1. Art and Diseño

J. Energy Power Sources
Vol. 2, No. 1, 2015, pp. 1-5
Received: September 8, 2014, Published: January 30, 2015
Journal of Energy
and Power Sources
www.ethanpublishing.com
Fill Factor Enhancement of Amorphous Silicon Solar
Cells for High Efficiency Tandem Structures
Osama Tobail
Egypt Nanotechnology Center (EGNC), Cairo University, 12588 Giza, Egypt
Corresponding author: Osama Tobail ([email protected])
Abstract: We focus on three types of material grown by Plasma Enhanced Chemical Vapour Deposition (PECVD) for fabricating the
absorber of the bottom cell in tandem structures: (1) Hydrogenated microcrystalline silicon (µc-Si:H); (2) Hydrogenated amorphous silicon
germanium a-SiGe:H alloy; (3) Hydrogenated microcrystalline silicon germanium (µc-SiGe:H) alloy. A baseline of reproducible and
homogeneous germanium based low band gab absorber bottom cell has been achieved. Solar cell characterization showed that the
fabricated bottom solar cell performance is limited by small fill factor due to high series resistance. AMPS-1D simulation of the fabricated
cells show that the fill factor limitation of the low band gap cells is due to a 0.5 eV barrier at the p/i interface. This barrier was reduced to
0.1 eV by adding Ge to the p-type window layer. The fabricated cells have an increased fill factor from 53.2 to 58.8% due to a reduction of
cell series resistance from 8.3 to 4.8 cm-2.
Keywords: Solar cells, hydrogenated amorphous silicon, tandem cells, band gap engineering.
Nomenclature:




d
DR
EQE
TCE
Ec
Eg
Ev
FF
IQE
J
Jsc
k
L
Lc
n
V
Voc
Rs
T
Absorption coefficient (1 cm-1)
Energy conversion efficiency (%)
Standard deviation
Wavelength of light (nm)
Absorber thickness (nm)
Dilution ratio (%)
External quantum efficiency (%)
Transparent conductive electrode
Conduction band edge (eV)
Energy gap (eV)
Valence band edge (eV)
Fill factor (%)
Internal quantum efficiency (%)
Measured current density (mA cm-2)
Short circuit current density (mA cm-2)
Imaginary part of the refractive index
Absorption length (cm)
Carrier collection length (nm)
Real part of the refractive index
Measured cell voltage (V)
Open circuit voltage (V)
Series resistance ( cm-2)
Temperature (C)
1. Introduction
The cost structure analysis of the supply chain for
crystalline silicon solar cell indicates high potential of
transferring certain industries to countries like Egypt
more than others [1]. Basically, energy and labor
intensive industries are more suitable for Egypt.
However, labor training and know-how localization is
a critical issue for sustainable Photovoltaic (PV)
industry and development. In the frame of localizing
different PV basic technologies, we are developing PV
technologies in the field of wafer based and thin film
silicon solar cells [2-5].
For thin film silicon solar cells tandem structures are
necessary to enable absorption of the whole spectrum.
We found that although the good collection length of
our solar cells in the range of 150 nm [4], which is
larger than our absorber thickness (120 nm), the test
cells suffer from limited fill factor. We ascribe the
limited fill factor to the elevated series resistance of the
test cells. Performing AMPS-1D simulation resulted in
2
Fill Fa
actor Enhance
ement of Amo
orphous Silico
on Solar Cells
s for High Effiiciency Tande
em Structures
s
that the reasoon of the elevvated series reesistance is thhe
potential barrrier between the high ban
nd gap window
w
p-layer and thhe low band ggap absorber i-layer.
i
Addinng
germanium to
t the p-type window layer reduces thhe
potential barrrier from 0.55 eV to 0.1 eV and hencce
reduced the cell series resisstance from 8.3 to 4.8 cm
m-2.
The impact off the new winddow layer on the
t fill factor is
that it was enhanced
e
from
m 53.2 to 58
8.8% due to a
reduction of cell
c series resiistance.
2. Experim
mental Detaiils
Multi-Cham
mber Plasma Enhanced Ch
hemical Vapoor
Deposition (P
PECVD) servves to depositt hydrogenateed
amorphous/m
microcrystallinne silicon an
nd germanium
m
alloys (a-Si::H/µc-Si:H aa-SiGe:H and
d µc-SiGe:H
H).
Different chaambers are used to avoid doping
d
memorry
and cross coontamination. Silane and germane
g
gasees
are used to deposit
d
siliconn and germaniium layers annd
phosphine annd diborane aare used for n- and p-typpe
doping, respeectively. The gas are generrally diluted iin
hydrogen witth a dilution rratio defined as the ratio oof
the flow ratees, DR = H2/((GeH4+SiH4).. The substratte
temperature is
i fixed at T = 250 C by heeating from thhe
bottom via ceramic
c
heateer. The pressu
ure and poweer
used at PECV
VD processing is 4 Torr (5
5.33 mbar) annd
-22
80 mW cm . Optimized deposition paarameters succh
as hydrogen dilution and Ge content are performeed
systematicallyy and publiished elsewh
here [6]. Foor
microcrystalline material, the plasm
ma power annd
pressure in addition
a
to tthe hydrogen dilution ratiio
were varied in ranges too optimize th
he crystallinityy,
which was inndicated by the open circuitt voltage of thhe
cell in additioon to Raman sshift.
Fig. 1 scheematically deppict the cell structure
s
of thhe
p-i-n solar celll on a glass suubstrate. The light penetratees
the glass and the ZnO:Al llayer and is ab
bsorbed mainlly
in the intrinsicc a-SiGe:H abbsorber layer. First,
F
aluminum
m
doped zinc oxxide (ZnO:All) was sputtereed on the glasss
substrate to foorm the Transsparent Condu
uctive Electrodde
(TCE). Surfacce texture was formed on the TCE by shoort
dip in 5% HC
CL solution. Thhe textured fron
nt ZnO togetheer
Fig. 1 A schematic of
o the a-SiGe based (p-i-n) ssolar cell
depositeed by PECVD on a glass supeerstrate.
with thhe back side reeflector compoosed of a stackk of ZnO
and metals
m
(Ag/Al)) enhance thee light trapping. The
TCE deposition
d
parameters aree adjusted tto form
ZnO:A
Al with sheet reesistance of 5 /Sq. The abbsorption
length L = 1/, wh
hile the absorpption coefficieent  of
the layer was determ
mined from thee imaginary paart (k) of
the measured compleex refractive inndex by the reelation 
= 2/kk. The complex refractive inndex (n + ik) and the
layer thickness
t
weere determined by modellling the
measurred white lightt interferometrry [7].
3. Ressults and Diiscussions
Our first results th
he hydrogen dilution
d
ratio eeffect on
the drifft length and hence
h
cell effiiciency was puublished
in Ref. [4]. We repeaated the cell onn large area too test the
statisticcs and homog
geneity of our process. Fig. 2 shows
the bacck side of the solar cells witth the metallicc points.
Each 6 cells are inccluded in a dotted
d
square and the
average av and sttandard deviaation  of the cell
efficienncy are shown
n near to the square. The sstatistics
of the 48
4 cells show
ws a standard deviation
d
of 0.15 over
2
10 cm . Table 1 su
ummarizes thee parameters of solar
cells inn Fig. 2 and its statistics.
The absorber material of the soolar cells in Fig. 2 was
analyzeed by our method publisheed in [4]. The method
determ
mines the drift length from the
t measured Internal
Quantuum Efficiency
y (IQE) withh the analogyy of the
well-knnown method to determine the
t diffusion llength of
Fill Fa
actor Enhance
ement of Amo
orphous Silico
on Solar Cells
s for High Effiiciency Tande
em Structures
s
3
ndard deviation
n for the 48 cellls
Table 1 The mean and stan
output parameeters shown in F
Fig. 2.
Parameter
Mean
Standard deviiation
Voc
522.665
69.7772
Jsc
15.88
0.797
FF
51.44
1.840

4.4
0.319
(a)
Fig. 2 Photoggraph of fabriccated array off 48 solar on aan
area of 10 cm2.
crystalline Si solar cells [8]. This metho
od determined a
carrier collecction length Lc = 190 nm, larger than thhe
absorber layyer thickness of d = 12
20 nm. As a
consequence of the relativvely large co
ollection lengtth
compared to cell
c thickness, the electroniic quality is noot
responsible foor the limited cell performan
nce. Therefore,
the small fill factor (51%) is most probaably limited bby
series resistaance and not material quaality, while w
we
determined a series resistannce in the ord
der of 8  cm--2.
While the celll area is smalll and no laterral resistance is
existing especcially becausee of the low sheet
s
resistancce
of our TCE (/Sq),
(
we asscribe the resiistive losses tto
the device itself especiaally the potential barrierrs
between diffferent layers. To better understand
u
thhe
potential barrriers resultingg from the baand diagram oof
our cells, we modeled the ssolar cell by AMPS-1D
A
[9].
Fig. 3a shhows the enerrgy band diag
grams resulteed
from AMPS--1D for two cells with diifferent p-layeer
band gap Eg = 1.8 eV for ppure a-Si:H an
nd Eg = 1.6 eV
V
for p-type a-SiGe:H. Holees are represeented as circlees
beneath the valence
v
band edge (Ev) and
d its motion is
indicated by a left arrow, w
while electronss (circles abovve
the conductioon band edge Ec) injection iss indicated by a
right arrow. It
I is obvious thhat the photog
generated holees
at the i-layer is
i facing a largger barrier if injected
i
to purre
+
a-Si:H than a--SiGe:H p -laayer. Fig. 3b-3
3c compares thhe
barrier heightt of both casess. It is clear fro
om Fig. 3b thaat
(b)
(c)
Fig. 3 Energy band
d diagrams resulted from A
AMPS-1D
ar cells: (a) with
h different p+-tyype band
simulattion for the sola
gaps; (b
b) with band gap of p+-layer a-Si:H Eg = 1..8 eV; (c)
with baand gap Eg = 1.6
6 eV for the Gee containing p+--layer.
the holles are facing 0.5 eV barrieer to flow intoo a-Si:H,
while only
o
0.1 eV potential
p
barriier is to be croossed in
the casse of a-SiGe:H
H in Fig. 3c. The Ge conttent was
adjusteed according to our optim
mization publiished in
Ref. [66] to adjust the band gap off the p+-layerr to Eg =
1.6 eV
V. To achieve the
t desired baand gap, a Gee content
of 20%
% was incorporated at 5% germane to silaane ratio
[6].
Fig. 4a-4b com
mpares the current/voltag
c
ge (J/V)
charactteristics under simulatedd AM1.5G aand the
Externnal Quantum Efficiency
E
(EQ
QE), respectivvely, for
solar cells
c
with a-Sii:H p-layer annd a-SiGe:H p+-layer.
The im
mproved cell efficiency duue to incorporration of
Ge into the p+-typee window layyer is ascribedd to the
enhancced series resistance from Rs = 8.3 to 4.8  cm-2,
while the
t absorber quality
q
was noot changed. Thhis result
confirm
ms that the red
duction of the p/i interface ppotential
barrierr height by inccorporating Gee into the p+-laayer and
reducinng its band gaap from 1.8 eV
V to 1.6 eV ennhanced
the chaarge carrier collection from
m the absorbeer to the
+
p -layeer and hence increased thee cell perform
mance by
series resistance
r
deccrease to the half.
h
It is knoown that
the p-layer should have
h
a wide baand gap to ennsure the
4
Fill Factor Enhancement of Amorphous Silicon Solar Cells for High Efficiency Tandem Structures
(a)
(a)
(b)
Fig. 4 (a) Measured current/voltage characteristics of the
cells with and without Ge incorporated in the p+-type
window; (b) Comparison of measured quantum efficiency of
cells with and without Ge incorporation into the p-layer.
(b)
Fig. 5 (a) Raman-Shift measurement for c-Si:H and
c-SiGe:H; (b) Measured absorption length as function of
wavelength for c-Si1-xGex:H for x = 0.1, 0.15 and 0.2 and
compared to standard c-Si:H (x = 0).
light transmission to the i-layer where the photogenerated
carriers can be collected by drift. Therefore, quantum
efficiency measurement is critical to investigate the
influence of the p-layer band gap reduction on the cell
spectral behavior. Fig. 4b compares the EQE for the
cells with a-Si:H and a-SiGe:H p-layers. Apparently,
the increase in the carrier collection of short
wavelengths, which are photons absorbed near to the
p/i interface, overcomes the slight decrease of photon
penetration into the i-layer due to the decreased band
gap of the a-SiGe:H p-layer. Consequently, the
overall behavior of the cell with reduced p-layer band
gap showed an improved spectral response beside the
improved electronic and electrical responses. A further
enhancement of the cell performance can be achieved
by reducing the i-layer thickness further to enable more
efficient carrier collection. However, reducing the
layer thickness will reduce the total absorbed photons.
Therefore, we are currently optimizing microcrystalline
silicon germanium alloy (c-SiGe:H) to achieve higher
absorption and at the same time enhanced electronic
quality in addition to elevated growth rate.
Fig. 5a compares between the measured Raman shift
of excitation laser of 633 nm for c-SiGe:H and
Fill Factor Enhancement of Amorphous Silicon Solar Cells for High Efficiency Tandem Structures
5
c-Si:H. For the microcrystalline Si, we observed the
Si-Si peak at 520 cm-1 in addition to the wide peak for
the amorphous Si at 480 cm-1. Both peaks are also
absorption length.
observed in the case of c-SiGe:H, in addition to
another clear peaks at 280 cm-1 and 380 cm-1. The peak
at 280 cm-1 was observed also in the literature [10]. We
ascribe the peak at 380 cm-1 to the Si-Ge interaction,
while its sharpness is related to the crystallinity level.
This work was performed under a joint development
agreement between IBM Research and the Government
of the Arab Republic of Egypt through Egypt
Nanotechnology Center (EGNC). I am very grateful to
A. Abou-Kandil, J. Kim, D. Sadana, M. El-Ashry for
the fruitful discussions and support.
The deposited c-SiGe:H layers shows also very high
absorption compared to c-Si:H. Fig. 5b compares the
measured absorption length of three different
c-SiGe:H layers deposited with variable Ge atomic
ratios to our reference c-Si:H. Increasing the Ge ratio
slightly decreases the absorption length at the
mid-wavelengths, however, it increases the absorption
at long wavelengths due to the mid-gap defects.
Therefore, considering the absorber electronic quality
should be considered together with the absorption
enhancement.
4. Conclusions
The origin of the fill factor limitation of the
a-SiGe:H bottom cell was determined experimentally
and by AMPS-1D simulation and found that not the
material quality, but the series resistance is responsible
for the reduced cell performance. The simulation
suggests decreasing the p-layer band gap from 1.8 to
1.6 eV to reduce the potential barrier from 0.5 to 0.1 eV.
Fabrication of identical cells except with p+-layers with
1.8 eV for one type and 1.6 eV for the other type
proved that the series resistance was reduced from 8.3
to 4.8  cm-2. The reduced series resistance resulted in
an increase in the carrier collection especially carriers
generated near to the p/i interface, as was obvious in
the measure quantum efficiency beside the J/V
characteristics and hence an 5.6% abs. enhancement of
the fill factor. For further enhancement of the bottom
cell, we are suggesting microcrystalline silicon
germanium alloy as an absorber with 5 times enhanced
Acknowledgment
References
[1]
O. Tobail, Potenial of photovoltaic industry in Egypt, J.
Energy Power Engineering 7 (2013) 1844-1851.
[2] K.E. Fogel, J.H. Kim, D.K. Sadana, J.S. Tulevski, A.I.
Abou-Kandil, H.S. Mohamed, M. Saad, O. Tobail,
Photovoltaic devices with an interfacial band-gap
modifying structure and methods for forming the same,
US patent Pub. No.: US 2012/0031477 A1, Feb. 9, 2012.
[3] O. Tobail, A.I. Abou-Kandil, M. El-Ashry, J. Kim, P.
Kozlowski, M. Saad, D. Sadana, Deposition of
hydrogenated thin film, US patent Pub. No.: US
2012/0156393 A1, Jun. 21, 2012.
[4] O. Tobail, Method to determine the collection length in
field-driven a-SiGe:H solar cells, Energy Procedia 10
(2011) 213-219.
[5] K. Wang, O. Gunawan, N. Moumen, G. Tulevski, H.
Mohamed, B. Fallahazad, E. Tutuc, S. Guha, Wire
textured, multi-crystalline Si solar cells created using
self-assembled masks, Opt Express 8 (2010) A568-A574.
[6] O. Tobail, Optimization of a-SiGe solar cells for tandem
structures, in: Proc. 3rd IEEE Thermal Issues in Emerging
Technologies Theory and Applications (ThETA 3) 2010,
pp.309-312.
[7] O. Tobail, Porous Silicon for Thin solar Cell Fabrication,
Shaker Verlag, 2009, p. 38.
[8] P.A. Basore, Extended spectral analysis of internal
quantum efficiency, in: Proc. 23rd IEEE Photovoltaic
Specialists Conference, Piscataway, New York, 1993, pp.
147-152.
[9] AMPS-1D Manual, The Center for Nanotechnology
Education and Utilization, The Pennsylvania State
University, http://www.ampsmodeling.org/.
[10] K.V. Maydell, K. Grundewald, M. Kellermann, O.
Sergeev, P. Klement, N. Reinighaus, T. Kilper,
Microcrystalline SiGe absorber layer in thin-film silicon
solar cells, Energy Procedia 44 (2014) 209-215.