See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/255001649

Simulation of Device Parameters of High Efficiency Multicrystalline Silicon

Solar Cells

Article · January 2011

DOI: 10.1680/emr.11.00007

CITATIONS READS

0 151

3 authors:

Vinay Budhraja D. Misra

Institute of Technology and Management New Jersey Institute of Technology
20 PUBLICATIONS   59 CITATIONS    190 PUBLICATIONS   1,070 CITATIONS   

SEE PROFILE SEE PROFILE

Nuggehalli M Ravindra
New Jersey Institute of Technology
419 PUBLICATIONS   2,710 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Semiconductors - Synthesis, Properties, and Applications View project

Semiconductors - Synthesis, Properties, and Applications View project

All content following this page was uploaded by Nuggehalli M Ravindra on 23 January 2014.

The user has requested enhancement of the downloaded file.

Emerging Materials Research Pages 25–32 http://dx.doi.org/10.1680/emr.11.00007
Research Article
Volume 1 Issue EMR1
Received 06/09/2011 Accepted 08/11/2011
Simulation of device parameters of Published online 10/11/2011
high efficiency multicrystalline silicon Keywords: solar cells/device simulation/silicon

solar cells
Budhraja, Misra and Ravindra

ice | science ICE Publishing: All rights reserved

Simulation of device parameters of high

efficiency multicrystalline silicon solar cells
1 Vinay Budhraja PhD 3 Nuggehalli M. Ravindra PhD
Research Assistant, National Renewable Energy Laboratory, Golden, USA; Professor and Chair, Department of Physics, New Jersey Institute of
New Jersey Institute of Technology, Newark, USA Technology, Newark, USA
2 Durgamadhab Misra PhD
Professor, Electrical and Computer Engineering Department, New Jersey
Institute of Technology, Newark, USA

1 2 3

The results of the simulation of the reported experimental results of high efficiency multicrystalline silicon (mc-Si) solar
cells, using PC1D software, are reported in this study. Results obtained by various groups have been incorporated and
compared in this study. The highest efficiency reported so far for mc-Si solar cells is 20·4% and 17–18% by research
laboratories and commercial houses, respectively. The efficiency can be further enhanced if passivation characteristics
on both the front and back surface are improved. The role of back surface recombination has become more significant in
light of the use of thin mc-Si wafers by the solar cell industry. Based on the passivation characteristics and considering
the understanding of the past three decades of studies, the authors have proposed and simulated a structure for mc-Si
solar cells to improve the performance of the same. The results of our modeled structure of mc-Si solar cell show an
efficiency of 21·88% with short-circuit current density, Jsc = 39·39 mA/cm2, and open circuit voltage, Voc = 0·666 V.

1. Introduction the understanding of defects and characterization techniques have

Multicrystalline silicon (mc-Si) solar cells account for around 50% also been discussed (Budhraja et al., 2011a).
of the total solar cell production. According to the January 2011
issue of Solar Industry (Magazine Article, 2011), mc-Si technology In the past three decades, the useful advancements in mc-Si solar
represents roughly 80% of the c-Si market on a dollar per watt basis cell technology were aimed at improving the efficiency from 6·14%
against three challengers: thin film silicon, CdTe and Copper indium in 1980 to 20·4% in 2004. Here, the authors focus mainly on the
gallium (di) selenide (CIGS). The mc-Si technology will continue modeling of world record cells. Besides the research groups that
to be profitable throughout the value chain from $1·45/W in 2009 to achieved world record efficiencies in mc-Si solar cells, there were
$0·93/W in 2015, assuming polycrystalline silicon (pc-Si) pricing some other groups that reported several new techniques in the
at $70/kg. The cost of mc-Si solar cells is less than monocrystalline processing steps of their mc-Si solar cells. These research groups
silicon solar cells because it is produced from lower grade feedstock could not achieve the highest efficiency but their contributions were
and out of spec material from the microelectronics industry. The well accepted in the advancements of mc-Si solar cell technology.
mc-Si solar cells are the better alternative to monocrystalline silicon
because of the tradeoff between cost and efficiency (Tobias et al., 1.1 Brief review of experimental work previously
2004). The reasons for the low performance of mc-Si solar cells reported by other groups
in comparison to monocrystalline silicon solar cells are: (i) less The cells listed below are those which have been fabricated by
capability for uniform texturing in mc-Si, (ii) surface passivation various research groups. Most of the processing steps are similar
of front and back surfaces is still not greatly improved, (iii) defects except for several variations used by these groups in process
such as dislocations, grain boundaries, defect clusters etc. are conditions at some steps. The only difference between the
present in mc-Si and (iv) the mechanical strength of mc-Si wafer is multicrystalline silicon (mc-Si) and polycrystalline silicon (pc-Si)
less than that of monocrystalline silicon wafer. The advancements is the grain size. The grain size of mc-Si is between 1 mm and
in mc-Si solar cell technology for the past three decades, in 10 cm and the size of grains in pc-Si is between 1 µm and 1 mm,
chronological order, have been reviewed and the advancements in respectively (Basore, 1994).

25
Emerging Materials Research Simulation of device parameters of
high efficiency multicrystalline silicon
solar cells
Budhraja, Misra and Ravindra

Cell 1: In 1980, T. Saitoh (Saitoh et al., 1980) achieved 6·14% cell efficiency was improved due to enshrouding cell surfaces in
efficiency in pc-Si solar cells by annealing in N2 and the efficiency thermally grown oxide and isotropic etching to form a hexagonally
was further increased by increasing the annealing time. It was symmetric honeycomb surface texture. This group achieved 18·2%
found that impurity gettering which begins at the wafer surface efficiency in HEM material using a honeycomb surface texture.
caused a decrease in junction leakage and enhanced open circuit This value was still less than the efficiency obtained by Rohatgi
voltage and fill factor. et al. in HEM material. This study shows that the cell efficiencies
mainly depend on the quality of mc-Si wafer.
Cell 2: In 1984, S. M. Johnson (Johnson et al., 1984) achieved
15·7% efficiency in their pc-Si solar cells. The grain size of the Cell 9: In 2004, O. Schultz (Schultz et al., 2004a) reported an
pc-Si wafer was around 0·5 cm. To obtain high efficiency in solar efficiency of 20·4%, which is the highest efficiency reported so far
cells, the polycrystalline material was casted in such a way that the in mc-Si solar cells. Schultz used wet SiO2 passivation instead of
base resistivity and other properties, similar to Czochralski silicon, thermally grown SiO2 because thermally grown oxide decreases the
were achieved. Because of this processing step, they called their lifetime which further reduces the cell parameters (Macdonald and
material semi-crystalline. In this case, two layers (MgF2 and Ta2O5) Cuevas , 2000; Schultz et al., 2004b) The physical cause of thermal
of AR coating were used. These layers provided an additional degradation is due to the dissociation of impurity precipitates which
advantage for reducing reflectance. results in a greater concentration of interstitial impurities.

Cell 3: In 1985, S. R. Wenham (Wenham et al., 1985) used In this study, the authors have modeled the cell structures 1–9. Based
plasma treatment in the processing of pc-Si solar cells. They also on the understanding of processing steps used in the fabrication
attempted the same processing steps on their float zone silicon of the cells discussed earlier, they have modeled a cell structure
wafers. Plasma processes were performed in three successive steps. so that high efficiency can be achieved in mc-Si solar cells. The
Plasma hydrogenation, plasma etching and plasma deposition of aim is to move one step beyond what has been achieved so far in
AR coating provided advantages such as the increase in short- mc-Si solar cell technology so that the performance of mc-Si solar
circuit current density (Jsc), contact resistance reduction and an cells can approach that of solar cells made on single crystal silicon.
overall improvement in cell parameters, respectively. This process The solar cell performance characteristics were modeled in PC1D
was called passivated emitter solar cell (PESC) because of the (Clugston and Basore, 1996). PC1D is a software package that uses
passivation of the n-region by SiO2. finite-element, drift-diffusion analysis to solve the fully coupled,
two-carrier semiconductor transport equations in one dimension
Cell 4: One year later, in 1986, S. Narayanan (Narayanan et al., (Sana et al., 1993).
1986) used phosphorous pretreatment in PESC. Phosphorous
pretreatment steps resulted in the improvement in diffusion length
and cell parameters. 2. Process steps: our modeled structure
The general fabrication technique for making solar cells from silicon
Cell 5: M.A. Green (Green et al., 1984) were able to increase material is well known. The particular processing steps which will
the efficiency by 1%. They developed buried contact technology be used in our modeled structure of a solar cell are as follows:
in their cell and used laser technique for texturing. The structural (i) starting material: p-type silicon with resistivity 0·6 Ω-cm and
rounding in their cell reduced stress concentration and dislocation thickness 150 µm, (ii) gettering process performed to remove both
generation during subsequent oxidation. areal and in-depth non-uniformities, (iii) plasma texturing, (iv) P
(phosphorous) diffusion (typically 900–950°C for 5–15 min), (v)
Cell 6: In 1993, P. Sana (Sana et al., 1993) reported a 1% increase junction isolation to remove n region from the wafer edges, (vi)
in efficiency in their cells in which the oxide passivation was hydrogenation for defect passivation, (vii) wet SiO2 (Schultz et al.,
performed after phosphorous gettering. Oxide passivation results 2004a) on front and back, (viii) AR coating of silicon nitride or
in an increase in Jsc and a decrease in saturation current (I01). TiO2 to reduce reflection losses, (ix) photolithography on front side
to make contact, (x) metallization on front side to make contact,
Cell 7: In 1996, A. Rohatgi (Rohatgi et al., 1996) reported 18·6% (xi) photolithography on back side to make contact and (xii) firing
efficiency in their cells in which the mc-Si was grown by heat to achieve proper back surface field (BSF).
exchanger method (HEM). But their process steps were the same
as used in OTC (Osaka Titanium Corporation) for mc-Si material. In the modeled structure of mc-Si solar cell, the authors choose the
In both OTC and HEM material, the effect of grain boundaries can above discussed processing steps. These are summarized in Figure 1.
be neglected because of the large grain size.
Figure 2 shows the layout of the modeled structure. Selective
Cell 8: In 1998, J. Zhao (Zhao et al., 1998) reported 19·8% metallization on the front and back help to improve passivation.
efficiency in cells fabricated on honeycomb textured mc-Si. The The role of front and back surface passivation is important in the

26
Emerging Materials Research Simulation of device parameters of
high efficiency multicrystalline silicon
solar cells
Budhraja, Misra and Ravindra

Parameters for PCID

DEVICE EXCITATION
Device Area: 1 cm2 Excitation modified from one-sun.exc
Front surface Texture Angle: 54.74 Excitation mode: Transient, 16 timesteps
Front surface barrer: 0.05 eV 0
Temperature : 25 C
Exterior Front reflectance: 0.05 Base circuit Sweep from -0.8 to 0.8 V
Internal optical reflectance enabled -2
Front surface optically rough Constant intensity: 0.1 W cm
Spectrum from am15g.spc
Base contact: 1.5 X 10-3 ohm
Internal conductor: 3 X 10-5 S RESULTS
Short-circuit Ib : 39.39 amps
REGION 1 Max base power out : 2.194 watts
Thickness : 150μm Open-circuit Vb: 0.6659 volts
Material used: Silicon
Carrier mobilities from internal model
Dielectric constant: 11.9
Bandgap: 1.1 eV
10 -3
Intrinsic conc. at 300 K : 1X 10 cm
Refractive index: Use standard file for silicon
Absorption coefficient: Use standard file for silicon
No free carrier absorption used
P-type background doping: 2.652 X 1016cm-3
Front diffusion: N-type, 1019cm-3
Rear diffusion: P-type, 1020cm-3
Bulk recombination life time : 50 μs
Front surface recombination velocity: 1000 cm/s
Back surface recombination velocity: 1000 cm/s

Figure 3. Summary of parameters for PCID model.

For the bulk substrate, the standard properties of silicon were used.
The standard data for the absorption coefficient and refractive index
Figure 1. Processing steps in our structure.
was used at 300K. The resistivity of the base substrate was taken to
be 0·65 Ω cm with boron doping concentration of 2·65 ×·1016 cm−3.
The bulk lifetime was taken to be 50 µs. In the world record cells,
Metal grid the research groups (Rohatgi et al., 1996; Schultz et al., 2004a;
SiNx Zhao et al., 1998) have used good quality bulk material made from
Wet SiO2 several methods such as HEM, electromagnetic casting (EMC),
n direct solidification system (DSS) and so forth. The bulk lifetime
in these materials is more than 100 µs but the authors choose the
p type p+ bulk lifetime to be 50 µs to show the modeling results for the worst
case. Phosphorous diffusion was used on the front side to make an
AI
n-p junction. The peak doping concentration on the front side was
taken to be 1019 cm−3. The internal series resistance was chosen
Figure 2. Final layout of our modeled structure. to be 1·5 mΩ on base contact. These values of resistances were
taken from 16% efficiency mc-Si solar cells. The experimental data
of these cells has been reported (Sopori, 2010). The front surface
present technology of mc-Si solar cells as the thickness of silicon
recombination velocity in n-region was varied from 103–106 cm/s.
wafer is reduced by the solar cell industry to reduce material costs.
Similarly, the back surface recombination velocity in p-region
The effect of front and back surface recombination velocity on the
was varied from 103–106 cm/s. The parameters used in PC1D are
performance on mc-Si solar cells has been discussed elsewhere
summarized in Figure 3.
(Budhraja et al., 2011b).

3.2 Comparison of current density (J) –

3. Results: modeling
voltage (V) characteristics of our modeled
3.1 Details of modeling parameters structure with other world record cells
used in our modeled structure For the light J-V characteristics, the spectrum of air mass 1·5
The modeling of structure as well as of the other cells was (Website 1) was used in the PC1D modeling tool. This corresponds
performed (i.e., cell 1–cell 9) using the PC1D modeling tool. The to a power density of 0·1 W/cm2. The light J-V characteristic of
model parameters used for simulation of cell 1–9 was taken from our modeled structure (Figure 2) was compared with the light J-V
the experimental data published by various groups. The details of characteristics of other cells discussed in section 1. This comparison
parameters used for the modeling of the structure are given in this is shown in Figure 4.
section. Since the solar cell industries have started to use thinner
wafers to reduce cost, the thickness of p-type wafer was chosen to The values of cell parameters including short circuit current density
be 150 µm. The internal reflectance of the cell was taken to be 0·05. (Jsc), open circuit voltage (Voc), fill factor (FF) and efficiency for

27
Emerging Materials Research Simulation of device parameters of
high efficiency multicrystalline silicon
solar cells
Budhraja, Misra and Ravindra

Cell (Year) Efficiency 3.2.1 Discussion on Table 1

Cell 1(1980) 6.14%
40 Cell 2(1984) 15.7% There is increment in all cell parameters from cell 1 to cell 2.
Cell 3(1985) 15.8% Insufficient passivation was used in cell 1. The high surface
35 Cell 4(1986) 15.9%
Cell 5(1989) 16.7% recombination limits high efficiency and other cell parameters in
Cell 6(1993) 17.7% cell 1. The use of a low grade polycrystalline wafer was another
30
Cell 6(1996) 18.6%
J (mA/cm2)

Cell 6(1998) 19.8% reason for reduced performance in the fabrication of cell 1. There
25 Cell 6(2004) 20.4% was a sudden jump in efficiency from cell 1 to cell 2. The pc-Si
Our modeled
20 structure, 21.88% wafer with a grain size of 1–2 mm was used in the fabrication of cell
15 2 which increases the efficiency of cell 2. The effective diffusion
length of the wafer used in the processing of cell 2 was more than
10
the diffusion length of the wafer used in the processing of cell 1.
5
From cell 2 to cell 4, there is not too much change in cell parameters.
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 The wafer quality and process steps used in the fabrication of cell
Voltage (V) 2, cell 3 and cell 4 were generally similar. The plasma treatments
used in the fabrication of cell 3 helps to improve its efficiency
Figure 4. Simulated results of J-V characteristics of the modeled
from 15·7 to 15·8%. The plasma treatment step consists of plasma
structure and other cells.
hydrogenation, plasma etching of cell emitter and plasma silicon
nitride deposition which helps to increase cell performance
various cells and our modeled structure are shown in Table 1. parameters. The process steps in cells 3 and 4 were exactly the same
Moreover, Table 1 summarizes the values of surface recombination except there is an additional step of phosphorous pretreatment in
velocity on the front surface (S1) and on the back surface (S2) used in cell 4. Phosphorous pretreatment increases the effective diffusion
the modeling. The values of cell thickness, resistivity, bulk lifetime, length by the diffusion of phosphorous in the poor crystallographic
front doping concentration and sheet resistance for cells 1–9 were regions of the wafer. Increase in diffusion length increases the
taken from the details of their references. There is negligible efficiency by 0·1% in cell 4. There is not much difference in front
difference between the modeling results and experimental results and back surface recombination in cells 2, 3 and 4.
of cell 1–9. The simulations based on PC1D are in excellent
agreement with the experimental results. Based on these results, There was another increase in efficiency from cell 4 to 5. Laser
the modeling of our structure was done in PC1D. texturing was used in the processing of cell 5 which increases

Jsc Voc Efficiency S1 S2

Cell no. Year (mA/cm2) (V) Fill factor (%) (cm/s) (cm/s)

1 1980 16·000 0·573 0·760 6·14 106 1·× 106

2 1984 33·675 0·609 0·779 15·70 104 1·× 106
3 1985 33·925 0·605 0·775 15·80 104 1·× 105
4 1986 34·032 0·605 0·781 15·90 104 1·× 105
5 1989 36·347 0·602 0·760 16·70 105 1·× 104
6 1993 35·170 0·623 0·794 17·70 104 1·× 104
7 1996 36·902 0·641 0·812 18·60 104 2·× 103
8 1998 38·178 0·656 0·802 19·80 103 1·× 104
9 2004 38·020 0·664 0·809 20·40 103 1·× 103
Our structure 2011 39·390 0·666 0·834 21·88 103 1·× 103

Table 1. Comparison of cell parameters of proposed structure with

other cells.

28
Emerging Materials Research Simulation of device parameters of
high efficiency multicrystalline silicon
solar cells
Budhraja, Misra and Ravindra

optical absorption on the front surface. Instead of screen printing,

patterning was used in cell 5 to make the front contact. Solar Rs Rsh J01 (mA/ J02 (mA/
cell (Ω.cm2) (kΩ) cm2) cm2)
There is a 1% increment from cell 5 to 6. An emitter etch back
technique was used to increase sheet resistance in the processing of Cell 1 2·87 0·30 2·68 × e–09 2·03 × e–07
cell 6. Moreover, the controlled back surface field helps to increase Cell 2 2·15 3·01 5·96 × e–09 3·40 × e–07
open circuit voltage in cell 6 resulting in significant improvement
Cell 3 2·13 3·04 5·71 × e–09 3·43 × e–07
in Voc and fill factor.
–09
Cell 4 2·12 3·84 1·73 × e 4·60 × e–07
But for the bulk material, there were the same process steps used in Cell 5 1·51 25·40 1·64 × e–09 3·48 × e–07
the fabrication of cell 6 and cell 7. The bulk life time of the wafer Cell 6 1·52 34·53 7·69 × e–09 2·45 × e–07
used in cell 7 was larger than that of the wafer used in cell 6. –09
Cell 7 1·52 54·56 8·83 × e 5·85 × e–07

From cell 8 to 9, there was 1·2% increment in efficiency. Patterning Cell 8 1·51 54·78 4·44 × e–09 4·10 × e–06
was used on both the front and back surface in the processing of Cell 9 1·48 55·95 2·01 × e–09 1·26 × e–06
cell 9. The honeycomb texture used in the fabrication of cell 9 Our 1·37 60·00 1·89 × e –09
1·20 × e–06
provides uniform texturing to increase multiple reflections on the structure
front surface.
Table 2. Parameters obtained from dark J-V characteristics.
The processing of cell 9 provides good surface passivation on the
front as well on the back surface. The wafer used in the fabrication
of cell 9 was of better quality than that used for cell 8. The research
group which processed cell 9 also attempted to fabricate a cell where, J01 and J02 are the saturation current densities, k is
using the wafer (i.e., HEM; used in the fabrication of cell 8) and Boltzman’s constant, T is the temperature, q is the electronic
achieved efficiency of 18·2%. This indicates that the performance charge, Rs is the series resistance, Rsh is the shunt resistance and
of the cell is mainly affected by the quality of the wafer. JL is the light induced current density and it equals to zero in dark
conditions.
There is very little change in efficiency and other performance
parameters from cell 8 to 9. Instead of thermal oxidation, wet SiO2 Solar cell performance parameters such as J01, J02, Rs, Rsh were
was used in the fabrication of cell 9. Thermal oxidation reduces calculated from the dark I–V curves and are shown in Table 2.
the bulk lifetime which further affects the efficiency and other cell The first term in Equation 1 is due to the recombination current
parameters. in the quasi neutral region (from first diode) and the second term
is due to the recombination current in the depletion region (from
In the proposed structure, the authors mainly focus on the the second diode). The value of series resistance (Rs) decreases
importance of front and back surface recombination velocities. and shunt resistance (Rsh) increases from top to bottom in Table 2.
They have assumed the front and back surface recombination This is in agreement with the explanation that high value of series
velocities to be 103 cm/s which are expected to be achieved resistance and low value of shunt resistance corresponds to the low
experimentally. On the front, they selected uniform texture which performance of the cell.
helps to increase multiple reflections. Large grain size which
reduces the grain boundary effects was selected (Ghitani and
Martinuzzi, 1989; Halder and Williams, 1983). Features such as 3.3 Effect of front surface passivation
uniform texturing, high diffusion length, low lifetime, high sheet in our modeled structure
resistance, low surface recombination on front and back etc. were The possible processing techniques for front surface passivation
used in the modeling of the structure. The results of the modeled that have been followed by various research groups are thermal
structure showed highest efficiency. oxidation, selective metallization, SiNx:H passivation, etc. Front
surface recombination velocity (S1) is an important parameter to
The dark current density (J) - voltage (V) curve fitting of the solar measure the passivation at the front surface. Figure 5(a) shows
cells was performed using a two-diode model. The equation is as the variation in J-V characteristics with respect to S1. For most
follows (Hussein et al., 2001): semiconductors, the surface recombination velocity is of the order
of 107 cm/s (Website 2) but, experimentally, it has become possible
1. ⎪⎧ ⎡ q (V − Rs I ) ⎤ ⎫⎪ ⎪⎧ ⎡ q (V − Rs I ) ⎤ ⎪⎫ ⎛ V − Rs I ⎞ to reduce the value of S1 to 103 cm/s (Rohatgi et al., 1996; Sana et
J = J 01 ⎨exp ⎢ ⎥ − 1⎬ + J 01 ⎨exp ⎢ ⎥ − 1⎬ + ⎜ ⎟ − JL ,
⎩⎪ ⎣ kT ⎦ ⎭⎪ ⎩⎪ ⎣ 2 kT ⎦ ⎭⎪ ⎝ Rsh ⎠ al., 1993; Schultz et al., 2004a). As the front surface recombination
velocity increases, the value of cell performance parameters such

29
Emerging Materials Research Simulation of device parameters of
high efficiency multicrystalline silicon
solar cells
Budhraja, Misra and Ravindra

(a) 40
Back surface Jsc
35 recombination (mA/ Voc Fill Efficiency
velocity (cm/s) cm2) (V) Factor (%)
30
S2=103 cm/sec S2 = 103 39·39 0·666 0·834 21·88
25
J (mA/cm2)

S1=103 cm/sec S2 = 104 39·35 0·665 0·833 21·83

20
S1=104 cm/sec S2 = 105 39·2 0·662 0·833 21·62
15
S1=105 cm/sec S2 = 106 39·1 0·659 0·831 21·47
10
S1=106 cm/sec
Table 3. The values of cell parameters in the modeled structure for
5
S1=107 cm/sec different values of S1; S2 = 103 cm/s in all cases.

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
40
Voltage (V)
(b) 35
25
S2=103 cm/s 30

S1=103 cm/s S1=103 cm/s

20 25
S2=103 cm/s
S1=104 cm/s
J (mA/cm2)

20
S2=104 cm/s
15 S1=105 cm/s
15
J (mW /cm2)

S2=105 cm/s
S1=106 cm/s
10
10 S2=106 cm/s
S1=107 cm/s
5

5 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Voltage (V)
0
Figure 6. J-V characteristics for various values of S2.
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Voltage (V)
recombination on the front surface. SiNx, along with SiO2, helps in
Figure 5. (a) Simulated J-V characteristics and (b) Power density
passivation as well as providing a suitable AR coating to increase
against voltage for various values of S1.
absorption.

3.4 Effect of back surface passivation

as Jsc, Voc, fill factor and efficiency decreases. Table 3 shows the in the modeled structure
calculated values of the cell parameters. The passivation techniques used on the back surface are back
surface field, oxidation, etc. In this structure, the authors select wet
The power density also reduces as S1 increases (as shown in oxidation followed by photolithography. Photolithography is used
Figure 5(b)). The maximum power density is shifted towards for selective opening so that the metallization can be performed to
higher voltages. At lower values of front surface recombination obtain a suitable back surface field. The parameter which measures
velocity, the value of power density is higher at all voltages. This the back surface passivation is the back surface recombination
is because, as S1 increases, the carriers start recombining at the velocity (S2). Figure 6 shows the variation in J-V characteristics
front surface which reduces the current and other associated cell with respect to S2. The variation in current density is very small
parameters. In this structure, the authors select oxidation and with respect to S2. The variation in spectral response with respect
metallization followed by photolithography on the front surface. to S2 has been shown elsewhere (Budhraja et al., 2011b). This also
The metal contact helps to collect carriers and SiO2 reduces the confirms that the difference in cell parameters with respect to S2 is

30
Emerging Materials Research Simulation of device parameters of
high efficiency multicrystalline silicon
solar cells
Budhraja, Misra and Ravindra

The modeling of the proposed structure of mc-Si solar cell gives Jsc
Front surface Jsc = 39·39 mA/cm2, Voc = 0·666 V, fill factor = 0·834 and efficiency =
recombination (mA/ Voc Efficiency 21·88%. Other parameters such as series resistance, shunt resistance,
velocity (cm/s) cm2) (V) Fill factor (%) saturation current densities, etc. were also calculated from the dark
J-V characteristics for various cells from the two-diode model.
S1 = 103 39·39 0·666 0·834 21·88
S1 = 104 39·35 0·662 0·833 21·71 It is important to have the proper front and back surface passivation,
especially when the solar cell industries are moving towards the use
S1 = 105 38·18 0·640 0·824 20·07
of thinner wafers to reduce cost. Parameters such as front and back
S1 = 106 37·22 0·605 0·822 18·53 surface recombination velocities were used to quantify the front and
back surface passivation respectively. In mc-Si solar cells, the front and
S1 = 107 35·56 0·591 0·819 17·22
back surface recombination velocities are very high, i.e. ≥ 106 cm/s, but
Table 4. The values of cell parameters in the proposed structure for it can be reduced to ≤ 103 cm/s by using passivation techniques such as
various values of S2, S1 = 103 cm/s in all cases. SiNx:H deposition, oxidation and back surface field. There was more
effect of front surface recombination velocity on the performance of
solar cells than back surface recombination velocity. The role of back
surface recombination velocity will be significant if the wafer gets
thinner. The change in spectral response with respect to S2 will be
very small. However, its significance at the front surface determines more pronounced (Budhraja et al., 2011c) when the thickness of the
the large variation in current density with respect to front surface cell is further reduced by the solar cell industry. This is because of the
recombination velocity. fact that, in thinner wafers, the generation rate will be significant at the
back surface and can no longer be neglected.
As the back surface recombination velocity increases, the value of cell
performance parameters decreases but the difference in magnitude The modeling results of the proposed structure of the mc-Si solar
is very small in comparison to those that correspond to the front cell are encouraging and should lead to improvement in design,
surface recombination velocity. The values of the cell parameters fabrication, performance characteristics and efficiencies of mc-Si
for various values of S2 are summarized in Table 4. There is a small cells.
change in efficiency from 21·88 to 21·47% as S2 changes from 103 to
106 cm/s. Similarly, there is a small change in other cell parameters
with respect to the change in S2. The effect of S2 in thick solar cells Acknowledgements
is very small because the generation rate becomes very small at the The authors thank Dr. Ana Kanevce and Dr. Bhushan Sopori of
back surface. The minority carriers which are generated close to the National Renewable Energy Laboratory for their valuable input. The
back surface are not able to reach the junction if the wafer is thick authors would also like to gratefully acknowledge the financial support
enough. The role of back surface passivation will be more significant rendered by the US Department of Energy for this contribution.
when the solar cell industries start to utilize thin wafers to reduce
cost (Ravi, 2011) because, in thin wafers, the diffusion length of
carriers is enough to reach the junction. The carriers will recombine REFERENCES
at the back surface instead of reaching the junction if the back surface Basore P (1994) Defining terms for crystalline silicon solar
recombination velocity is high. The open-circuit voltage will also cells, progress in photovoltaics: research and applications 2:
increase in thin wafers if the proper back surface field is available on 177–179.
the back surface. The back surface field has a net effect of passivating Budhraja V, Misra D and Ravindra NM (2011a) Advancements in
the back surface, especially when the wafer is thin. PV multicrystalline silicon solar cells from 1980 to 2010 — An
overview. 37th IEEE PVSC, Seattle, Washington, pp. 19–24.
Budhraja V, Sopori B, Ravindra NM and Misra D (2011b) An
4. Conclusions improved model of dislocations in a silicon solar cell. 21st
J-V characteristics of various mc-Si solar cells were modeled and Workshop on crystalline silicon solar cells & modules:
compared in PC1D software. It was confirmed that PC1D was able Materials and processes, Breckenridge, Colorado, pp.158-159.
to simulate the experimental results for various mc-Si solar cells that Budhraja V, Sopori B, Ravindra NM and Misra D (2011c)
have been reported by a number of research groups. The modeling Improved dislocation model of silicon solar cells with the effect
of the proposed structure of the mc-Si solar cell was performed in of front and back surface recombination velocity. (in press).
PC1D. Based on the understanding and the literature survey, the Clugston D and Basore P (1996) PC1D Version 5: 32-bit solar
authors chose processing steps which are able to yield high efficiency cell modeling on personal computers. 26th IEEE Photovoltaic
in mc-Si solar cells. Specialists Conference Anaheim CA, pp. 207–210.

31
Emerging Materials Research Simulation of device parameters of
high efficiency multicrystalline silicon
solar cells
Budhraja, Misra and Ravindra

Ghitani HEl and Martinuzzi S (1989) Influence of dislocations on metallurgical-grade silicon. IEEE Transaction on Electron
large grained polycrystalline silicon cells. I.Model. Journal of Devices 27: 671–677.
Applied Physics 66: 1717–1722. Sana P, Salami J and Rohatgi A (1993) Fabrication and
Green MA, Blakers AW, Jiqun S, Keller EM and Wenham SR analysis of high-efficiency polycrystalline silicon
(1984) 19.1% efficient silicon solar cell. Applied Physics solar cells. IEEE Transaction on Electron Devices 40:
Letters 44: 1163–1164. 1461–1468.
Halder NC and Williams TR (1983) Grain boundary effects in Schultz O, Gluntz SW and Willeke GP (2004a) Multicrystalline
polycrystalline silicon solar cells I: Solution of the three silicon solar cells exceeding 20% efficiency. Progress in
dimensional diffusion equation by the green’s function Photovoltaics Research and Applications 12: 553–558.
method. Solar Cells 8: 201–223. Schultz O, Riepe S and Glunz SW (2004b) Influence of high
Hussein R, Borchert D, Grabosch G and Fahrner WR (2001) Dark temperature processes on multicrystalline silicon. Solid State
I-V-T measurements and characteristics of (n) a-Si/(p) c-Si Phenomena 95/96: 235–240.
heterojunction solar cells. Solar Energy Materials and Solar Sopori B, Rupnowski P, Shet S et al. (2010) Influence of defects
Cells 69: 123–129. and defect distributions in multicrystalline silicon on solar cell
Johnson SM and Winter C (1984) High efficiency large area performance. 35th IEEE PVSC, Honolulu, Hawai, June 20–25,
polysilicon solar cells. IEEE 17th PVSC 1121–1126. pp. 2233–2237.
Macdonald D and Cuevas A (2000) The trade-off between Tobias I,Canizo CD and Alonso J (2004) Handbook of Photovoltaic
phosphorous gettering and thermal degradation in Science and Engineering, Chapter 7, pp. 254–307. Website 1:
multicrystalline silicon. 16th European Photovoltaic Solar http://rredc.nrel.gov/solar/spectra/am1.5/; Website 2:
Energy Conference. pp. 1707–1710. http://pveducation.org/pvcdrom/pn-junction/surface-
Magazine Article (2011) “C-Si Putting Pressure On Thin Films”, recombination.
January. Solar Industry: 4. Wenham SR, Willison MR, Narayanan S and Green MA (1985)
Narayanan S, Wenham SR and Green MA (1986) High efficiency Efficiency improvement in screen printed polycrystalline
polycrystalline silicon solar cells using phosphorous silicon solar cells by plasma treatment. 18th IEEE PVSC
pretreatment. Applied Physics Letters 48: 873–875. 1008–1013.
Ravi KV (2011) Poly-less, ingot-less, kerfless production of very Wu B, Stoddard N, Ma R and Clark R. Bulk multicrystalline
thin (<50 microns) single crystal single wafers, solar cells and silicon growth for photovoltaic (PV) application. Journal of
modules. 21st Workshop on crystalline silicon solar cells & Crystal Growth 310: 2178–2184.
modules: materials and processes, Breckenridge, Colorado, Zhao J, Wang A, Green MA and Ferrazza F (1998) 19.8%
July 31–August 3. efficient honeycomb textured multicrystalline and 24.4%
Rohatgi A, Narasimha S , Kamra S et al. (1996) Record high monocrystalline silicon solar cells. Applied Physics Letters
18.6% efficiency solar cell on HEM multicrystalline material. 73: 1991–1993.
25th IEEE PVSC 741–744. Zolper JC, Narayanan S, Wenham SR and Green MA (1989)
Saitoh T, Warabisako T, Kuroda E et al. (1980) Impurity 16.7% efficient, laser textured, buried contact polycrystalline
gettering of polycrystalline solar cells fabricated from refined silicon solar cell. Applied Physics Letters 55: 2363–2365.

WHAT DO YOU THINK?

To discuss this paper, please email up to 500 words to the
managing editor at [email protected]

Your contribution will be forwarded to the author(s) for

a reply and, if considered appropriate by the editor-in-
chief, will be published as a discussion in a future issue of
the journal.

ICE Science journals rely entirely on contributions sent in

by professionals, academics and students coming from the
field of materials science and engineering. Articles should
be within 5000-7000 words long (short communications and
opinion articles should be within 2000 words long), with
adequate illustrations and references. To access our author
guidelines and how to submit your paper, please refer to
the journal website at www.icevirtuallibrary.com/emr

32

View publication stats