MATERIALS

SCIENCE &

ELSEVIER Materials Science and Engineering B102 (2003) 238—246

www.elsevier.com/locate/mseb

Recent advanced applications of AAS and ICP-MS in the

semiconductor industry
Mohammad B. Shabani a'* , Y. Shiina a, F.G. Kirscht b , Y. Shimanuki a
d
Sumitomo Mitsubishi Silicon Co, 314, Kaneuchi, Noda-shi, Chiba-ken 278-0015, Japan b
Sumitomo Müsubishi Silicon Croup, 1351 Tandem Avenue N.E., Salem, OR 97303, USA

Abstract

We report on instrumentation-related challenges of applying graphite furnace atomic absorption spectroscopy (GF-AAS)
and inductively coupled plasma mass spectrometry (ICP-MS). We show that a significant amount of polyatomic species
derived from silicon sample solution in the plasma, such as SiO, SiOH, SiOHrSiOH3, Si02, Si02H, Si02H2 and Si02H3, can
hamper the detection limits of many elements of interest. This paper describes a method for eliminating these polyatomic
ions. We discuss the advantages and disadvantages of vapor phase decomposition method (VPD), drop etching method (DE)
and drop sandwich-etching method (DSE) for the recovery of metal impurities from a silicon wafer surface. We report the
application of the DSE method for the evaluation of near-surface metal impurities, used for gettering studies. We describe the
direct acid bulk decomposition (DABD) and the room temperature acid vapor phase decomposition method (RT-AVPD) for
the determination of metal impurities in bulk silicon. Finally, we report concentration of trace metal contamination in several
chemical reagent solutions. 0 2003 Elsevier B.V. All rights reserved.
Keywords: Inductively coupled plasma mass spectrometry; Graphite furnace atomic absorption spectroscopy; Vapor phase decomposition; Drop etching;
Drop sandwich etching; Direct acid bulk decomposition; Room-temperature acid vapor phase decomposition

1. Introduction In the course of silicon manufacturing, very fast

and highly sensitive trace metal characterization
techniques are essential for the determination of
metal impurities in chemical reagents and on the
surface or in the bulk of silicon wafers. Until recently,
the tools mainly used for quality control were
total reflection X-ray fluorescence (TXRF) [l ,2],
inductively coupled plasma mass spectrometry (ICPMS) [5 Ca and Fe, preventing their determination at low
—7]. Performing extremely sensitive analysis at steadily levels. Cold plasma conditions can be used to reduce
increasing demands is a challenge for every analytical
technique [8]. the formation of molecular species in the plasma and
It is beyond dispute that ICP-MS is a powerful tool in this way, permit the measurement of 39K, 4oca and
56
for elemental analysis. The ICP-MS is capable of Fe and improve detection limits of these elements.
producing pg ml ¯l and sub pg ml ¯l detection limits, The highresolution magnetic sector ICP-MS is
has an extended dynamic range (six orders of magni- another alternative method that can be used to
separate these molecular species' interferences from
elements of interest. Matrix interferences in the ICP-
MS resulting from the silicon and from solvent of
sample preparation are investigated
graphite- furnace atomic absorption spectrometry
(GF-AAS) [3,43 and, in the last few years, tude), very
 239 Shabani et al.

broad elemental Nebulizer (l min ¯ l) 0.9

Chamber gas (kPa) 215 (hot plasma)
coverage, good 325 (cold plasma)
precision when an
internal standard is
employed, and
outstanding
semiquantitative-
analysis
performance.
Although the
resulting mass
spectrum of
quadrupole mass
spectrometer is
relatively simple,
consideration must
be given to
background
molecular ions formed between the plasma gas,
entrained air and the solvent. In many instances, the
level of interference is small. However, the 40Ar+ O-mass unit
38
ArH and 40Ar 160+ ions occurring at masses 39, 40 introduced by sample matrix. The ICP-MS operating
and 56 interfere strongly with the major isotopes of conditions are listed in Table 1. A SIMA 6000 Perkin
39
K, — Elmer GF-AAS was used.
/
Sample wafers were 150 and 200 mm CZ P-type
boron-doped 10 Qcm. The wafers were cleaned with
determination of metal impurities in chemicals and modified SCI and SC2 solutions and contaminated by
surface or bulk of silicon wafers. ICP-MS is widely spin coating or by SCI contaminated solutions. High
appfied for quality control through all customary purity water with less than I pg ml ¯l of each metal
device fabrication steps such as wafering, polishing, impurity and high purity HF, HN03 and H202 with
cleaning, oxidation, epitaxy, ion implantation, and less than 10 pg g¯ of each metal impurity
evaluation of gettering capability. (TAMAPUREAA-10) were used for sample
preparation in this study. A clean room class 100 with
a clean draft, class 10, was used for the measurements
2. Experimental and sample preparation, respectively.

An ICP-MS (SPQ 9000, Seiko Instruments Inc.,

Japan) is used with a hydrofluoric acid resistance 3. Results and discussion
torch and Pt sampler and skimmer cones. A PFA
free aspirating nebulizer was used (microflow-100 3.1. Inductively coupled plasma mass spectrometry
PFA; Elemental Scientific, Omaha, Ne, USA). Cold
plasma or hot plasma conditions are applied to ICP-MS is a relatively new technique for trace
reduce the formation of moleC.ular species in the element analysis. The research and development
plasma and those started in 1975 and the first commercial instruments
became available in 1983. The method has found a
Table 1 wide range of applications in research and industries.
TCP-MS instrumental operating parameters The ICP is a very high temperature source. The ICP
has a high electron density of around 10—15—10 16
ICP-MS Seiko Instrument SPQ9000
cm and the electronic temperature is around 8000—
SPQSeikoICP-MS
instrument
Nebulizer Microflow-100 PFA 10 000 K. Most elements introduced into the plasma
Spray chamber Teflon chamber are therefore very efficiently ionized and singly
Torch Hydrofluoric acid resistance charged positive ions predominate. Ions which are
RF forward power (W) 1200 extracted from the plasma
Gas flow rates
Plasma gas (l min ¯ l ) 16
Auxiliary gas (l min 1.0 O deflecter unit
 M.B. Shabani et al. Materials Science and Engineering B102 (2003) 238—246 240

Plasma
Carrier Gas
Auxiliary
Table 2 and pass through the
sampler and skimmer
cone must
Molecular interferenace of Si matrix effect on the different mass be directed into a mass
numbers Mass no.
filter. This is achieved
with a series of electrical
lenses. To separate ions
o Si from electrons, atoms and
molecules before entering
a mass filter, a Q deflector
is located between lenses
and mass filter. By rapidly
sweeping the voltages, the
mass filter scans the mass
range, producing a typical
mass spectrum in
Sio + 44(44Ca) 45(43Sc) 46(46Ti) the region 5—240 atomic
mass units (AMU).
Alterna-
SiOH 45(45Sc) 46(46Ti) 47(47Ti) tively, the voltages may be
preset to allow a specific
Si0H2+ 46(46Ti) 47(47Ti) 48(48Ti) mass to pass through the
quadrupole. The
quadrupole
Si0H3+ SiF 47(47Ti) 48(48Ti) 49(49Ti) mass filter can be
47(47Ti) 48(48Ti) 49(49Ti)
operated at a nominal
resolution of 1
SiFH 49(49Ti) 50(50Ti)
AMU. The ion detection
system consists of an
electron
SiFH3 49(49Ti) 50(50Ti) 51
multiplier of the
continuous or discrete
dynode type.
60(60Ni) 61(61Ni) 62(62Ni) Very high dynamic linear
range ion intensities up to
108
Si02H 61(61Ni) 62(62Ni) 63(63cu) can be handled. A diagram of SPQ 9000 Seiko ICP-
MS
SiO,H2+ 62(62Ni) 63(63cu) 64(64Zn)
instrument is shown in Fig. 1.
Si0,H3+ 63(63cu) 64(64Zn) 65(65cu)
SiOF 63(63cu) 64(64Zn) 65(65cu) The high degree of ionization for most elements in the
SiOFH 64(64zn) 65(65cu) 66(66Zn)
plasma results in extremely high sensitivity with detec-
SiOFH,+ 65(65cu) 66(66Zn) 67(67Zn)
tion limits in the range 10—0.01 pg ml Of particular
68(68Zn) 69(69Ga) 70(70Ge) advantage is the multi-element capability of the
techni-
241 Shabani et al.

/ Materials Science and Engineering B102 (2003) 238—246

que, which is based on the relative simplicity of the resulting mass

spectrum, even from very complex

Si 28 Ni 60 Ni 61 Ni 62 cu 63 Zn 64 cu 65
(Si02) (Si02H) (Si02H2) (SiOF) (SiOFH) (SiFH2)

Fig. 2. Relative intensity of Si and Si spectral interferences on the mass numbers of 60, 61, 62, 63, 64 and 65 (hot plasma).

100

Si 28 Ni 60 Ni 61 Ni 62 cu 63 Zn 64 cu 65
(Si02) (Si02H) (Si02H2) (Si0F) (SiOFH) (SiFH2)

Fig. 3. Relative intensity of Si and Si spectral interferences on the mass numbers of 60, 61, 62, 63, 64 and 65 (cold plasma).
I plasma gas, entrained air and the sample solution.
Table 2 lists some of the species present in the
background mass spectrum of silicon matrix solution
and the interferences occurring at specific masses.
The levels of these interferences are strongly
dependent on the cold or hot plasma operating
conditions. For example, at hot plasma operating
condition, the argon and argon molecular ions occur
at masses 39ArH+ , 40Ar and 56ArO , which coincide
with 39 K , 4o ca and 56Fe+ isotopes. These
50 100 150 200 interferences can be completely removed at cold
Si concentration (ppm)
plasma operating condition. In contrast, the Si
Fig. 4. Relation between the amount of Si matrix in the blank solution and
molecular ions, which strongly interfere with several
ppt amounts of interferences on the intensity of different isotopes of elements of interest at cold plasma operating
element. condition, can be reduced at hot plasma condition.
Figs. 2 and 3 show the relative intensity of Si and Si
matrices. Although the resulting mass spectrum is spectral interferences on the mass numbers of 60, 61,
relatively simple, •consideration must be given to
background molecular ions formed between the 62, 63, 64 and 65 at hot and cold plasma conditions,
 M.B. Shabani et al. Materials Science and Engineering B102 (2003) 238—246 242

respectively. It can be clearly seen that at hot plasma of magnitude lower than in cold plasma condition.
condition Si spectral interferences are several orders Fig. 4 shows the relation between the amount

DE(one-drop etching method)

Wafer

Wafer Wafer

Teflon plate

Dropl
et

Fig. 5. Schematic diagram of the collection of trace metal impurities on the surface of silicon wafer by VPD and DE methods.
243 M.B. Shabani et al. /
Materials Science and Engineering B102 (2003) 238—246

100%

HF/HN03 solution
Surface

Teflon plate

Etching solution
Fig. 6. Collection efficiency of Cu by of 2% HF and
VPD and DE methods using different
solutions for recovery.
2% 1--1202
solution
of Si matrix in the blank directly on
solution and ppt amounts of the surface of
interferences on the intensity the wafer.
of different metals using cold Metal
plasma. These levels of Si impurities in
matrix interferences will be the etching
reduced severely if we use hot solution can be measured
plasma conditions. after recovery. Fig. 5 shows
the schematic diagram of the
3.2. Analysis of silicon wafer
collection of trace metal
surfaces impurities on the surface of a
silicon wafer by VPD and DE
Generally, the vapor phase methods. To compare the
decomposition method (VPD) metallic removal efficiency of
and the drop etching method these two methods, silicon
(DE) have been used for the wafers were contaminated
recovery of metal impurities with known amounts of trace
from the surface of wafers [10 metals in SCI solution and
—14]. In the VPD method, a then measured after recovery
silicon wafer is placed on a by using GF-AAS. In this
stand in a closed box study, the recovery of Na,
containing fresh, concentrated Mg, Al, K, Ca, Cr, Fe, Ni, Cu
HF solution. After and Zn were examined by the
decomposition of native oxide VPD and DE methods. It was
or thermal oxide in the found that except for Cu all
presence of the HF vapor, a metals investigated give
drop of mixture of 2% HF and similar, basically complete
2% H202 is placed on the recovery using the VPD or
surface of the wafer and DE methods. In the case of
rolled throughout the entire Cu, poorer recovery was
wafer surface in a found using the VPD method.
reproducible manner in order Fig. 6 shows collection
to collect the metal impurities efficiency of Cu by the VPD
for analysis. In the DE and DE methods using
method, surface oxide etching different solutions for
can occur naturally by an recovery. Poorer Cu recovery
addition of one-drop mixture by VPD or by HF alone,
 M.B. Shabani et al. Materials Science and Engineering B102 (2003) 238—246 244

without addition of H202 as Al 1.3E+07 1.3E+07

an oxidizing agent, is due to 4.8E+07
the fact that Cu metal formed 5.9E+07
Table 4
Typical concentration of metal contamination in chemical reagent solutions

Mg cu Zn

TAMAPURE-AA-10(H202)
TAMAPURE-AA-IO(HF) 1-2
TAMAPURE-AA-10(HN03)
1-2 1-2 1-5 1-2 1-2
TAMAPURE-AA-IO(HCI) 1-2
1-2 1-2
High purity H20 < 0.1 1-2
EL-grade (NH40H) 5 160 15 10 100 12 15 6 10
El.-grade (H202) 5 5 30 30 4 15 2 7 4
with HF vapor or HF solution 2.0B +07
does not dissolve in a HF 1.4E+07
solution alone. In 2.0E+07
GF-AAS 60Ni 8.7E+07
63 5.OE+08
Fig. 7. Schematic diagram of the 65 1.5E+08
collection of trace metal impurities on 64zn 1.9E+08
the surface of silicon wafer by DSE 5.5E+07
methods. 68zn 1.0E+08
5.0E+07

DE, drop etching; VPD, vapor phase

addition, another advantage decomposition.
of DE over VPD is its faster Unit: ppt.
procedure for sample
preparation. from a 200 mm wafer by
There are obvious VPD and DE are 2 and 20 pg
advantages and ml • l , respectively. As earlier
disadvantages using ICP- indicated, a major
MS for the measurement of disadvantage arises from
metal impurities after sample solution containing
collection by VPD and DE higher Si matrix. Table 3
methods. Total Si matrix shows a comparison between
concentration in 1 ml detection limits of VPD and
solution after surface native DE methods at cold plasma
oxide (oxide thickness less condition. As can.. be seen, at
than 1 nm) recovery lower mass numbers no
difference in detecfion limits
can be seen using VPD or DE
Table 3 Comparison between method. However, at higher
detection limits of ICP-MS by
DE and VPD method
masses, because of more Si in
the DE solution increased
detection limits are reached
Elements DE (atoms cm for 60Ni, 63 Cu, 65Cu, 64Zn and
68
(atoms cm ¯ 2) Zn isotopes.

4.4E+06 3.3. Analysis of near-surface

4.4E+06
bulk silicon by depth profiling
Mg 7.0E+06 7.0E+06
245 M.B. Shabani et al. / Materials Science and Engineering B102 (2003) 238—246

Both VPD and DE concentration in this

method can collect only solution is approximately
metal impurities on the top 40 mg ml ¯ 1 . This very
surface of a wafer. Since high concentration of Si
predetermined etching of and high concentration of
silicon is not applied here, acids solution cannot be
bulk silicon close to wafer directly introduced into the
surfaces cannot be ICP-MS. The presence of
measured. For this purpose, Si and acids at high
the one-drop sandwich concentrations results in a
etching method (DSE) is matrix suppression effect
used allowing depth that changes the measured
profiling of metal ion signals of the isotopes
contamination [11,12]. of interest. Therefore, the
Using this method, one solution was evaporated in
droplet a small beaker, to remove
(typically 1 ml) of a the Si matrix and acids
mixture of 38% HF and solution, and then
68% HN03 solution from redissolved in diluted acid
high purity acids is placed solution for the
on the surface of a clean measurement by ICP-MS.
Teflon plate. The droplet is In the case of GF-AAS
sandwiched by the wafer instrument, this very high
and the Teflon plate. The concentration of Si and
etching is completed within acids solution can directly
2 or 3 min and the etching be measured without any
depth can be controlled by noticeable matrix effect.
the amount of HF in the This method is very useful
mixture solution. Metal for the evaluation of gettering
impurities in the etching efficiency and control of
solution can be measured contamination from furnace
after recovery. Fig. 7 shows annealing and ion
the schematic diagram of implantation. Figs. 8 and 9
the collection of trace show two examples of
metals by the DSE method. evaluation of internal
In the case of 150 mm gettering and PIP + boron
wafers, for decomposition gettering using the DSE depth
of 1 gm silicon by I ml profiling method. More data
solution, a ratio of 1/1.5 of are contained in the following
mixture of HF/HN03 is Refs.
required. The Si
I.OE+15

of

Depth surface ( m)
M.B. Shabani et al. Materials Science and Engineering B102 (2003) 238—246 246
Fig. 8. Depth profiles of Cu contamination in Si wafer with internal gettering.

I .OE+15
0 I .0E+14

I.OE+13

I.OE+12

I ,OE+I I
0

5 5 0
Depth from surface( U m)
Fig. 9. Depth profiles of Cu contamination in p/p+ epitaxial Si wafer,
3.4. Analysis of bulk silicon will increase the total blank
levels in the final solution. As
Generally, the direct acid the result, the detection limits
bulk decomposition method of most elements will be
(DABD) and room increased using this method.
temperature acid vapor phase To reduce total procedures
decomposition method (RT- blank levels and to improve
AVPD) have been used for detection limits, a RT-AVPD
recovery of metal impurities was developed. For
from the bulk of the silicon decomposition of bulk silicon
wafers [19,201. For the wafers by the RTAVPD
decomposition of bulk method, a closed box
silicon, a mixture of 38% HF containing a mixture of
and 68% HN03 solution from concentrated
high purity acids is directly HF/HNOdH2S04 (3/1/3 by
used for decomposition of weight) is used. In this
silicon. After recovery by method, the silicon bulk can
vaporization of silicon and be decomposed within 12 h
acid solutions, the metal by the vapor of the HF and
impurities can be measured HN03. Finally, the metal
by ICP„MS. Fig. 10 shows impurities in the residue of
the schematic diagram of the the decomposed silicon, after
decomposition of Si bulk by dissolution and removal of
DABD method. In this additional silicon can be
method, decomposition of 1 g measured by ICP-MS. Fig. 11
silicon requires 20 ml mixture shows the schematic diagram
of 2/1 of HF/HN03. This of the decomposition of Si
amount of high purities acids bulk by RT-AVPD method.
247 M.B. Shabani et al. / Materials Science and Engineering B102 (2003) 238—246

Using this method, the Si is ICP-MS measurement

dissolved by clean acid technique, the detection limits
vapors and the resulting of IE9—1E10 atoms cm¯ can
solution has less be achieved for most
contamination in comparison elements. This method is used
with DABD method. By for determination of Fe
applying this method with solubility at 600,
2/1 20g

Hot plate
HF/HN03

Hot plate
4 x I .0g sample

ICP-
MS
GF-
AAS

Fig. 10. Schematic diagram of decomposition of Si bulk by DABD methods.

 M.B. Shabani et al. / Materials Science and Engineering B102 (2003) 238—246 248

Teflon Closed Box

vapor HF+HN03

Wafer

liqui
HF+HN03+H2S04
d
Fig. I l. Schematic diagram of decomposition of Si bulk by RT-AVPD methods.
700, 800, 900 and 1000 o c in removed by evaporation in order to
silicon doped with 1.3E15, 4.5E18, achieve good instrument performance
IE19 and 6E19 boron atoms cm -3 and eliminate the problem of mass
[21]. Fig. 12 shows measured Fe spectra interferences which occur in
solubility in lightly and highly the argon plasma, or are resultant
boron-doped silicon as a function of from the plasma sampling process.
temperature. The combination of RT- One major source of metal
AVPD and ICP-MS provide reliable impurities on the surface of silicon
solubility data, specifically in low
wafer is cleaning solutions.
temperature range. To the best of
our knowledge, this is a new result. Recently, the quality of cleaning
solutions has been improved and, as
the result, the level of metal
3.5. Analysis of chemicals impurities on the surface of silicon
wafers has been reduced to the level
Analysis of chemical reagents, of the limits of detection of most
used for cleaning of silicon wafers instruments. To measure such an
in different steps of wafer
249 M.B. Shabani et al. /
manufacturing, is of critical extremely low concentration of
importance in determining the final metals on the surface or in bulk of
quality of the performance of silicon wafers, very high purity of
semiconductor devices. Several of chemical solution for sample
these chemicals require that the preparation is required. Table 4
chemical matrix is shows concentration level of several
metal impurities in
1.3E15 (100) —A—4.5E18
(0.0160)
1.0E19 (0.00880) 6.0M 9 (0.002 Q)

I.O.E+I
600 700 800 900 1000 Diffusön
ten»eratue (oc)

Fig. 12. Measured Fe solubility in lightly and highly boron-doped silicon as a function of
témperature by combination of RT-AVPD and ICP-MS.

Materials Science and Engineering B102 (2003) 238—246

high purity water, in EL grades H202 and NH40H for molecular ions occur at masses 39ArH , 40Ar and
56
SCI cleaning solution and in TAMAPURE-AA-IO ArO , which coincide with 39 K + , 40Ca+ and 56 Fe+
high purity H202, HF, HN03 and HCI for chemical isotopes were completely removed at cold plasma
analysis. As it can be seen, the level of contamination operating conditions.
in TAMAPURE-AA-IO are less than 1 pg ml -I for It was found that except for Cu all metals
many elements. For the analysis of the above investigated give similar and basically complete
solutions, 10—50 ml of the solution were evaporated recovery using the VPD or DE methods. Poorer
in a Teflon beaker and the residue was re-dissolved in I recovery for Cu was observed when using the VPD
ml of dilute method. It is shown that in DE method, Si matrix in
acid solution, which was then measured by ICP-MS. the solution increased detection limits reached for
60
The metal contamination concentration in the Ni, 63 Cu, 65 Cu, 64Zn and 68 Zn isotopes.
chemical reagents was less than 10 pg ml ¯ 1 . This A DSE method was presented for the evaluation of
very low level of metals does not significantly effect gettering efficiency and control of contamination
the determination of trace metals recovered from the from furnace annealing and ion implantation. Silicon
surface of the silicon wafer. bulk dissolved by the clean acid vapors of the RT-
AVPD method and the resulting solution has less
contamination in comparison with the DABD
method. By applying this method with ICP-MS
4. Conclusions measurement technique, the detection limits of IE9—
lE10 atoms cm were achieved for most elements.
The ICP-MS spectral and matrix effect study
showed that the presence of high concentration of Si
interfere with the analysis of several elements of
interest when operating under cold plasma conditions. Acknowledgements
It was shown that operating hot plasma conditions
can reduce the interference. The argon and argon
 M.B. Shabani et al. / Materials Science and Engineering B102 (2003) 238—246 250
The Authors are grateful to Patrick Taylor for the [21] M.B. Shabani, Y. Shiina, Y, Shimanuki, F.G. Kirschit, Solid
critical reading of the manuscript. State Phenomena, 82-84 (Gadest 2001) (2001) 331.
References

H. Kondo, J. Ryuta, E. Morita, T. Yoshimi, Y. Shimanuki,

Jpn. J. Appl. Phys. 31 (1992) LN.
[21 C. Neumann, P. Eichinger, Spectrochim. Acta 46B (10) (1991) 1369.
[3] M.B. Shabani, H. Suga, M. Fukuda, K. Miyomoto, Y.
Sayama, Extended Abstract 1 D 06, 54 meeting of The Japan
Society of Analytical Chemistry, Gunma, 1993, p. 299.
14] M.B. Shabani, Y. Shiina, Extended Abstract 2 A 09, 50 meeting of
The Japan Society of Analytical Chemistry, Fukuoka, 2001, pp. 17,
299.
[5] J. Fucska, S.H. Tan, M.K. Balazs, J. Electrochem. soc. 140 (1993)

[6] A Krushevska, S. Tan, M. Passer, X.R. Liu, J. Anal. At. Spectrom.

15 (2000) 1211.
[7] A. Walsh, D. Potter, E. McCurdy, R.C. Hutton, in: G. Holland, A.N.
Eaton (Eds.), Application of Plasma Source Mass Spectrometry,
Royal Society of Chemistry, 1991, p. 12.
[8] L. Fabry, S. Pahlke, L. Kotz, G. Tolg, Fresenius J. Anal.
Chem. 271 (1994) 349.
[9] K.E. Jarvis, A.L. Gray, R.S. Houk, Handbook of Inductively
Coupled Plasma Mass Spectrometry, Blackie, London,
1992, pp. 129-143.
[10] A. Shimazaki, H. Hiratsuka, Y. Matsushita, S. Yoshii, Extended
Abstract on Solid State Devices and Materials, 1984, p. 281.
M.B. Shabani, T. Yoshimi H. Abe, M. Fukuda, Y. Sayama, Abstract
396, The Electrochemical Society, vol. 94-2, Florida, 1994, p. 608.
[12] M.B. Shabani, T. Yoshi_nli, S. Okuuchi, H. Abe, Solid State

Phenomena 57-58 (Gadest 97) (1997) 81.

[13] M.B. Shabani, T. Yoshimi, H. Abe, T. Nakai, B. Cordts, in: P.
Hemment, S. Cristoloveanu, K. Izumi, T. Houston, S. Wilson (Eds.),
Silicon-on-Insulator Technology and Devices VIl, PV 96-3 The
Electrochem. Soc., Proceeding Series, Pennington, NJ, 1996,

[14] M.B. Shabani, T. Yoshimi, H. Abe, J. Electrochem. Soc. 143 (1996)

2025.
M.B. Shabani, T. Yoshimi, S. Okuuchi, T. Shingyoji, in: B.O.
Kolbesen, P. Stallhofer, C. Claeys, F. Tardiff (Eds.), Crystalline
Defect and Contamination, PV 97-22 The Electrochem. Soc.,
Proceeding Series, Pennington, NJ, 1997, p. 318.
[1 61 M.B. Shabani, S. Okuuchi, T. Yoshimi, T. Shingyoji, F.G. Fritz, in:
C. Claeys, M. Wtanabe, P. Stallhofer, H.J. Dawson (Eds.), High
Purity Silicon, PV 98-13 The Electrochem. Soc., Proceeding Series,
Pennington, NJ, 1998, p. 313.
M.B. Shabani, T. Yoshimi, S. Okuuchi, T. Shingyoji, F.G. Fritz, in: C.
Claeys, P. Rai-Choudhury, M. Wtanabe, P. Stallhofer, H.J. Dawson (Eds.),
High Purity Silicon, PV 2000-17 The Electrochem. Soc., Proceeding
Series, Pennington, NJ, 2000, p. 305. M.B. Shabani, S. Okuuchi, Y.
Shimanuki, in: B.O. Kolbesen, C. Claeys, P. Stallhofer, F. Tardiff, J.
Benton, T.J. Shaffner, D. Schroder, S. Kishino, P. Rai-Choudhury (Eds.),
Analytical and Diagnostic Techniques for Semiconductor Materials
Devices, and Processes, PV 99-16 The Electrochem. Soc., Proceeding
Series, Pennington, NJ, 1999, p. 510.
M.B. Shabani, T. Yoshimi, H. Abe, M. Fukuda, Y. Sayama,
2H09 44 Annual Meeting of The Japan Society of Analytical
Chemistry, Hokkaido, 1995, p. 405.
[20] M.B. Shabani, Y. Shiina, Y, Shimanuki, Extended Abstract 3pZC-2,
61 Autumn Meeting of The Japan Society of Applied Physics, 2000,
p. 687.