Control of the Nanostructures in Amorphous Silicon Germanium Thin Films for Improved Photovoltaic Device

ADVANCED NANOMATERIALS AND ITS APPLICATIONS
8-9th August, 2008 School of Materials Science & Nanotechnology,Jadavpur University



Control of the Nanostructures in Amorphous Silicon Germanium Thin Films
for Improved Photovoltaic Device
Ayana Bhaduri* and Partha Chaudhuri
Indian Association for the Cultivation of Science, Jadavpur, Kolkata – 700032,
India
Abstract:
Incorporation of the fine particles or clusters (size range 2-10nm) formed in the
plasma into the amorphous silicon based films and devices deposited by conventional
13.56 MHz rf plasma enhanced chemical vapour deposition (PECVD) method has strong
influence on their performance. The main idea of this study is to regulate the
incorporation of the fine particles/clusters into the a-SiGe:H films by applying square
wave pulse modulation (SWPM) of the rf amplitude. Using SiH4-GeH4 mixture as source
gases with high H2 dilution two series of samples were deposited. For the first one named
as #DC series, where duty cycle was varied from 100% to 50% keeping total time period
(Ton+Toff) at 737msec. The second #Ton series samples were deposited with Ton time
varied from 300msec to 30msec keeping Ton = Toff. In both the cases we have observed a
clear variation in the incorporated particle density for various films. With #DC series
samples as intrinsic layer for single junction pin a-SiGe:H solar cell, we have observed a
significant improvement in ‘Quantum efficiency’ in long wavelength region with 75%
DC sample as i-layer.
* email: ayana.iacs@gmail.com
1. Introduction
Deterministic synthesis of functional nanoassemblies ranging from nanostructures
to intricate nanopatterns and nanodevices is a current demand of modern nanoscience and
nanotechnology [1]. In this regard, the nanostructured Si:H and nanostructured SiGe:H
usually referred as polymorphous hydrogenated silicon (pm-Si:H) and polymorphous
hydrogenated silicon germanium (pm-SiGe:H) respectively are promising materials for
application in single and multijunction solar cells [2]. Incorporation of the fine particles
or clusters (size range 2-10nm) formed in the plasma into the amorphous silicon based
films and devices have strong influence on their performance [3].
2. Experimental Details
The main idea of this study is to regulate the incorporation of the fine
particles/clusters into the a-SiGe:H films by applying SWPM of the rf amplitude of
conventional 13.56 MHz RF PECVD. Using SiH4-GeH4 mixture as source gases with
high H2 dilution two series of samples were deposited. In the first series (#DC) we have
kept the total pulse time period fixed at 737 ìsec while gradually changed the “duty
cycle” (DC) defined by, ´100%
+
=
on off
on
T T
T
DC from continuous mode (DC=100%) to
50% by controlling the ‘on’ and ‘off” times. In the second series (#Ton) we have varied
2
the ‘on’ time (Ton) from 300 msec to 30 msec keeping Ton equal to Toff. Other deposition
parameters are Tdep =250 °C, P = 0.5 Torr, Prf = 25 W. The gas flows were 2.8 sccm of
SiH4, 0.8 sccm of GeH4 and 95 sccm of H2. The deposition parameters were chosen to
keep the plasma conditions close to the powder formation. For HRTEM study the films
of thickness of 300  were deposited on C-coated copper grids placed on the grounded
anode. We have applied #DC series sample as intrinsic layer for single junction solar cell.
The powder or particle incorporation in the films as well as the film
microstructures linked to these powders were investigated by high resolution
transmission electron microscopy (HRTEM). We have counted the particle density from
the TEM micrograph with standard Image software. Surface roughness of the various
layers was measured by atomic force microscopy (AFM). The ambipolar diffusion length
(Ld) measured by steady state photocarrier grating (SSPG) technique. Quantum efficiency
of the solar cells was measured in a calibrated measurement system.
3. Results
3.1. HRTEM study
For the #DC series samples the particle density is highest for the continuous mode
(DC=100%) deposited sample and it gradually decreased with the decrease in duty cycle
Fig. 1(a). For example, the sample deposited with 75% duty cycle has incorporated
particles within size range of 4 – 6 nm. In the Fig. 1(b) we have shown the micrograph of
the sample deposited with DC at 75%. The micrograph depicts a two-phase material
where isolated nanocrystallites and /or nanoclusters are embedded in an amorphous
tissue. Fringe widths of .319nm corresponding to the (111) plane of c-Si.5Ge.5 as well as
.325nm corresponding to the (111) plane of c-Ge have been measured from the
micrograph.
100 90 80 70 60 50
0.0
0.2
0.4
0.6
0.8
1.0
1.2
IPN(mm
-2)
Duty Cycle (%)
(a)
x104
Fig. 1(a) Evolution of particle
density with change in duty cycle
Fig.1(b) HRTEM fringe patterns
of a-SiGe:H for 75% DC
3
In the Figure 2 it is shown that for the
#Ton series we have observed gradual
increase of number density of fine
particles with increasing Ton time. With
increase in Ton time, the particle size also
increases. For example, at Ton of 1 msec,
particle size is ~ 4nm, at Ton of 20 msec,
it is ~ 7-8 nm and for continuous mode it
is above 10 nm. The standard deviation
of the size of the particles in the #Ton
series is ~1 nm.
With the increase in Ton time, the size of the particles increases (Fig. 3 (a-c)).
Moreover there is a decrease in the nanocrystallites as observed from the gradual
decrease in sharpness of the FFT from (a) to (c) of Fig. 3. From Fig. 3 we notice a
gradual increase of the fringe width from .194nm to .198nm as the Ton time is decreased
from continuous (Fig.3(c)) to 1 msec (Fig. 3(a)). The fringe patterns of the
nanocrystallites correspond to (022) plane of crystalline SiGe with increasing Ge
concentration on lowering of the Ton time.
3.2. Surface morphology study by AFM:
Surface morphology of the films was characterized by AFM. AFM images (with
dimensions 1mm ´1 mm) were taken for randomly selected parts of the sample. AFM 3D
0 5 10 15 20 25 30
0
2
4
6
8
10
IPN(mm
-2)
Ton(msec)
x10
3
Fig. 2 Evolution of the particle
density with Ton time
(a) (b) (c)
Fig. 3. HRTEM micrographs of a-SiGe:H samples obtained with Ton times of (a) 1
msec, (b) 20 msec and (c) continuous. FFT corresponding to (022) planes of SiGe
nanocrystallites are shown in the insets.
4
surface topography images of #DC series samples of different duty cycles are shown in
Figs. 4(a) and (b). Statistical analysis of the surface roughness of the images was
performed and described by the root mean square (rms) roughness RRMS. There is a
significant increase in RRMS from 1 to 12 nm with the lowering of the duty cycle from
100% (continuous mode) 50% (Fig.5).
Fig. 4. AFM 3D surface topography image of (a) continuous mode and (b) DC of 75%
3.3. Diffusion length measurement by SSPG
The ambipolar diffusion length (Ld) of the carriers in #DC series of samples
measured by SSPG method [4] at room temperature is shown in Fig. 6. Ld becomes
maximum for DC of 75 %. Here we must underline that this value of Ld is one of the
largest ever reported for a-SiGe:H material exhibiting a bandgap of 1.44 eV [5].
3.4. QE of the solar cell
Continuous 90 80 70 60 50
0
2
4
6
8
10
12
RRMS(nm)
Duty cycle (%)
50 60 70 80 90 100
50
60
70
80
90
100
110
Diffusion length Ld (nm)
Duty cycle (%)
400 600 800 1000
0
20
40
60
80
Quantum Efficiency (%)
Wavelength (nm)
a-SiGe:H
a-Si:H
750nm
14%
1%
(a) (b)
Fig 5.The RMS surface roughness of the a-SiGe:H
films deposited with various duty cycles
Fig. 6. Evolution of the ambipolar
diffusion length Ld with the duty
cycle
Fig. 7 Comparison of QE of an a-
SiGe:H solar cell with an a-Si:H solar
cell
5
We have applied #DC series sample as intrinsic layer for single junction p- i-n
structure solar cell. We have observed a significant improvement in ‘Quantum efficiency’
in long wavelength region with 75% DC sample as i-layer than a standard a-Si:H based
pin solar cell from the Fig 7.
4. Discussion
The SWPM of the RF plasma controlled the selective incorporation of
nanoparticles into the deposited a-SiGe:H films. During the plasma-on time of the
modulation cycle, an amorphous SiGe film is deposited onto the substrate and, at the
same time, nanoparticles nucleate and grow in the ionized gas phase. During the
afterglow periods, the particles leave the plasma and are deposited onto the a-SiGe:H
film. Therefore, after a number of cycles, the final structure will consist of nanoparticles
of SiGe and Ge embedded in the a-SiGe:H matrix .
With lowering of the Ton time the growth of the particles within the plasma is
inhibited. Thus for the #DC series the particle incorporation decreases with decrease in
DC (i.e. lowering of Ton time) and the size of the particles also decreases (Fig. 1(a)). The
sample deposited with DC of 75% consists of the a-SiGe:H material with embedded
nanocrystallites of isotropic shape and nearly uniform size distribution (Fig. 1(b)).
Powder growth in SiH4 discharge takes place in the time scale of the order of seconds [6]
whereas for GeH4 discharge there are reports of faster growth (order of tens of
milliseconds) of the observable powders within the plasma [7]. Our earlier study revealed
a quite important observation that the powder formation within a discharge containing
GeH4 is of the order of hundreds of msec [5]. So fine particles formed in #DC series are
mainly due to Ge-Ge clustering as the time scale is lowered. For the #Ton series we have
observed the effect of Ton more precisely on the particle growth. As the Ton increases,
powder growth from both GeH4 and SiH4 become plausible. For Ton 20 msec,
reasonable fraction of the particle formation comes from SiH4 also. The number and size
of the incorporated fine particles increases with the increase in both Ton (Fig.2 and Figs.
3(a-c)) and duty cycle (Fig. 1(a)). In the Figs. 3(a-c) the gradual increase of the (022)
plane fringe width indicates a corresponding increase in Ge concentration in the
nanocrystallites of SiGe indicating an enhanced Ge clustering with the decrease of Ton
time. A preferential growth in the <022> direction also supports dominance of Ge
nucleation in this time scale range [8].
Surface roughness increases for #DC series samples with the decreasing duty
cycle due to the presence of isolated grains or nanocrystallites as supported by HRTEM
(Figs. 4(a) & (b)). Ld is almost doubled for DC=75% and DC=50% where the surface
roughness is also high (Fig.6). The presence of isolated nanocrystallites improves the
transport properties as reflected from the high Ld values and high photo response of the
samples with embedded nanostructures. The nanostructural features within the SiGe:H
films have very strong influence on their transport properties. The improvement of the
long wavelength QE of the solar cells incorporating the 75% DC deposited SiGe:H as the
i-layer (Fig. 7) clearly indicates the good effect of the incorporated particles from the
plasma.
6
5.Conclusion
Optimum particle size and ion bombardment gives rise to nanocrystallization. So
intelligent control of the fine particle’s size and number density may give improved
transport properties of the materials. In this study we have controlled the density and the
size of the particles incorporated in the SiGe films deposited by rf PECVD from silane -
germane mixture by playing with the plasma on time. Significant changes in the
nanostructure in these samples and the consequential improvement in the transport
properties of the material resulted in the improved quality solar cells.
References
1 Nanotechnology: Shaping the World Atom by Atom, National Science and Technology
Council, 1999; http://www.nano.gov
2 Y. Poissant, P. Chatterjee and Pere Roca i Cabarrocas, JAP, 94 (2003) 7305, M. E.
Gueunier , J. P. Kleider, R. Bruggemann, S. Lebib, P. Roca i Cabarrocas, R. Meaudre and
B. Canut, JAP, 92 (2002) 4959
3 A. Hadjadi, A. Beorchia, P. Roca i Cabarrocas, L. Boufendi, S. Heut, and J. L.
Bubendorff, J. Phys. D 34 (2001) 690
4 D. Ritter, E. Zeldov, and K. Weiser, Appl. Phys. Lett. 49 (1986) 791
5 P. Chaudhuri, A. Bhaduri, A. Bandypadhyay, S. Vignoli, P. P. Ray, C. Longeaud, J. non
Cryst Solids 354 (2008) 2105
6 M. Shiratani, T. Fukuzawa, Y. Watanabe , H. Kawasaki, Jpn. J. Appl. Phys. 38 (1998)
4542
7 H. Kawasaki, J. Kida, K. Sakamoto, T. Fukuzawa, M. Shiratani and Y. Watanabe, J.
Appl. Phys. 83 (1998) 5665
8 X. Niu and V.I. Dalal, JAP, 98 (2005) 96103










Custom Search