- Email: [email protected]
Some studies on chemically synthesized antimony-doped CdSe thin films
Materials Chemistry and Physics 77 (2002) 669–676
Some studies on chemically synthesized antimony-doped CdSe thin films E.U. Masumdar a , V.B. Gaikwad b , V.B. Pujari b , P.D. More b , L.P. Deshmukh b,∗ a Rajarshi Shahu Mahavidyalaya, Latur 413512, Maharashtra, India Thin Film and Solar Studies Research Laboratory, Department of Physics (Applied Electronics), Shivaji University Centre for P.G. Studies, Solapur-Pune Road, Kegaon, Solapur 413002, Maharashtra, India b
Received 9 August 2001; received in revised form 5 January 2002; accepted 25 January 2002
Abstract A solution growth process is employed for deposition of the pure and antimony-doped CdSe thin films with Sb3+ doping concentration from 0.005 to 5 mol%. Cadmium sulphate, sodium selenosulphite (refluxed) and antimony trichloride were the basic starting materials. The samples were deposited at 60 ◦ C in an aqueous alkaline medium and were analysed spectrophotometrically, before characterizing them through the structural, microscopic, optical, and transport characterization techniques. The terminal thickness was found to increase with the Sb3+ content from 0 to 0.1 mol% and for further increase in Sb3+ concentration up to 5 mol%, the thickness decreased. The as-deposited films were found to be polycrystalline with the hexagonal wurtzite structure. The optical absorption studies gave a high coefficient of absorption (α = 104 cm−1 ) with an allowed direct type of transitions. The optical energy gap (Eg ) decreased typically from 1.79 to 1.61 eV as the doping concentration (Sb3+ ) was increased from 0 to 0.1 mol% and then it increased at higher doping levels. Electrical conductivity measurements revealed two types of conduction mechanisms, namely grain boundary scattering limited and a variable range hopping conduction. These studies showed that electrical conductivity increased with antimony content in CdSe from 0 to 0.1 mol% and then decreased for higher values of the Sb3+ contents. The thermoelectric power measurements showed that the thermally generated voltage was of the order of several microvolts and samples exhibited n-type conduction. The carrier concentration (n), mobility (µ) and intergrain barrier potentials (Φ B ’s) were computed and it was found that the carrier concentration has a poor variation with Sb3+ concentration and temperature, whereas the carrier mobility is a sensitive function of both. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Chemical deposition process; Bandgap; Electrical conductivity; Thermoemf
1. Introduction Cadmium selenide with some additives is nowadays attracting a great deal of attention owing to its potential, fundamental, experimental and applied interests in a variety of thin film devices such as laser screen materials, projection colour TVs, nuclear radiation detectors, light emitting diodes, etc. [1–6]. It is indeed worth mentioning that it has an important place in photovoltaic applications and that it can be synthesized very easily and cheaply [4,7–10]. Both binary and doped and/or ternary CdSe in bulk or in thin film form were used in electrochemical conversion devices (photoelectrochemical cells) [7–12]. However, the conversion efficiency so far reported is low and one of the major reasons is the higher resistance of the active electrode material. Doping the photoelectrode material with a suitable ∗ Corresponding author. E-mail address: [email protected] (E.U. Masumdar).
impurity concentration could effectively reduce the materials resistivity. Few of the materials (trace amounts) are tried in this direction [13–18]. Attempts are therefore made in this study to prepare thin CdSe films of varying Sb3+ doping concentration (0.005–5 mol%) by a simple, inexpensive solution growth process [13,14,19,20] and to characterize them through the structural, microscopic, optical and electrical transport properties with an ultimate intention of their possible usage in electrochemical conversion.
2. Experimental details The undoped and doped CdSe thin films with Sb3+ doping concentration from 0.005 to 5 mol% were obtained onto the ultrasonically cleaned glass substrates by a simple chemical deposition technique [13,14,19,20]. For deposition of the samples, cadmium sulphate, sodium selenosulphite and antimony trichloride (all AR grade) were used as the basic
0254-0584/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 4 - 0 5 8 4 ( 0 2 ) 0 0 1 2 2 - 0
670
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
starting materials. Sodium selenosulphite was obtained by refluxing selenium powder (5 g) with anhydrous sodium sulphite (12 g) in 200 ml double distilled water at 80 ◦ C for 9 h and was added drop wise in the reaction mixture by a separate arrangement made to the deposition system. The calculated amount of Sb3+ dopant was added to the reaction mixture directly for each composition. Triethanolamine was used as a complexing agent. Sodium hydroxide and aqueous ammonia were used to adjust the pH value and to increase the film adherence, respectively. The deposition parameters such as pH, deposition time, temperature, and speed of the substrate rotation were optimized to yield good quality films. The deposition was carried out onto the thoroughly cleaned glass substrates by changing only the Sb3+ concentration, keeping the other conditions unchanged. The film thickness was determined with the Fizeau interference technique. The spectrophotometric technique was used to determine the composition of the doped and undoped films. For this, the films were dissolved in HCl and the resulting solution was treated with the appropriate reagents to determine Cd, Se and Sb quantitatively by measuring the optical absorption of the coloured complexes at the appropriate wavelengths [21]. The X-ray diffractograms were obtained for these samples with Philips-PW-1710 X-ray diffractometer (Cu K␣ line) over the range of 2θ angles from 20◦ to 80◦ . The microscopic features were observed through a scanning electron microscope (CEMECA, SU-30, France). A spectrophotometer Hitachi-330 (Japan) was used to measure the optical absorbance of the samples in the wavelength range from 350 to 950 nm. The samples were then characterized by the electrical conductivity and thermoelectric power measurement techniques in the temperature range between 300 and 600 K and 300 and 500 K, respectively. Silver paint was used as the contact material. A chromel–alumel thermocouple was used to sense the working temperature. The current passing through the samples, voltages and thermally generated voltages were recorded as usual.
3. Results and discussion 3.1. Growth mechanism and physical properties An attempt is made to deposit antimony-doped cadmium selenide thin films by a chemical deposition technique [19,20]. The resulting pH of the reaction mixture, deposition temperature, speed of the substrate rotation and time of deposition, etc., control the deposition rate. We tested the deposition of CdSe and CdSe:Sb (0.01 mol%) at various deposition temperatures from 30 to 80 ◦ C. It was found that no film formation was seen at room temperature, although the reaction mixture was kept for a very long time. The rate of deposition was found to be enhanced significantly as temperature was increased and is shown in Fig. 1(a). At about 60 ◦ C, the thickness of the film is maximum at our
Fig. 1. Growth kinetics for CdSe:Sb thin films. (a) Film thickness versus temperature for 0 mol% (䊊) and 0.01 mol% (䊉). (b) Film thickness versus time of deposition for 0 mol% (䊊) and 0.01 mol% (䊉).
experimental conditions and above this temperature CdSe gets immediately precipitated and settled down in the container. This can be explained as usual [22]: i.e. at increasingly higher temperatures, the rates of release of Cd2+ and Se2− ions are higher, this enhances the rate of film formation and increases the film thickness up to 60 ◦ C. Above this temperature, Cd2+ and Se2− ions are released at a relatively faster rate and therefore the ions may not have sufficient time to condense on the substrate surface leading to precipitation of CdSe molecules. On the similar lines the enhancement in the thickness of the CdSe:Sb (0.01 mol%) film may also be given. The optimization of the deposition time is presented in Fig. 1(b). It was found that initially the layer thickness increased with the deposition time and later on it saturates.
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
671
Fig. 2. Variation in film thickness with the Sb3+ doping concentration in CdSe:Sb thin films.
This may be explained as follows: First, as the deposition time increased more and more ions get deposited on the substrate resulting in increased layer thickness. Secondly, for higher deposition time (>90 min), the process of depletion of number of ions in the solution caused growth rate to decrease, resulting in saturation of the layer thickness. In our case, maximum film thickness is obtained for the deposition time of 90 min. The other parameters pH and speed of the mechanical churning were also optimized and were found to be 10 ± 0.2 and 70 rpm, respectively. Then the CdSe:Sb thin film samples were deposited on to the microslide glasses at the optimized deposition conditions (i.e. at temperature = 60 ◦ C, time = 90 min, pH = 10 ± 0.2 and speed = 70 rpm). The as-deposited samples are thin, relatively uniform, smooth, and tightly adherent to the substrate support and diffusely reflecting. The colour of the deposits changed gradually from dark orangish red to yellowish orange as Sb3+ doping concentration was varied from 0 to 5 mol%. The terminal thickness of the as-grown layers increased nonlinearly with increasing Sb3+ content from 0 to 0.1 mol% and then decreased with further increase in the doping concentration. This is shown in Fig. 2. The behaviour may be explained as follows: First, the role of an antimony atom as a nucleation centre i.e., up to 0.1 mol% doping level, enhances the growth process and therefore the thickness. Secondly, at higher doping levels (> 0.1 mol%), antimony may occupy the interstitial sites causing an impurity scattering and thereby preventing further film growth [23]. 3.2. Chemical analysis All the film samples were dissolved in a 25 ml 5.5 M HCl and analysed quantitatively for Cd, Se and Sb by a
spectrophotometric technique [24]. The as-taken bath contents were also analysed by this technique. The observed contents of Cd, Se and Sb tally with that of the as-taken bath within the 5% error limit. It appeared that Sb3+ replaced Cd2+ in the deposition process. 3.3. Crystallographic and microscopic observations The XRD studies revealed that the samples are polycrystalline in nature. A careful examination of the experimental data (d-values, intensities of reflections, etc.) showed a good fit with the JCPD data [25,26]. These studies showed that CdSe exhibited hexagonal wurtzite and cubic zinc blende structures. This is in excellent agreement with the earlier reports [27,28]. The average lattice parameters of the hexagonal phase were found to be a = 4.2998 Å and c = 7.0422 Å which are in good agreement with the standard results (a = 4.30 Å and c = 7.010 Å). The intensities of reflections went on increasing with the Sb3+ content from 0 to 0.1 mol% and then decreased for higher Sb3+ contents. The average crystal size was calculated using Debye–Scherrer’s relation [29], D¯ =
kλ B cos θ
where D is the crystal size; k = 0.94, a shape and size constant; B the full width at half maximum; θ the diffracting angle and λ is the wavelength. It is found that the average crystallite size increased with the Sb3+ doping concentration from 0 to 0.1 mol% and then decreased for further increase in Sb3+ content. The SEM micrographs also supported these observations. The SEM micrographs of few of the typical films are shown in Fig. 3. It seems that the crystallite size is improved considerably with the increase in Sb3+
672
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
Fig. 3. SEM micrographs of CdSe:Sb thin films: (a) 0 mol%, (b) 0.01 mol%, (c) 0.05 mol%, (d) 0.1 mol%, (e) 2 mol% and (f) 5 mol%.
concentration. The CdSe sample doped with 0.1 mol% Sb3+ is more, crystalline with the intercrystalline spacing reduced. For higher Sb3+ doping levels, the crystallite size is reduced and the intercrystalline spacing is further increased. The overall observation is that for all the doped structures, two or more crystallites seem to be fused together. This may cause a considerable error in the measurement of the grain size [30]. The typical values of the grain sizes are documented in Table 1. 3.4. Optical studies The optical spectra of these samples were obtained and studied to evaluate the absorption coefficient (α), optical gap
(Eg ) and the nature of transitions involved. Fig. 4 represents the spectra for few of the typical samples. It is found that the optical absorbance of the doped and undoped samples is of the order of 104 –105 cm−1 . Fig. 5 represents variation of (αhν)2 versus hν to determine the optical gap. The straight line portions of these plots are extended on the hν-axis to obtain the actual values of the optical gap for each of the doping concentrations. The optical gaps for various doping concentrations are tabulated in Table 1. The optical gap for pure CdSe is 1.79 eV and it decreased to 1.61 eV as Sb3+ content was varied from 0 to 0.1 mol% and then increased for higher values of Sb3+ content in CdSe. The decrease in band gap could be ascribed to the improved grain structure of the film due to the segregation of the impurity atoms
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
673
Table 1 Effect of Sb3+ doping concentration on various properties of CdSe:Sb thin films Serial number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sb3+ (mol%) 0.0 0.005 0.01 0.025 0.05 0.075 0.1 0.15 0.20 0.30 0.50 1.0 2.0 3.0 5.0
Optical gap, Eg (eV) 1.79 1.84 1.78 1.77 1.76 1.67 1.61 1.8 1.9 1.93 1.95 1.99 2.1 2.2 2.3
Power factor, m 0.50 0.58 0.51 0.52 0.58 0.57 0.53 0.53 0.56 0.55 0.52 0.57 0.56 0.54 0.58
Activation energy (eV) HT
LT
0.67 0.66 0.65 0.64 0.63 0.62 0.57 0.59 0.63 0.67 0.70 0.71 0.73 0.85 0.86
0.240 0.230 0.232 0.220 0.220 0.200 0.140 0.170 0.180 0.190 0.196 0.201 0.206 0.210 0.211
along the grain boundaries [30,31], whereas the increase in the band gap at higher doping levels can be attributed to the increased amount of disorder caused by addition of impurity atoms [30]. As suggested by the value of the absorption coefficient (α) and straight line nature of (αhν)2 versus hν plots, especially on the higher energy side, the mode of the optical transition in these films is of band-to-band direct type. This has also been confirmed by plotting ln(αhν) versus ln(hν − Eg ). For a direct allowed type transition the
Ean (eV) 0.052 0.052 0.048 0.051 0.059 0.051 0.059 0.059 0.051 0.057 0.057 0.049 0.048 0.054 0.052
Barrier potential, Φ B (eV)
Donor level, Ed = 2Ea σ
Grain size (Å) From XRD
From SEM
0.650 0.645 0.620 0.615 0.590 0.566 0.476 0.536 0.595 0.623 0.655 0.680 0.695 0.804 0.864
0.48 0.46 0.46 0.448 0.44 0.40 0.28 0.34 0.36 0.38 0.392 0.40 0.412 0.42 0.42
194 – 202 – 219 224 231 218 202 – 188 – 166 – 153
452 – 471 – 510 – 538 – 469 – 438 – 392 – 322
above variation should yield a straight line with slope (power factor) equal to 0.5 [5]. In our case the values of slopes are nearly 0.5 (Table 1) confirming the band-to-band direct type of transition involved. 3.5. The electrical transport studies The dc dark electrical conductivities of all these samples were measured in the temperature range 300–600 K. The
Fig. 4. Wavelength dependence of the optical absorption coefficient for CdSe:Sb thin films: 0 mol% (䊊), 0.01 mol% (䊉), 0.05 mol% (), 0.1 mol% (䉱), 0.2 mol% (䊐), 0.5 mol% (䊏), 2 mol% (×), and 5 mol% ( ).
674
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
Fig. 5. Variation of (αhν)2 versus hν for different CdSe:Sb films of varying Sb3+ concentrations: 0 mol% (䊊), 0.01 mol% (䊉), 0.05 mol% (), 0.1 mol% (䉱), 0.2 mol% (䊐), 0.5 mol% (䊏), 2 mol% (×), and 5 mol% ( ).
variations of log σ versus inverse absolute temperature for eight typical CdSe:Sb films are shown in Fig. 6 and from the figure the activation energies in the low and high temperature regions were determined for each of the doping concentrations. These are listed in Table 1. The temperature dependence of an electrical conductivity for both pure and doped samples showed an usual Arrhenius behaviour consisting of high and low temperature regions with two distinct conduction mechanisms: first, a grain boundary
Fig. 6. Temperature dependence of an electrical conductivity for CdSe:Sb thin films: 0 mol% (䊊), 0.01 mol% (䊉), 0.05 mol% (), 0.1 mol% (䉱), 0.2 mol% (䊐), 0.5 mol% (䊏), 2 mol% (×), and 5 mol% ( ).
scattering limited and second, a variable range hopping [32,33]. It is seen from Fig. 7 that the room temperature electrical conductivity enhanced with an increase in the Sb3+ concentration from 0 to 0.1 mol% and then decreased
Fig. 7. Variation of an electrical conductivity with Sb3+ concentration in CdSe:Sb thin films: 298 K (䊉) and 418 K (䊊).
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
for further increase in Sb3+ content in CdSe. A discussion of the conductivity modulation is given by Bube [33] for bismuth-doped CdS thin films. A similar trend can be followed in our case. Thus an increase in conductivity with Sb3+ content up to 0.1 mol% is due to the reason that Sb3+ atoms may be doped into the vacancies of CdSe crystal substitutionally and acting as donors. At the higher Sb3+ doping levels (>0.1 mol%), the CdSe lattice may become greatly distorted and substitutional Sb3+ does not now contribute as a donor and thus decreases the conductivity. Thermoemf measurements showed n-type conduction in these films. From the experimental observations it appeared that the temperature dependence of thermoelectric power is approximately linear in the low temperature region, whereas it deviated from the linear behaviour at higher temperatures. This is shown in Fig. 8. The nonlinearity of the plots indicated nondegeneracy of the material whose thermoelectric power is proportional to the nth power of the absolute temperature and is a weak function of the temperature itself. The carrier densities (n) at different temperatures were calculated for all the samples. The mobilities (µ) were, therefore, determined. Fig. 9 is a sketch of the carrier density (n) and mobility (µ) as a function of the Sb3+ doping concentration in CdSe. It is seen that the carrier density and mobility vary with the Sb3+ content in almost the same fashion as that of the conductivity. The increase in carrier concentration up to 0.1 mol% doping level can be understood from the role of antimony as a donor. Antimony is reported to replace the cadmium substitutionally and produce shallow donor levels in the band gap within the allowable range of doping concentration. However, at higher doping levels, the carrier concentration decreased owing to the decreased grain size that resulted into the trapping of the charge carriers
675
Fig. 8. Variation of thermoelectric power with temperature for CdSe:Sb thin films: 0 mol% (䊊), 0.01 mol% (䊉), 0.05 mol% (), 0.1 mol% (䉱), 0.2 mol% (䊐), 0.5 mol% (䊏), 2 mol% (×), and 5 mol% ( ).
at the grain boundaries making them immobile [34]. The variations in carrier mobility are more sensitive to antimony doping concentration compared to the carrier density. Further, the mobility increased with the applied temperature suggesting the presence of scattering mechanism associated with the intergrain barrier [32]. The intergrain barrier potentials were then determined for all the samples and are listed in Table 1. It is seen that intergrain barrier potentials (Φ B ) decreased with Sb doping concentration,
Fig. 9. Variation of the carrier concentration n: (䊊) and mobility, µ (䊉) with the Sb3+ concentration in CdSe:Sb thin films.
676
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
attained a minimum at 0.1 mol% and afterwards increased with further increase in Sb doping concentration.
4. Conclusions A simple chemical growth process is described for the synthesis of antimony-doped cadmium selenide thin films. The physical, structural, microscopic, optical, and electrical transport characteristics of these films have been examined. XRD studies revealed that these samples are polycrystalline in nature exhibiting the hexagonal wurtzite and cubic zinc blende structures. The crystallite size depends on the doping mol% of antimony in CdSe and is maximum at 0.1 mol% Sb concentration in CdSe. The polycrystallinity of the films is well supported by SEM micrographs. The optical absorption studies showed that the transitions involved in these films are of the band-to-band direct type. The optical gap decreased typically from 1.79 to 1.61 eV with the increase in Sb content from 0 to 0.1 mol% and then increased further on the higher doping side. The electrical conductivity, mobility and carrier concentration enhanced with the doping concentration up to 0.1 mol% and then decreased with further increase in doping concentration. From these studies we may conclude that doped CdSe (Sb3+ , 0.1 mol%) will be the most suitable photoelectrode in photoelectrochemical cell.
Acknowledgements One of the authors (EUM) is thankful to the University Grants Commission, Government of India, New Delhi for awarding him fellowship under the Faculty Improvement Programme (FIP) and to the authorities of Shri Shiv Chattrapati Shikshan Sanstha and Principal, Rajarshi Shahu Mahavidyalya, Latur for kind permission to avail the fellowship. References [1] A.H. Burger, M.J. Roth, Cryst. Growth 70 (1984) 386. [2] A.A. Basam, A.W. Brinkman, G.J. Russell, K.J. Woods, J. Cryst. Growth 8 (1988) 667. [3] K.C. Sharma, J.C. Garg, Ind. J. Pure & Appl. Phys. 6 (1988) 480. [4] V. Krishnan, K.K. Mishra, K. Rejeshwar, Electrochem. Soc. 139 (1992) 23.
[5] P. Gupta, B. Maity, A.B. Maity, S. Choudhari, A.K. Pal, Thin Solid Films 260 (1995) 75. [6] N. Samarth, J.K. Luth Fardyan, R.G. Alonso, Appl. Phys. Lett. 56 (1990) 1163. [7] L.P. Deshmukh, G.S. Shahane, J. Int. Electron. 83 (1997) 341. [8] L.P. Deshmukh, G.S. Shahane, K.M. Garadkar, Ind. J. Pure & Appl. Phys. 35 (1997) 560. [9] D.S. Sutrave, G.S. Shahane, V.B. Patil, L.P. Deshmukh, Turk. J. Phys. 23 (1999) 1. [10] D.S. Sutrave, G.S. Shahane, L.P. Deshmukh, Int. J. Electron. 87 (2000) 1177. [11] K. Tsevtkova, K. Kochev, Sol. Ener. Mater. Sol. Cells 31 (1993) 429. [12] K.R. Murali, V. Subramanian, N. Rangarajan, Bull. Electrochem. 9 (1993) 249. [13] G.S. Shahane, L.P. Deshmukh, Mater. Chem. Phys. 70 (2001) 112. [14] L.P. Deshmukh, S.G. Holikatti, J. Phys. D: Appl. Phys. 27 (1994) 1786. [15] S.H. Pawar, L.P. Deshmukh, Mater. Chem. Phys. 10 (1984) 83. [16] S. Roy, R. Banerjee, A.K. Barua, Thin Solid Films. 79 (1981) 155. [17] G.K. Padam, G.L. Malhotra, S. Rao, J. Phys. Appl. Phys. 63 (1988) 770. [18] K. Subaramaiah, V.S. Raja, Sol. Energ. Mater. Sol. Cells 32 (1994) 1. [19] E.U. Masumdar, V.B. Pujari, V.B. Gaikwad, V.B. Patil, L.P. Deshmukh, Proceedings of the 11th International Workshop on the Physics of Semiconductor Devices, IIT, New Delhi, India, December 11–15, 2001. [20] E.U. Masumdar, V.B. Pujari, V.B. Gaikwad, P.D. More, L.P. Deshmukh, Proceedings of the National Symposium on Science and Technology of Vacuum and Thin Films, IISc, Bangalore, India, September 5–7, 2001. [21] P.D. More, G.S. Shahane, L.P. Deshmukh, Ind. J. Pure & Appl. Phys. 40 (2002) 62. [22] G.S. Shahane, D.S. Sutrave, L.P. Deshmukh, Ind. J. Pure & Appl. Phys. 34 (1996) 153. [23] G.S. Shahane, K.M. Garadkar, L.P. Deshmukh, Mater. Chem. Phys. 51 (1997) 246. [24] A.R. Patil, V.N. Patil, P.N. Bhosle, Thin Solid Films, in press. [25] ASTM Data Card for X-ray Diffraction, 8-459. [26] ASTM Data Card for X-ray Diffraction, 19-191. [27] P.J. Sebastin, Thin Solid Films 245 (1995) 132. [28] H.P. Sarma, N. Rangarajan, K.R. Murthi, Phys. Stat. Sol. (a) 148 (1995) k77. [29] B.D. Cullity, Elements of X-ray Diffraction, Addison Wesley, Reading, MA, 1978, p. 102. [30] G.S. Shahne, L.P. Deshmukh, Mater. Chem. Phys. 51 (1997) 246. [31] U. Pal, S. Saha, A.K. Choudhuri, V.V. Rao, H.D. Banarjee, J. Phys. D 22 (1989) 965. [32] G. Micocci, A. Tepore, R. Rella, P. Siciliano, Phys. Stat. Sol. (a) 148 (1995) 431. [33] R.H. Bube, Photoconductivity of Solids, Wiley, New York, 1963, p. 189. [34] L.P. Deshmukh, S.G. Holllikatti, B.M. More, Mater. Chem. Phys. 39 (1995) 743.
Some studies on chemically synthesized antimony-doped CdSe thin films E.U. Masumdar a , V.B. Gaikwad b , V.B. Pujari b , P.D. More b , L.P. Deshmukh b,∗ a Rajarshi Shahu Mahavidyalaya, Latur 413512, Maharashtra, India Thin Film and Solar Studies Research Laboratory, Department of Physics (Applied Electronics), Shivaji University Centre for P.G. Studies, Solapur-Pune Road, Kegaon, Solapur 413002, Maharashtra, India b
Received 9 August 2001; received in revised form 5 January 2002; accepted 25 January 2002
Abstract A solution growth process is employed for deposition of the pure and antimony-doped CdSe thin films with Sb3+ doping concentration from 0.005 to 5 mol%. Cadmium sulphate, sodium selenosulphite (refluxed) and antimony trichloride were the basic starting materials. The samples were deposited at 60 ◦ C in an aqueous alkaline medium and were analysed spectrophotometrically, before characterizing them through the structural, microscopic, optical, and transport characterization techniques. The terminal thickness was found to increase with the Sb3+ content from 0 to 0.1 mol% and for further increase in Sb3+ concentration up to 5 mol%, the thickness decreased. The as-deposited films were found to be polycrystalline with the hexagonal wurtzite structure. The optical absorption studies gave a high coefficient of absorption (α = 104 cm−1 ) with an allowed direct type of transitions. The optical energy gap (Eg ) decreased typically from 1.79 to 1.61 eV as the doping concentration (Sb3+ ) was increased from 0 to 0.1 mol% and then it increased at higher doping levels. Electrical conductivity measurements revealed two types of conduction mechanisms, namely grain boundary scattering limited and a variable range hopping conduction. These studies showed that electrical conductivity increased with antimony content in CdSe from 0 to 0.1 mol% and then decreased for higher values of the Sb3+ contents. The thermoelectric power measurements showed that the thermally generated voltage was of the order of several microvolts and samples exhibited n-type conduction. The carrier concentration (n), mobility (µ) and intergrain barrier potentials (Φ B ’s) were computed and it was found that the carrier concentration has a poor variation with Sb3+ concentration and temperature, whereas the carrier mobility is a sensitive function of both. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Chemical deposition process; Bandgap; Electrical conductivity; Thermoemf
1. Introduction Cadmium selenide with some additives is nowadays attracting a great deal of attention owing to its potential, fundamental, experimental and applied interests in a variety of thin film devices such as laser screen materials, projection colour TVs, nuclear radiation detectors, light emitting diodes, etc. [1–6]. It is indeed worth mentioning that it has an important place in photovoltaic applications and that it can be synthesized very easily and cheaply [4,7–10]. Both binary and doped and/or ternary CdSe in bulk or in thin film form were used in electrochemical conversion devices (photoelectrochemical cells) [7–12]. However, the conversion efficiency so far reported is low and one of the major reasons is the higher resistance of the active electrode material. Doping the photoelectrode material with a suitable ∗ Corresponding author. E-mail address: [email protected] (E.U. Masumdar).
impurity concentration could effectively reduce the materials resistivity. Few of the materials (trace amounts) are tried in this direction [13–18]. Attempts are therefore made in this study to prepare thin CdSe films of varying Sb3+ doping concentration (0.005–5 mol%) by a simple, inexpensive solution growth process [13,14,19,20] and to characterize them through the structural, microscopic, optical and electrical transport properties with an ultimate intention of their possible usage in electrochemical conversion.
2. Experimental details The undoped and doped CdSe thin films with Sb3+ doping concentration from 0.005 to 5 mol% were obtained onto the ultrasonically cleaned glass substrates by a simple chemical deposition technique [13,14,19,20]. For deposition of the samples, cadmium sulphate, sodium selenosulphite and antimony trichloride (all AR grade) were used as the basic
0254-0584/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 4 - 0 5 8 4 ( 0 2 ) 0 0 1 2 2 - 0
670
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
starting materials. Sodium selenosulphite was obtained by refluxing selenium powder (5 g) with anhydrous sodium sulphite (12 g) in 200 ml double distilled water at 80 ◦ C for 9 h and was added drop wise in the reaction mixture by a separate arrangement made to the deposition system. The calculated amount of Sb3+ dopant was added to the reaction mixture directly for each composition. Triethanolamine was used as a complexing agent. Sodium hydroxide and aqueous ammonia were used to adjust the pH value and to increase the film adherence, respectively. The deposition parameters such as pH, deposition time, temperature, and speed of the substrate rotation were optimized to yield good quality films. The deposition was carried out onto the thoroughly cleaned glass substrates by changing only the Sb3+ concentration, keeping the other conditions unchanged. The film thickness was determined with the Fizeau interference technique. The spectrophotometric technique was used to determine the composition of the doped and undoped films. For this, the films were dissolved in HCl and the resulting solution was treated with the appropriate reagents to determine Cd, Se and Sb quantitatively by measuring the optical absorption of the coloured complexes at the appropriate wavelengths [21]. The X-ray diffractograms were obtained for these samples with Philips-PW-1710 X-ray diffractometer (Cu K␣ line) over the range of 2θ angles from 20◦ to 80◦ . The microscopic features were observed through a scanning electron microscope (CEMECA, SU-30, France). A spectrophotometer Hitachi-330 (Japan) was used to measure the optical absorbance of the samples in the wavelength range from 350 to 950 nm. The samples were then characterized by the electrical conductivity and thermoelectric power measurement techniques in the temperature range between 300 and 600 K and 300 and 500 K, respectively. Silver paint was used as the contact material. A chromel–alumel thermocouple was used to sense the working temperature. The current passing through the samples, voltages and thermally generated voltages were recorded as usual.
3. Results and discussion 3.1. Growth mechanism and physical properties An attempt is made to deposit antimony-doped cadmium selenide thin films by a chemical deposition technique [19,20]. The resulting pH of the reaction mixture, deposition temperature, speed of the substrate rotation and time of deposition, etc., control the deposition rate. We tested the deposition of CdSe and CdSe:Sb (0.01 mol%) at various deposition temperatures from 30 to 80 ◦ C. It was found that no film formation was seen at room temperature, although the reaction mixture was kept for a very long time. The rate of deposition was found to be enhanced significantly as temperature was increased and is shown in Fig. 1(a). At about 60 ◦ C, the thickness of the film is maximum at our
Fig. 1. Growth kinetics for CdSe:Sb thin films. (a) Film thickness versus temperature for 0 mol% (䊊) and 0.01 mol% (䊉). (b) Film thickness versus time of deposition for 0 mol% (䊊) and 0.01 mol% (䊉).
experimental conditions and above this temperature CdSe gets immediately precipitated and settled down in the container. This can be explained as usual [22]: i.e. at increasingly higher temperatures, the rates of release of Cd2+ and Se2− ions are higher, this enhances the rate of film formation and increases the film thickness up to 60 ◦ C. Above this temperature, Cd2+ and Se2− ions are released at a relatively faster rate and therefore the ions may not have sufficient time to condense on the substrate surface leading to precipitation of CdSe molecules. On the similar lines the enhancement in the thickness of the CdSe:Sb (0.01 mol%) film may also be given. The optimization of the deposition time is presented in Fig. 1(b). It was found that initially the layer thickness increased with the deposition time and later on it saturates.
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
671
Fig. 2. Variation in film thickness with the Sb3+ doping concentration in CdSe:Sb thin films.
This may be explained as follows: First, as the deposition time increased more and more ions get deposited on the substrate resulting in increased layer thickness. Secondly, for higher deposition time (>90 min), the process of depletion of number of ions in the solution caused growth rate to decrease, resulting in saturation of the layer thickness. In our case, maximum film thickness is obtained for the deposition time of 90 min. The other parameters pH and speed of the mechanical churning were also optimized and were found to be 10 ± 0.2 and 70 rpm, respectively. Then the CdSe:Sb thin film samples were deposited on to the microslide glasses at the optimized deposition conditions (i.e. at temperature = 60 ◦ C, time = 90 min, pH = 10 ± 0.2 and speed = 70 rpm). The as-deposited samples are thin, relatively uniform, smooth, and tightly adherent to the substrate support and diffusely reflecting. The colour of the deposits changed gradually from dark orangish red to yellowish orange as Sb3+ doping concentration was varied from 0 to 5 mol%. The terminal thickness of the as-grown layers increased nonlinearly with increasing Sb3+ content from 0 to 0.1 mol% and then decreased with further increase in the doping concentration. This is shown in Fig. 2. The behaviour may be explained as follows: First, the role of an antimony atom as a nucleation centre i.e., up to 0.1 mol% doping level, enhances the growth process and therefore the thickness. Secondly, at higher doping levels (> 0.1 mol%), antimony may occupy the interstitial sites causing an impurity scattering and thereby preventing further film growth [23]. 3.2. Chemical analysis All the film samples were dissolved in a 25 ml 5.5 M HCl and analysed quantitatively for Cd, Se and Sb by a
spectrophotometric technique [24]. The as-taken bath contents were also analysed by this technique. The observed contents of Cd, Se and Sb tally with that of the as-taken bath within the 5% error limit. It appeared that Sb3+ replaced Cd2+ in the deposition process. 3.3. Crystallographic and microscopic observations The XRD studies revealed that the samples are polycrystalline in nature. A careful examination of the experimental data (d-values, intensities of reflections, etc.) showed a good fit with the JCPD data [25,26]. These studies showed that CdSe exhibited hexagonal wurtzite and cubic zinc blende structures. This is in excellent agreement with the earlier reports [27,28]. The average lattice parameters of the hexagonal phase were found to be a = 4.2998 Å and c = 7.0422 Å which are in good agreement with the standard results (a = 4.30 Å and c = 7.010 Å). The intensities of reflections went on increasing with the Sb3+ content from 0 to 0.1 mol% and then decreased for higher Sb3+ contents. The average crystal size was calculated using Debye–Scherrer’s relation [29], D¯ =
kλ B cos θ
where D is the crystal size; k = 0.94, a shape and size constant; B the full width at half maximum; θ the diffracting angle and λ is the wavelength. It is found that the average crystallite size increased with the Sb3+ doping concentration from 0 to 0.1 mol% and then decreased for further increase in Sb3+ content. The SEM micrographs also supported these observations. The SEM micrographs of few of the typical films are shown in Fig. 3. It seems that the crystallite size is improved considerably with the increase in Sb3+
672
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
Fig. 3. SEM micrographs of CdSe:Sb thin films: (a) 0 mol%, (b) 0.01 mol%, (c) 0.05 mol%, (d) 0.1 mol%, (e) 2 mol% and (f) 5 mol%.
concentration. The CdSe sample doped with 0.1 mol% Sb3+ is more, crystalline with the intercrystalline spacing reduced. For higher Sb3+ doping levels, the crystallite size is reduced and the intercrystalline spacing is further increased. The overall observation is that for all the doped structures, two or more crystallites seem to be fused together. This may cause a considerable error in the measurement of the grain size [30]. The typical values of the grain sizes are documented in Table 1. 3.4. Optical studies The optical spectra of these samples were obtained and studied to evaluate the absorption coefficient (α), optical gap
(Eg ) and the nature of transitions involved. Fig. 4 represents the spectra for few of the typical samples. It is found that the optical absorbance of the doped and undoped samples is of the order of 104 –105 cm−1 . Fig. 5 represents variation of (αhν)2 versus hν to determine the optical gap. The straight line portions of these plots are extended on the hν-axis to obtain the actual values of the optical gap for each of the doping concentrations. The optical gaps for various doping concentrations are tabulated in Table 1. The optical gap for pure CdSe is 1.79 eV and it decreased to 1.61 eV as Sb3+ content was varied from 0 to 0.1 mol% and then increased for higher values of Sb3+ content in CdSe. The decrease in band gap could be ascribed to the improved grain structure of the film due to the segregation of the impurity atoms
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
673
Table 1 Effect of Sb3+ doping concentration on various properties of CdSe:Sb thin films Serial number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sb3+ (mol%) 0.0 0.005 0.01 0.025 0.05 0.075 0.1 0.15 0.20 0.30 0.50 1.0 2.0 3.0 5.0
Optical gap, Eg (eV) 1.79 1.84 1.78 1.77 1.76 1.67 1.61 1.8 1.9 1.93 1.95 1.99 2.1 2.2 2.3
Power factor, m 0.50 0.58 0.51 0.52 0.58 0.57 0.53 0.53 0.56 0.55 0.52 0.57 0.56 0.54 0.58
Activation energy (eV) HT
LT
0.67 0.66 0.65 0.64 0.63 0.62 0.57 0.59 0.63 0.67 0.70 0.71 0.73 0.85 0.86
0.240 0.230 0.232 0.220 0.220 0.200 0.140 0.170 0.180 0.190 0.196 0.201 0.206 0.210 0.211
along the grain boundaries [30,31], whereas the increase in the band gap at higher doping levels can be attributed to the increased amount of disorder caused by addition of impurity atoms [30]. As suggested by the value of the absorption coefficient (α) and straight line nature of (αhν)2 versus hν plots, especially on the higher energy side, the mode of the optical transition in these films is of band-to-band direct type. This has also been confirmed by plotting ln(αhν) versus ln(hν − Eg ). For a direct allowed type transition the
Ean (eV) 0.052 0.052 0.048 0.051 0.059 0.051 0.059 0.059 0.051 0.057 0.057 0.049 0.048 0.054 0.052
Barrier potential, Φ B (eV)
Donor level, Ed = 2Ea σ
Grain size (Å) From XRD
From SEM
0.650 0.645 0.620 0.615 0.590 0.566 0.476 0.536 0.595 0.623 0.655 0.680 0.695 0.804 0.864
0.48 0.46 0.46 0.448 0.44 0.40 0.28 0.34 0.36 0.38 0.392 0.40 0.412 0.42 0.42
194 – 202 – 219 224 231 218 202 – 188 – 166 – 153
452 – 471 – 510 – 538 – 469 – 438 – 392 – 322
above variation should yield a straight line with slope (power factor) equal to 0.5 [5]. In our case the values of slopes are nearly 0.5 (Table 1) confirming the band-to-band direct type of transition involved. 3.5. The electrical transport studies The dc dark electrical conductivities of all these samples were measured in the temperature range 300–600 K. The
Fig. 4. Wavelength dependence of the optical absorption coefficient for CdSe:Sb thin films: 0 mol% (䊊), 0.01 mol% (䊉), 0.05 mol% (), 0.1 mol% (䉱), 0.2 mol% (䊐), 0.5 mol% (䊏), 2 mol% (×), and 5 mol% ( ).
674
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
Fig. 5. Variation of (αhν)2 versus hν for different CdSe:Sb films of varying Sb3+ concentrations: 0 mol% (䊊), 0.01 mol% (䊉), 0.05 mol% (), 0.1 mol% (䉱), 0.2 mol% (䊐), 0.5 mol% (䊏), 2 mol% (×), and 5 mol% ( ).
variations of log σ versus inverse absolute temperature for eight typical CdSe:Sb films are shown in Fig. 6 and from the figure the activation energies in the low and high temperature regions were determined for each of the doping concentrations. These are listed in Table 1. The temperature dependence of an electrical conductivity for both pure and doped samples showed an usual Arrhenius behaviour consisting of high and low temperature regions with two distinct conduction mechanisms: first, a grain boundary
Fig. 6. Temperature dependence of an electrical conductivity for CdSe:Sb thin films: 0 mol% (䊊), 0.01 mol% (䊉), 0.05 mol% (), 0.1 mol% (䉱), 0.2 mol% (䊐), 0.5 mol% (䊏), 2 mol% (×), and 5 mol% ( ).
scattering limited and second, a variable range hopping [32,33]. It is seen from Fig. 7 that the room temperature electrical conductivity enhanced with an increase in the Sb3+ concentration from 0 to 0.1 mol% and then decreased
Fig. 7. Variation of an electrical conductivity with Sb3+ concentration in CdSe:Sb thin films: 298 K (䊉) and 418 K (䊊).
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
for further increase in Sb3+ content in CdSe. A discussion of the conductivity modulation is given by Bube [33] for bismuth-doped CdS thin films. A similar trend can be followed in our case. Thus an increase in conductivity with Sb3+ content up to 0.1 mol% is due to the reason that Sb3+ atoms may be doped into the vacancies of CdSe crystal substitutionally and acting as donors. At the higher Sb3+ doping levels (>0.1 mol%), the CdSe lattice may become greatly distorted and substitutional Sb3+ does not now contribute as a donor and thus decreases the conductivity. Thermoemf measurements showed n-type conduction in these films. From the experimental observations it appeared that the temperature dependence of thermoelectric power is approximately linear in the low temperature region, whereas it deviated from the linear behaviour at higher temperatures. This is shown in Fig. 8. The nonlinearity of the plots indicated nondegeneracy of the material whose thermoelectric power is proportional to the nth power of the absolute temperature and is a weak function of the temperature itself. The carrier densities (n) at different temperatures were calculated for all the samples. The mobilities (µ) were, therefore, determined. Fig. 9 is a sketch of the carrier density (n) and mobility (µ) as a function of the Sb3+ doping concentration in CdSe. It is seen that the carrier density and mobility vary with the Sb3+ content in almost the same fashion as that of the conductivity. The increase in carrier concentration up to 0.1 mol% doping level can be understood from the role of antimony as a donor. Antimony is reported to replace the cadmium substitutionally and produce shallow donor levels in the band gap within the allowable range of doping concentration. However, at higher doping levels, the carrier concentration decreased owing to the decreased grain size that resulted into the trapping of the charge carriers
675
Fig. 8. Variation of thermoelectric power with temperature for CdSe:Sb thin films: 0 mol% (䊊), 0.01 mol% (䊉), 0.05 mol% (), 0.1 mol% (䉱), 0.2 mol% (䊐), 0.5 mol% (䊏), 2 mol% (×), and 5 mol% ( ).
at the grain boundaries making them immobile [34]. The variations in carrier mobility are more sensitive to antimony doping concentration compared to the carrier density. Further, the mobility increased with the applied temperature suggesting the presence of scattering mechanism associated with the intergrain barrier [32]. The intergrain barrier potentials were then determined for all the samples and are listed in Table 1. It is seen that intergrain barrier potentials (Φ B ) decreased with Sb doping concentration,
Fig. 9. Variation of the carrier concentration n: (䊊) and mobility, µ (䊉) with the Sb3+ concentration in CdSe:Sb thin films.
676
E.U. Masumdar et al. / Materials Chemistry and Physics 77 (2002) 669–676
attained a minimum at 0.1 mol% and afterwards increased with further increase in Sb doping concentration.
4. Conclusions A simple chemical growth process is described for the synthesis of antimony-doped cadmium selenide thin films. The physical, structural, microscopic, optical, and electrical transport characteristics of these films have been examined. XRD studies revealed that these samples are polycrystalline in nature exhibiting the hexagonal wurtzite and cubic zinc blende structures. The crystallite size depends on the doping mol% of antimony in CdSe and is maximum at 0.1 mol% Sb concentration in CdSe. The polycrystallinity of the films is well supported by SEM micrographs. The optical absorption studies showed that the transitions involved in these films are of the band-to-band direct type. The optical gap decreased typically from 1.79 to 1.61 eV with the increase in Sb content from 0 to 0.1 mol% and then increased further on the higher doping side. The electrical conductivity, mobility and carrier concentration enhanced with the doping concentration up to 0.1 mol% and then decreased with further increase in doping concentration. From these studies we may conclude that doped CdSe (Sb3+ , 0.1 mol%) will be the most suitable photoelectrode in photoelectrochemical cell.
Acknowledgements One of the authors (EUM) is thankful to the University Grants Commission, Government of India, New Delhi for awarding him fellowship under the Faculty Improvement Programme (FIP) and to the authorities of Shri Shiv Chattrapati Shikshan Sanstha and Principal, Rajarshi Shahu Mahavidyalya, Latur for kind permission to avail the fellowship. References [1] A.H. Burger, M.J. Roth, Cryst. Growth 70 (1984) 386. [2] A.A. Basam, A.W. Brinkman, G.J. Russell, K.J. Woods, J. Cryst. Growth 8 (1988) 667. [3] K.C. Sharma, J.C. Garg, Ind. J. Pure & Appl. Phys. 6 (1988) 480. [4] V. Krishnan, K.K. Mishra, K. Rejeshwar, Electrochem. Soc. 139 (1992) 23.
[5] P. Gupta, B. Maity, A.B. Maity, S. Choudhari, A.K. Pal, Thin Solid Films 260 (1995) 75. [6] N. Samarth, J.K. Luth Fardyan, R.G. Alonso, Appl. Phys. Lett. 56 (1990) 1163. [7] L.P. Deshmukh, G.S. Shahane, J. Int. Electron. 83 (1997) 341. [8] L.P. Deshmukh, G.S. Shahane, K.M. Garadkar, Ind. J. Pure & Appl. Phys. 35 (1997) 560. [9] D.S. Sutrave, G.S. Shahane, V.B. Patil, L.P. Deshmukh, Turk. J. Phys. 23 (1999) 1. [10] D.S. Sutrave, G.S. Shahane, L.P. Deshmukh, Int. J. Electron. 87 (2000) 1177. [11] K. Tsevtkova, K. Kochev, Sol. Ener. Mater. Sol. Cells 31 (1993) 429. [12] K.R. Murali, V. Subramanian, N. Rangarajan, Bull. Electrochem. 9 (1993) 249. [13] G.S. Shahane, L.P. Deshmukh, Mater. Chem. Phys. 70 (2001) 112. [14] L.P. Deshmukh, S.G. Holikatti, J. Phys. D: Appl. Phys. 27 (1994) 1786. [15] S.H. Pawar, L.P. Deshmukh, Mater. Chem. Phys. 10 (1984) 83. [16] S. Roy, R. Banerjee, A.K. Barua, Thin Solid Films. 79 (1981) 155. [17] G.K. Padam, G.L. Malhotra, S. Rao, J. Phys. Appl. Phys. 63 (1988) 770. [18] K. Subaramaiah, V.S. Raja, Sol. Energ. Mater. Sol. Cells 32 (1994) 1. [19] E.U. Masumdar, V.B. Pujari, V.B. Gaikwad, V.B. Patil, L.P. Deshmukh, Proceedings of the 11th International Workshop on the Physics of Semiconductor Devices, IIT, New Delhi, India, December 11–15, 2001. [20] E.U. Masumdar, V.B. Pujari, V.B. Gaikwad, P.D. More, L.P. Deshmukh, Proceedings of the National Symposium on Science and Technology of Vacuum and Thin Films, IISc, Bangalore, India, September 5–7, 2001. [21] P.D. More, G.S. Shahane, L.P. Deshmukh, Ind. J. Pure & Appl. Phys. 40 (2002) 62. [22] G.S. Shahane, D.S. Sutrave, L.P. Deshmukh, Ind. J. Pure & Appl. Phys. 34 (1996) 153. [23] G.S. Shahane, K.M. Garadkar, L.P. Deshmukh, Mater. Chem. Phys. 51 (1997) 246. [24] A.R. Patil, V.N. Patil, P.N. Bhosle, Thin Solid Films, in press. [25] ASTM Data Card for X-ray Diffraction, 8-459. [26] ASTM Data Card for X-ray Diffraction, 19-191. [27] P.J. Sebastin, Thin Solid Films 245 (1995) 132. [28] H.P. Sarma, N. Rangarajan, K.R. Murthi, Phys. Stat. Sol. (a) 148 (1995) k77. [29] B.D. Cullity, Elements of X-ray Diffraction, Addison Wesley, Reading, MA, 1978, p. 102. [30] G.S. Shahne, L.P. Deshmukh, Mater. Chem. Phys. 51 (1997) 246. [31] U. Pal, S. Saha, A.K. Choudhuri, V.V. Rao, H.D. Banarjee, J. Phys. D 22 (1989) 965. [32] G. Micocci, A. Tepore, R. Rella, P. Siciliano, Phys. Stat. Sol. (a) 148 (1995) 431. [33] R.H. Bube, Photoconductivity of Solids, Wiley, New York, 1963, p. 189. [34] L.P. Deshmukh, S.G. Holllikatti, B.M. More, Mater. Chem. Phys. 39 (1995) 743.