Possibility of the Reverse Monte Carlo Modeling for the Amorphous Si Deposited on the Reactive Ion E

1
Possibility of the Reverse Monte Carlo Modeling
for the Amorphous Si Deposited on the Reactive
Ion Etched Si Substrate

• The Institute of Scientific and Industrial
  Research, Osaka University
• School of Materials Science, Japan Advanced
  Institute of Science and Technology
• Toshio Kawahara, Yoshinori Matsui, Seiichi
  Tagawa, Tomoji Kawai and Hideki Matsumura

2
Contents

• Introduction
• Cat-CVD and PECVD
• Comparison of structures of a-Si deposited by
  catCVD and PECVD
• Amorphous Si deposited on the reactive ion etched
  Si substrate

3
Introduction

• Devices made as a thin film form
• Solar cell, Thin film transistor
• Hydrogenated amorphous silicon (a-SiH)

• Deposition methods
• Plasma enhanced CVD (PECVD)
• Sputtering in hydrogen gas
• Catalytic CVD (cat-CVD)

Cat-CVD Stable and high quality Low hydrogen
concentration
Structure? Comparison to other methods? Experiment
in thin films?
4
Plasma enhanced CVD (PECVD)
5
Catalytic CVD (cat-CVD)
6
PECVD and cat-CVD
7
Purpose of experiments

• Comparison of a-Si deposited by PECVD and cat-CVD
  using RMC modeling
• Discussion about the possibility to RMC modeling
  for thin films.

8
Parameters for RMC modeling

• Coordination number constraint
• Fourier transform IR (FTIR) spectra
• Density of a-Si
• Rutherford Back Scattering (RBS)

9
Coordination number constraint

• Unconstraint RMC modelling has many freedom.
• Then, the coordination number constraint seems to
  be important.

We used the FTIR spectra for the estimation of
hydrogen concentration used as the constraint.
10
Fourier transform IR spectrometer
11
Density of a-Si
Back scattering of He ions
Number of Si atoms
Rutherford Back Scattering (RBS)
12
Relation between density and H contents
Density of a-Si was estimated by the RBS.
Estimated density has linear relation to the
hydrogen concentration.
13
Comparison of structure of a-Si deposited by cat
CVD and PECVD

• Current devices PECVD
• Improved devices cat-CVD

14
X-ray configuration
Seemann-Bohlin method 1. Low incident angle 2.
2?diffraction
2?
f
1 mm
0.4 deg.
a-Si
Si (100)
Q range 0.6 to 17.1 Å-1
Rigaku Rint 2000
Mo Ka
For safety, we used q range up to around 10 Å-1.
4 150 degrees
15
Structure factor S(Q)
Structure factor for a-Si deposited by cat-CVD
and PECVD
Different peak structures
3D structure?
Jpn. J. Appl. Phys. 44 (2005) 3808.
16
FTIR spectra for PECVD a-Si
Area ratio of two peaks Si-H Si-H2 5.371
Using A-values for the stretching mode (Phys.
Rev. B109, 13367) Si-H Si-H2 10.18 1.36
17
RMC modelling for a-Si by PECVD
Starting from 4096 atoms of c-Si structure
Si-H Si-H2 (RMC) 10.13 1.34
18
3D structure
Raman spectra for a-Si
TA/TO ratio is large in a-Si (PECVD).
Defective voids is small.
Large stress in a-Si
19
Bond angles
Main peak in a-Si (PECVD) 103.7 deg. a-Si
(cat-CVD) 105.8
Larger sifts from tetrahedral values of 109.5
generates stress in a-Si.
20
Conclusion

• We simulate the a-Si structures by RMC.
• a-Si deposited by PECVD has little void in 3D
  structures.
• a-Si (PECVD) has larger stress than that in
  cat-CVD.

21
Amorphous Si deposited on the reactive ion etched
Si substrate
Nano fabrication and structural Analysis
22
Transmission geometry for XRD
Obtained spectra are constructed both from the
a-Si and Si- substrate.
Detector
XRD
X-ray Focusing
a-Si Si substrate
To reduce the contribution of c-Si,
etching
23
Wet and Dry process
Wet chemical etching
Dry etching
Ion, atom
RF power
KOH
80 oC (Si) 250 oC (GaN)
Physical and chemical etching in process gas
Controllability is better in dry etching.
24
Plasma etching
Electron cyclotron resonance
III-V etc.
Conventional reactive ion etcher (RIE)
Inductively coupled plasma
Si etc.
25
Process for XRD samples
3. Sample deposition by cat-CVD.
1. Drilling by the diamond drill.
a-Si
Si (100)
Si (100)
2. Reactive ion etching for backside of the
substrate.
4. XRD measurements at Spring-8.
Ar CF4
Si (100)
26
XRD for a-Si on Si substrate
Si(220)
Si(220)
Thickness less than 10 mm
Substrate Thickness 500 mm
Thickness of a-Si layer was 1.6 mm.
When the substrate was thin, amorphous peaks
could also be observed.
27
Si substrate and a-Si
S(Q) for a-Si
Structure factor for substrate
Si(220)
Large contribution from c-Si should be removed.
28
RMC modelling for a-Si

• We try to model a-Si structure from the data that
  includes both substrate peaks and amorphous
  peaks.
• Substrate peaks subtraction was done by the
  direct calculation of the ratio of the first peak
  or the normalization calculation using mcgr.

29
FTIR spectra for a-Si
Area ratio of two peaks Si-H Si-H2 9.021
Using A-values for the stretching mode Si-H
Si-H2 1.56 0.4
30
Direct subtraction
The ratio was calculated from the first peak of
c-Si Si (110).
2nd peak of c-Si
31
Direct subtraction without second peak
There is a large second peak in the original
data. It seems to be difficult to be subtracted
by the ratio calculation. Then, the second peak
was deleted.
2nd peak of c-Si
Too small r in g(r) can also be observed.
32
MCGR Normalization
r1
2nd peak of c-Si
r2
First peak in g(r) is around r 2.3. Consistent
to stable a-Si by cat-CVD.
33
MCGR Normalization (without 2nd peak)
r1
2nd peak of c-Si
r2
Fitting of S(Q) was improved by the deleting of
2nd peak.
34
Improvements by mcgr

• First and second large peak in g(r) is around 2.3
  Å and 3.9 Å.
• Coordination number constraint can help the
  convergence.
• Remaining peak in the subtracted data generates
  the error of the modelling.

35
Conclusion

• We can obtain the amorphous peaks from a-Si
  deposited on the Si substrate in the transmission
  geometry.
• MCGR normalization improves to pull out the
  amorphous peaks.

36
Future improvements of estimation

• Using the XRD pattern for the same substrate used
  for deposition.
• The backside layer thickness dependence can help
  the estimation of the substrate patterns.
• Different materials such as SiO2 could help to
  pull out the amorphous peaks.

37
Acknowledgements

• Dr. Kohara of JASRI for the XRD experiments at
  Spring-8 (BL04B2)
• Prof. Kobayashi of ISIR, Osaka Univ. for sample
  preparation of PECVD
• Dr. Tabuchi and Mr. Fukuda of JAIST for sample
  preparation and help of experiments