oxide layer.

The size of the Au clusters can be estimated fkom the peak position and the width of the resonance through therelation r = Amvf/Acofi[1] where Amis a constant, taken to be 1.5 for Au, vf is the Fermi velocity of electrons in gold andAcoxis the full width at half maximum of the absorption peak. From Fig. 9 the Au cluster diameter is estimated at 5 nm.The cluster diameters in Fig. 10 are estimated to be at least 4 nm for the 5 x 1017ion/cm2 implant and 6 nm for the 3 x1017ion/cm2 implant. -

iA

A A

A3e17A~ llOO°C {

0.9- 6b — Difference

‘“w“ virgin

A0.8-

0.7-

0

0.6-

0.5-

$0

0.2 8 1 I400 500 600 700 800

VVavdength (rim)

Figure 9. Optical Density vs. Wavelength for 1MeV Au implanted at room temperature into6a N-doped SiC. Ion fluence was 3 x 10i7/cm2.Sample was heated in Ar to 1100”C.

2.8 * I i I I 1 IJ

Dkmeter of Au clusters, d-4 mu -

?.

2.2-

2.0-.2 “8tn

● 3 x 10’7/cm<

A 5 x 10’7/cm<g

1.4-8

■ 1 x lo’’krn:1.2-

1.o-

50.250.20

o.15m-700

Wavelength (rim)

Figure 10. Optical Density vs. Wavelength for 1MeV Au implanted at 500°C into 4m Al-doped SiCthen heated in air to 1100”C for 1 hr. Bottom trace

‘ is optical absorption of the unirnplanted sample.

Figure 11 shows the absorption band obtained from a SiC crystal bombarded by 2.0 MeV Ag at 500”C at 1.1 x1017/cm2 and annealed at 1100”C in argon for one hour. The optical absorption spectra shown has had the baselinesubtracted. Using the centroid of the peak, 493 nm, and equation (2), the calculated index of refraction for SiC at theimplanted volume is 2.15. It is important to subtract the absorption due to bombardment damage using a bombarded SiCas a refmence sample in the experiment. In figures 11 and 12, for Ag and Cu implanted at 500”C, the baseline is a sampleimplanted with the same ion at room temperature. A negative value can result at wavelengths where the increase in theopt ical density induced by the ion beam damage for room temperature implantation is greater than that created by elevatedtemperature implantation.

Figure 12 shows the absorption band obtained from a SiC crystal bombarded by 2.0 MeV Cu at 500”C, at 3 x10]7/cm2 and annealed at 900°C, in argon for ninety minutes. The optical absorption band for Cu implanted SiC is weakand falls at the same vicinity as the natural optical absorption of the nitrogen doped SiC, 630 nm. Using the centroid ofthe peak, 659 nm, and equation (2), the calculated index of refraction for SiC at the implanted volume is 2.65, which is

7

.,.

1.9-

.&m~l.8-

;

&“’-

1.6-

1.5460 500 7b0

Wavelength~m)

Figure 11. Optical density vs. wavelength for2 MeV Ag implanted into 4~ p-type SiC at500°C; after baseline subtraction.

0.5- n 1 I 1659 urn

o.4- !

,&,3 -g

~,.,.

&

o.1-

0.0 t I I I400 500 600 7@ ~o ~

Wavelength(ryn) *

Figure 12. Optical density vs. wavelength fm 2.0MeV Cu implanted into 6fi n-type SiC at 500°C;ailer baseline subtraction.

the same as the theoretical value [12]. The measured location of the absorption band for Cu implanted SiC is at the upperlimits of where it should be and the calculated index of refraction for the SiC in the implanted volume is higher than thatobserved for the virgin crystal. Since the measured absorption band is red shitle~ this anomaly can not be attributed tothe natural absorption band of the nitrogen-doped SiC.

Figure 13 shows typical optical absorption spectra for SiC implanted with 160 keV Sn at room temperature andannealed at 200°C in argon for one hour. This figure also shows that as implantation fluence increases the absorptionbaseline increases. Using the measured absorption band, 406 nm, and equation (2) the calculated index of refi-action forSiC at the implanted volume is 2.22. Increasing the Sn fluence produced an increase in the height of the opticalabsorption band due to formation of Sn nanoclusters. Observation of the fo~ation of Sn nanodusters in SiC usingvisible optical spectrometry was made possible because of the high index of refraction of the host material. tiherwise,the absorption band would have been located in the W region.

I

Figure 13. Optical density vs._ 5e16,200C, 1~ Ar wavelength for 160 keV Sn

implanted into 6fi n-type SiCat 28°C and annealed at200”C for 1 h, no baselinesubtraction.

2- ;: 406nm

1 I I 8 I u 1 i400 405 410 415 420 425 430 435 440

Wavelength (rim)

.

Implantation at room temperature results in a large increase in optical absorption that can mask the surt?iceplasrnon resonance absorption band. This difficulty is alleviated by implanting at elevated temperature (500”C), whichinhibits the formation of defects. ,The broad plasmon resonance spectra indicate that the nanockt.ers are very small.Further work is necessary at higher implantation and annealing temperatures.

CONCLUSIONS

In this manuscript we discussed only two of the applications of the SiC crystals: 1) application for fabrication ofelectronic sensors 2) application for fabrication of optical devices. 13ecause of its hardness, SiC is typically used forsurface coating and recently for fabrication of w@dows and mirrors used in the harsh environment of space.

The latter application, fabrication of optical devices, is mtliex new: ‘~i>’high index of refraction in SiC, 2.655,., ........ ,.”.,

gave us the edge to observe the absorption spectra due to Sn nanoclusters. The strength of Si-C band as well as thediilisivity of metals in SiC should be used to study the possibility of formation of epi-layer of nanoclusters for variousmaterials with high volume &action, which will enable us to use much lower implantation dose and use lower powerlasers to operate optical devices.

ACKNOWLEDGMENTS

This project was supported by the Center for Irradiation of Materials at Alabama A&M University and NASA-GRCContract No. NAG3-2 123. The work at ORNL was.sponsored by the Dsvision of Materials Science~ U.S. D@rtrnent ofEnergy, under Contract DE-AC05-960R22464 with Lockheed Mmlin Energy Research Corp.

REFERENCES

1.2.3.4.5.

6.

7.

8.

9.10.11.12.

13.

14.

15.

16.17.18.19.

J. R O’Connor, J. Smiltens, Silicon Carbide, A High-Temperature Semiconductor (Pergamon, NY, 1960).G. L. Harris Properties of Silicon Carbide (INSPEC, London, 1995).K. Shenai, R S. Scott and B. J. Balig~ IEEE Trans. Elec. Dev. 36 (1989) 1811.G.W. Hunter, P.G. Neudeck, G.D. Jefferson, G.C. Madzsar, C.C. Liu and Q.H. Wu, Report E-7773 NASA, (1993).G. W. Hunter, P.G. Neudeck, C.C. Liu and Q.H. Wu, Conference on Advanced Earth-To-Orbit PropulsionTechnology, (1994).L–Y Chen, G.W. Hunter, P.G. Neudeck, D. Knight, C.C. Liu and Q.H. Wu, Proceedings 19@ Meeting ofElectrochemical Socie~ (1996).L. A. Spetz A. Baranzahi, P. Tobias, I. Lundstrom, High temperature sensors based on metal-insulator-siliconcarbide devices, Physics Status Solidi (A) Applied Research, v 162, 1 (1997) 493-511.G. Muller, G. Krotz, E. Niemann, SiC for sensors and high-temperature electronics, Sensors and Actuators, A, 43,1-3 (1994) 259-268M. A. George, M. A. Ayoub, D. Ila, D. J. Larkin, Mat. Res. SW. Symp. Proc. Vol. 572, 123-128 (1999). -J. F. Ziegler, J. P. Biersack, U, Littmark, Z%eStopping and Range of Ions in Solids, Pergamon Press, NY, 1985.G. W. fbOIL J. Appl. Phys. 46 (1975) 4466.

C. W. White, D. S. Zhou, J. D. Budai, R A. Zuhr, R. H. Magruder and D. H. Osborne, Mat. Res. SW. Symp. Proc.Vol. 316 (1994) 499.E. K. Williams, D. Ila, A. DarWish, D. B. Poker, S. S. Sarkisov, M. J. Curley, J-C. Wang, V. L. Svetchnikov, H.W. Zandbergen, Nucl. Instr. and Meth. in Phys. Res. B148 (1998)1074 .D. Ila, E. K. Williams, S. Sarkisov, C. C. Smith, D. B. Poker, and D. K. Hensley, Nucl. Instr. and Meth. in Phys.Res. B141 (1998) 289.G. Fuchs, G. Abouchacra, M. Treilleux, P. Th6venar~ and J. Serughetti, Nucl. Instr. and Meth. in Phys. Res. B32(1988) 100.G. Abouchacra, G. Chassagne, and J. Serughetti, Radiation Effects 64 (1982) 189.G. Mie, Ann. Physik 25 (1908) 377.D. R. Lide, Ed., CRC Handbook of C%ernist~ and Physics, 76th Edition (CRC Press, Boat Raton, 1987).T. Friessenegg, M. Boudreau, J. Brown, P. Mascher, P. J. Simpson, and W. pu~ J. Appl. Phys. 80 (4), (1996)2216.

9