Solar Energy Materials & Solar Cells 136 (2015) 113–119 Contents lists available at ScienceDirect Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat Influence of sulfate residue on Cu2ZnSnS4 thin films prepared by direct solution method Viet Tuyen Nguyen a,c, Dahyun Nam a, Mungunshagai Gansukh a, Si-Nae Park b, Shi-Joon Sung b, Dae-Hwan Kim b, Jin-Kyu Kang b, Cong Doanh Sai c, Thi Ha Tran c,d, Hyeonsik Cheong a,n a Department of Physics, Sogang University, Seoul 121-742, Korea Advanced Convergence Research Center, Daegu Gyeongbuk Institute of Science and Technology, Daegu 711-873, Korea c Vietnam National University, College of Science, 334 Nguyen Trai, Hanoi, Vietnam d Hanoi University of Mining and Geology, Co Nhue, Hanoi, Vietnam b art ic l e i nf o a b s t r a c t Article history: Received 11 September 2014 Received in revised form 11 December 2014 Accepted January 2015 Raman scattering and atomic force microscopy measurements on Cu2ZnSnS4 thin films prepared by a direct solution method revealed that metal sulfates of various morphologies (dense clusters or separated particles) were partially embedded on the surface of the Cu2ZnSnS4 layer This residue was removed during the subsequent chemical bath deposition of the CdS buffer layer However, the removal of the residue led to poor crystallinity and reduced photocurrent near the location of the residue, which suggests that controlling the formation of the sulfates during the fabrication of the absorber layer would be critical for obtaining high efficiency solar cells by the solution method & 2015 Elsevier B.V All rights reserved Keywords: Copper–zinc–tin sulfide Thin films Raman spectroscopy Laser beam induced current Residue Sulfurization Introduction Recently, Cu2ZnSnS4 (CZTS) and Cu2ZnSn(S,Se)4 (CZTSSe) have emerged as a promising candidate to replace CuIn1À xGaxSe2 (CIGS) in photovoltaic applications As a p-type semiconductor, CZTS closely resembles CIGS, the leading material for absorber in thin film solar cells to date CZTS has many advantages for conversion of solar radiation, such as a near optimum direct band gap energy of Eg ¼(1.4$ 1.6) eV, a high absorption coefficient ($ 104 сm À 1), and p-type conductivity [1–3] Using only low cost, abundant elements instead of rare and expensive elements such as Ga and In, CZTS is obviously a good choice to reduce the cost and for sustainable development of the photovoltaics technology However, since the cost of absorber is only a part of the total cost of solar cell panels, a cheap and highly effective preparation process, other than vacuum methods, is needed in order to realize commercially viable CZTS solar cells It is encouraging that during the last several years, the efficiency of CZTS solar cells has significantly increased The record efficiency achieved for CZTS solar cells prepared by vacuum and nonvacuum methods are not much different Recently, a record efciency n Corresponding author Tel./fax: ỵ 82 717 8434 E-mail address: hcheong@sogang.ac.kr (H Cheong) http://dx.doi.org/10.1016/j.solmat.2015.01.003 0927-0248/& 2015 Elsevier B.V All rights reserved of 12.6% [4] for the glass/Mo/CZTSSе/CdS/ZnO/ITO/Ni-Al structure has been achieved by a solution-based process However, in order to achieve efficiency close to 20%, which is necessary for this absorber to become commercially viable, various factors that adversely affect the solar cell performance must be identified and controlled One common problem for solution based processes is the formation of residues in the absorber layer Carbon is the best studied one among detected residues Many researchers reported that carbon could be formed as layers resulting in serious problems and drop the efficiency of the cell because they not only increase the series resistance but could also give rise to low adhesion between the CZTS layer and the Mo substrate [5–7] This seems to be a serious problem for CZTS layers prepared from solution-based methods where most solvents, binders or stabilizers are organic and carbon-rich in nature Apart from carbon, special attention is necessary to other residues such as oxygen and sulfur in order to raise the quality of solution based absorber layer because they are also abundant in precursor materials However, unlike carbon, the effects of oxygen or sulfur have not been investigated thoroughly It is not even clear whether these residues are harmful or beneficial for the solar cell performances [8–15] In this article, we study the formation of sulfate (SO2– ) residues on CZTS thin films prepared by a solution method Understanding the transformation of such sulfate residues during 114 V.T Nguyen et al / Solar Energy Materials & Solar Cells 136 (2015) 113–119 a solution-based synthesis process is important because it could provide useful information to avoid the formation of this residue and hence give us a possibility to further improve the efficiency of CZTS solar cells Experiment 2.1 Samples Two sets of CZTS, CZTS/CdS and CZTS/CdS/ZnO/Al full cell samples were prepared by a solution method on soda-lime glass substrates coated with molybdenum of thickness 500 nm and 1000 nm Metal chlorides (CuCl2, ZnCl2, SnCl2) and thiourea were used as source materials These compounds were completely dissolved in a mixture of de-ionized water and ethanol to produce a precursor solution The concentrations of CuCl2, ZnCl2, SnCl2 and thiourea in the solution were 0.9, 0.7, 0.5, and M, respectively The precursor solution was subsequently spin-coated onto molybdenum-coated soda-lime glass substrates at 5000 rpm for 30 s The spin-coated CZTS precursor thin films were then baked on a hotplate at 350 1C for for preannealing The spin coating process was repeated several times to obtain the desired thickness The precursor films were then sulfurized in a two-zone tubular quartz furnace In order to maintain proper sulfur concentration, sulfur was evaporated at 300 1C, and the sample zone was heated to different temperatures in the range of 520 1C to 570 1C for 30 Argon carrier gas with 200 sccm at atmospheric pressure was used in the quartz furnace to facilitate the flow of sulfur vapor, and both heating zones were naturally cooled to room temperature The composition, examined by energy-dispersive spectroscopy (EDS), was copper-poor (Cu/ZnỵSn$ 0.7; Zn/Sn$ 1) More details on the preparation process can be found elsewhere [16] A CdS buffer layer ( $ 80 nm) was deposited onto CZTS films by chemical bath deposition (CBD) using CdS and thiourea solution A solution of CdSO4 and thiourea in a mixture of ammonia and deionized water was heated to 651C before dipping the absorber samples inside During the chemical bath process, the solution was stirred at a constant rate The CdS thickness was controlled by changing the deposition time A 80-nm CdS was obtained after 13 when preparing the solar cells To study the effect of the CBD process on the residues, we used a shorter deposition time of for the CZTS/CdS samples After depositing the buffer layer, RF sputtering was used to sequentially deposit a 50-nm-thick intrinsic ZnO layer and a 300-nm-thick Al-doped ZnO layer on top of the buffer layer in order to fabricate solar cells Finally, a 500-nm-thick Al collection grid was thermally evaporated on top of the device At each step of the fabrication process a set of samples were kept aside in order to monitor the effect of the processes Table summarizes the samples used in this study S#-1 refers to bare absorber films, S#-2 films with CdS buffer layers, and S#-3 full cells Table summarizes photovoltaic properties of the solar cells fabricated on 500-nm Mo-coated substrates, characterized by using a Keithley 2400 source meter unit and a solar simulator (Newport 69907) to simulate AM 1.5 solar irradiation Solar cells fabricated on 1000-nm Mo-coated substrates did not show decent photovoltaic performances and their data are not presented here 2.2 Measurements For secondary phase detection, macro-Raman measurements were carried out in the quasi-backscattering geometry by using several excitation wavelengths to take advantages of quasiresonance conditions as well as the different penetration depths of the lasers The excitation sources were the 632.8-nm line of a He– Ne laser, the 514.5-nm line of an Ar ion laser, and the 325.0-nm line of a He-Cd laser The penetration depth of the laser ranges from Table Sample information Sample name S1-1 S2-1 S3-1 S4-1 S5-1 S6-1 S7-1 S8-1 S1-2 S2-2 S3-2 S4-2 S5-2 S6-2 S7-2 S8-2 S5-3 S6-3 S7-3 S8-3 Mothickness (nm) Absorberthickness 520 1C 540 1C 560 1C 570 1C 520 1C 540 1C 560 1C 570 1C 520 1C 540 1C 560 1C 570 1C 520 1C 540 1C 560 1C 570 1C 520 1C 1000 1000 1000 1000 500 500 500 500 1000 1000 1000 1000 500 500 500 500 500 1.60 1.55 1.52 1.43 2.22 1.94 1.81 1.67 1.60 1.55 1.52 1.43 2.22 1.94 1.81 1.67 2.22 8 8 8 8 13 540 1C 500 1.94 13 560 1C 500 1.81 13 570 1C 500 1.67 13 Stack Sulfurization sequences temperature CZTS CZTS CZTS CZTS CZTS CZTS CZTS CZTS CZTS/CdS CZTS/CdS CZTS/CdS CZTS/CdS CZTS/CdS CZTS/CdS CZTS/CdS CZTS/CdS CZTS/CdS/ ZnO/Al CZTS/CdS/ ZnO/Al CZTS/CdS/ ZnO/Al CZTS/CdS/ ZnO/Al Duration of CBD (min) (μm) Table Photovoltaic properties of the solar cells with absorber layer sulfurized at different temperatures Sample name VOC (V) JSC (mA/cm2) FF (%) Efficiency (%) S5-3 S6-3 S7-3 S8-3 0.42 0.45 0.48 0.34 13.35 16.29 1.36 1.33 54.34 52.95 27.67 32.22 3.08 3.95 0.18 0.15 140 nm for the 488-nm line to 170 nm for the 633-nm line [17] The measurements were performed on several different parts of the samples in order to confirm that there is no macroscopic inhomogeneity The laser power was controlled by a tunable neutraldensity filter to $20 mW The excitation laser beam was linefocused with a cylindrical lens to an area of $ 100 μm  mm The scattered light was filtered by a pair of holographic edge filters and then dispersed by an iHR-550 or a TR-550 spectrometer (JY Horiba) Finally, the signals were detected with a liquid nitrogen cooled back-illuminated charge-coupled-device (CCD) detector array Micro-Raman imaging measurements were performed with a 50  microscope objective (0.8 N.A.) using the 632.8-nm line of a He–Ne laser as the excitation source The size of the laser spot on the sample was estimated to be approximately mm in diameter This excitation wavelength was chosen in order to avoid optical absorption in the CdS buffer layer (Eg ¼2.42 eV at room temperature) [18] and the power was kept at 0.5mW to prevent the sample from being damaged by laser heating Laser beam induced current (LBIC) measurements were performed on the same system by using the 632.8-nm laser as the excitation source The laser power for the LBIC measurements was nW ($ 1.3 kW/m2) in order to simulate the AM 1.5 condition (1 kW/m2) The photocurrent was measured in the ac mode by using a chopper to modulate the excitation light at a frequency of 410 Hz The images were obtained by raster-scanning the sample with a computer-controlled translation stages in 1-μm steps for imaging large areas and 0.5-μm steps for small areas The microRaman and AFM images were obtained in subsequent V.T Nguyen et al / Solar Energy Materials & Solar Cells 136 (2015) 113–119 measurements on the same sample, and from the same area All measurements were carried out in an ambient condition Results and discussion It is well known that the secondary phases in kesterite compounds are difficult to distinguish by the X-ray diffraction technique [17] Raman scattering with different excitation wavelengths to take advantages of the quasi-resonant conditions could help resolve this problem For CZTS, two most likely secondary phases are Cu2SnS3 115 (CTS) and ZnS Whereas the former has a small band gap (0.9 eV) and can be detected by using long-wavelength excitations such as 633 or 785 nm [19], the latter can be detected easily with ultraviolet excitation due to the resonance with the ZnS band gap (Eg ¼3.84 eV) with the 325-nm line (3.82 eV) of a He-Cd laser [20] Firstly, we took macro Raman spectra of the samples with the 514.5 nm excitation wavelength which is far from the resonant conditions of CTS and ZnS in order to study the influence of the processing conditions on the development of the main CZTS phase Fig 1(a) shows that at all sulfurization temperature, the main peaks of CZTS at 287 and 337 cm À and their second order Fig Macro Raman spectra of samples prepared on 500 nm-thick Mo substrates (S5-1 to S8-1) measured with different excitation wavelengths: (a) 514.5 nm, (b) 632.8 nm, and (c) 325 nm Fig (a) Optical microscope and (b) AFM images of sample S2-1; and (c) optical microscope and (d) AFM images of sample S6-1 The red (normal region) and blue (dark region) dots in (a) indicate the positions where the Raman spectra in Fig were taken The blue square in (d) represents the area for Raman imaging in Fig 4(b) 116 V.T Nguyen et al / Solar Energy Materials & Solar Cells 136 (2015) 113–119 peaks in the region from 600 to 750 cm À are detected clearly As the annealing temperature is increased, the intensity of the main CZTS peaks at 287 and 337 cm À (A1 peak) [21] increases prominently; indicating that the crystallinity of the samples further improved at higher sulfurization temperatures For samples S1-1 through S4-1, the same trend was observed (not shown here) From the Raman spectra taken with red laser as an excitation source (Fig 1b), it is noted that at higher sulfurization temperatures (560 and 570 1C), the signal due to the CTS secondary phase becomes clear at 264, 302, and 366 cm À [17] Even though the crystal quality increases at higher temperatures, the detrimental effects of secondary phases seem to be more significant and so the efficiency of the cell drops abruptly at temperatures higher than 540 1C Since sample S6-3, sulfurized at 540 1C, showed the highest efficiency, we will focus on the samples sulfurized at this temperature in the following Optical microscope images show that on all the samples there were a number of dark regions of different morphologies ranging from dense clusters of several tens of micrometers to particles of several micrometers Fig 2(a) and (c) show optical images of two typical morphologies of these dark regions on samples prepared at 540 1C The corresponding AFM images [Fig 2(b) and (d)] reveal that the dark regions are segregated as dense clusters with the height of several micrometers Fig shows the micro Raman spectra taken from the normal and dark regions of sample S2-1 sulfurized at 540 1C The Raman spectrum taken from the dark region is clearly different from the one taken from the normal region Composition analysis by EDS (not shown) revealed that the dark regions are rich in oxygen Oxygen could be included Fig Micro Raman spectra taken from the normal and dark regions on sample S21 as indicated by the red and blue dots in Fig 2(a); and Raman spectra of CuSO4 and ZnSO4 for reference into the CZTS absorber layer because some metal precursors could be hydrolyzed in water, which was used as solvent [15] Thermal gravimetric analysis studied by Madarászet al [22] and M Krunks et al [23] show that the reaction of metal chlorides and thiourea at high temperatures could lead to the formation of the sulfate group (SO2– ) In order to check this possibility, we took the Raman spectra of CuSO4 and ZnSO4 powders and compared them with that from the dark regions in our samples Fig clearly shows that many features of the spectrum taken from the dark region match with the characteristic peaks in the Raman spectra of CuSO4 and ZnSO4, and so it is reasonable to identify the residues on the surface of our samples as metal sulfates X-ray photoemission spectroscopy measurements also confirmed this interpretation In the EDS analysis, some carbon signals were detected But the fact that no carbon-related peaks were observed in the Raman spectra indicates that carbon impurities not aggregate in the film to form cluster-like residues No other impurities such as chlorine were observed Fig shows the images of the Raman intensity at 988 cm À 1, which show the distribution of metal sulfates on the surface of the samples S2-1 and S6-1 Note that the image areas match with corresponding optical and AFM images in Fig These images reconfirmed that the morphology of the dark regions match with the distribution of sulfates In order to study the effects of CBD, small pieces of the samples were cut and a thin CdS layer was deposited by CBD The CdS layer was thinner than what is usually used in full solar cell structures because we wanted to probe the impact of CBD without the interference from the CdS layer After CBD, the optical images in Fig 5(a) and (c) show that the dark regions still remain The AFM images in Fig 5(b) and (d) show that the remaining dark regions are pits with depths of several hundreds of nanometers, not clusters Micro Raman measurements were used again to investigate the pit regions formed after CBD of CdS In order to probe the CZTS under the CdS layer, we used the 632.8 nm-line of He-Ne laser for its energy is smaller than the band gap of CdS Fig shows that the characteristic peaks of sulfates disappeared Instead, the MoS2 signals are enhanced in the pit regions We interpret that metal sulfates were removed during CBD, which is justifiable because metal sulfates are polar compounds of ionic bonding and can be easily dissolved and washed away by polar solvent used in the CBD process This was further verified by dipping a sample in DI water: the residue was removed after water-dipping Because the sulfates are partially embedded in the CZTS layer, removal of the sulfates would leave pits in the CZTS film Since the CZTS layer is thinner in the pit regions, the laser can reach the MoS2 layer underneath This would explain the enhanced MoS2 signal from the pit regions Fig shows the MoS2 Raman intensity images of typical CZTS/CdS samples and confirm that the morphology of the pits matches with the MoS2 Raman signal image Fig Raman images of sulfates (a) sample S2-1and (b) sample S6-1 by tracing the intensity of the sulfate peak at 988 cm À The scan area of (b) is indicated by a small blue square in Fig 2(d) V.T Nguyen et al / Solar Energy Materials & Solar Cells 136 (2015) 113–119 117 Fig (a) Optical microscope and (b) AFM images of sample S2-2, and (c) optical microscope and (d) AFM images of sample S6-2 after depositing CdS by chemical bath process In (a), the red and blue dots indicate the positions where the Raman spectra in Fig were taken The blue square in (d) represents the area for Raman imaging in Fig 7(b) Fig Micro Raman spectra taken in normal and pit (dark) regions of sample S2-2 after CBD In Fig 6, the main CZTS peak at 337 cm À appears broader and redshifted in the spectrum taken from the pit region The sharpness of Raman peak could be used as a qualitative indicator for the crystallinity The peak position, on the other hand, reflects local stoichiometry or strain of the film Fig shows the images of the position and the width of the most intense A1 mode peak of CZTS after CBD The high contrast between the pit and normal regions clearly demonstrate the difference between the two regions It is possible that the CZTS film underneath sulfate clusters has different crystallinity, composition, and/or strain It is also possible that such difference was induced during the CBD process During the sulfurization process, the formation of sulfates may result in Cu or Zn vacancies in the CZTS film as the metal ions are taken by the sulfates High density of vacancies would degrade the crystallinity and result in broader Raman peaks The chemical composition would also be affected, resulting in relative shift of the Raman peaks Such effects have been observed in Raman scattering of disordered kesterite phase of CZTS and CZTSe [24–28] Also, local strain distribution could contribute to the redshift of the Raman peak As the film is thinner in the pit regions, the strain could be locally different from the normal region Fig To demonstrate the negative impact of the sulfate residues on the solar cell efficiency, we conducted an LBIC measurement on sample S6-3 The photocurrent in the dark region in the optical image is much smaller than in the normal region The anti-correlation between the MoS2 Raman intensity and the LBIC signal clearly demonstrates the detrimental effect of the sulfate residues even after being removed by the CBD process Pits formed by the removal of metal sulfates during CBD certainly should have some negative impact on the efficiency of the cell As the thickness of the absorber layer is reduced, such regions cannot absorb as many photons as normal regions If the CZTS layer is totally removed in some regions, the cell could be shunted as the front and back contacts are short-circuited Furthermore, the decrease in the crystallinity of CZTS may well contribute to a lower efficiency Conclusion Metal sulfates were found in CZTS films prepared by a solution-based method, regardless of the sulfurization temperature The sulfates were removed by the CBD process for the CdS buffer layer, but left pits in the CZTS films The CZTS film in the pit regions seems to have lower 118 V.T Nguyen et al / Solar Energy Materials & Solar Cells 136 (2015) 113–119 Fig Images of the integrated intensities of the MoS2 Raman signals from 400 to 460 cm À from (a) S2-2 and (b) S6-2 after CBD Fig Images of (a) position and (b) width of the CZTS A1 mode peak at 337 cm À from S2-2 Fig (a) Optical microscope, (b) AFM, (c) MoS2 Raman intensity, and (d) corresponding LBIC images of a CZTS solar cell (S6-3) crystallinity, different stoichiometry and/or local strain After a solar cell is fabricated, such pit regions give significantly low photocurrents, reducing the solar cell efficiency Therefore, preventing the formation of such residues in the CZTS films would be critical to achieve high efficiency CZTS solar cells by solution-based methods The origins of such sulfates and the mechanism for their formation are subjects of further studies Acknowledgments This work was supported by the New & Renewable Energy of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Korea government Ministry of Trade, Industry and Energy (No 20123010010130) V.T Nguyen et al / Solar Energy Materials & Solar Cells 136 (2015) 113–119 References [1] W.C Liu, B.L Guo, X.S Wu, F.M Zhang, C.L Mak, K.H Wong, Facile hydrothermal synthesis of hydrotropic Cu2ZnSnS4 nanocrystal quantum dots: bandgap engineering and phonon confinement effect, J Mater Chem A (2013) 3182–3186 [2] K Woo, Y Kim, J Moon, A non-toxic, solution-processed, earth abundant absorbing layer for thin-film solar cells, Energy Environ Sci (2012) 5340–5345 [3] S Chen, J.-H Yang, X.G Gong, A Walsh, S.-H Wei, Intrinsic point defects and complexes in the quaternary kesterite semiconductor Cu2ZnSnS4, Phys Rev B 81 (2010) 245204(1)–245204(10) [4] W Wang, M.T Winkler, O Gunawan, T Gokmen, T.K Todorov, Y Zhu, D.B Mitzi, Device characteristics of CZTSSe thin film solar cells with 12.6% efficiency, Adv, Energy Mater (2014) 1301465–1301470 [5] G.M Ilari, C.M Fella, C Ziegler, A.R Uhl, Y.E Romanyuk, A.N Tiwari, Cu2ZnSnSe4 solar cell absorbers spin-coated from amine-containing ether solutions, Sol Energy Mater Sol Cells 104 (2012) 125–130 [6] E.J Lee, S.J Park, J.W Cho, J.H Gwak, M.K Oh, B.K Min, Nearly carbon-free printable CIGS thin films for solar cell applications, Sol Energy Mater Sol Cells 95 (2011) 2928–2932 [7] S.J Ahn, C.W Kim, J.H Yun, J.H Gwak, S.H Jeong, B.H Ryu, K.H Yoon, CuInSe2 (CIS) Thin film solar cells by direct coating and selenization of solution precursors, J Phys Chem C 114 (2010) 8108–8113 [8] C Dun, N.A.W Holzwarth, Y Li, W Huang, D.L Carroll, Cu2ZnSnSxO4 À x and Cu2ZnSnSxSe4 À x: first principles simulations of optimal alloy configurations and their energies, J Appl Phys 115 (2014) 193513(1)–193513(12) [9] C Tablero, Effect of the oxygen isoelectronic substitution in Cu2ZnSnS4 and its photovoltaic application, Thin Solid Films 520 (2012) 5011–5013 [10] Z Su, K Sun, Z Han, H Cui, F Liu, Y Lai, J Li, X Hao, Y Liu, M.A Green, Fabrication of Cu2ZnSnS4 solar cells with 5.1% efficiency via thermal decomposition and reaction using a non-toxic sol–gel route, J Mater Chem A (2014) 500–509 [11] J Zhong, Z Xia, C Zhang, B Li, X Liu, Y.-B Cheng, J Tang, One-pot synthesis of self-stabilized aqueous nanoinks for Cu2ZnSn(S,Se)4 solar cells, Chem Mater 26 (2014) 3573–3578 [12] J.W Cho, A Ismail, S.J Park, W Kim, S Yoon, B.K Min, Synthesis of Cu2ZnSnS4 thin films by a precursor solution paste for thin film solar cell applications, ACS Appl Mater Interfaces (2013) 4162–4165 [13] T Washio, T Shinji, S Tajima, T Fukano, T Motohiro, K Jimbo, H Katagiri, 6% Efficiency Cu2ZnSnS4-based thin film solar cells using oxide precursors by open atmosphere type CVD, J Mater Chem 22 (2012) 4021–4025 [14] M Bär, B.-A Schubert, B Marsen, S Krause, S Pookpanratana, T Unold, L Weinhardt, C Heske, H.-W Schock, Native oxidation and Cu-poor surface structure of thin film Cu2ZnSnS4solar cell absorbers, Appl Phys Lett 99 (2011) 112103(1)–112103(4) [15] H Katagiri, K Jimbo, S Yamada, T Kamimura, W.S Maw, T Fukano, T Ito, T Motohiro, Enhanced conversion efficiencies of Cu2ZnSnS4-based thin film [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] 119 solar cells by using preferential etching technique, Appl Phys Express (2008) 041201(1)–041201(2) S.N Park, S.J Sung, D.H Son, D.H Kim, M Gansukh, H Cheong, J.K Kang, Solution-processed Cu2ZnSnS4 absorbers prepared by appropriate inclusion and removal of thiourea for thin film solar cells, RSC Adv (2014) 9118–9125 P.A Fernandes, P.M.P Salomé, A.F da Cunha, Study of polycrystalline Cu2ZnSnS4 films by Raman scattering, J Alloys Compd 509 (2011) 7600–7606 J.H Lee, W.C Song, J.S Yi, K.J Yang, W.D Hanc, J Hwang, Growth and properties of the Cd1 À xZnxS thin films for solar cell applications, Thin Solid Films 431–432 (2003) 349–353 Y.T Zhai, S Chen, J.H Yang, H.J Xiang, X.G Gong, A Walsh, J Kang, S.H Wei, Structural diversity and electronic properties of Cu2SnX3 (X¼ S, Se): a firstprinciples investigation, Phys Rev B 84 (2011) 075213(1)–075213(6) X Fontané, L Calvo-Barrio, V Izquierdo-Roca, E Saucedo, A Pérez-Rodriguez, J.R Morante, D.M Berg, P.J Dale, S Siebentritt, In-depth resolved Raman scattering analysis for the identification of secondary phases: characterization of Cu2ZnSnS4 layers for solar cell applications, Appl Phys Lett 98 (2011) 181905 D Dumcenco, Y.-S Huang, The vibrational properties study of kesterite Cu2ZnSnS4 single crystals by using polarization dependent Raman spectroscopy, Opt Mater 35 (2013) 419–425 J Madarász, P Bombicz, M Okuya, S Kaneko, Thermal decomposition of thiourea complexes of Cu(I), Zn(II), and Sn(II) chlorides as precursors for the spray pyrolysis deposition of sulfide thin films, Solid State Ionics 141–142 (2001) 439–446 M Krunks, T Leskelä, I Mutikainen, L Niinistö, A thermoanalytical study of copper(I) thiocarbamide compounds, J Therm Anal Calorim 56 (1999) 479–484 R Caballero, E Garcia-Llamas, J.M Merino, M León, I Babichuk, V Dzhagan, V Strelchuk, M Valakh, Non-stoichiometry effect and disorder in Cu2ZnSnS4 thin films obtained by flash evaporation: Raman scattering investigation, Acta Mater 65 (2014) 412–417 M.E Rodriguez, M Placidi, O Vigil-Galán, V Izquierdo-Roca, X Fontané, A Fairbrother, D Sylla, E Saucedo, A Pérez-Rodríguez, Compositional optimization of photovoltaic grade Cu2ZnSnS4 films grown by pneumatic spray pyrolysis, Thin Solid Films 535 (2013) 67–72 R Djemour, A Redinger, M Mousel, L Gütay, X Fontané, V Izquierdo-Roca, A Pérez-Rodríguez, S Siebentritt, The three A symmetry Raman modes of kesterite in Cu2ZnSnSe4, Op Express 21 (2013) A695–A703 M.Y Valakh, O.F Kolomys, S.S Ponomaryov, V.O Yukhymchuk, I.S Babichuk, V Izquierdo-Roca, E Saucedo, A Perez-Rodriguez, J.R Morante, S Schorr, I.V Bodnar, Raman scattering and disorder effect in Cu2ZnSnS4, Phys Status Solidi RRL (2013) 1–4 X Fontané, V Izquierdo-Roca, E Saucedo, S Schorr, V.O Yukhymchuk, M.Y Valakh, A Pérez-Rodríguez, J.R Morante, Vibrational properties of stannite and kesterite type compounds: Raman scattering analysis of Cu2(Fe, Zn)SnS4, J Alloy Comp 539 (2012) 190–194 ... was deposited onto CZTS films by chemical bath deposition (CBD) using CdS and thiourea solution A solution of CdSO4 and thiourea in a mixture of ammonia and deionized water was heated to 651C before... solution The concentrations of CuCl2, ZnCl2, SnCl2 and thiourea in the solution were 0.9, 0.7, 0.5, and M, respectively The precursor solution was subsequently spin-coated onto molybdenum-coated... quasi-backscattering geometry by using several excitation wavelengths to take advantages of quasiresonance conditions as well as the different penetration depths of the lasers The excitation sources