Photoconductivity in Inverse Silicon Opals Enhanced by Slow Photon Effect- Yet
Another Step Towards Optically Amplified Silicon Photonic Crystal Solar Cells
Takashi Suezaki,1,2 Hiroshi Yano2, Tomoaki Hatayama,2 Geoffrey A. Ozin3 and Takashi
Fuyuki.2
1
Photovoltaic & Thin Film Research Laboratories, Kaneka Corporation
5-1-1, Torikai-Nishi, Settsu, Osaka 566-0072 (Japan)
2
Microelectronic Device Science Laboratory, Graduate School of Materials Science, Nara
Institute of Science and Technology
8916-5 Takayama, Ikoma, Nara 630-0192 (Japan)
3
Materials Chemistry Research Group, Department of Chemistry, University of Toronto
80 St. George Street, Toronto, ON M5S 3H6 (Canada)
Details for preparation of inverse silicon opals
We fabricated i-Si-o by infiltrating the voids of the silica opal template with Si using
chemical vapor deposition (CVD) with disilane and subsequently etching the silica spheres
with hydrofluoric acid (HF) [1]. Silica spheres with diameter 460 nm were synthesized by a
modified Stöber process involving the base catalyzed hydrolytic poly-condensation of
tetraethylorthosilicate. We chose silica opals with 460 nm sphere diameter for templates
because our previous work showed little dependence of σd on the photonic crystal lattice
constant of i-Si-o. In order to form the silica opal templates, spheres were dispersed in ethanol
for several hours and then crystallized as silica opal film by evaporation induced
self-assembly (EISA) method on sapphire substrates. These substrates were cleaned by
piranha before silica opal film formation. The template opal deposition conditions, such as
concentration of spheres in ethanol (2.5 wt. % sphere) and rate of evaporation (0.4 ml/Hour)
on sapphire substrates were optimized in order to achieve well-ordered opal film. Infiltration
of amorphous silicon into the interstitial voids of silica opals was subsequently performed
using a CVD apparatus and disilane as a silicon precursor. For the CVD we used a disilane
pressure of 1.0 torr and deposition temperature of 480 ˚C. Then reactive ion etching was
performed to remove silicon over-layers that were deposited during the silicon infiltration
process. The as-deposited Si in the interstitial voids of silica opals by CVD is intrinsically
amorphous. To obtain inverse structure, we etched silica opal with 2 wt. % HF. We prepared
all samples on sapphire substrates instead of glasses due to their high stability against HF
etching and that made it much easier to form intact electrodes for accurate measurements of
the dc electrical dark conductivities (σd) for the i-Si-o films. [2]
Measurements details
Reflectance spectra of each sample at every stage of processing were measured using a Perkin
Elmer UV/VIS/near-IR spectrometer Lambda 900. The probe light diameter was about 2 mm.
Scanning Electron Microscopy (SEM) images were obtained using Hitachi, S-5200.
To measure the dc electrical dark conductivity (σd), the current was recorded while the bias
across the electrode pair was increased from -100 V to +100 V in increments of 20 V.
Aluminum co-planar electrodes with a separation of 1 mm were evaporated on the surface of
the samples by a physical vapor deposition apparatus. The σd was determined from the
resulting Current – Voltage (I-V) curves, which were linear implying a good ohmic contact
between the aluminum and the i-cSi-o, with the equation σd =IL/VA, where L is the gap
length between two electrodes and A is the cross sectional area of the films. The area includes
not only the Si framework but also the void spaces. Low temperature conductivity
measurements were performed with a low temperature measurement system (Nagase Co.) and
the activation energy (Ea) was determined from a plot of measured σd as function of reciprocal
temperature using the same electrode configuration under a bias of 100 V. A Keithley 6517A
High Resistance Meter was used for all electrical measurements.
Cross Sectional image of i-cSi-o
Figure S1. Cross sectional SEM image of i-cSi-o obtained from 600 nm spheres (a) and 280
nm (b) on sapphire substrates. The bottom side of images is sapphire substrate.
The cross sectional SEM images of i-cSi-o obtained from 600 nm spheres (a) and 280 nm (b)
shows good face centered cubic (fcc) structural stacking to [1 1 1] direction in spite of some
defects probably originated by the sample preparation for SEM cross sectional imaging. The
stacking layer numbers are 11 for 600 nm and 26 for 280 nm, respectively and those numbers
are enough for possessing photonic band structures.
Comparison of optical property of i-cSi-o from theoretical calculation and experimental
result
Figure S2. Comparison of optical photonic properties of i-cSi-o from theoretical calculation
and experimental result for that of obtained from 600 nm spheres (a) and 280 nm spheres (b).
The left side is a photonic band diagram for i-cSi-o from theoretical calculation. The right
side is the corresponding experimental reflectance spectrum of i-cSi-o.
The Г-L direction; corresponding to the [1 1 1] direction for a face centered cubic (fcc) lattice;
has three photonic band gaps (depicted with shading) from 1000 nm to 2000 nm for i-cSi-o
obtained by 600 nm spheres (a) and from 400 nm to 1000 nm for i-cSi-o obtained by 280 nm
spheres (b) in a theoretical calculation[3, 4]. The light with wavelengths at photonic band gaps
could not penetrate into the photonic crystal lattice and therefore those wavelengths should be
reflected. The experimental reflection spectrum of i-cSi-o is well matched to the theoretical
calculation in spite of the finite lattice and surface morphology that should cause a deviation
of intensity from perfect reflection.
The IQE spectra after HPP treatment at 133 K for i-cSi-o and the corresponding
reflectance spectra
(a)
100 diameter: 280 nm
133 K
50
2.0
630 nm
after HPP
4.0
3.0
0.15
0.10
50
IQE [%]
Reflectance [%]
133 K
after HPP
5.0
Reflectance [%]
diameter: 600 nm
IQE [%]
100
(b)
0.05
1.0
0
400 600 800
Wavelength [nm]
0.0
0
400 600 800
Wavelength [nm]
0.00
Figure S3. (a) The IQE after HPP treatment at 133 K for i-cSi-o obtained from 600 nm
spheres (red) and the corresponding reflectance spectrum (black). (b) The IQE after HPP
treatment at 133 K for i-cSi-o obtained from 280 nm spheres (red) and the corresponding
reflectance spectrum (black). The arrows with numbers point to the peak positions of the IQE
spectra.
Figure S3a shows the IQE after HPP treatment at 133 K for i-cSi-o made from silica spheres
with a 600 nm diameter and the corresponding reflectance spectrum. The IQE after HPP for
i-cSi-o from 600 nm showed increasing trend via long wavelength, however no specific
correlation with reflectance was observed, that effect should be originated from a mechanism
for transportation of electrical carriers in the semiconductor with carrier trap states as it was
discussed in the text. At Ultara-Violet range was different and the ratio was slightly higher
than visible range, which might come from complex surface conduction or influence of
contamination in Si matrix on i-cSi-o. On the contrary, the IQE after HPP for i-cSi-o from
280 nm kept a relatively narrow peak at 630 nm as well as before HPP; however a shoulder at
380 nm was almost disappeared; which almost coincide with its reflection spectrum as well.
(Fig. S3b).
Enhancement ratio of the IQE spectrum for i-cSi-o obtained from 600 nm spheres
Figure S4. The enhancement ratio of the IQE spectrum for i-cSi-o obtained from 600 nm
spheres at 133 K after HPP with that of the corresponding reflectance spectrum.
Figure 5 shows the enhancement ratio of the IQE spectrum for i-cSi-o obtained from 280 nm
spheres at 133 K after HPP with that of the corresponding reflectance spectrum; the sign of
slow photon effect and increasing trend via long wavelength are observed. In contrast, the
case of i-cSi-o obtained from 600 nm spheres does not seem to be shown slow photon
correlation.(Figure S4) However increasing trend via long wavelength are observed similarly
because that effect should be originated from a mechanism for transportation of electrical
carriers in the semiconductor with carrier trap states as it was discussed in the text.
The behaviour of enhancement ratio for i-cSi-o from 600 nm spheres at Ultara-Violet range
was different and the ratio was slightly higher than visible range, which might come from
complex surface conduction or influence of contamination in Si matrix on i-cSi-o. However,
at least it is less important at the UV range behaviour for photovoltaic application due to
existence of absorption loss by front TCO or emitter.
References
[1]
N. Tétreault, H. Míguez, G. A. Ozin, Adv. Mater., 16, 1471, 2004.
[2]
T. Suezaki, P. G. O’Brien, J. I. L. Chen, E. Loso, N. P. Kherani and G. A. Ozin, Adv.
Mater., 21, 559, 2009
[3]
S. John, Phys. Rev. Lett., 58, 2486, 1987.
[4]
E. Yablonovitch, Phys. Rev. Lett., 58, 2059, 1987.