Solar Cells, 18 (1986) 25 - 30
SURFACE-MODIFIED CdTe PEC S O L A R CELLS K. C. M A N D A L ,
S. B A S U and D. N. B O S E
Semiconductor Division, Materials Science Centre, Indian Institute of Technology, Kharagpur 721302 (India) (Received August 15, 1984; accepted in revised form August 28, 1985)
Summary The effect of ruthenium surface modification is shown to improve considerably the properties of p-CdTe PEC solar cells. The dark I - V characteristic shows a decrease in J0 from 2.8 × 10 -7 A cm -2 to 5.1 × 10 -s A cm -2 and a decrease in ideality factor n from 2.72 to 1.54. Under AM 1 illumination Voc increased from 0.63 to 0.92 V vs. NHE, Jsc from 2.17 mA cm -2 to 3.23 mA cm -2 and fill factor F F from 0.38 to 0.48. The minority carrier diffusion length Ln is shown to increase from 0.72 #m to 0.99 /zm after ruthenium modification. The maximum current density obtainable is limited b y material resistivity.
CdTe has become of considerable interest in photovoltaic solar energy conversion as a semiconductor with the near optimum bandgap of 1.46 eV and as the only m e m b e r of the I I - V I family that exhibits b o t h types of conductivity . It has also been used as a photoelectrode in PEC systems and stable operation was reported  for n-type material with Te2-/Te22- redox systems at pH 10 - 12. It has recently been shown b y Parkinson e t al.  and others [4, 5] that modification o f the surface properties of the I I I - V c o m p o u n d s GaAs, InP and GaP b y ruthenium ions has considerable beneficial effects on the properties of PEC solar cells. In this report we demonstrate the application of similar surface modification techniques to p-CdTe which results in increases in short-circuit density Jsc, open-circuit voltage Voc and fill factor FF.
2. Experimental detaila CdTe was synthesized from 5 Ns purity elements at 1100 °C and polycrystalline materials of grain size ~ 2 m m were obtained. Doping was carried out by incorporating phosphorus atoms as acceptors. Resistivity and Hall effect measurements showed, at 306 K, a hole concentration of 3 × 1013 0379-6787/86/$3.50
© Elsevier Sequoia/Printed in The Netherlands
26 cm -3, and a hole mobility of 159 cm 2 V - ' s-1 giving a resistivity of 3 × 103 ~2 cm. Undoped CdTe prepared by a similar m e t h o d had hole concentrations o f 3 × 10 t' cm -3 and resistivities of 2 × 107 ~2 cm. Although t he resistivity even after doping was rather high, the experiments nevertheless demonstrated the effect o f surface modification in this I I - V I c o m p o u n d . Samples were cut by a low-speed wire saw to reduce surface damage, lapped, polished and etched in a solution of 3 parts concent rat ed H2SO4 and 7 parts saturated K2Cr207 solution and were finally washed with 2% brominated methanol. Samples were 0.03 cm 2 in area and 0.04 cm in thickness. Ohmic c o n t a c t was made by electroless gold deposition where gold chloride (HAuC14.H20) solution of concentration 0.05 g cm -3 in triple-distilled water was used. The sample was kept in the solution with stirring. A thin gold film was obtained which was t h o r o u g h l y washed with triple-distilled water and annealed at 300 °C for 3 min under a hydrogen atmosphere. The electrochemical cell had a single c o m p a r t m e n t consisting of the semicond u c t o r p h o t o c a t h o d e and a Pt-wire c o u n t e r electrode. All experiments employed 0.1 M Sn 4+/2+ in 0.1 M HC1 and 0.1 M KC1 at pH = 1.2. Solutions were kept under positive argon pressure during experiments due to the air sensitivity o f the electrolyte. Experiments conduct ed bot h in t he dark and with AM 1 illumination showed no measurable weight loss over 336 h. The current remained constant over these periods. The ext rem el y low dark currents and voltages were f ur t her indications of electrode stability. Moreover, the solution in the cell was analysed for cadmium before and after illumination using atomic absorption spectroscopy. No change in the cadmium cont e nt o f the solution ( < 2 X 10 -6 M) was observed. R u t h e n i u m modification was carried o u t as report ed earlier for InP . Samples o f p-CdTe after etching carefully were rinsed with distilled water. Samples were etched by 8 M HC1 for 30 s - 1 min to obtain a m at t e black surface and were then immersed in 0.01 M RuC13.3H20 in 0.1 M HNO3 solution f or 14 h. Samples were t he n washed in triple-distilled water prior to use. The surface t o p o g r a p h y before and after r u t h e n i u m modification is shown in Fig. 1.
3. Results and discussion The samples exhibited good rectification in the dark. F r o m the forward In J-V characteristic, the ideality factor was f o u n d to be 2.72 and the saturation current density to be 2.8 × 10 -7 A cm -2. The actual dark current density J0 in the PEC system was 5.0 × 10 -6 A cm -2 and the dark voltage was 14 mV. The dark voltage before modification is mainly due t o the presence o f ionic impurities in triple-distilled water. This has been proved by further experiments using deionized water when the dark voltage came dow n to 2 - 3 mV. The (1/C2)-V plot measured at 1 kHz showed a linear variation giving a carrier concent r a t i on of 6 × 1013 cm -3 and a fiat band voltage of +1.02 V (NHE).
Fig. 1. SEM pictures of: (a) "shiny" etched surface; (b) " m a t t e " etched surface; (c) ruthenium-modified surface. 120
c~ l.O 20 0 0.0
Cell voltoge, V
Fig. 2. Solar cell plots for modified and unmodified surfaces: curve A, unmodified surface, F F = 0.38; curve B, ruthenium-modified surface, F F = 0 . 4 8 . W i t h A M 1 i l l u m i n a t i o n t h e s o l a r c e l l p l o t s h o w n in F i g . 2 g a v e Jsc = 2 . 1 7 m A c m - 2 , Vo¢ = 0 . 6 3 V v s . N H E w i t h a fill f a c t o r o f 0 . 3 8 . T h e l o w fill f a c t o r is p a r t l y d u e t o a r e l a t i v e l y h i g h s e r i e s r e s i s t a n c e w h i c h w a s e s t i m a t e d
Wavelength, micron Fig. 3. Spectral response of p-CdTe in Sn 4+/2+ redox before (o) and after ($) Ru-modification.
to be 3.2 k ~ . The normalised spectral response of the p-CdTe-electrolyte system before and after surface modification was directly determined using a Jarrel-Ash m o n o c h r o m a t o r as shown in Fig. 3. The results showed a decrease in both the magnitude and extent of the sub-bandgap response. Further, the peak response which occurred at 0.8 pm in both cases was found to increase in magnitude by a factor of 1.5. The area under the spectral response curve which is proportional to J,c was thus found to increase by a factor of 1.46 on modification. After ruthenium modification the surface properties underwent a dramatic change as shown in Fig. l(c). The forward In J - V characteristic showed a reduction in J0 to 5.1 × 10 -s A cm -2 and in ideality factor n to 1.54. The (1/C2)-V plot after modification gave V f b = +1.20 V (NHE) and N A = 5.6 × 1013 cm -3. The dark current density in the electrolyte was also reduced to 1.8 X 10 -6 A cm -2, though the dark voltage increased to 28 mV. This may be explained in terms of metallic ruthenium deposition on the CdTe surface as was evident from ESCA studies. Macroscopic metal-electrolyte junctions m a y cause a slight increase in dark voltage. On AM 1 illumination the solar cell plot shown in Fig. 2 gave Voc = 0.92 V vs. NHE, Jsc = 3.23 mA cm -2 with a fill factor of 0.48. The results before and after modification are summarized in Table 1. The minority carrier diffusion length Ln in p-CdTe was determined in the electrolyte by measuring the q u a n t u m efficiency W of carrier collection vs. applied cathodic bias Vb with He-Ne laser illumination. The relation used is ln(1 - - ~7) = - - a ( 2 e e o / q N a ) l / 2 V b 1/2 - - ln(1 + czLn) as described by Russak e t al. , where the optical absorption coefficient c~ is 4.45 X 104
29 TABLE 1
Effect of ruthenium-modification on the properties of p-CdTe-electrolyte interfaces Surface condition
X 10 - 7
5.1 X 10 - s
(mA cm -2)
(A c m - 2 )
cm -1 at 632.8 nm as determined experimentally, e is the semiconductor dielectric constant and other terms have their usual significance. In Fig. 4 the intercept of the plot o f l n ( 1 - ~?) vs. Vb I n gives Ln while the slope gives the acceptor concentration Na. The values of Na thus determined were found to agree very well with those obtained from M o t t - S c h o t t k y experiments. Ln was thus found to increase from 0.72 to 0.99 # m on modification, due to a reduction of surface recombination velocity associated with removal of surface states as discussed later. Parkinson e t al.  first demonstrated the effect of surface modification on n-GaAs. They found an increase in Voc which they ascribed to removal of surface state pinning. While fill factor was also found to increase, Jsc remained unaltered for single-crystal n-GaAs but was found to increase for polycrystalline GaAs . Similar results were also obtained by Ramprakash e t al.  for n-InP, the decrease of surface state density in the band-
h-l.Z, c -1.~ -1.2 -1./.
017 0'8 0'.9 VbI/z (volts) 1/2
Fig. 4. Quantum efficiency vs. potential plots at • = 632.8 nm for unmodified (o) and modified ( e ) p-CdTe.
gap having been directly demonstrated using sub-band spectroscopy by Bose et al. . In the present experiments it was found that an increase in Voc of 50% was accompanied by an almost equal increase in Jsc. S c h o t t k y barrier studies on n- and p-CdTe have shown an insensitivity of barrier height to metal work function , the interface index being estimated to be between 0.25 and 0.4. Thus surface states dominate interface behaviour in CdTe and hence ruthenium interaction is responsible for removal of surface states as in the case of GaAs and InP. This is also the case for p-CdTe, the increase in J ~ of 50% being attributed to surface and grain boundary passivation with consequent increase in effective diffusion length of carriers as directly demonstrated here. Since the absorption depth for CdTe is also small, being 1 pm, any decrease in surface recombination velocity would result in an increase in Ln and hence in carrier collection efficiency leading to an increase in Js¢ and fill factor.
4. Conclusions Polycrystalline p-CdTe with phosphorus doping (Na = 3 × 1013 cm -3) was synthesized and used as photoelectrode in a PEC cell containing Sn 4+/2÷ redox electrolyte (pH = 1.2). The CdTe surface was modified by chemical treatment with ruthenium ions. Subsequent experiments showed an increase in both the efficiency and the stability of the PEC cells. This was attributed to a decrease in surface recombination velocity, which was demonstrated by an increase in minority carrier diffusion lengths Ln from 0.72 pm to 0.99 gm, and hence a higher collection efficiency. Reduction of surface states within the bandgap was also demonstrated from spectral response measurements.
References 1 K. Zanio, in R. F. Willardson and A. C. Beer (eds.), Cadmium Telluride: Semiconductors and Semimetals, Vol. 13, Academic Press, New York, 1978. 2 A. B. Ellis, S. W. Kaiser and M. S. Wrighton, J. Am. Chem. Soc., 98 (1978) 6418. 3 B. A. Parkinson, A. Heller and B. Miller, J. Electrochem. Soc., 126 (1979) 954. 4 Y. Ramprakash, J. N. Roy, S. Basu and D. N. Bose, Proc. Solar World Congr., Perth, Australia, 1983, Vol. 3, Pergamon, Sydney, p. 1706. 5 M. A. Butler and D. S. Ginley, Appl. Phys. Lett., 42 (1983) 582. 6 M. A. Russak, J. Reichman, H. Witzke, S. K. Deb and S. N. Chen, J. Electrochem. Soc., 127 (1980) 725. 7 W. D. Johnston, Jr., J. Electrochem. Soc., 127 (1980) 90. 8 D. N. Bose, Y. Ramprakash and S. Basu, J. Electrochem. Soc., 131 (1984) 850. 9 T. P. Ponpon, M. Saraphy, E. Buttung and P. Siffert, Phys. Status SolidiA, 59 (1980) 259.