Time of flight spectrometry in heavy ion backscattering analysis

Nuclear Instruments and Methods in Physics Research 218 (1983) 1-5 North,Holland, Amsterdam

l

TIME OF FLIGHT SPECTROMETRY IN HEAVY ION BACKSCATTERING ANALYSIS A. C H E V A R I E R

and N. CHEVARIER

Institut de Physique Nucl~aire (et IN2P3), Universit~ Claude Bernard Lyon- 1, 43, bd du 11 Novembre 1918, 69622 Villeurbanne Cedex, France

Time' of flight spectrometry for the backscattering analysis of MeV heavy ions is proposed. The capabilities and limitations of this method are investigated. Depth and mass resolutions obtained in measurements of oxide film thickness as well as in GaAs layers analysis are presented. The importance of minimizing pile-up without significant loss of resolution by the use of an adequate absorber set just in front of the rear detector is underlined.

1. Introduction

2. Experimental device

The advantages of heavy ion backscattering techniques for the non-destructive analysis of thin solid films have been extensively discussed. The classical relationships of backscattering give theoretical improvements of sensitivity, selectivity, and depth resolution as the masses of the incident projectiles are increased. In order to take advantage of such characteristics, we need an experimental device which can give a good resolution but also an optimized efficiency in order to minimize radiation damage. We propose to detect the backscattered heavy ions using a high resolution time of flight spectrometer. The performance and limitations of this method are studied in the case of 2 to 7 MeV nitrogen incident beams, available from the 4 MV Van de Graaff of the Institut de Physique Nucl6aire de Lyon.

The experimental set-up is given in fig. 1. The time of flight detection is performed at 150 ° laboratory angle using a one meter flight path. The start detector is made from a thin foil associated with channel plate electron multipliers [1] (see insert fig. 1). Because of the energy range of the detected particles, great care must be paid to the thickness and homogeneity of these foils. To compromise between maximum secondary electron emission and minimum energy straggling we generally use a 5 # g / c m 2 formvar film. For the rear detector two possibilities have been studied. First a detector of the same type as the start detector has been used. This device allows a time resolution of 150 ps, measured using a 7 MeV nitrogen beam scattered on a gold target [2], but there is a larger solid angle of 0.1 msr for one meter flight path length. In order to reduce the incident

[I

~ ../~.

' .

':

InbCeiadent

Faraday cup Backscattered rticles

".. . = ~ ~ I/4 Turn

Path Fad

: FLIGHT PATH LENGTH

{ = 1

METER

Fig. 1. Experimental setup (SED: secondary electron device). O167-5087/83/$03.00 © Elsevier Science Publishers B.V. (North-Holland Physics Publishing Division)

~

v Acceleralion Harp

2

A. Chevarier, N. Chevarier / TOF spectrometry in backscattering analysis

Counts

®

® 16

"SJ 0

•

•

•

I

200 T.O.F.(ns)

w

•

v

w

-

•

•

' Iw!

I

1

I

I

I

I

175

150

2 E (MeV)

3

4

5

Fig. 2. Time spectra of 6 MeV nitrogen scattered on a Ag-Au sample. The spectrum denoted by (~) corresponds to the same conditions as (~) but an absorber has been set just in front of the rear time detector.

dose, we have recently built a new type of rear time detector which allows for a larger solid angle. In the following experiments a 0.2 msr solid angle has been used. The scattered particles impinge directly o n the 8 x 25 m m 2 useful detection area of channel plate. The efficiency of this detector is limited to 50% due to the intrinsic characteristics of the channel plate. The anode o u t p u t of each time detector is fed directly into a c o n s t a n t fraction discriminator [3]. Time of flight is given by a t i m e - a m p l i t u d e converter. In order to minimize dead time, the rear detector is used as a zero time detector. F r o n t channel plate detector pulses are delayed as necessary and give the stop signal to the t i m e - a m p l i t u d e convertor. Time spectra are registered o n a m u l t i c h a n n e l analyzer and routed to a c o m p u t e r for data processing. This spectrometer is a fast timing device. In order to minimize m u l t i c h a n n e l analyser dead time one can adj u s t the delay to select a n adequate spectral range. As shown in fig. 2, it is possible to remove one part of the spectra using an a b s o r b e r placed just in front of the rear detector. In fig. 2A, the RBS time spectra of 6 MeV nitrogen scattered o n a silver sample covered with sharp gold film is shown. The time scale is 86 p s / c h a n n e l and the 86 ns analysis range allows observation of nitrogen scattered on b o t h silver a n d gold. The spectra shown in fig. 2B corresponds to similar condition but a 810 / ~ g / c m 2 a l u m i n i u m foil has been put just in front of the

rear time detector. N i t r o g e n ions up to 3.5 MeV are stopped in the a l u m i n i u m foil so that only the gold spectra is registered. The m a i n p o i n t is that resolution o n the gold surface (300 ps which correspond to 20 keV) is not affected while pile up is minimized. Such a property should be invaluable when using time of flight in elastic recoil [4] or nuclear reaction analysis techniques.

3. Depth and mass resolution The d e p t h resolution depends on the total collision losses a n d of energy resolution which in our case is deduced from the time resolution R (t):

R(E) E

2 R(t) t '

R(E)=KE3/2R(t).

(1)

F r o m relation (1) R ( E ) will be smaller if the energy is decreased a n d R ( t ) stays constant. At lower energy the increase of the R u t h e r f o r d cross section could allow the integrated dose to the sample to be minimized. We measured the resolution in the case of a 2 to 7 MeV nitrogen b e a m for perpendicular incidence on a gold target and angular aperture of + 0.25 °. The front detector formvar foil was 5 / ~ g / c m 2 thick. Measured time resolution for different detected energies are shown in fig. 3. A drastic worsening of the time resolution up to 1

A. Chevarier, N. Chet)arier / TOF spectrometry in backscattering analysis

ns)

R

R (keV3

2

20

Rc,, SX

3O

10

\Xx\ 2O I

I

I

I

!

i

I

I

I

I

1

2

3

4

5

1

2

3

4

5

.10

E (MeV)

Fig. 3. Time, energy and depth resolution as a function of nitrogen backscattered energy on a gold target• The contribution to straggling in the front detector electron-production foil is drawn in dashed lines.

ns in the case of 1.5 MeV detected nitrogen is observed which is mainly due to straggling in the electron production foil (dashed line). Energy a n d depth resolution on gold are presented in the right part of fig. 3. The b a l a n c e between different factors, resolution, time of flight and stopping power, leads to small variations in d e p t h resolution. One must keep in mind that such m e a s u r e m e n t s include b e a m energy resolution (10 -3 ) a n d angular dispersion. In order to underline the performance and limitations of the technique two examples will be given. The first one is shown in fig. 4. A l p h a a n d 15N spectra for

anodically oxidised films [5] of Nb205 (2000 A and 500 respectively) are compared. The alpha spectrum was

lO3;,?"..?--.',,,

Counts

.I

::'.-,, . .'~.-~:,'-...""~:." ".-";'v"' ." !..:

i 0

E

500

0

•

|

e

i

•

w

e

~ ..., o

o,.

%.

0

• --

"~p,

° .O°o

-i •.~ .~.,~'..':,.~. "'C) '~" • ,,;.

.

.I

I

I

I

I

I

"

J

I .7.-

0

~--

1

Channel number

Channel number

Fig. 4. Typical 4He and 15N spectra for a Nb205 oxide on Nb. Spectrum @ corresponds to 1.8 MeV alpha backscattered on a 2000 ,h, thick oxide film and spectrum (~) to 5 MeV I~N backscattered on a 400 ,~ oxide film (upper trace, zero sb~fted by 1 vertical division).

!

Fig. 5. Typical ]~N time of flight spectra for a 370 ,~ Ta205 oxide on Ta. The upper spectrum corresponds to a 2 MeV nitrogen incident beam and the lower to 7 MeV. Each horizontal division corresponds to 20 channels and the time scale is 86 ps/channel.

A. Cheuarier, N. Chevarier / TOF spectrometry in backscattering analysis

4

•GSlE+e~

Counts 69

Ga

~Ga

• 00E+0~

e.e

~. Channel number

i ~,.~

I

4 expended, part fig. b

• 6SIE ÷0~

(b)

400A

4

D

".;..:'?...:. A

•

R

.OOE+O

5oA

=

-

i 449.0

330.e

I

I

I

I

I

146

144

142

140

138 TOF (ns)

I

I

1

1

3.0

3.1

3.2

3.3

E (MeV)

Fig. 6. Time of flight spectrum of 7 MeV 15N2+ scattered on a GaAs layer. In (b) an extended part of the spectra is presented.

taken from ref. 6 and was measured under the following experimental condition: alpha incident energy 1.8 MeV, 10/~C-particle, solid angle 0.39 msr, resolution 15 keV, energy calibration 2.5 keV/channel. The nitrogen spec-

trum was measured by time of flight spectrometry with a 100 /~C integrated particle current of 5 MeV tSN2+ incident beam. In the oxide plateau range the nitrogen spectrum corresponds to 2.5 keV/channel, the resolu-

A. Chevarier, N. Chevarier / TOF spectrometry in backscattering analysis

tion and solid angle are equal to 15 keV and 0.2 msr respectively allowing a comparison of the two spectra. The widths of the oxide plateaux are similar corresponding to an improvement in depth resolution by a factor of 4 in the nitrogen backscattering case. But for comparable counting rate the 5 MeV nitrogen backscattering analysis implies about 15 times more energy deposition per unit length than the 1.8 MeV alpha particle. As far as depth resolution alone is concerned it would be better to reduce the energy to about 2 MeV. At this energy taking into account the increase of cross section ( × 6) and decrease of stopping power ( x 0.5) will lead to more realistic energy deposition. This is shown in fig. 5 in which RBS spectra of 7 and 2 MeV nitrogen beams on a 370 A Ta205 anodic oxide film on tantalum are presented. In both cases the time calibration 86ps/channel, the corresponding width to resolution ratios are 3.9 and 5.3, the energy resolution being respectively 26 and 15 keV. Evermore at 2 MeV a dose of 15 times less is required. The specific advantage of high energy heavy ion scattering is to associate high mass resolution with depth resolution. An example is given in the case of amorphous GaAs layer analysis performed using a 7 MeV nitrogen incident beam. The GaAs: H films are the same as those considered in ref. 7. They were obtained from the rf sputtering of a monocrystalline G a A s wafer in H 2 - A r atmosphere. The argon partial pressure is maintained constant, the H 2 / ( A r + H2) ratio being adjusted to 0, 1, 5, 10, 20 and 40%, respectively. As shown in fig. 6 the mass separation between 75As, 71Ga, 69Ga is clear. A 50 A depth resolution allows a detailed arsenic depth profile to be probed over the first 400 ,&. Such results are to be compared with the 600 depth resolution obtained in alpha backscattering anal-

5

ysis [7] performed at 8 MeV incident energy necessary in order to obtain mass separation. 4. Conclusion Time of flight detection of backscattered heavy ions appears to be a suitable method for high elements depth profiling in that very surface. The main problem to be overcome is to minimize the integrated dose to the sample. To achieve this a large area rear detector has been developed and the next step is to increase the efficiency of the device using both directly impinging particles and electrons produced in a good efficiency electron production foil placed just in front of the channel plate. The authors would like to thank Dr G. Amsel for fruitful and stimulating discussions.

References [1] A.M. Zebelman, W.B. Meyer, K. Halbach, A.M. Poskanzer, R.B. Sextro, B. Gabor and D.A. Landis, Nucl. Instr. and Meth. 141 (1977) 439. [2] A. Chevarier, N. Chevarier and S. Chiodelli, Nucl. Instr. and Meth. 189 (1981) 525. [3] J. Pouthas and M. Engrand, Nucl. Instr. and Meth. 161 (1979) 331. [4] J.P. Thomas, M. Fallavier, D. Ramdane, N. Chevarier and A. Chevarier, these Proceedings (IBA-6), p. 125. [5] M. Croset, E. Petreanu, D. Samuel, B. Amsel and J.P. Nadai, J. Electrochem. Soc. 118 (1971) 717. [6] F. Abel, G. Amsel, M. Bruneaux, C. Cohen, A. Maurel, S. Rigo and J. Roussel, J. Radioanal. Chem. 16 (1973) 587. [7] J.P. Thomas, M. Fallavier, H. Carchano and L. Alimoussa, these Proceedings (IBA-6), p. 579.