Formation mechanism of Cu2ZnSnSe4 absorber layers during selenization of solution deposited metal precursors

Journal of Alloys and Compounds 567 (2013) 102–106

Contents lists available at SciVerse ScienceDirect

Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom

Formation mechanism of Cu2ZnSnSe4 absorber layers during selenization of solution deposited metal precursors Carolin M. Fella a,⇑, Alexander R. Uhl a, Ceri Hammond b, Ive Hermans b, Yaroslav E. Romanyuk a, Ayodhya N. Tiwari a a b

Empa – Swiss Federal Laboratories for Materials Science and Technology, Laboratory for Thin Films and Photovoltaics, Überlandstrasse 129, 8600 Dübendorf, Switzerland ETH Zürich, Department of Chemistry and Applied Biosciences, Wolfgang-Pauli-Str. 10, 8093 Zürich, Switzerland

a r t i c l e

i n f o

Article history: Received 10 January 2013 Received in revised form 27 February 2013 Accepted 2 March 2013 Available online 20 March 2013 Keywords: Cu2ZnSnSe4 Kesterite Thin film solar cell Formation mechanism Selenization Secondary phase

a b s t r a c t Phase-pure Cu2ZnSnSe4 (CZTSe) layers are necessary for achieving efficient thin film solar cells. This requires the knowledge of intermediate phases and their existence regions during the evolution of the CZTSe phase within its homogeneity range. Here we investigate the growth mechanism of different phases when solution deposited metal salt precursors are selenized into CZTSe layers. A combination of in situ and ex situ X-ray diffraction, Raman spectroscopy, energy dispersive X-ray spectroscopy (EDX) and scanning electron microscopy at successively increasing substrate temperatures is used to track evolving crystal phases. The growth starts with the fast formation of binary Cu–Se phases that are present between 190 °C and 320 °C. Overlapping diffraction patterns of CZTSe/Cu2SnSe3/ZnSe phases evolve from 280 °C onwards and remain until a final temperature of 550 °C. The ternary Cu2SnSe3 phase co-existing with CZTSe between 340 °C and 370 °C is confirmed by Raman spectroscopy and point EDX measurements. No individual zinc or tin binary phases can be detected. We propose the kinetically driven formation mechanism, which starts with the selenization of Cu requiring the lowest activation energy for reaction, and then proceeds via the gradual incorporation of Sn and Zn to yield the final CZTSe phase. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction The family of kesterite compounds, including Cu2ZnSnS4 (CZTS), Cu2ZnSnSe4 (CZTSe) and the sulfo-selenide Cu2ZnSn(S,Se)4 (CZTSSe), draws increased attention for the production of low cost thin film solar cells since they mainly consist of earth abundant or readily available elements. Kesterite can be derived from the wellknown chalcopyrite absorber material Cu(In,Ga)Se2 by substituting In and Ga (both group III elements) by Zn (group II) and Sn (group IV) atoms [1]. The currently highest conversion efficiency of 11.1% has been reported for a CZTSSe solar cell absorber prepared by non-vacuum deposition of a hydrazine based solution-particle ink [2], whereas co-evaporated pure CZTSe based cells yield up to 9.15% [3]. Other deposition approaches with notable conversion efficiencies include: co-sputtering of Cux(S,Se)y, Znx(S,Se)y, and Snx(S,Se)y followed by high temperature annealing in SnS and S2 gas resulting in 9.3% efficiency [4], electroplating metal stacks converted into CZTS by high temperature sulfurization with 7.3% efficiency [5], and sintering CZTS and CZTGeS nanocrystal films in Se vapor yielding 7.2% and 8.4% efficiencies for CZTSSe and CZTGeSSe solar cells, ⇑ Corresponding author. Tel.: +41 58 765 61 01; fax: +41 58 765 40 42. E-mail address: [email protected] (C.M. Fella). 0925-8388/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jallcom.2013.03.056

respectively [6,7]. One of the simplest non-vacuum methods is the so-called sol–gel or solution technique, in which precursor solutions of metal salts are knife-coated or spin-coated and then annealed in elemental selenium vapor, resulting in a conversion efficiency of up to 4.3% [8]. Almost all of the aforementioned deposition techniques involve two steps. First, metal containing precursors with or without chalcogen addition are deposited, and second, the precursors are annealed under optimized conditions in a sulfur/selenium containing atmosphere to yield the final kesterite phase. In this respect, the knowledge of the formation and conversion reactions that occur during the annealing step is essential for obtaining phase-pure CZTS(Se) absorbers with targeted metal ratios. The reaction paths during selenization depend on annealing conditions and on the initial precursor, which can be either a stack of Cu, Zn, and Sn-containing layers or a homogeneous mixture of all metals with or without chalcogen elements. In the case of selenization of sequentially evaporated Sn–Zn–Cu metal stacks, the preferential formation of different binary copper selenides was observed, which were converted into a mixture of CZTSe and ZnSe phases above 370 °C [9]. For mixed hydrazine based precursor inks with ZnS(e)(N2H4) nanocrystals and a bimetallic Cu2Sn(Se,S)x framework no binary Cu or Sn chalcogenides were detected during the formation of CZTSSe absorbers [10]. For co-electroplated

C.M. Fella et al. / Journal of Alloys and Compounds 567 (2013) 102–106

Cu–Zn–Sn precursors two different reaction pathways via binary sulfides were observed for Cu-poor and Cu-rich precursors during their sulfurization; both leading to the ternary Cu2SnS3 (CTS) phase that reacted later with ZnS to form CZTS [11]. For CZTS thin films grown by a sol–gel sulfurization of spin-coated metal salt precursors, CuxS phase was formed first and then it got converted by further sulfurization at more than 450 °C [12]. These studies indicate that the order in which individual metals react with chalcogen does not depend only on the metal sequence in the initial stack, but strongly depends on relative reactivity of each metal, and Cu exhibits typically higher reaction rates than Sn and Zn. Although the influence of remaining secondary phases, like Cu2SnSe3(CTSe), ZnSe or MoSe2, on the performance of kesterite absorber based solar cells is widely debated, most studies agree that the secondary phases should be avoided [13,14]. The CTSe phase is the only ternary compound in the Cu2Se–SnSe2 system [15,16]. Though CTSe is a p-type semiconductor with an optical band gap of 0.84 eV and a high absorption coefficient a  104 cm1 [17], its co-existence with CZTSe may lead to a reduced open circuit voltage in the finished solar cell device. Presence of ZnSe (Eg = 2.8 eV [18]) is probably the origin of high series resistance in CZTSSe cells [19]. The presence of its analog ZnS at the CZTS/ MoSx interface is nevertheless not an obstacle for achieving an 8.4%-efficient CZTS cell [20]. Binary copper phases, Cu–Se, can promote the grain growth during the CZTSe evolution but typically result in shunted devices because of their high conductivity [21]. This study presents a detailed insight into the phase formation processes of CZTSe during the selenization of solution-deposited metal salt precursors. Several complementary characterization techniques are employed to identify intermediate phases. The proposed formation mechanism is discussed taking into account thermodynamic and kinetic considerations.

2. Experimental Non-vacuum deposition of copper (II) nitrate hemipentahydrate, zinc (II) nitrate hexahydrate and tin (IV) chloride hydrate with a Cu-poor (Cu/(Zn + Sn) = 0.65) and Zn-rich (Zn/Sn = 1.48) composition dissolved in ethanol and 1,2-propanediol was done by knife-coating on Mo coated soda lime glass substrates. Precursor solution rheology was adjusted by adding ethyl cellulose dissolved in 1-pentanol to the previously described metal-containing solution. The precursor films were dried at a maximum temperature of 230 °C under a lamp and on a hot plate in air. Selenization was performed in a two-zone tube furnace, unless otherwise stated. Further details of the process are described elsewhere [8]. In situ X-ray diffraction (XRD) was carried out in detector scan mode with fixed incident angle of 10° with a PANalytical X’Pert Pro MPD system. Substrate heating at 20 °C/min was applied from room temperature to 550 °C. XRD patterns were recorded every 10 °C and took about 13 min. The sample was held isothermally at 550 °C for 60 min, and subsequently cooled naturally, whereupon XRD patterns were collected every 50 °C. Selenization without excessive metal losses was ensured by a Se top capping layer on the precursor, and additional surrounding Se pellets placed inside a graphite dome. The graphite dome led, however, to carbon reflections in the diffractograms (PDF-number 26-1079). It has to be noted that this experimental configuration was not identical to our standard selenization set-up because the temperature was kept constant during each intermediate XRD scan, and therefore the selenization duration was much longer. Ex situ experiments were performed in a constantly pumped two-zone tube furnace, in which the selenium and substrate zone temperatures were controlled independently. Elemental selenium vapors were transported by nitrogen carrier gas into the substrate zone. The standard heating program with a ramp of 30 °C/ min was interrupted at various substrate temperatures in order to obtain individual samples, which were naturally cooled. This procedure is similar to selenization conditions for processing solar cells [8]. Obtained layers were characterized by ex situ XRD on a Siemens D5000 diffractometer with incomplete Ni-filtered Cu Ka radiation, Raman spectroscopy using HeNe (633 nm) and Ar ion (514 nm) excitation sources with spot size of about 30 lm in backscatter configuration connected to a CCD camera (both Renishaw, Ramascope 2000 and inVia, respectively). Additionally, scanning electron microscopy (SEM) equipped with an energy dispersive X-ray (EDX) analyzer for compositional point measurements at low acceleration voltage of 7 kV allowed morphology inspections and point compositional measurements (done on a FEI Nova NanoSEM230).

103

3. Results and discussion 3.1. In situ XRD The evolution of various crystal phases as a function of annealing temperature was first investigated through in situ XRD, as presented in Fig. 1. The most intense Bragg reflection at 2h  26.5° stems from the graphite dome. The identified crystal phases are listed on the right Y-axis. At low temperatures of 100–220 °C peaks of crystalline selenium are visible. Above 220 °C selenium is molten, and Cu converts into the orthorhombic CuSe2 phase (denoted in Fig. 1 by °) [PDFnumber 01-071-0046], which persists until 300 °C. In parallel, at 280 °C the overlapping reflections of CZTSe/CTSe/ZnSe [PDFnumbers: 01-070-8930/01-072-8034/37-1463] evolve at 2h  27.2°, 45.1° and 53.6°. Because of their very similar XRD patterns, it is not possible to distinguish between CZTSe, CTSe and ZnSe phases, and therefore these three phases are denoted with () in Fig. 1. The evidence for the tetragonal CZTSe and/or monoclinic CTSe phases is the weak reflection at 2h  17.3°, which becomes visible at 310 °C (indicated with an arrow in Fig. 1). The definitive proof for the presence of CZTSe appears at 370 °C, when (2 1 1) and (1 1 4) reflections evolve at 2h  36.2° and  38.9°, respectively. These reflections are absent for ternary CTSe and binary ZnSe phases. The formation of MoSe2 starts above 420 °C. The sample is held for 60 min at 550 °C and cooled down afterwards without any compositional changes (not shown). 3.2. Ex situ XRD The precursor layer exhibits only cubic Cu peaks (PDF-number 01-089-2838) while Zn and Sn are embedded in an amorphous carbon matrix in the form of organo-metallic complexes, see Fig. 2. The possibility of copper segregations in solution-derived precursor layers for Cu(In,Ga)Se2 was reported previously by Uhl et al. [22]. Ex situ XRD measurements show preferential selenization of copper, confirming the in situ XRD results. At low substrate temperatures of 230–320 °C, ex situ XRD mainly shows a hexagonal CuSe phase (PDF-number 01-072-8417) and some CuSe2 phase observed only at Tsub = 320 °C. The narrow existence range for the selenium-rich CuSe2 phase can be explained by the lower Se supply in the ex situ experiment as compared to the in situ XRD, where a large excess of selenium is ensured. It is worth mentioning that Cu2xSe with expected XRD peaks at 2h  44.5° and 26.8° (PDFnumber 01-071-0044) is not seen. Cu2ZnSnSe4/Cu2SnSe3/ZnSe phases (summarized by ) evolve from 340 °C on. 3.3. Ex situ Raman More conclusive information on phases can be obtained from Raman measurements shown in Fig. 3. Only results for the excitation wavelength of 633 nm are shown, because the measurements at excitation wavelength of 514 nm provide identical information. Raman spectroscopy is sensitive to the local chemical environment of the atoms in a crystal, meaning that slight compositional changes can cause Raman peaks to shift in frequency [23]. The probing depth of Raman spectroscopy depends on the absorption coefficient (a) of the material for the respective wavelength of the excitation source, and is roughly given by d  1/a (or d  1/(2a) in the case of backscattering configuration). For CZTSe with an absorption coefficient >104 cm1, this results in a probing depth of around 100 nm. Whilst the observed Raman peaks are in agreement with the phases detected by ex situ XRD, Raman analysis of the same samples allows CZTSe and CTSe to be distinguished (alongside several elemental EDX point analyses, see Section 3.4). For a low annealing

C.M. Fella et al. / Journal of Alloys and Compounds 567 (2013) 102–106

60

Se Se

55

45 40 35 30

CuCl

25

Se

Se

20

Cu2ZnSnSe4, Cu2SnSe3, ZnSe = * Mo MoSe2

° °° ° ° °° ° ° (114) ° (211) °° °°

Se Se Se Se

50

2 Theta (°)

CuSe2

Se

*

400 500 600 700

* Carbon

800

Mo CZTSe CZTSe

900 1000

MoSe2

* Carbon

°

(101,111)

15

300

Intensity (arb. units)

104

CZTSe, CTSe MoSe2

10

100

200

300

Tsubstrate (°C)

400

500

Fig. 1. In situ XRD of a Cu–Zn–Sn-precursor with Se top layer. Selenium melts at 220 °C where it starts first to react with Cu to form the CuSe2 phase, which persists until 300 °C. At around 280 °C, the first peaks of Cu2ZnSnSe4/Cu2SnSe3/ZnSe evolve (labeled by  on the right Y-axis). Peaks of CZTSe/CTSe at 2h = 17.3° and two weak peaks belonging only to CZTSe phase are visible at 36.2° and 38.9°, indicated by arrows.

20

25

30

35

40

45

50

55

60

2 theta (°) Fig. 2. Ex situ XRD of metal precursor layers after interrupted selenization experiments. CuSe and CuSe2 phases are observed before the growth of Cu2ZnSnSe4/Cu2SnSe3/ZnSe starts at 340 °C.

temperature of 230 °C, the dominating Raman peak at 262 cm1 can be attributed to binary Cu–Se i.e. CuSe, due to a Se–Se stretching mode [24] and Cu2Se [25]. Considering also the XRD results it can be attributed to the CuSe phase. Additionally, a peak at 242 cm1, belonging to elemental Se, is identified [26]. At Tsub = 320 °C, a multiphase structure is observed. At this temperature, the CuSe intensity is seen to decrease, whilst a broad peak at around 193 cm1, and a second weak peak at 233 cm1 evolve. It is well-known that the major peak in the CZTSe Raman spectra at 194 cm1 is the A1 vibrational mode, which arises from Se vibrations [27,28]. Thus, the broad peak at 193 cm1 is the first indication of CZTSe crystallization, where the peak broadening may arise from a high defect density and local compositional fluctuations. The local disorder in the cation sub-lattice could also be a reason for the peak shoulder [23]. Similar Raman spectra were also reported by Volobujeva et al. [9]. The weak Raman mode at 233 cm1 most likely belongs to Cu2ZnSnSe4 [28]. A temperature increase >320 °C clearly shows the A1 vibrational mode of CZTSe. A broad shoulder at 180 cm1 could indicate the ternary CTSe phase in the region Tsub = 320–

Intensity (arb. units)

Intensity (arb. units)

Mo Kβ

Intensity (arb. units)

Tsub (°C)

510 460 420 370 340 320 300 275 230 180 prec. Cu CuSe CuSe2 ZnSe Cu2SnSe3 Cu2ZnSnSe4

15

194 *

194 *

Mo

172

*

233

* Cu2ZnSnSe4 172

*

233

*

180

*

150

231

420 °C 370 °C 340 °C 300

200 250 Raman shift (cm-1) 510 °C

460 °C 420 °C 370 °C 340 °C 320 °C 230 °C

CZTSe CTSe CuSe Se ZnSe

100

150

200

250

300

350

400

Raman shift (cm-1) Fig. 3. Raman spectra (k = 633 nm) for samples selenized at temperatures of 230– 510 °C. Between about 340–370 °C, there is a co-existence of CZTSe and CTSe (highlighted in the inset). Upon reaching substrate temperatures P420 °C, all three CZTSe Raman modes at 172, 194 and 233 cm1 are visible.

370 °C, however, the peak is hidden between the CZTSe main modes at 194 and 172 cm1. An additional proof for ternary CTSe are numerous point EDX measurements as described in the next section, pointing out a very low Zn content in the crystalline material of the selenized surface layer. The inset of Fig. 3 highlights the transition from mixed CZTSe and CTSe to an apparently more phase pure CZTSe material. At higher substrate temperatures, P420 °C the shoulder at 180 cm1 significantly decreases, resulting in the typical footprint of kesterite (i.e. Raman peaks at 172 cm1, 194 cm1 and 233 cm1). Raman spectra at k = 633 nm do not show the ZnSe phase characterized by a main Raman mode at 250 cm1 [29]. Measurements with a shorter excitation wavelength of 514 nm were performed to improve a potential response from ZnSe (not shown), but again, no ZnSe could be found. We cannot exclude that even shorter excitation wavelength would be needed to detect small amounts of ZnSe because of its high band gap of 2.8 eV [18], or ZnSe is simply segregated towards the Mo contact away from the layer surface.

105

C.M. Fella et al. / Journal of Alloys and Compounds 567 (2013) 102–106

Fig. 4. SEM images of samples obtained after interrupting the selenization process at Tsub = 320 °C (a), 340 °C (b), 420 °C (c) and 510 °C (d) with EDX point measurements demonstrating the initially evolving CuSe phase and subsequent incorporation of Sn and later Zn.

3.4. Ex situ SEM and EDX In and ex situ XRD data are supported by direct quantitative elemental analysis from point EDX measurements (see Fig. 4). The SEM top view images show the samples obtained after interrupting the selenization process at Tsub = 320 °C (a), 340 °C (b), 420 °C (c) and 510 °C (d). Hexagonal crystals assigned to the klockmannite CuSe phase are visible for Tsub = 320 °C. The Zn-poor composition close to the CTSe phase is measured for Tsub = 340 °C. Main incorporation of Zn occurs at Tsub = 420 °C although there is still not enough Se available for the full conversion into Cu2ZnSnSe4. Small crystallites with Zn-rich composition are detected at Tsub = 510 °C (Fig. 4d). Table 1 summarizes XRD, Raman and SEM/EDX results. Depending on the Se vapor pressure, CuSe2 or CuSe phases are formed first and persist until 320 °C. There is a co-existence of Cu2SnSe3 and Cu2ZnSnSe4 between 340 and 370 °C, and the main CZTSe phase is dominant above 420 °C. The temperature offset between in and ex situ XRD for the evolution of different phases can be explained by the different experimental setups, so that crystal phases for in situ XRD are observed at somewhat lower temperatures than

Table 1 Evolving phases with respective characterization method and temperature range. Method

Phases

Temperature (°C)

in situ XRD

CuCl Se CuSe2 ZnSe/CZTSe/CTSe CZTSe (1 0 1)/CTSe (11–1) CZTSe (2 1 1) and (1 1 4)

RT-210 RT-220 190–300 280–550 310–550 370–550

ex situ XRD

Cu CuSe CuSe2 ZnSe/CZTSe/CTSe CZTSe (1 0 1)/CTSe (11–1) CZTSe (2 1 1)

RT-280 230–320 320 340–510 420 420

Raman 633 nm

Se CuSe CTSe CZTSe

230 230–320 320–370 320–510

SEM & EDX at low kV

CuSe CTSe CZTSe ZnSe?

320 340 420 510

in ex situ experiments where the selenization is interrupted 2 min after reaching the desired substrate temperature. It is worth mentioning that a further increase of the substrate temperature can lead to the decomposition of CZTSe into Cu2xSe and ZnSe via the loss of Se and SnSe, especially if no excess of Se and/or Sn is provided [30,31]. The phase evolution during annealing in elemental Se can be written in the following sequence: 190  C

340  C

ðCu; Zn; SnÞ ƒƒƒƒ! CuSex ƒƒƒƒ! Cu2 SnSe3 þ Cu2 ZnSnSe4 precursor



420  C

! Cu2 ZnSnSe4

The reaction pathway is clearly governed by kinetic factors, such as the fast reaction of copper with selenium at low temperatures, due to a low activation energy for the initial reaction and a high diffusion of copper atoms from the precursor layer. The standard enthalpy of formation DfH0298K for binary metal selenides increases as follows: MoSe2 (196 kJ/mol) < ZnSe (159 kJ/mol) < SnSe2 (121.3 kJ/mol) < SnSe (88.7 kJ/mol) < Cu2Se (65.3 kJ/ mol) < CuSe2 (48.1 kJ/mol) < CuSe (41.8 kJ/mol) [32]. This row presents the same trend as the more rigorous approach with full free energy of binary selenides calculated at 550 °C [33] and clearly indicates that binaries like ZnSe and MoSe2 are more thermodynamically stable than any copper selenides. Nevertheless, the kinetically favored Cu–Se phases initiate the later formation of quaternary Cu2ZnSnSe4 during annealing of homogeneous solution-deposited, metal containing precursor layers in elemental Se vapor. 4. Conclusions We have shown that the evolution of the CZTSe phase during the selenization of metal salt precursors starts with the fast formation of copper selenide phases and the subsequent incorporation of Sn and finally Zn. The CZTSe phase appears above 280 °C, and it coexists with the ternary CTSe. Substrate temperatures higher than 420 °C should be applied in order to maximize the conversion into the main CZTSe phase. Acknowledgments The authors thank S. Yoon for providing and helping to plan the in situ XRD measurements. Y.E.R. acknowledges the support of the Swiss National Science Foundation, Project Nr. PZ00P2_126435/1.

106

C.M. Fella et al. / Journal of Alloys and Compounds 567 (2013) 102–106

References [1] S. Schorr, Sol. Energy Mater. Sol. Cells 95 (2011) 1482–1488. [2] T.K. Todorov, J. Tang, S. Bag, O. Gunawan, T. Gokmen, Y. Zhu, D.B. Mitzi, Adv. Energy Mater. 3 (2013) 34–38. [3] I. Repins, C. Beall, N. Vora, C. DeHart, D. Kuciauskas, P. Dippo, B. To, J. Mann, W.C. Hsu, A. Goodrich, R. Noufi, Sol. Energy Mater. Sol. Cells 101 (2012) 154– 159. [4] V. Chawla, B. Clemens, 38th IEEE Photovoltaic Specialists Conference (PVSC38), Austin, Texas, 2012. [5] S. Ahmed, K.B. Reuter, O. Gunawan, L. Guo, L.T. Romankiw, H. Deligianni, Adv. Energy Mater. 2 (2012) 253–259. [6] Q. Guo, G.M. Ford, W.-C. Yang, B.C. Walker, E.A. Stach, H.W. Hillhouse, R. Agrawal, J. Am. Chem. Soc. 132 (2010) 17384–17386. [7] Q. Guo, G.M. Ford, H.W. Hillhouse, R. Agrawal, in: Proc. IEEE Photovoltaic Specialists Conference 37 (2011) 003522–003526. [8] C.M. Fella, A.R. Uhl, Y.E. Romanyuk, A.N. Tiwari, Phys. Status Solidi A 209 (2012) 1043–1048. [9] O. Volobujeva, J. Raudoja, E. Mellikov, M. Grossberg, S. Bereznev, R. Traksmaa, J. Phys. Chem. Solids 70 (2009) 567–570. [10] W.-C. Hsu, B. Bob, W. Yang, C.-H. Chung, Y. Yang, Energy Environ. Sci. 5 (2012) 8564–8571. [11] R. Schurr, A. Hölzing, S. Jost, R. Hock, T. Voss, J. Schulze, A. Kirbs, A. Ennaoui, M. Lux-Steiner, A. Weber, I. Kötschau, H.-W. Schock, Thin Solid Films 517 (2009) 2465–2468. [12] K. Maeda, K. Tanaka, Y. Nakano, H. Uchiki, Jpn. J. Appl. Phys. 50 (2011) 05FB081–05FB084. [13] D.B. Mitzi, O. Gunawan, T.K. Todorov, K. Wang, S. Guha, Sol. Energy Mater. Sol. Cells 95 (2011) 1421–1436. [14] J. Just, D. Lützenkirchen-Hecht, R. Frahm, S. Schorr, T. Unold, Appl. Phys. Lett 99 (1053) (2011) 2621051–2621262. [15] I.D. Olekseyuk, I.V. Dudchak, L.V. Piskach, Phys. Chem. Solid State 1 (2001) 195–200. [16] I.V. Dudchak, L.V. Piskach, J. Alloys Comp. 351 (2003) 145–150.

[17] G. Marcano, C. Rincón, L.M. de Chalbaud, D.B. Bracho, G. Sánchez Pérez, J. Appl. Phys. 90 (2001) 1847–1853. [18] K. Shahzad, D.J. Olego, C.G. Van de Walle, Phys. Rev. B 38 (1988) 1417– 1426. [19] A. Redinger, M. Mousel, M.H. Wolter, N. Valle, S. Siebentritt, Thin Solid Films (2013), http://dx.doi.org/10.1016/j.tsf.2012.11.111. [20] B. Shin, O. Gunawan, Y. Zhu, N.A. Bojarczuk, S.J. Chey, S. Guha, Prog. Photovolt: Res. Appl. 21 (2013) 72–76. [21] I. Repins, N. Vora, C. Beall, S.-H. Wei, Y. Yan, M. Romero, G. Teeter, H. Du, B. To, M. Young, R. Noufi, Mater. Res. Soc. Symp. Proc. 1324 (2011) 97–108. [22] A.R. Uhl, C. Fella, A. Chirila, M.R. Kaelin, L. Karvonen, A. Weidenkaff, C.N. Borca, D. Grolimund, Y.E. Romanyuk, A.N. Tiwari, Prog. Photovolt.: Res. Appl. 20 (2012) 526–533. [23] J. Álvarez-García, V. Izquierdo-Roca, A. Pérez-Rodríguez, Raman spectroscopy on thin films for solar cells, in: D. Abou-Ras, T. Kirchartz, U. Rau (Eds.), Advanced Characterization Techniques for Thin Film Solar Cells, Wiley-VCH, Weinheim, Germany, 2011, pp. 365–386. [24] M. Ishii, K. Shibata, H. Nozaki, J. Solid State Chem. 105 (1993) 504–511. [25] B. Minceva-Sukarova, M. Najdoski, I. Grozdanov, C.J. Chunnilall, J. Mol. Struct. 410 (1997) 267–270. [26] P. Nagels, E. Sleeckx, R. Callaerts, L. Tichy, Solid State Commun. 94 (1995) 49– 52. [27] M. Himmrich, H. Haeuseler, Spectrochim. Acta 47A (1991) 933–942. [28] M. Ganchev, L. Kaupmees, J. Iliyna, J. Raudoja, O. Volobujeva, H. Dikov, M. Altosaar, E. Mellikov, T. Varema, Energy Procedia 2 (2010) 65–70. [29] D. Nesheva, M.J. Scepanovic, S. Askrabic, Z. Levi, I. Bineva, Z.V. Popovic, Acta Phys. Pol. A 116 (2009) 75–77. [30] A. Weber, R. Mainz, H.W. Schock, J. Appl. Phys. 107 (2010) 0135161–0135166. [31] A. Redinger, D.M. Berg, P.J. Dale, S. Siebentritt, J. Am. Chem. Soc. 133 (2011) 3320–3323. [32] K.C. Mills, Thermodynamic Data for Inorganic Sulphides, Selenides and Tellurides, Butterworth, London, 1974. [33] J.J. Scragg, P.J. Dale, D. Colombara, L.M. Peter, Chem. Phys. Chem. 13 (2012) 3035–3046.