Abstract
CuSbS2 thin films were obtained by spray pyrolysis deposition, using polymeric additives for controlling the surface properties and film’s composition. Ternary crystalline chalcostibite compounds have been obtained without any postdeposition treatments. XRD spectra and IR spectroscopy were used to characterize films composition and interactions between components. Films morphology and surface energy were investigated using AFM microscopy and contact angle measurements. Hydrophobic and hydrophilic polymers strongly influence the composition and film morphology.
1. Introduction
In heterojunction devices (solar cells), copper-based absorbers have been intensively studied mainly for their p-type conductivity due to copper deficiency [1]. More recently, ternary CIS/CIGS compounds with Cu(In/Ga/Sb/Bi)(S/Se) structures were investigated as candidates in non-silicon solar cell application. The In3+ or Ga3+ ions in the lattice reduce the Cu+ ions mobility thus increasing the stability comparing to the copper sulfides (Cu2−xS). Most of the studies are focused on Cu(In, Ga)(S/Se)2 while antimony-based compounds are less reported. In this paper CuSbS2 (chalcostibite) thin films are investigated as an alternative absorber, because of the similar properties required for photovoltaic materials, with the advantage that antimony has a lower purchase price than indium. Also, copper-gold structures that reduce the CIS/CIGS efficiency were not reported for the antimony compound.
The first report on chalcostibite occurrence came from Rhar-el-Anz in Morocco/Tunisia in 1942 as a mineral of dark gray color with metallic luster, [2], with an orthorhombic [JCPDS 44-1417] crystal structure. In terms of structural properties, CuSbS2 is often formed as a mixture of different phases also including Sb2S3 [3], or Cu3SbS3, and Cu3SbS4 [1, 4].
Literature reports on its synthesis using different techniques, both in crystalline and amorphous state. Amorphous CuSbS2 was obtained from vacuum postgrowth treatments of single-source thermal evaporation films, with optical transition in the range 1.8–2 eV [3]. Crystalline CuSbS2 was obtained by hydrothermal synthesis at 200°C [5] with similar structure as the natural mineral (JCPDS 35-413) and congruent crystallization at 552°C [1]. There are other various deposition techniques reported for obtaining crystalline CuSbS2 thin films, with variable properties in terms of crystallinity, phase composition, and morphologies: chemical bath deposition, [6], single-source thermal evaporation method [3, 7], solvothermal growth [8], and sputtering [9]. Thin films of CuSbS2 with different morphologies, varying from brick-like to plank-like, fibbers [10], and spherical morphology are reported to be obtained using spray pyrolysis deposition [8, 10, 11].
Crystalline CuSbS2 thin films (direct band-gap 1.52 eV and p-type conductivity) have been used in solid state solar cells with the structure: FTO/CdS/Sb2S3/CuSbS2 [12], FTO/TiO2/CuSbS2/C [11], and TCO/CdS/Sb2S3/Sb2S3-CuS(CuSbS2)/Ag [6]. One of the main limitations in the cells efficiency is imposed by the contact at the interface between the n-type and p-type semiconductors. Therefore, there is a strong need for tailoring the thin films morphology, supporting a good reciprocal infiltration, and avoiding shunts.
The morphology of the thin layers depends on the deposition conditions, with influence on the ratio between the nucleation and growth rates (e.g., high nucleation/growth ratio results in small grains, dense films). These rates are correlated with the reaction affinity of the precursors’ species which, in turn, depends on their stability. The control of the reactants stability in the precursors’ solution can be reached using additives able to form physical or chemical bonds with the precursor ions. The usually mentioned additives are low-molecular organic compounds (with OH and C=O groups). Their use, although efficient, can rise toxicity and cause problems. The use of polymers, with variable hydrophobic degree, represents a novel alternative investigated in this paper.
Considering the multilayered structure of a solid state solar cell, another important aspect governing the reciprocal infiltration is a good adherence of the precursor on the host layer, developed as a high energy surface. Therefore, the influence of polymeric additives on the surface properties of the CuSbS2 films is also reported.
2. Experimental
2.1. Deposition Part
CuSbS2 thin films were deposited by spray pyrolysis deposition on an automatic installation, on FTO (fluorine-doped thin oxide, TEC 8/3 mm Pilkington) conductive glass. The glass was washed in an ultrasonic bath, in several immersions in alcohol and then dried with compressed air. The glass was heated at the deposition temperature on a thermostatic heating plate. As precursors, aqueous solutions of copper(II)-chloride dehydrate, (Scharlau, CuCl2·2H2O, min 99%), concentration between 16.2–17.4 mM, antimony(III) acetate, (Alfa Aesar, (CH3COO)3Sb, 99.99%), concentration between 24–76.2 mM, and thiourea, concentration 210 mM, (Scharlau, H2NCSNH2, 99%) were prepared, with a small addition of HCl for insuring the solubility of the antimony acetate. Two polymeric additives, sodium maleate-methyl metacrylate (hydrophilic polymer, HFL) and sodium maleate-vinyl acetate (hydrophobic polymer, HFB) were used, completing the complexion effect of thiourea.
Thin films, with different composition and morphology were obtained by varying the precursors’ concentrations and molar ratio between components, as presented in Table 1.
For samples G and H polymeric additives were used in 100 ppm concentration.
Previously optimized deposition parameters [11] were used for all the samples: temperature of 240°C, carrier gas (air) pressure of 1.5 bar, spraying height 25 cm, with 60 seconds pause between two spraying sequences.
2.2. Film Characterization
The CuSbS2 films deposited on FTO were analyzed using X-ray diffraction (XRD, Bruker-AXS-D8, CuKα1 radiation) and atomic force microscopy (AFM/STM, NT-MDT model NTEGRA Probe Nanolaboratory), in semicontact mode, with an Si-tip (CSG10, force constant 0.15 N/m, tip radius10 nm); surface energy and wetting properties were analyzed using the Sessile drop technique for contact angle measurements with a OCA-20 Contact Angle-meter (DataPhysics Instruments). The IR reflectance spectra were recorded with a Spectrum BX FT-IR System (Perkin Elmer).
3. Results and Discussions
At the selected deposition parameters, all films presented a smooth homogenous aspect.
The diffraction analysis of the films show the formation of crystalline CuSbS2 (PDF 00-009-0143) in orthorhombic polymorph, along with famatinite Cu3SbS4 (PDF 00-035-0581), proving that spray pyrolysis can be used as deposition method for crystalline ternary Cu-Sb-S thin films. The Cu3SbS4 pattern in the XRD spectra of chalcostibite films was also obtained and explained by Rodríguez-Lazcano et al. as the result of copper sulfide excess which leads to the formation of copper-rich Cu-Sb-S compounds [1]. Literature [1] mentions that antimony in excess can further form binary sulphides either crystalline or amorphous. Still, the results show that, in our experimental conditions, even in Sb-excess precursors only ternary stable compounds are formed and not mixtures of secondary sulphides.
The deposition temperature assures enough thermal energy for the two main processes that take place on the substrate surface: first, the evaporation of solvents and breaking of the chemical bounds in the precursors, followed by nucleation and crystal growth.
Because evaporation of solvents takes place/very/close to the substrate surface, regions with lower temperatures are temporarily created. Thus, an important factor is the pause between the spraying sequences that must be long enough to allow the substrate reheating and the formation of nucleation centers.
The XRD spectra of the C, D, and E samples are presented in Figure 1. Since the films are very thin, the XRD spectrum of tin dioxide SnO2 (PDF 01-077-0451) from the FTO substrate is also visible.
Figure 2 presents the XRD pattern for F, G, and H samples, having the same precursor solution composition in which polymeric additives were used. At this precursor molar ratio in the solution, a new phase is detected in the XRD pattern, copper antimony sulphide, Cu3SbS3 (PDF 01-088-1335). The XRD spectra of these films do not indicate carbon in crystalline form as a residue from the polymers, while amorphous carbon based compounds are in small amounts, as shown in the FTIR spectra. This indicates that the deposition temperature is high enough for the almost total degradation of the organic polymers, forming volatile compounds.
Using the Scherrer’s formula, the crystallite dimensions were estimated and are presented for the Cu-Sb-S phases identified in the A–H samples in Table 2.
3.1. Samples without Polymeric Additives
For samples A, B, and F, where the Cu : Sb precursor ratio is under 1 : 2, the Sb content influences both the nucleation and growth rate, this can be deduced form the large differences in the crystallite size and their proportion in the two phases. For sample, F it was registered also a new phase Cu3SbS3. For samples C, D, and E, where the Cu : Sb precursor ratio is between 1 : 2.48–1 : 4.4, the Sb content variation influences only the nucleation rate. The data shows that the crystallite size of the two phases is almost equal for each sample. The nucleation speed for sample D is the largest of the samples presented, resulting in a large number of crystals with the smaller dimensions.
3.2. Samples with Polymeric Additives
The crystallite size of the Cu3SbS3 phase varies in a large range as result of the polymer complexion effect on the film formation rate, (films F, G, and H). This influence is registered only for the Cu3SbS3 phase, the CuSbS2 and Cu3SbS4 crystallite sizes remaining almost constant for the three samples. This could be the result of a more sensitive synthesis reaction of Cu3SbS3, when the influence of the polymer-precursor interactions becomes significant. The crystallite sizes of this compound are much larger than the corresponding values for CuSbS2, and Cu3SbS4, therefore, we can conclude that there is a lower nucleation rate (more stable polymer-precursors intermediates) and/or a fast growth rate for Cu3SbS3. The samples with polymers contain a smaller amount of CuSbS2 and a larger amount of Cu3SbS4 than the sample with the same composition and no additive, indicating that competitive reactions are running and the reaction affinity is influenced by the polymers.
Infra-red reflectance spectra were recorded for samples F, G, H, Figure 3, looking for residual carbon (crystalline or amorphous) from the polymer that can be additionally found in the films.
(a)
(b)
The spectra give few clues to the reaction mechanism and the processes that take place in the solutions. The functional groups from the urea compounds, NH2, C–N, C–N–H, N–H can be found at 1400, 1500, 1620, 2068, 3350, and 3800 nm [13]. Peaks observed at 702 and 1200 nm are attributed to C=S groups from the thiourea. The vibration at 2350 nm, attributed to O=C=O, can be found only for samples G and H, being related to the polymers composition as trapped or adsorbed compound. The peaks present at 1711 and 1768 nm are assigned to the COO− and C=O groups from the maleic anhydride. At 2880 a peak can be attributed to an S–CH3 group.
The peak at 620 nm, was assigned to Cu2O [13], in the solution spectra and is found both in the precursors and in the thin films. In the film, the copper oxide however is believed to be in amorphous form because it does not appear in the XRD spectra of the samples.
In the films, the IR peaks positioned at 1989 nm were attributed to C–C groups, and the vibrations found at 1040 and 1110 nm are associated with copper thiourea complexes [14].
The carbon traces revealed in the films by the FTIR analysis, as CO2 and other C–C groups were not detected by the XRD analysis, confirming the amorphous structures.
Thus, the films contain very low amount of amorphous inorganic phases (Cu2O) and traces of organic compounds resulted in the thermal decomposition of the polymeric additives.
The morphology of the deposited samples was investigated and the AFM images (Figures 4(a)–4(e)) of the films show the effect of the precursor solution composition and of the polymeric additives.
(a)
(b)
(c)
(d)
(e)
In Figures 4(a) and 4(b), the AFM images recorded for samples B and D are presented. Sample B presents a multigrain structure with open, micropores, that can have a positive effect on the reciprocal infiltration of the n and p-type semiconductors in the solid state solar cells. The increase in the antimony content leads to the extension of the grains into fibber-like aggregates.
The AFM images recorded for samples F, G, and H are presented in Figures 4(c)–4(e). The use of polymeric additives increases the density of the films. The worm-like structures obtained for sample F tend to bind together and form denser structures in the films were polymeric additives were used. Also a decrease in grain size is observed in these samples. Correlating these results with the XRD data we can assume that the larger crystallites, with regular geometry, are able to (slowly) grow in dense structures, this effect being more evident when the hydrophilic polymer is used.
Contact angle measurements were used to calculate the surface energy of the samples. Because solid state solar cells are constructed by the infiltration of the n and p-type semiconductors, a good contact between the semiconductors is essential. During SPD, deposition takes place from liquid, aqueous solutions; this is why substrates with high surface energy and good wetting capabilities are required. The contact angle measurements were carried out using two liquids: water and glycerol. With the value of the initial contact angles, energy surface was calculated using Owens, Wendt and Kaelble method [15]. The data are presented in Table 3.
When using glycerol, with low polarity, all the substrates show large and similar contact angle values, corresponding to polar, partially charged, substrates.
The contact angles measured for water (with much higher polarity) allow us to sense the differences among the ionic degree of the samples. By using polymeric additives (samples G and H), a decrease in the initial contact angle is observed, in comparison with sample F. This gets the conclusion that polymers also have influence on the surface charge distribution, as result of changing the nucleation and growth ratios. With a contact angle value below 90°, sample H has the best wetting capacity and the largest surface energy from the three samples. The surface roughness can also influence the contact angle (being lower for rough surfaces). It is interesting also to notice the very low dispersive component of the surface energy in sample G (registered as “0”) proving that the hydrophilic polymer supports the formation of rather dense, ionic layers.
4. Conclusions
Thin films of crystalline Cu-Sb-S ternary compounds were obtained on FTO substrate using spray pyrolysis deposition technique, without any postdeposition treatments. Films containing CuSbS2, Cu3SbS4, and Cu3SbS3 phases were prepared by varying the precursors’ composition. Evidences of traces of amorphous carbon-based compounds were recorded in the FTIR spectra, for the samples using polymer additives.
The samples morphology can be varied from multigrain, open pores up to long fiber-like aggregates, and to smaller worm-like morphologies. The use of polymeric additives results in a densification of the films and a decrease in the grain size, containing large crystallites of Cu3SbS3.
The polymeric additives represent a powerful tool also in tuning the surface energy, thus increasing the wetting behavior, as a prerequisite for developing efficient n-p junctions in solid state solar cells.
Acknowledgment
This work was supported by CNCSIS_UEFISCDI Project no. PNII-IDEI 840/2008, Romania.