3256
J. Phys. Chem. C 2010, 114, 3256–3259
Photovoltaic Behavior of Nanocrystalline SnS/TiO2 Yu Wang, Hao Gong,* Benhu Fan, and Guangxia Hu Department of Materials Science and Engineering, National UniVersity of Singapore, 10 Kent Ridge Crescent, Singapore 119260, Singapore ReceiVed: August 6, 2009; ReVised Manuscript ReceiVed: December 3, 2009
Nanocrystalline tin sulfide (SnS) was prepared by chemical bath deposition, and the photovoltaic behavior of SnS/TiO2 was studied. The X-ray diffraction pattern and transmission electron microscopy revealed an ∼6 nm SnS polycrystalline orthorhombic structure. The SnS film exhibited a band gap of 1.3 eV, and its absorption coefficient was more than 1 × 104 cm-1 in the visible light range. The electrical conductivity activation energy of the SnS film was 0.22 eV, determined when the sample was heated in the temperature range of 111-144 °C. Although the sample was insulating at room temperature, photovoltaic behavior was found in a SnS/TiO2 structure, with an open-circuit voltage (Voc) of 471 mV, a short-circuit current density (Jsc) of 0.3 mA/cm2, and the conversion efficiency (η) of 0.1% under 1 sun illumination. The properties of SnS and the reasons behind the photovoltaic phenomenon of SnS/TiO2 are discussed. Introduction Tin sulfide (SnS), a group IV-VI chalcogenide semiconductor compound, has attracted much attention for photovoltaic devices because of its p-type conduction, high absorption in the visible range, little toxicity, and inexpensiveness. There are various techniques to prepare SnS films, such as spray pyrolysis,1,2 thermal evaporation,3-6 sputtering,7 electrochemical and photochemical deposition,8-11 and chemical bath deposition (CBD).12-14 The chemical bath deposition is simple, low-cost, and suitable for producing large-area thin films for solar energy related applications. Gunasekaran et al.11 reported a cadmium sulfide (CdS)/SnS solar cell with an efficiency of 0.22%, obtained by chemical bath deposition. The window layer of several solar cells usually used CdS; however, the Cd element is very toxic to humans. Miyawaki et al.15 suggested that zinc sulfide (ZnS) is a good substitute for CdS. They prepared ZnS/ SnS cells by the photochemical and electrochemical deposition methods, but this type of cell has a very low short-circuit current density of 0.95 µm/cm2 and a low open-circuit voltage of 135 mV. In recent years, titanium dioxide (TiO2) as a wide band gap semiconductor for dye-sensitized solar cell has been developed extensively and advanced greatly. However, no study on SnS/TiO2 photovoltaic properties has been reported to the best of our knowledge. In this paper, SnS film was deposited on glass substrate by chemical bath deposition for the study of SnS properties, and SnS film was also deposited on a TiO2 layer to investigate the photovoltaic behavior. Experimental Section SnS films were deposited by chemical bath deposition. The CBD solution for preparing SnS thin films was constituted of 0.95 g of tin chloride (SnCl2) dissolved in 5 mL of acetone, 8 mL of 98% triethanolamine (TEA), 8 mL of 0.1 M thioacetamide (CH3CSNH2), 6 mL of 24% ammonia solution, and deionized water, making the volume to be 100 mL. The cleaned glass substrates were vertically placed into this alkaline solution. The SnS thin film deposition was carried out at 75 °C for 1 h. * To whom correspondence should be addressed. Tel: +65 65164632. E-mail:
[email protected].
To investigate the photovoltaic behavior, SnS was deposited on the TiO2/FTO/glass by chemical bath deposition. The photovoltaic device consisted of the SnS layer deposited on a porous TiO2 layer, the FTO (F-doped tin dioxide) conduction layer covered with the TiO2 porous layer as the front electrode, and a Pt layer as the back electrode. The single TiO2 porous layer was made from TiO2 paste (Solaronix) by the “doctorblade” method. The TiO2 device assembling was reported by Wang et al.16 The crystal structure of samples was characterized by using a Bruker X-ray diffractometer (XRD) with Cu KR radiation, λKR ) 0.15406 nm, step size ) 0.05°, and the time/step size ) 3 s/step. High-resolution transmission electron microscopy (HRTEM) images and selected area diffraction (SAED) patterns were recorded by using a JEM 2010F TEM. Sn and S contents were determined by two methods for double confirmation. One method was EDX (energy-dispersive spectroscopy). The other was ICP-OES (inductively coupled plasma-optical emission spectrometry, PerkinElmer Dual-view Optima DV5300) with an Elementar Vario Micro Cube. The optical properties were investigated by Varian Cary 5000 UV-vis-NIR spectroscopy. The light source used in the photoconduction measurement was the Newport 1000 W Oriel Solar Simulator. The I-V curve measurement and the critical parameter calculations, such as short-circuit current (Isc), short-circuit current density (Jsc), opencircuit voltage (Voc), fill factor (FF), cell efficiency (η), shortcircuit resistance (Rsc), and open-circuit resistance (Roc) were obtained by Newport’s new Oriel I-V test station with Keithly 2420. Results and Discussion Tin sulfide film and powder of a brown-black color were synthesized by CBD. The SnS crystal structure of the thin-film samples on glass substrate was reported in our previous paper.14 The CBD-prepared SnS film was similar to that electrodeposited by Mishra et al.8 and the film via evaporating a SnS source at Ts ) 275 °C by Devika et al.17 In this paper, the thin films obtained were similar to what we have reported. To find whether the film and the precipitates are both SnS, XRD on the powder precipitates was performed. The powder was obtained by
10.1021/jp9075756 2010 American Chemical Society Published on Web 02/04/2010
Photovoltaic Behavior of Nanocrystalline SnS/TiO2
J. Phys. Chem. C, Vol. 114, No. 7, 2010 3257
Figure 3. Plot of ln(R/R0) vs 1000/T for the SnS grown on glass to obtain the electrical conductivity activation energy.
Figure 1. XRD pattern of SnS powder obtained by chemical bath deposition. The insets a-c show the individual peaks after Gaussian peak fitting the three broad peaks at 2θ ) 26.7, 31.7, and 39.1°, respectively.
atomic ratio was found to be 1.06. These results indicated that SnS synthesized by CBD was close to stoichiometric with slight sulfur rich. Koteeswara Reddy et al.18 reported a nearly stichiometric SnS with a Sn/S atomic ratio of 1.01 by thermal evaporation technique in ultrahigh vacuum. Albers et al.19 reported that the pure SnS crystal showed p-type conduction and this conductivity is caused by an excess of sulfur. To find the electrical conductivity activation energy, the thermal electronic property of SnS is investigated. From the temperature-dependent resistance studies, the electrical conductivity activation energy Ea of SnS can be evaluated using the Arrhenius equation20
( )
σ ) σ0 exp Figure 2. (a) The selected area electron diffraction pattern of the SnS by TEM. (b) The TEM image of SnS nanoparticles.
centrifuging in the order of the precursor solution, the dionized water solution containing the precipitate, the ethanol solution containing the precipitate. Figure 1 shows the XRD pattern of the SnS powder that exhibits a few peaks, and the peaks are broad. The main broad peaks are at 2θ ) 26.7, 31.7, and 39.1°, which can be indexed to planes (120)/(021), (101)/(111)/(040), and (131)/(041) of herzenbergite SnS (JCPDS card no. 39-0354). From XRD data, some peaks appeared overlapping, which was from the small crystallites, which led to the broadening and overlapping of the nearby peaks. The two peaks of 2θ(120) ) 26.1° and 2θ(021) ) 27.0° overlapped and appeared to be the 26.7° broad peak; the three peaks of 2θ(101) ) 30.8°, 2θ(111) ) 31.7°, and 2θ(040) ) 32.1° to be the 31.7° broad peak; and the peaks of 2θ(131) ) 39.0° and 2θ(041) ) 39.3° to be the 39.1° broad peak. The individual peaks were distinguished after Gaussian fitting, as shown in insets a-c of Figure 1. Grain size D can generally be determined from the Sherrer equation, D ) (K · λ)/(β · cos θ), where K is the shape factor, λKR is the incidence X-ray wavelength (0.15406 nm), β is the fwhm (full width at half-maximum) in radians, and θ is the Bragg peak position. K is taken as 0.89, as this value is normally used for unknown materials. fwhm values for peaks of (120), (021), (101), (111), (040), (131), and (041) planes are 1.30, 1.30, 0.70, 0.70, 0.70, 1.32, and 1.32°, respectively. The grain sizes are then 6.2, 6.2, 11.6, 11.7, 11.7, 6.3, and 6.3 nm for these peaks, respectively. The electron diffraction pattern, Figure 2a, confirms the structure of SnS. The HRTEM image of SnS, Figure 2b, shows SnS nanocrystals of ∼6 nm. The grain size of SnS synthesized by our CBD was much smaller than the literature reported ∼180 nm of SnS obtained using an evaporation method at a substrate temperature of 275 °C by Devika et al.17 The composition of the sample was estimated by using EDX, and the S/Sn atomic ratio was 1.03. Because EDX is not a good tool for the characterization of precise chemical composition, another chemical composition analysis, ICP-OES with an Elementar Vario Micro Cube, was further carried out. The S/Sn
Ea kBT
(1)
where σ conductivity is given by
σ)
1 A 1 ) ) ·g F R·l R
(2)
R is bulk resistance and g is a geometric parameter that is assumed to be constant. Therefore, the Arrhenius equation can be expressed as
( )
ln
Ea R ) R0 kBT
(3)
Ea is the electrical conductivity activation energy, kB is the Boltzmann constant, 8.6 × 10-5 eV/K, and T is the temperature in K. The resistance of the SnS film on glass at different temperatures was measured by a multimeter. The resistance of the SnS film on glass below 110 °C is very large, more than 2 GΩ, which is beyond the limitation of the multimeter. When the temperature increased to 111 °C, the resistance of the SnS film can be measured. The resistance decreased with the increasing temperature. The plot of ln(R/R0) and 1000/T,according to eq 3, is shown in Figure 3, in the 1000/T range of 2.6-2.4 K-1, corresponding to the temperature range of 111-144 °C. The electrical conductivity activation energy is found to be 0.22 eV. Because Sn ionic radii21 of VI coordination is 0.69 Å and S ionic radii21 of VI coordination is 1.84 Å, it is easier to form Sn vacancies than S interstitial. In this paper, slightly S-rich SnS might have Sn vacancies rather than S interstitials. Therefore, the electrical conductivity activation energy of 0.22 eV is due to acceptor levels of Sn vacancies. This agrees with Ristov et al.22 and Devika et al.:17 they obtained SnS film by thermal evaporation, which had 0.23 and 0.28-0.34 eV electrical conductivity activation energies, respectively. Lopez et al.23 and Koteswara Reddy et al.24 reported electrical conductivity activation energies of 0.54 and 0.45 eV using spray
3258
J. Phys. Chem. C, Vol. 114, No. 7, 2010
Wang et al.
Figure 4. The variation of the absorption coefficient with photon energy of the SnS film, the inset shows typical plot of (Rhν)1/2 against photon energy.
pyrolysis. The Ea ) 0.22 eV is less than Eg/2, indicating an extrinsic semiconductor property of the SnS film. The conduction type of SnS was tested by the Seebeck method. The two ends of the SnS sample were connected to a multimeter. One end of the SnS was connected to the COM port and the other end to the mV-Ω port of the multimeter. The multimeter showed 0.00 mV. When the sample end connecting to the V-Ω port was heated, the multimeter showed a negative value. Furthermore, the voltage becomes more negative with a temperature increase. This indicated p-type conduction of the SnS sample. To further confirm SnS p-type, the end of the SnS sample connecting to the COM port of the multimeter was heated; the mulitmeter showed a positive voltage, and this positive voltage increases with temperature increase. This Seebeck experiment confirmed that the SnS was p-type. The absorption ability of SnS is important in photovoltaic applications, which can be evaluated by the absorption coefficient. The absorption coefficient (R) was estimated from the measured transmittance (Tλ), absorbance (Abs), and thickness (t), following the formula
Rλ )
ln(1/Tλ) Abs · ln 10 2.303 · Abs ) ) t(cm) t(cm) t(cm)
(4)
The absorbance (Abs) of the film was measured in the wavelength range of 200-1200 nm at room temperature. The absorption coefficient of SnS film is high, more than 104 cm-1 at 800 nm, as shown in Figure 4, which suggests that SnS absorbs visible-range sunlight significantly. The optical band gap of SnS can be estimated according to Tauc’s law
(Rhν)n ) A(Eg - hν)
(5)
where n ) 1/2 for an indirect transmission. The Tauc plot of 1 (Rhν) /2 and photon energy (hν) is shown in the inset in Figure 4. The band-gap energy was determined by extrapolating the straight line portion of the plot to intersect the (hν) axis to be 1.3 eV, which is close to the optimal band gap of 1.5 eV as absorber for the solar cell. However, Koktysh et al.25 reported an indirect energy band gap of 1.6 eV for sub-10 nm nanocrystalline particles and an indirect band gap of 1.06 eV for sub-200 nm SnS particles. The quantum confinement due to particle size was considered to result in an enlargement of the band gap. Avellaneda et al.12 reported an indirect bandgap of 1.12 eV for the orthorhombic SnS film and an indirect bandgap of 1.6 eV for zinc blende SnS.
Figure 5. Current density-voltage (J-V) plot of the SnS/TiO2 cell in the dark and under 100 mW/cm2 illumination. The inset by SEM shows the cross-sectional image of the structure based on SnS on TiO2.
The highest efficiency of 1.3% for SnS-based solar cells, a CdS/SnS heterostructure, ever reported was by Ramakrishna Reddy et al.2 However, CdS is not a good choice for solar cells due to the high toxicity of Cd. Therefore, it is necessary to fabricate Cd-free SnS-based solar cells. Ichimura et al.26 reported a ZnO/SnS heterostrucutre solar cell, which showed rectification properties and a low conversion efficiency 0.01%. In this paper, wide-gap TiO2, an abundant and nontoxic material, was used to combine with SnS. A heterostructure SnS/TiO2 was fabricated. For this structure (FTO/Pt + SnS/TiO2/FTO), one FTO substrate slide coated with the porous TiO2 layer served as the front electrode, the SnS layer was deposited on a porous TiO2 layer, and a Pt coating layer on the other FTO substrate slide was faced to the SnS layer and served as the back electrode. FTO/Pt + SnS/TiO2/FTO under illumination had no apparent photovoltaic properties, and the conversion efficiency was lower than 0.02%. This poor efficiency was possibly due to the poor conductivity of SnS and TiO2, which resulted in the poor rates in the separation and collection of the electron-hole pairs generated. To improve the conductivity, electrolyte was added, and the cell structure was FTO/Pt + electrolyte + SnS/TiO2/ FTO. Strong photovoltaic behavior was observed for this cell structure. Figure 5 presents the current density-voltage characteristic of the FTO/Pt + electrolyte + SnS/TiO2/FTO structure in the dark and under 1 sun (100 mW/cm2) light illumination. The active area measured from the film dimension was ∼0.52 cm2. The rectifying behavior appeared in both the dark and under illumination conditions. Under the illumination, the cell exhibited a high open-circuit voltage (Voc) of 471 mV, a short-circuit current density (Jsc) of 0.30 mA/cm2, a conversion efficiency (η) of 0.10%, and a fill factor (FF) of 0.71. Roc and Rsc, which were equal to dV/dI at I ) 0 and V ) 0 (see Figure 5), were 270 and 34 936 Ω, respectively. A small value of Roc and a big value of Rsc indicate a high working powder of the solar cell. In this work, Roc and Rsc were automatically calculated from the I-V by the system of Newport’s new Oriel I-V test station with Keithly 2420. The value of Voc was higher than 260 mV of the CdS/SnS photovoltaic structure reported by Avellaneda et al.27 and 405-460 mV of the TiO2/In2S3/CIS (CIS ) copper indium selenide) solar cell reported by O’Hayre et al.28 Although the conduction of this cell with electrolyte is still not so good, possibly due to high Roc, it is better than the structure without electrolyte. Some electrons and holes generated under illumination are separated and collected by the electrodes, and this test cell showed a conversion efficiency of 0.10%. The efficiency of 0.10% is close to 0.21% of the 9 nm TiO2/In2S3/CIS solar cell28 where expensive indium and selenium were used. To
Photovoltaic Behavior of Nanocrystalline SnS/TiO2 confirm that the SnS layer plays a role in the photovoltaic behavior for the cell structure of FTO/Pt + electrolyte + SnS/ TiO2/FTO, the FTO/Pt + electrolyte + TiO2/FTO structure, a cell structure without a SnS layer, was fabricated. Its active area was 0.46 cm2. Under illumination, there was a conversion efficiency (η) of 0.03%. With the presence of the SnS layer, higher conversion efficiency indicated better photovoltaic performance. It is better to admit that the efficiency of our SnS/ TiO2 cell is much lower than those of the well-developed solar cells, such as GaAs, CIGS (copper indium gallium selenide), and CdTe solar cells. However, this is the first demonstration of the photovoltaic behavior of the SnS/TiO2 structure, and there is much space for further improvement. Furthermore, SnS is nontoxic and cheap, and the processing cost we employed is low, which have been the targets of a new generation of photovoltaic materials. Conclusions Nanocrystalline SnS films and powders were deposited by using chemical bath deposition at 75 °C. The XRD pattern and SAED rings showed that the samples were polycrystalline herzenbergite SnS. The SnS showed slightly S-rich. The SnS obtained had p-type conduction and 0.22 eV of electrical conductivity activation energy. The band gap of SnS is 1.3 eV, and the absorption coefficient is high, 104 cm-1 at 800 nm, suggesting that SnS can absorb visible-range sunlight strongly. A photovoltaic structure based SnS/TiO2 exhibited photovoltaic behavior, which has an open-circuit voltage of 471 mV, a current density of 0.3 mA/cm2, and the conversion efficiency of 0.1% under 100 mW/cm2 illuminations. SnS is shown to be a promising absorber material. Acknowledgment. This work was supported by the MOE Tier 1 Research Fund (R-284-000-068-112 and R-284-000-071112). References and Notes (1) Koteeswara Reddy, N.; Ramakrishna Reddy, K. T. Mater. Chem. Phys. 2007, 102, 13.
J. Phys. Chem. C, Vol. 114, No. 7, 2010 3259 (2) Ramakrishna Reddy, K. T.; Koteswara Reddy, N.; Miles, R. W. Sol. Energy Mater. Sol. Cells 2006, 90, 3041. (3) Noguchi, H.; Setiyadi, A.; Tanamura, H.; Nagatomo, T.; Omoto, O. Sol. Energy Mater. Sol. Cells 1994, 35, 325. (4) El-Nahass, M. M.; Zeyada, H. M.; Aziz, M. S.; El-Ghamaz, N. A. Opt. Mater. 2002, 20, 159. (5) Tanusevski, A.; Poelman, D. Sol. Energy Mater. Sol. Cells 2003, 80, 297. (6) Devika, M.; Reddy, N. K.; Reddy, S.; Ahsanulhaq, Q.; Ramesh, K.; Gopal, E. S. R.; Gunasekhar, K. R.; Hahn, Y. B. J. Electrochem. Soc. 2008, 155, H130. (7) Kamoshida, S.; Suzuki, S. Japan Patent JP 08176814, A2, 1996. (8) Mishra, K.; Rajeshwar, K.; Weiss, A.; Murley, M.; Engelken, R. D.; Slayton, M. J. Electrochem. Soc. 1989, 136, 1915. (9) Brownson, J. R. S.; Georges, C.; Levy-Clement, C. Chem. Mater. 2006, 18, 6397. (10) Boonsalee, S.; Gudavarthy, R. V.; Bohannan, E. W.; Switzer, J. A. Chem. Mater. 2008, 20, 5737. (11) Gunasekaran, M.; Ichimura, M. Sol. Energy Mater. Sol. Cells 2007, 91, 774. (12) Avellaneda, D.; Delgado, G.; Nair, M. T. S.; Nair, P. K. J. Electrochem. Soc. 2008, 155, D517. (13) Nair, M. T. S.; Nair, P. K. Semicond. Sci. Technol. 1991, 6, 132. (14) Wang, Y.; Reddy, Y. B. K.; Gong, H. J. Electrochem. Soc. 2009, 156, H157. (15) Miyawaki, T.; Ichimura, M. Mater. Lett. 2007, 61, 4683. (16) Wang, Q.; Ito, S.; Gra¨tzel, M. J. Phys. Chem. B 2006, 110, 25210. (17) Devika, M.; Reddy, K. T. R.; Reddy, N. K.; Ramesh, K.; Ganesan, R.; Gopal, E. S. R.; Gunasekhar, K. R. J. Appl. Phys. 2006, 100, 023518. (18) Koteeswara Reddy, N.; Hahna, Y. B.; Devika, M.; Sumana, H. R.; Gunasekhar, K. R. J. Appl. Phys. 2007, 101, 093522. (19) Albers, W.; Haas, C.; Vink, H. J.; Wasscher, J. D. J. Appl. Phys. 1961, 32, 2220. (20) Hu, G.; Gong, H. Acta Mater. 2008, 56, 5066. (21) Pauling, L. The Nature of the Chemical Bond; Cornell University Press: Ithaca, NY, 1960. (22) Ristov, M.; Sinadinovski, G.; Grozdanov, I.; Mitreski, M. Thin Solid Films 1989, 173, 53. (23) Lopez, S.; Ortiz, A. Semicond. Sci. Technol. 1994, 9, 2130. (24) Koteswara Reddy, N.; Ramakrishna Reddy, K. T. Thin Solid Films 1998, 325, 4. (25) Koktysh, D. S.; McBride, J. R.; Rosenthal, S. J. Nanoscale Res. Lett. 2007, 2, 144. (26) Ichimura, M.; Takagi, H. Jpn. J. Appl. Phys. 2008, 47, 7845. (27) Avellaneda, D.; Nair, M. T. S.; Nair, P. K. Thin Solid Films 2009, 517, 2500. (28) O’Hayre, R.; Nanu, M.; Schoonman, J.; Goossens, A.; Wang, Q.; Gra¨tzel, M. AdV. Funct. Mater. 2006, 16, 1566.
JP9075756
