ARTICLE IN PRESS

Microelectronics Journal 40 (2009) 120– 125

Contents lists available at ScienceDirect

Microelectronics Journal
journal homepage: www.elsevier.com/locate/mejo

A simple analytical model of thin films crystalline silicon solar cell

with quasi-monocrystalline porous silicon at the backside
Monem Krichen , Abdelaziz Zouari, Adel Ben Arab
Laboratoire de Physique Appliquée, Equipe de modélisation des Composants Semi-conducteurs, Faculté des Sciences de Sfax, Université de Sfax, B.P. 1171, 3000, Sfax, Tunisia

a r t i c l e in f o a b s t r a c t

Article history: An analytical model that simulates the performance of an elementary thin silicon solar cell with a thin
Received 10 April 2008 film quasi-monocrystalline porous silicon (QMPS) at the backside reflector is developed. A complete set
Accepted 14 June 2008 of equations for the photocurrent generated under the effect of the reflected light is solved analytically
Available online 12 August 2008
in each region. The collection of the light absorbed by the QMPS layer has been discussed and the
Keywords: analytical solution of the light-generated current in this layer is derived. The maximum of the
Solar cells photocurrent density calculated in the present study is in accordance with the numerical values
Quasi-monocrystalline porous silicon established by Bergmann et al. Furthermore, the influence that the layer’s number of double porosities
Photocurrent density and high porosity have on the photovoltaic parameters is studied. It is demonstrated that the
Cell efficiency
photovoltaic parameters increase with the number of double porosities that the layer might have in a
given structure. When the QMPS layer is formed by three double-porosity layers 20%/80% and for a 5-
mm-thick film c-Si, the backside reflector gives a total improvement of about 6 mA/cm2 for the
photocurrent density and 3.2% for the cell efficiency.
& 2008 Elsevier Ltd. All rights reserved.

1. Introduction successfully introduce PS in silicon solar cells as an antireflective

and passivating coating [6,7].
The use of thin crystalline silicon solar cells, manufactured More recently, QMPS layers have been reported to have a
through various versions of large transfer processes, is one of the significantly high absorption coefficient compared to crystalline
most promising approaches to achieve both performance silicon within the whole range of the solar spectrum of
improvement and cost reduction. This is partially due to the photovoltaic solar cell applications [8,9]. A semi-empirical model
inexpensiveness of the material and the easiness of the manu- has been developed to account for the high absorption coefficient
facturing process [1–4]. Being a perfect light diffuser, the porous of QMPS layer which predicts the absorption coefficient of the
silicon (PS) or quasi-monocrystalline PS (QMPS) layer could be QMPS layer in terms of different thickness, porosity and void size
used as a backside reflector that enhances the effective thickness [8]. The transport parameters such as those of the minority carrier
of the film c-Si through the light being reflected [2–4]. mobility and lifetime in QMPS layer have also been reported [10].
The reflectance of a multi-layered structure increases when the Several simulation models for thin film solar cells with QMPS
number of the layers in such a structure is increased. The gain in layer or PS on the rear side have been developed to account for the
reflectance is intricately related to the absorption losses of QMPS improvement of photocurrent density and cell efficiency [2,3,11].
which are increased by the total thickness of the multi-layered Most of these models, however, seem to focus solely on light
structure. It has been recently reported that for thin epitaxial reflected but paid little attention to the contribution of the QMPS
silicon films (i.e. below 5 mm), the advantage of having a large to photovoltaic current generation coming from the light
number of repetitions reduces. This is because in the multi- absorbed. An analytical study is deemed necessary for determin-
layered PS there is more light absorption than in single-layered ing the potential uses of this material, particularly the number of
ones [5]. It is worth noting that most of the current experimental layers it contains, as a backside layer in thin silicon solar cells. It is
studies have employed a QMPS layer formed by four or six layers. the purpose of this paper to present a simple analytical solution
Few experimental studies, however, have been reported to for the contribution that this material (QMPS) has in the
improvement of photocurrent generation and the enhancement
of the cell photovoltaic parameters, when acting as a backside
reflector.
 Corresponding author. Tel.: +216 97 678 174. In the present study, we consider an elementary n+pp+ solar
E-mail address: [email protected] (M. Krichen). cell consisting of thin film crystalline silicon (regions n+ and p)

0026-2692/$ - see front matter & 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mejo.2008.06.041
 ARTICLE IN PRESS

M. Krichen et al. / Microelectronics Journal 40 (2009) 120–125 121

Nomenclature f(l) Incident photon flux

R(l) reflection coefficient at the front surface
q electron charge Wn+ (Wp) emitter (base) thickness
VT thermal voltage Wc (Wp+) film c-Si (QMPS) thickness
Ln (Lp) diffusion length of minority carriers in the Sp ðSþ
nÞ recombination velocity at the front (back) contact
base (emitter) region Jph,0 total photocurrent density of the conventional
Lnn effective diffusion length of minority carriers BSF silicon solar cell
in the QMPS layer JRph;1 ðJ Rph;2 Þ reflected photocurrent density by the back
Dn (Dp) diffusion constant of minority carriers in the (front) contact
base (emitter) region JQMPS
ph
QMPS photocurrent density
Dnn
effective diffusion constant of minority carriers DJph total increase in photocurrent density
in the QMPS layer DZ total increase in cell efficiency
tn minority carrier lifetime in the base region J0E (J0B) reverse saturation current density in the
tnn effective minority carrier lifetime in the QMPS layer emitter (base) region
Na (Nd) dopant concentrations in the base (emitter) region Pin power density output (AM 1.5)
Nþ
a dopant concentrations in the back region Pm maximum power output of solar cell
aðlÞðan Þ absorption coefficient in the c-Si (QMPS) Vm (Jm) voltage (current density) at maximum power
at a wavelength l output

with a thin film QMPS back region (p+ type). The increase in the where Jph,0(l) is the total photocurrent density collected from
photocurrent generated under the effect of the reflected light by different regions in a conventional BSF silicon solar cell [12]. As for
the rear surface in the base region J Rph;B and the emitter region JRph;E the other components they stand for the following: J Rph;1 ðlÞ is the
are solved analytically, including the QMPS photocurrent density photocurrent density contributed to by the photo-generated
J QMPS
ph
contributed to by the light absorbed when the reflected light carriers within the base and the emitter regions due to reflected
is not fully total. In addition, in our survey we have taken into light by the rear surface, J Rph;2 ðlÞ is the photocurrent density
account the photocurrent generated effect on the second reflec- contributed to by the presence of a second optically rough front
tion that occurs at the level of the front surface. The QMPS layer is surface, and J QMPS
ph
ðlÞ is the photocurrent density contributed by
modeled by an effective recombination velocity at the back the absorbed light in the QMPS layer. The solution of the current
surface in the base reverse saturation current density. First, we continuity equation for JRph;2 ðlÞ is similar to the n+ and p regions
compared our theoretical results pertaining to the maximum except for the generation rate of minority carriers that decreases
photocurrent generated under the effect of the optimum reflected the level Rd exp(2aWc) where Rd is the reflectance at the back
light to the numerical results established by Bergmann et al. in region.
order to enhance the reliability of our theory. The effect of the
reflectance of the multi-layered structure (QMPS layer) on the 2.1. Photocurrent density contributed by the reflected light
increase of the photocurrent density DJph, as well as the cell
efficiency DZ is also studied.
The increase in photocurrent density that results from the light
reflected by the rear surface is expressed by
2. Theoretical model
JRph;1 ðlÞ ¼ JRph;B ðlÞ þ J Rph;E ðlÞ (2)
+ +
We consider an n pp solar cell structure with a QMPS at the
back (p+ type), as illustrated in Fig. 1. The total photocurrent where JRph;B ðlÞ and J Rph;E ðlÞ are the photocurrent densities collected
density Jph(l) at a wavelength of an elementary cell under from the base and the emitter regions, respectively.
illumination can be written as follows: Under illumination, the diffusion equation in the base region
affects the reflected light, and may be written as follows:
J ph ðlÞ ¼ Jph;0 ðlÞ þ J Rph;1 ðlÞ þ J Rph;2 þ J QMPS
ph
ðlÞ (1)
2
d ðDnÞ Dn
Dn  ¼ g ref ðx; lÞ (3)
dx2 tn
A B C D E
where Dn represents the concentration of the excess minority
carriers (electrons) and Dn and tn are the minority electron
diffusion constant and lifetime in the p-region, respectively. The
generation rate of minority carriers (electron) for the reflected
hν
n+ p+ light is given by
glass

g ref ðx; lÞ ¼ ð1  RÞRd fa expðað2W c  xÞÞ (4)

The boundary conditions for the base photocurrent case are as

follows:
SiNx
Dnðx ¼ W 1 Þ ¼ 0 ðat x ¼ CÞ (5)
x
0 We W1 Wc H

dDn Se;ppþ
Fig. 1. One-dimensional schematic model of an elementary solar cell with a QMPS ¼ Dnðx ¼ W c Þ ðat x ¼ DÞ (6)
layer on the rear side.
dx x¼W c Dn
 ARTICLE IN PRESS

122 M. Krichen et al. / Microelectronics Journal 40 (2009) 120–125

The photocurrent density that results from the light reflected in

the base region (p) is then expressed by

Ln af
J Rph;B ðlÞ ¼ qRd ðR  1Þ expðaðW c þ W p ÞÞ
ða2 L2n  1Þ
 
ðSe;ppþ Ln =Dn Þ½ coshðW p =Ln Þ  expðaW p Þ þ sinhðW p =Ln Þ  aLn expðaW p Þ
 aLn þ (7)
ðSe;ppþ Ln =Dn Þ sinhðW p =Ln Þ þ coshðW p =Ln Þ

where Se;ppþ is the effective surface recombination velocity at

x ¼ D [13].
Similarly, the photocurrent generated by the light reflected in
the emitter region is given by

Lp af
J Rph;E ¼ qRd ðR  1Þ expðað2W c  W nþ ÞÞ
ða2 L2  1Þ
p
 
ðSp Lp =Dp Þ coshðW nþ =Lp Þ þ sinhðW nþ =Lp Þ  ðaLp þ ðSp Lp =Dp ÞÞ expðaW nþ Þ
aLp þ (8)
ðSp Lp =Dp Þ sinhðW nþ =Lp Þ þ coshðW nþ =Lp Þ

where Sp is the recombination velocity at the front contact (at 2.3. Photovoltaic parameters
x ¼ A).
The increase in photocurrent density in a thin silicon solar cell
2.2. Photocurrent density contributed in QMPS layer with a thin film QMPS at the backside is given by

DJph ¼ JRph;1 þ J Rph;2 þ J QMPS

ph
(15)
In cases when the reflection by the backside reflector is not
total, we have also determined the contribution that the The increase in conversion efficiency of an elementary cell is
photocurrent density has for the QMPS layer due to the absorbed easily computed by
light. In these conditions, the diffusion equation in the back region
may be written as follows: DZ ¼ ZR1 þ ZR2 þ ZQMPS (16)
2
d Dnþ Dnþ where ZR1 and ZR2 represent the improvement of the cell efficiency
Dnn  n ¼ g n ðx; lÞ (9)
dx2 tn due to light reflected by the back and front surfaces, respectively.
Whereas, ZQMPS is the improvement carried by the QMPS layer.
where Dn+ represents the concentration of excess minority carrier
Every term of the conversion efficiency is determined from this
electrons in this region and Dnn and tnn are the minority electron
equation,
diffusion constant and lifetime, respectively [10]. The generation
rate of minority carriers in the QMPS layer is expressed by
V im J im
n n n Zi ¼ (17)
g ðx; lÞ ¼ ð1  RÞð1  Rd Þa f expðaW c Þ expða ðx  W c ÞÞ (10) Pin
where an is the effective absorption coefficient of QMPS layer However, the maximum output power Pm of the cell can be found
which depends on the void radii and the porosity [8]. only by maximizing the power P ¼ VJ(V) with respect to V. This
The boundary conditions for the QMPS photocurrent case are brings about the well-known transcendental equation [15]
as follows:
! !
Dnþ ðx ¼ W c Þ ¼ 0 ðat x ¼ DÞ (11) V im V im Jiph
1þ exp ¼ þ1 (18)
VT VT J0

dDnþ  Sþ
 ¼  n Dnþ ðx ¼ HÞ ðat x ¼ EÞ (12) with
dx x¼H Dn
where Sþ
n and H are the recombination velocity at the back surface J0 ¼ J 0E þ J 0B (19)
and the cell thickness, respectively.
The photocurrent density (at x ¼ D) due to the absorption light where J0E and J0B are the conventional reverse dark saturation
of wavelength l in the QMPS layer is then expressed by current density in the emitter and base regions, respectively, [13]

Lnn an f
J QMPS
ph
ðat DÞ ¼ qð1  RÞð1  Rd Þ 2 2
expðaW c Þ
ðan Lnn  1Þ
" #
ððSþ n n n n n n
n Ln =Dn Þð coshðW pþ =Ln Þ  expða W pþ ÞÞ þ sinhðW pþ =Ln Þ þ a Ln expða W pþ ÞÞ
n n
an Lnn  (13)
ðSþ n n n n
n Ln =Dn Þ sinhðW pþ =Ln Þ þ coshðW pþ =Ln Þ

Thereafter, the photocurrent that is generated in this region and

(at x ¼ C) is given by [14]
   ( ! )
Wp V im
J QMPS ðat CÞ ¼ J QMPS ðat DÞ sec h (14) Jim ¼ J 0 exp 1 þ J iph (20)
ph ph Ln VT
 ARTICLE IN PRESS

M. Krichen et al. / Microelectronics Journal 40 (2009) 120–125 123

In the current study, we have taken the different heavily doped low curve (curve 1) is for a conventional BSF solar cell. The
regions into account. The effective carrier concentration nie is maximum conversion efficiency curve reflects more or less the
related to both band-gap narrowing (BGN) and degeneracy by [13] same nature as Jphmax. The cell efficiency is improved from 12.07%
to 15.5% for Wc ¼ 5 mm, whereas it increases from 16.66% to
n2ie ¼ n2i0 expðDEg =V T Þ (21) 18.96% for Wc ¼ 25 mm.

where ni0 is the intrinsic silicon carrier concentration and DEg is

3.2. Effect of the high porosity on the increase in cell parameters
the apparent BGN [16,17].
In the previous graphs that indicate the dependence of cell
photovoltaic parameters, we suppose that the reflected light is
3. Results and discussion
optimum, and that the value of the reflection coefficient at the
rear surface is equal to unity (Rd ¼ 1). In this section, we discuss
The results of the photovoltaic parameters presented in
the effect of high porosity and multi-layered structure in the back
this paper are computed using the numerical values indicated in
region (QMPS layer) on the increase of the photocurrent density
Table 1.
DJph and cell efficiency DZ. The different values of reflectance Rd
used in this study are reported in [5].
3.1. Effect of the QMPS back region (p+) on the cell photovoltaic Fig. 3(a and b) shows the effect of high porosity on the DJph and
parameters DZ for two single double-porosity layers 20%/60% (Rd ¼ 18%) and
20%/80% (Rd ¼ 40%) as functions of the film thickness (Wc). The
In this section, we discuss the results of the calculation of the variations in the increase in calculated photocurrent density and
maximum photocurrent density Jphmax and the maximum cell cell efficiency, due to the reflected light, are shown in Fig. 3
efficiency Zmax of two types of BSF thin crystalline silicon solar (curves (1-a) and (1-b)). It can be seen from Fig. 3(a), curves (1-a)
cells for different values of film thickness (Wc). and (1-b), that the magnitude of the improvement in photocurrent
The first BSF solar cell is called the conventional BSF solar cell, density for small film thicknesses is greater than that for larger
where the backside region is a crystalline silicon layer (c-Si). In film thicknesses when the high porosity is about 80%. For
the second BSF solar cell the backside region is formed by a thin
film QMPS.
Fig. 2(a) shows the variation of Jphmax versus the film thickness
(Wc) for the two types of BSF solar cells. The conservative value of
the transport parameters and the surface recombination velocities 36
Maximum photocurrent density

are also considered in these calculations. The numerical results

[11] related to these models of BSF solar cell are also indicated in
the same figure.
Jphmax (mA/cm2)

For the purpose of comparison, the thus calculated Jphmax 32

values are plotted in Fig. 2(a) for different cases. The results 3
indicate that, in all cases, the tendencies of Jphmax dependence on
Wc are similar to those observed by Bergmann et al. [11]. Our 2 1: both interfaces flat
calculated values are in good agreement with his numerical 28 2: rough rear
3: both interfaces rough
results [11] over a wide range of cell thickness, particularly when
: Simulation B. [11]
the value of the simulated coefficient reflection Rd is equal to
unity (i.e. JQMPS
ph
¼ 0). This agreement reinforces the validity of our 1
24
approach.
In addition, the same figure shows that the computed value of 0 10 20 30 40 50
the maximum photocurrent density of the BSF solar cells with
Film thickness Wc (μm)
QMPS layer at the backside is significantly higher than those of
the conventional BSF solar cell (p+ is c-Si) for the entire cell
thicknesses. This enhancement in Jphmax is due to the first light 20
reflected by the QMPS layer (curve 2) and the addition of the
second light reflected by the front surface (curve 3). For Wc ¼ 5
Maximum cell efficiency max (%)

mm, the increase in photocurrent density is 27%, whereas it is 18

13.2% for Wc ¼ 25 mm.
Fig. 2(b) illustrates the variation of maximum conversion
efficiency Zmax versus the film thickness for different cases. The
16
Table 1
3
Physical parameter values 1: both interfaces flat
2
Parameter Value Reference 14 2: rough rear
3: both interfaces rough
We 0.3 mm
Nd 2  1019 cm3
Na 2  1017 cm3
1
12
Nþ 1019 cm3 [11]
a 0 10 20 30 40 50
Sp 103 cm/s [11]
Sþ 104 cm/s [11] Film thickness Wc (μm)
n
Pin 100 mW/cm2 [2]
R 5% Fig. 2. (a) Variation of the maximum photocurrent density versus film thickness Wc
and (b) variation of the maximum cell efficiency versus film thickness Wc.
 ARTICLE IN PRESS

124 M. Krichen et al. / Microelectronics Journal 40 (2009) 120–125

4 7
2-b

1 double-porosity layers 2-b 3 double-porosity layers

6
a: 20/60 % (Rd = 18 %) a: 20/60 % (Rd = 51 %)
1-b 1-b
b: 20/80 % (Rd = 40 %) b: 20/80 % (Rd = 80 %)
Δ Jph (mA/cm2)

Jph (mA/cm2)
5
2-a 2-a

4
2 1-a
1-a

1 2
0 10 20 30 40 50 0 10 20 30 40 50
Film thickness Wc (µm) Film thickness Wc (μm)

2.0
3.5
2-b 2-b
1 double-porosity layers
3 double-porosity layers
a: 20/60 % (Rd = 18 %) 3.0
1-b a: 20/60 % (Rd = 51 %)
1-b b: 20/80 % (Rd = 40 %)
1.6 b: 20/80 % (Rd = 80 %)
 (%)

2.5 2-a
η (%)

2-a

2.0 1-a
1.2

1-a 1.5

0.8
0 10 20 30 40 50 1.0
0 10 20 30 40 50
Film thickness Wc (μm)
Film thickness Wc (μm)
Fig. 3. (a) The effect of the high porosity on the DJph for different film thicknesses
Wc and (b) the effect of the high porosity on the DZ for different film thick- Fig. 4. (a) The effect of double-porosity layers on the DJph for different film
nesses Wc. thicknesses Wc and (b) the effect of double-porosity layers on the DZ for different
film thicknesses Wc.

Wc ¼ 5 mm the improvement in Jph by the light reflected is increases by increasing the number of layers in such a structure.
1.8 mA/cm2 for high porosity of 60%, and 3.06 mA/cm2 for porosity However, the improvement in QMPS photocurrent density for
of 80%. This enhancement is explained by the refractive index of different film thicknesses and for three double-porosity layers is
QMPS layer that decreases with the porosity [7]. Consequently, greater than that for single double-porosity layers. This enhance-
the value of reflectance Rd is increased. ment is explained by the fact that the increase in Rd is coupled
In addition to the QMPS photocurrent density (J QMPS
ph
), due to with an increase in the width of the backside reflector.
the light absorbed by this material, the total DJph (or DZ) is It can also be noticed from Fig. 4(a) that the photocurrent
represented by curves (2-a) and (2-b). It is clear from these two density contributed to by the reflected light (Rd ¼ 80%) (curve
figures that the light absorbed in the QMPS layer improves the cell (1-b)) is greater than the total photocurrent density contributed to
parameters significantly only if the cell thickness is small. This the value of reflectance, equal to 51% (curve (2-a)).
result is attributed to the fact that the carriers-generated rate in
this material decreases with Wc (Eq. (10)).
3.4. Effect of the recombination velocity Sþ
n on the increase in cell
parameters
3.3. Effect of the multi-layer in QMPS on the increase in
cell parameters Fig. 5(a and b) shows the effect of the recombination velocity at
the back surface on the increase of the cell parameters (DJph and
Fig. 4(a and b) shows the effect of a threefold repetition of DZ) of the elementary cell with a QMPS layer at the backside. Two
double-porosity layers (i.e. QMPS is formed by six layers) on the film thicknesses are considered Wc ¼ 5 and 20 mm, and the
increase in photocurrent density DJph and cell efficiency DZ versus QMPS film is formed by three double-porosity layers 20%/80%
film thickness Wc. A careful comparison of these figures with (Rd ¼ 80% [5]).
Fig. 3(a and b) shows that the enhancement of cell performance Fig. 5(a) presents the variation of the increase in photocurrent
occurs when the reflectance of the multi-layered structure density DJph calculated as a function of the recombination velocity
 ARTICLE IN PRESS

M. Krichen et al. / Microelectronics Journal 40 (2009) 120–125 125

12.3% with regard to the total improvement DZ (Eq. (16)) for

7 Wc ¼ 5 mm, whereas it is 8.6% for Wc ¼ 20 mm.

2-a
4. Conclusion
6
1-a
Jph (mA/cm2)

In this paper, an analytical model for the simulation of the

performance of QMPS backside reflector in an elementary thin
5 silicon solar cell has been developed. A complete set of equations
a: Wc = 5 μm
for the photocurrent generated under the effect of the reflected
2-b b: Wc = 20 μm
Sp = 103 cm/s light is solved analytically in the base and the emitter regions. It
1-b has been demonstrated that the QMPS layer has been contributed
4
to by the photovoltaic generation effect from the absorbed light.
This layer is modeled by an effective recombination velocity at the
back surface in the base reverse saturation current density. It has
3 been shown that the cell parameters (Jph and Z), compared to the
10 102 103 104 105 106 conventional BSF (p+ is c-Si layer) thin silicon solar cell, have been
+
Velocity Sn (cm/s) improved due to the high porosity, double-porosity layers, and
recombination velocity value at the back contact. In addition, we
studied the maximum photocurrent density effect resulting from
3.5
the optimum light reflector in a thin film silicon solar cell. The
2-a
results calculated in this study are in accordance with the
numerical values established by Bergmann et al.
3.0 1-a
Acknowledgment
 (%)

2.5 2-b Thanks are due to Mr. Smaoui Anwar, from the English
department at the Sfax Faculty of Science for carefully proof-
1-b reading and constructively revising this paper.

2.0 a: Wc = 5 μm
References
b: Wc = 20 μm
Sp = 103 cm/s
[1] K. Van Nieuwenhuysen, F. Duerinckx, I. Kuzma, D. van Gestel, G. Beaucarne,
J. Poortmans, J. Cryst. Growth 287 (2006) 438–441.
1.5 [2] R. Brendel, Sol. Energy 77 (2004) 969–982.
10 102 103 104 105 106 [3] R.B. Bergmann, C. Berge, T.J. Rinke, J. Schmidt, J.H. Werner, Sol. Energy Mater.
+
Velocity Sn (cm/s) Sol. Cells 74 (2002) 213–218.
[4] T.J. Rinke, R.B. Bergmann, J.H. Werner, J. Appl. Phys. A68 (1999) 705–707.
[5] R. Bilyalov, C.S. Solanki, J. Poortmans, O. Richard, H. Bender, M. Kummer,
Fig. 5. (a) Variation of the DJph versus recombination velocity at the back surface H. von Känel, Thin Solid Films 403–404 (2002) 170–174.
3
Sþ
n (QMPS film is formed by 3 double-porosity layers 20%/80% and Sp ¼ 10 cm/s) [6] R.J. Martin-Palma, L. Vazquez, P. Herrero, J.M. Martinez-Duart, M. Schnell,
and (b) variation of the DZ versus recombination velocity at the back surface Sþ n S. Schaefer, Opt. Mater. 17 (2001) 75–78.
(QMPS film is formed by three double-porosity layers 20%/80% and Sp ¼ 103 cm/s). [7] S. Strehlke, S. Bastide, J. Guillet, C. Lévy-Clément, Mater. Sci. Eng. 69–70B
(2000) 81–86.
[8] D. Majumdar, S. Chatterjee, M. Dhar, S.K. Dutta, H. Saha, Sol. Energy Mater. Sol.
Cells 77 (2003) 51–64.
[9] H. Seel, Rolf Brendel, Thin Solid Films 451–452 (2004) 608–611.
Sþ
n for devices with and without the contribution of light- [10] M. Banerjee, S.K. Dutta, U. Gangopadhyay, D. Majumdar, H. Saha, Solid-State
generated electron current in the QMPS layer. It can be seen that Electron. 49 (2005) 1282–1291.
[11] R.B. Bergmann, T.J. Rinke, R.M. Hausner, M. Grauvogl, M. Vetter, J.H. Werner,
the improvement in DJph decreases when the value of recombina- Int. J. Photoenergy 1 (1999) 83–87.
tion velocity at the back surface is increased. The decrease in DJph [12] S.M. Sze, Physics of Semiconductor Devices, Bell Laboratories Inc., Murray
is more significant in the case of Wc ¼ 5 mm, since the QMPS Hill, New Jersey, 1981.
[13] A. Rohatgi, P.R. Choudhury, IEEE Trans. Electron. Devices ED-31 (5) (1984)
photocurrent density is more important for small c-Si thickness.
596–601.
Fig. 5(b) illustrates the variation of the increase in cell [14] J.P. McKelvey, Solid State and Semiconductor Physics, Harper and Row, New
efficiency (DZ). We notice that these variations are similar to York, 1966.
those of the improvement in the photocurrent density (Fig. 5(a)). [15] Alexis De Vos, Sol. Cells 8 (1983) 283–296.
4 [16] J. Del Alamo, S. Swirhum, R.M. Swanson, Solid-State Electron. 28 (1985)
When the value of Sþ n is o10 cm/s, the improvement in cell 47–54.
efficiency due to the absorbed light in the QMPS layer (ZQMPS) is [17] J.W. Slotboom, H.C. Graaff, Solid-State Electron. 19 (1976) 857–863.