J O U R N A L O F M AT E R I A L S S C I E N C E : M AT E R I A L S I N E L E C T RO N I C S 1 5 ( 2 0 0 4 ) 3 3 5 ± 3 4 0

Photoelectrochemical solar (PES) cells with

Cd1 xPbxSe polycrystalline thin ®lm electrodes
R. N. MULIK
D.B.F. Dayanand College of Arts and Science, Solapur-413002, M.S., India
L. P. DESHMUKH, V. B. PATIL
Department of Applied Physics, Shivaji University Centre for P. G. Studies, Solapur-413255,
M.S., India

An electrochemical solar cell behavior of a mixed type n-Cd1 xPbxSe 0  x  0.5† thin ®lm
electrode is presented in this paper. A series of the electrodes were prepared for the values of
x in the above range and the electrochemical cells of the con®guration n-Cd1 xPbxSe/
K4Fe(CN)6/C were fabricated. The current±voltage and capacitance±voltage characteristics of
the as-fabricated cells in dark were examined to understand nature of the charge transfer
process across the electrode/electrolyte interface. These cells were then illuminated with a
light of 20 mW cm 2 intensity and the power output curves were obtained and analyzed.
Both open circuit voltage Voc† and short circuit current Isc† found boosted enhancing the
power conversion ef®ciency (Z%) from 0.065% to 0.28% at x ˆ 0.1. The ¯at band potentials
Vfb† were also determined for these cells and it is found to attain a maximum value at x ˆ 0.1.
The other cell parameters were estimated and showed improvement in the overall
performance of a cell. Attempts were made to explain the observed results through modi®ed
electrode properties.
# 2004 Kluwer Academic Publishers

1. Introduction polished stainless steel substrates using a chemical

A semiconductor/electrolyte interface employed for deposition process [7, 11, 12]. Cadmium sulphate, lead
conversion of the optical energy into an equivalent acetate and sodium selenosulphite were used as the
electrical energy has been the subject of intensive source materials. For deposition, appropriate quantities
research [1±3]. A large number of semiconductor of cadmium sulphate and lead acetate were mixed
materials, both binary and mixed/alloyed types, have together and complexed with a suf®cient quantity of
been tried [1±8]. In this respect, chalcogenides of triethanolamine. Aqueous ammonia was used to
cadmium, lead and zinc have proved their potential as increase ®lm adherence to the substrate support. The
ef®cient absorbers of electromagnetic radiation [9±11]. sodium selenosulphite was used as a source of Se2
Among these, CdSe and PbSe are highly sensitive to ions and was obtained by the reaction between the
incident light and in view of their practical applications a selenium powder and sodium sulphite solution [12, 13].
study of photoelectrochemical properties of their mixed/ It was added slowly into the reaction mixture at a
alloyed thin ®lm structures is of scienti®c and techno- constant rate 0:42 ml min 1 †. Mechanically cleaned
logical importance. Based on the fact that various and well-polished stainless steel substrates were
materials parameters, such as the optical gap, optical positioned vertically on a specially designed substrate
absorbance, lattice constants, electrical conductivity, holder and rotated with a 70 + 2 rpm speed in the
thermoelectric power, etc. can be best controlled by reaction bath. This provides uniform mechanical
controlling the materials composition, the mixed/alloyed churning of the reaction mixture. The reaction was
type composites of CdSe and PbSe appear to be very allowed to proceed at 55 + 2  C for 90 min. The ®lms
promising in photoelectrochemical conversion. It is, were then detached from the substrate holder and
therefore, thought to be proper to explore the possibility washed with doubly distilled water and preserved. The
of employing a (Cd,Pb)Se material in these electro- thickness of the electrode layers were measured by an
chemical conversion cells. interference fringes technique. The layer thickness
increased typically, from 0.32 to 0.5 mm as x was
varied from 0 to 0.5. The structure and crystallinity of
2. Experimental details all the samples were examined by an X-ray diffraction
2.1. Preparation of the electrodes technique. A Philips-PW3710, X-ray diffractometer
Cd1 x Pbx Se 0  x  0.5† thin ®lm electrodes of with CuKa line (1.5406 A Ê ) was used for this purpose.
varying composition were synthesized on the well- The range of 2y angles was from 10 to 80 .
0957±4522 # 2004 Kluwer Academic Publishers 335
2.2. Fabrication of the photoelectrochemical our case are of rectifying Schottky barrier type [5, 7, 14].
solar (PES) cells The ideality factor nd † of the junction which is a
The electrochemical photovoltaic cells were then measure of quality of the junction was therefore
fabricated out of these electrodes and an electrolyte determined for all the cells from variation of the log I
consisting of K4 Fe(CN)6 redox couple in an H-shaped versus V. These are listed in Table I. It is seen from the
curette ( pH ˆ 11.2 + 0.2). A CoS-treated graphite rod values of nd that all PES cells exhibit a recombination
was employed as a counter electrode and placed behind mechanism at the interfaces. Obviously, the smaller the
the active photoelectrode at a distance of about 2±3 mm. value of nd , the lesser is the recombination and
The arrangement was such that the counter electrode was consequently, lower is the trap density associated with
toward the bulk of the electrolyte whereas photoelectrode the electrode material. The space charge layer capaci-
was adjusted vertically in the near vicinity pf the glass tance versus voltage (versus SCE) measurements were
wall. The cell was then illuminated from the photoelec- made on these cells to determine the ¯at-band potential,
trode side (1 cm2 area) with a white light source of 250 W which determines the relative positions of the Fermi level
tungsten lamp (without ®fter). The cell characteristics of a semiconductor and that of the electrolyte and
such as dynamic I±V and C±V curves in dark, barrier consequently amount of band bending at the interface
height, photovoltaic output characteristics under constant [1, 2, 4, 5, 7]. The intrinsic band bending at the interface
illummation 20 mW cm 2 † and photoresponse mea- is therefore a measure of the depletion layer width inside
surements were carried out with a special reference to the the semiconductor and determines ability of the electrode
photoelectrode composition. The cell performance was material to operate under the short circuit condition. This
examined through the cell parameters. is equivalent to a measure of the maximum open circuit
voltage that can be obtainable from a cell. For a
semiconductor/electrolyte interface, the space charge
3. Results and discussion capacitance is related to the applied voltage as follows:
When surface of a semiconductor electrode dipped into a
2
redox electrolyte is illuminated by a light of photon C ˆ 2=qee0 ND V Vfb kT=q† 1†
energy greater than the optical gap of a semiconductor,
excess charge carriers generated at the cost of light where Vfb is the ¯at-band potential and other terms
absorption are separated at the electrode/electrolyte signify their usual meaning. The Mott±Schottky plots
interface; the charge carriers generated in the bulk of were then constructed for all the cells to evaluate the
electrode material being annihilated. For an n-type actual values of the ¯at-band potentials as shown in
semiconductor, the electrons move deep into the bulk Fig. 1. The extrapolation of linear regions of C 2 V
of the material whereas holes arrive at the surface of the plots gives the values of ¯at-band potential and are cited
semiconductor allowing redox reactions to occur. The in Table I, which shows that Vfb values lie in the
charge separation process results in a counter ®eld at the 535 mV to 975 mV (versus SCE) range and when
interface, which is maximum at the open circuit referred to the electrode composition x†, Vfb maximize
condition, called the ``open circuit photopotential''. at x ˆ 0.1 with a value approximately 0.28 V more
This photopotential draws the photogenerated electrons negative than pure CdSe. For other cells Vfb lies in the
from the semiconductor to the counter electrode while above two values. The built-in-potentials fB † for
the holes retain at the semiconductor surface where they different cell structures were also determined from
are captured by the electrolyte species. A net current is temperature dependence of the reverse saturation current
therefore said to ¯ow through the interface, which is I0 [5, 7, 8]. fB 's were calculated from the linear portion
in¯uenced by the speed of the redox reaction, the nature of log I0 =T 2 † versus 1/T plots and are listed in Table I.
of the electrode/electrolyte interface and properties of the To test the photosensitivity of the electrode material, a
electrode material. The current±voltage and capacitance± white light source of 20 mW cm 2 intensity was used for
voltage (versus SCE) characteristics were therefore illumination and the short circuit current Isc † and open
studied for each of the cell con®gurations. It appeared circuit voltage Voc † were noted for 1 cm2 active area of
that, for an electrode/electrolyte system, nature of the the photoclectrode. A semiconductor/electrolyte inter-
charge transfer is best controlled by a well-known Butler± face can be modeled as a Schottky barrier junction and it
Volmer relation [5, 7, 14]. The calculated values of the is therefore possible to represent the current±voltage
symmetry factors showed that the junctions formed in characteristic by a relation

T A B L E I Various performance parameters of a cell formed with Cd1 x Pbx Se photoelectrode 0  x  0.5†

Compositions x Voc (mV) Isc mA=cm2 † Rs (O) Rsh (KO) nL nd Vfb (mV) (versus SCE) fB

0.0 300 110 600 6.22 2.04 5.60 610 0.66

0.025 250 100 800 5.00 2.02 5.60 955 0.71
0.05 430 130 666.6 7.20 2.18 3.64 920 0.71
0.075 480 275 333.3 6.44 2.07 3.64 960 0.80
0.1 510 300 285.7 7.30 1.89 2.92 975 0.86
0.15 500 260 320 5.00 1.74 4.04 900 0.48
0.2 495 150 500 4.33 1.69 5.16 855 0.48
0.35 200 90 800 4.00 1.46 5.60 650 0.48
0.5 180 85 800 4.00 1.61 5.25 535 0.46

336
 Figure 2 The power output curves for different cells with x ˆ 0 (d),
x ˆ 0.025 (s), x ˆ 0.05 (n), x ˆ 0.075 (m), x ˆ 0.1 (s), x ˆ 0.15 (d),
x ˆ 0.2 (w), x ˆ 0.35 (6) and x ˆ 0.5 (s).

excitation, the maximum ef®ciency or the open circuit

Figure 1 Mott±Schottky plots for determination of the ¯at band
potential x ˆ 0 (n), x ˆ 0.025 (&), x ˆ 0.05 (s), x ˆ 0.075 (&), x ˆ 0.1 voltage ef®ciency is expressed as [15]:
(s), x ˆ 0.15 (d) and x ˆ 0.5 (6).
Zmax ˆ Vredox Vfb †e=Eg 5†
I ˆ Iph Id ˆ Iph I0 ‰exp qV=nd kT† 1Š 2†
Thus, both maximum. output voltage and ef®ciency
where, I is the net current density, Iph and Id are the photo depend on Vfb and Eg . In our case, the cell delivered a
and dark current densities, respectively, I0 is the reverse maximum power output at x ˆ 0.1 and can be ascribed to
saturation current density, V is the applied bias voltage the higher value of the ¯at-band potential of a cell. For a
and nd is the junction ideality factor. For bias voltages cell with an n-type semiconductor photoelectrode, the
exceeding 3 kT/q and in the equilibrium open circuit larger the value of the ¯at-band potential, the more
conditions Iph ˆ Id , V ˆ Voc †, Equation 2 can be negative is the photoelectrode material and greater is the
rearranged as possibility of the output photovoltage and consequently
the conversion ef®ciency. The enhancement in the ¯at-
Voc %nL kT=q ln Isc =I0 3†
band potential/open circuit photopotential is due to
where, Voc is the maximum voltage at the open circuit creation of the new donor levels by substitution of the
condition and Isc is the short circuit current density. Cd2 ‡ by Pb2 ‡ ions, which effectively shift the Fermi
Under this condition, because Isc ! FL (the incident light level increasing the amount of band bending and hence
intensity) and Isc 4 I0 , Equation 3 reduces to the open circuit voltage [12]. The incremental change in
the short circuit photocurrent is due to decreased band
Voc %nL kT=q ln FL 4†
gap, increased photosensitivity and a considerable
A plot of Voc against In FL would therefore be a straight
line and quality factor of the lighted junction nL † may
then be computed for the devices. An ideal device should
have a nL value of unity so that slope of the above plot
would be 60 mV. In our case nL 's were calculated by
plotting Voc versus log Isc and are shown in Table I. The
useful power output curves were obtained for all the cell
con®gurations 20 MW cm 2 †. Fig. 2 shows eight such
representative curves. The cell parameters such as open
circuit voltage Voc †, short circuit current Isc †, ef®ciency
(Z%), ®ll factor (ff%) and series and shunt resistances
(Rs and Rsh ) were determined from these curves. Both
open circuit voltage and short circuit current have been
found to be boosted signi®cantly at x ˆ 0.1 and decreased
thereafter. For the range of x-values between 0.15 and
0.5, the cell performance is poor. The ef®ciency and ®ll
factor vary with the electrode composition x† and are
shown in Fig. 3. It is therefore of interest to interpret, in
general, how the ef®ciency of a PES cell varies with the Figure 3 Dependence of ef®ciency Z% and ®ll factor ff% on
materials parameters such as Eg , Vfb , etc. Under light photoelectrode compositions, x.

337
Figure 4 Variation of log s versus electrode composition, x.

improvement in the grain structure causing the series

resistance of a cell to decrease. The room temperature
electrical conductivities of all these electrode materials
were therefore measured to support the above observa-
tions. Fig. 4 is a variation of log s versus electrode
composition x, which demonstrates that the conductivity
Figure 5 X ray diffractograms of few of the typical photoelectrode
increased by an order up to x ˆ 0.2 and then decreased a compositions.
little for the change of x from 0.2 to 0.5. The variation in
conductivity in the ®rst region of the conductivity-
composition spectra may be correlated to the improved
grain structure as can be seen from the XRD observations
(Fig. 5) and the decreased band gap [12, 16]. The second
region of the spectra is characterized by a little decrease
in the electrical conductivity and can directly be related
to the increased amount of disorder in the samples due to
excess incorporation of of Pb2 ‡ into the lattice of CdSe
[16]. This also decreases the crystallite size as revealed
from the XRD observations (Fig. 5). The lattice
parameters were then determined for all the electrode
structures. In Fig. 6, variation of c and c/a with x is shown
only for the solid solution region. The change in lattice
parameter c is characterized by a linear variation up to
x ˆ 0.1 and then it becomes quasi-linear for
0.1  x  0.2. The c/a ratio is independent of x (Fig. 6)
as expected. The scanning electron microscopy (SEM)
micrographs shown in Fig. 7 are self explanatory of grain
boundary improvement (especially at x ˆ 0.1). For higher Figure 6 Variation of c and c/a with composition, x.
values of x x 4 0.1†, the crystal dimensions decreased
causing disordering of the material, and consequently,
PES cell performance decreases [12, 16].
materials. A series of such materials can be built over
the whole range of the composition parameter x, thus
4. Conclusions offering a wide ¯exibility for their use in a variety of
The preparation of a variable gap Cd1 x Pbx Se thin applications. With an optimum composition xopt †,
semiconductor ®lm material is made possible by an material provides excellent electrical, optical and
extremely simple and easier method that involves least structural properties desired for photovoltaic applica-
consumption of both power and the basic starting tions. When utilized in a PES cell, it gives the best
338
Figure 7 SEM micrographs for six representative photoelectrode materials. (a) x ˆ 0, (b) x ˆ 0.05, (c) x ˆ 0.1, (d) x ˆ 0.2, (e) x ˆ 0.35 and (f ) x ˆ 0.5.

performance at x ˆ 0.1. Both conversion ef®ciency (Z) References

and the ®ll factor (ff ) found to be boosted at this x-value. 1. A. J. NOZIK, Annu. Rev. Phys. Chem. 29 (1978) 189.
In general, the ef®ciencies so far obtained with such cells 2. R . N . N O U F I , P. A . KO H L
and A . J . B A R D , J. Elect. Chem. Soc.
are low and is due to several factors such as 125 (1978) 375.
3. R . M E M M I N G , Electrochim. Acta 25 (1980) 189.
photoelectrode resistivity, light absorption losses in the 4. M . A . B U T L E R and D . S . G I N L E Y, J. Mater. Sci. 15 (1980) 1.
electrolyte solution and absence of the post-preparative 5. A . A R U C H A M I , G . A R AVA M U D A N and G . V. S U B B A R AO ,
treatments. Bull. Mater. Sci. 4 (1982) 483.

339
 6. D . CA H E N , P. J . I R E L A N D , L . L . K A Z M E R S K I and F. A . Devices'', Vol. II, edited by V. Kumar and S. K. Agarwal
THIEL, J. Appl. Phys. 57 (1985) 4761. (Allied Publishers Ltd., IIT Delhi, India, 1999) p. 1392.
7. L . P. D E S H M U K H , A . B . P A LW E and V. S . S AWA N T , Sol. Cells 13. P. K . M A H A P AT R A and A . R . D U B E Y, Sol. Ener. Mat. Sol. Cells
28 (1990) 1. 32 (1994) 29.
8. K . R A J E S H WA R , L . T H O M P S O N , P. S I N G H , R . C . 14. J . O . M . B O C K R I S and A . K . N . R E D DY, in ``Modern
K A I N T H L A and K . L . C H O R P R A , J. Electrochem. Soc. 128 Electrochemistry'', Vol. 2, edited by J. O. M. Bockris and
(1981) 1744. A. K. N. Reddy (Plenum Press, New York, 1973) p. 845.
9. P. K . M A H A P AT R A and C . B . R OY, Electrochim. Acta 29 (1984) 15. S . C H A N D R A , in ``Photoelectrochemical Cells'', edited by D. S.
1435. Campbell (Gorden and Breach Science Publishers, New York,
10. M . A . K E N AW Y, H . A . Z AY E D and A . M . I B R A H I M , Ind. J. 1985).
Pure Appl. Phys. 29 (1991) 624. 16. R . N . M U L I K , Ph.D. Thesis, Shivaji University, Kolhapur, M.S.
11. L . P. D E S H M U K H and S . G . H O L I K AT T I , Bull. Mater. Sci. 17 India (2001).
(1994) 455.
12. R . N . M U L I K , D . S . S U T R AV E , V. B . P AT I L , P. D . M O R E , G . S .
S H A H A N E and L . P. D E S H M U K H , in ``Proceedings of the 10th Received 3 December 2002
International Workshop on the Physics of Semiconductor and accepted 23 September 2003

340