CZTS thin film solar cells on flexible Molybdenum foil by electrodeposition-annealing route

Earth-abundant and non-toxic Kesterite-based Cu2ZnSnS4 (CZTS) thin film solar cells are successfully fabricated on flexible Molybdenum (Mo) foil substrates by an electrodeposition-annealing route. A well-adherent, densely packed, homogeneous, compact, and mirror-like CZT precursor is initially produced through electrodeposition by using a rotating working electrode. Subsequently, the co-electrodeposited CuZnSn (CZT) precursor is sulfurized in quartz tube furnace at 550 °C for 2 h in N2 atmosphere with the presence of elemental sulfur in order to form CZTS. Different characterization techniques like XRD, SEM, HR-TEM, Raman, and Photoluminescence demonstrate that almost phase-pure CZTS formed after sulfurization. A flexible Al/Al-ZnO/i-ZnO/CdS/CZTS/Mo foil solar cell is produced, where CdS is deposited by chemical bath deposition and transparent conducting oxide (TCO) is deposited by DC sputtering. The CZTS solar device shows a 0.55% power conversion efficiency on flexible Mo foil substrate and it constitutes the first prototype of this kind of solar cell produced by electrodeposition-annealing route without any surface modification of the Mo substrate.


Introduction
In order to make a real impact on worldwide energy demand (at the TW level), solar materials should be cheap, non-toxic, and earth-abundant, and they should provide high functional lifetime with high power conversion efficiency (PCE). Over the last few decades, thin film technologies based on direct band gap metal chalcogenides have shown the potential to produce photovoltaic (PV) cells with low cost along with easy fabrication routes. Kesterites-based chalcogenides like CZTS (Cu 2 ZnSnS 4 ), CZTSe (Cu 2 ZnSnSe 4 ), and mixed sulfo-selenide CZTSSe (Cu 2 ZnSn(S, Se) 4 ) are attracting much attention nowadays from researchers as promising alternatives to existing commercialized CIGS (Cu(In, Ga)Se 2 ) and CdTe absorber layers. This is due to the low cost of their mass production, as constituent elements in Kesterites are non-toxic and abundant in the earth's crust compared to CIGS and CdTe. Moreover, they have suitable direct band gap near to 1.45 eV (CZTS), around 1 eV (CZTSe), and 1.0-1.5 eV (CZTSSe, depending on the S/(S + Se) ratio) together with high absorption coefficient (> 1 × 10 4 cm −1 ), which make them ideal among all the second generation thin film solar cells [1][2][3][4]. Furthermore, theoretical conversion efficiencies of Kesterites-based solar cells are around 32% like chalcopyrite according to Shockley-Queisser photon balance calculations [5]. Due to the rapid progress of the kesterite material development, the conversion efficiency of the Kesterites-based thin films solar cells have already advanced to 9.4% (CZTS) [6], 11.6% (CZTSe) [7], and 12.6% (CZTSSe) [8]. They are still far behind the theoretical conversion efficiency, despite these remarkable developments.
A suitable fabrication method is important to achieve high-quality Kesterites without having any defects, which ultimately govern the conversion efficiency. The volatile nature of some of their elemental constituents and related binary chalcogenides make the synthesis of stoichiometric Kesterites difficult [9,10]. In addition, it is difficult to fabricate single-phase Kesterites as they exist within a narrow composition field which is normally known as a Cu-poor (Cu/(Zn + Sn) ≈ 0.7-0.9) and Zn-rich (Zn/Sn ≈ 1.1-1.4) range, according to the phase diagram [9][10][11][12]. Besides inappropriate chemical composition of the metallic precursors, binary/ternary chalcogenide phases could also be created due to decomposition of Kesterites at high temperature during sulfurization or selenization [13].Till now, different vacuum and non-vacuum deposition techniques have been deployed to fabricate Kesterites (CZTS, CZTSe and CZTSSe). Both vacuum and non-vacuum techniques are very much in competition on conversion efficiency, and in some cases, non-vacuum processes are even more efficient, contrary to the case of CIGS cells where vacuum-based techniques deliver higher conversion efficiencies [8]. Several non-vacuum techniques like sol-gel, electrodeposition, solution-based method etc. have been widely studied to fabricate Kesterites thin films [13][14][15].
Most of the Kesterites-based thin films solar cells have been fabricated till now on typical Mo-coated soda lime glass (SLG). There are only few reports on fabrication of Kesterites-based solar cells on the flexible substrate like metallic foils or polyimides, irrespective to the fabrication processes [16][17][18][19][20][21][22][23][24]. Fabrication of Kesterites-based thin film solar cells on the flexible substrate has huge potential in order to minimize production cost due to roll-to-roll manufacturing on the flexible substrate which enables the use of compact size deposition equipment with high throughput [25]. It is, however, important to point out that, in the case of flexible solar cells, efficiency depends on bending radius. Normally, an increase in efficiency for concave bending and a decrease in the case of convex bending is observed [26,27]. Flexible metal foil substrates like molybdenum (Mo), stainless steel, aluminum (Al), titanium (Ti) etc. are potential candidates of back contact as they are cheap, durable, lightweight, and sustainable at high-temperature sulfurization or selenization processes. In addition to this, a metallic back contact layer is not required when metal substrate is used. On the other hand, the maximum operating temperature of polyimide is typically below 500 °C which makes it unsuitable for those fabrication processes where high-temperature annealing is involved in the presence of sulfur or selenium. Mo foil can be a better choice among all the flexible metallic foil substrates considering its capability to withstand very high temperature along with high conductivity and compatible coefficient of thermal expansion (5.2·10 -6 K −1 ) [19]. Moreover, unlike stainless steel, it does not need any diffusion barrier layer to prevent migration of substrate atom during sulfurization or selenization. Till now, maximum power conversion efficiency of CZTS on flexible Mo foil substrate is 3.82% [19], whereas maximum conversion efficiency of CZTS on Mo-coated SLG is 9.4% [6]. On the other hand, maximum power conversion efficiency of CZTS solar cell on flexible stainless steel was reported to be 4.2% through sputtering-H 2 S reactive annealing process with the incorporation of alkali doping [18]. CZTS thin film solar cells on Mo foil achieving the efficiency of 3.82% have been fabricated through electrodeposition-annealing route. Point to be noted, before deposition of CZTS on Mo foil, 800 nm Mo back contact was deposited on Mo foil using DC sputtering in their work. Meanwhile, power conversion efficiencies of CZTS and CZTSe on Mo-coated SLG have already reached 8.1% [28] and 8.2% [29], respectively, through electrodeposition-annealing route. It has been demonstrated that alkali doping at the interface of CZTS and Mo has significant effects on the conversion efficiency of the solar cells which has already been observed in CIGS thin film solar cells [30]. On the other hand, formation of MoS 2 at the interface of CZTS has also substantial effects on conversion efficiency as it is an indirect bandgap semiconductor with a gap of 1.3 eV which creates a back contact blocking (Schottky) barrier that restricts hole transport across the CZTS/Mo interface [31]. Compared to a variety of vacuum-based techniques and wet chemical methods, the material utilization rate in electrodeposition is better than 90%, partly due to the selectivity and partly because there is extensive know-how on the reuse and recycle of electroplating bath [32].
In this work, we fabricated 0.55% efficient CZTS thin film solar cells on flexible Mo foil substrate through electrodeposition-annealing route. Initially, a Cu-Zn-Sn (CZT) precursor is deposited on flexible Mo foil through co-electrodeposition, without sputtering a preliminary Mo layer. Other works reporting precursors deposition on unmodified Mo substrates are available in the literature [33,34]. The present work, however, presents important elements of innovation and uniqueness. The most important is the approach followed for the deposition step, in which a single alkaline electrolyte is employed to directly deposit CZT. Kalinauskas et al. [33] and Stanchik et al. [34], on the contrary, report CZTS fabrication starting from multilayer precursors of the kind CuSn/Zn or Cu/Zn/Sn, all deposited from acidic electrolytes. Moreover, Stanchik et al. report CZTSe rather than CZTS fabrication. Finally, both Kalinauskas et al. and Stanchik et al. limit the investigation to material fabrication, without producing complete solar cells. Following electrodeposition, the CZT layer is sulfurized at high temperature to get CZTS. Good quality of CZTS depends on the quality of the CZT precursor, and for this reason, high-quality CZT precursor is prepared using rotating working electrode during co-electrodeposition. No sputtering steps are employed on Mo foil before co-electrodeposition of CZT unlike other works in the literature [19]. Following this approach, fabrication process can be considerably simplified and optimized in terms of cost and time.

Electrodeposition of the CZT precursor on Mo foil
Co-electrodeposition of Cu-Zn-Sn (CZT) was carried out under galvanostatic control in a conventional electrochemical cell assembly at room temperature. Formulation of electrolyte bath has already been described in the existing literature [35]. Briefly, such electrolyte contained 2.5 g/l CuSO 4 5H 2 O, 15 g/l ZnSO 4 7H 2 O, 10.5 g/l Na 2 SnO 3 3H 2 O, 200 g/l K 4 P 2 O 7 , and 0.4 g/l CH 3 (CH 2 ) 11 OSO 3 Na. A rotating working electrode was used in place of the simple vertical working electrode (at a rotation speed of 25-35 rpm). Its characteristics are detailed in the supporting materials (figure S1). Flexible Mo foil substrates (200 µm thick, Goodfellow) with an exposed area of 1.6 × 2.0 cm 2 and titanium mesh were used as a working electrode and counter inert electrode, respectively. Mo substrates were cleaned in acetone, immersed in 32% wt. HCl for 10 min to remove oxide, washed in distilled water, and finally dried under N 2 atmosphere. Mo foil was then mounted on the rotating electrode and the Mo surface was immersed horizontally in the electrolyte. A 30 rpm rotation speed was employed during electrodeposition. Electrodeposition experiments were accomplished by using an AMEL Model-549 Potentiostat. Electrodeposition runs have been performed by using 4.5 -5.5 mA/cm 2 current density for 7-8 min, while keeping the pH of the solution at 11. The electrolyte was dummied for 30 min at 5 mA/cm 2 using the same anode/cathode configuration in order to eliminate impurities.

CZT sulfurization process
After electroplating CZT precursor, the sample was sulfurized at 550 °C for 2 h in a quartz tube furnace with the presence of 25 mg of elemental sulfur powder (Sigma Aldrich; reagent grade, 100 mesh particle size). During sulfurization, a heating rate of 20 °C/minute was used and a very small flow of N 2 was maintained throughout the process [12,35]. At the end of the process, CZTS was allowed to naturally cooldown.

CZTS solar cell construction
In order to form the p-n junction, n-type semiconductor CdS (≈70 nm) was deposited by chemical bath deposition on top of CZTS. An 80 nm intrinsic i-ZnO buffer layer, which acts to prevent any shunts, was then deposited by RF sputtering. Later, a TCO layer consisting of 350 nm Al-doped ZnO (AZO) was grown by DC-pulsed (2 kHz) sputtering. Finally, cells were completed by evaporating an Al grid contact on top of it. In this way, Al/Al-ZnO/i-ZnO/CdS/CZTS/ Mo foil device has been prepared. 6 single cells, each characterized by a 0.15 cm 2 area, were fabricated on each CZTS coated Mo foil. No anti-reflection coating was used during the measurement of the cells.

Characterization techniques
Composition, morphology, crystalline phases, and roughness of the CZT precursors and sulfurized CZTS films were investigated by SEM, Laser Profilometry, XRD, SEM, TEM, and Raman spectroscopy. Material band gap was investigated by Photoluminescence (PL) spectroscopy. The formation of impurity and MoS 2 at the interface of CZTS and Mo was evaluated through HR-TEM and electron energy loss spectroscopy (EELS). The crystallographic phase of sulfurized films was analyzed by X-ray diffraction (XRD), using a Philips X-pert MPD instrument with CuK α1 radiation (1.5406 Å).The morphology and chemical compositions of the samples were studied by Scanning Electron Microscopy (Model Zeiss EVO 50) together with Energy Dispersive X-ray Spectroscopy (EDS) (Oxford instrument, Model 7060). UBM laser profilometer was used for assessing the roughness parameters. Raman measurements were carried out in the air at room temperature with a micro-Raman spectrometer (Horiba Jobin-Yvon Labram HR800) in the backscattering configuration. The 785 nm excitation wavelength has been used during the Raman characterization of the samples. To avoid the laser heating of the sample, the laser power density was kept as low as possible. The laser power was reduced from 100 to 1 mW at the sample passing through a suitable filtering system with spot size about 1 µm diameter. Here, the beam was focused on the sample using an objective lens with 50 × magnification. Photoluminescence (PL) spectra of the films have been acquired at 15 K temperature with above bandgap excitation (λ exc = 805 nm). All PL measurements were performed with a spectral resolution of 6.6 nm using a standard lock-in technique in conjunction with a single grating monochromator and an InGaAs detector. Selected Area Electron Diffraction (SAED) and TEM observation were performed with a JEOL 2100F (Schottky) microscope equipped with a high-resolution pole piece, operated at 200 keV. EELS was acquired with a GIF Quantum and direct-detection sensor with a collection angle of 30 mrad [36]. Solar cell J-V characteristics were taken under Air Mass 1.5 conditions (simulating terrestrial applications) with a Abet Lot-Oriel Solar simulator with the constant incident power density of 1 Sun (100 mW/cm 2 ) and with Keithley 2440 source meter. Representative data from each characterization technique are reported in the results section.

CZT deposition on Mo flexible foil
Due to the large reduction potential window among Cu, Zn, and Sn metal ions, it is very difficult to co-electrodeposit CZT from single electrolyte. In our previous work, we were able to electrodeposit them together from single electrolyte where pyrophosphate was used as an only single complexing agent [35]. In order to obtain better precursors in terms of homogeneous distribution of elements and minimum roughness of the film with respect to previous works, here improved bath with rotating working electrode was employed during electrodeposition. It is possible to provide constant mass flux to the cathode surface using rotating working electrode which is very important to get homogeneous and smooth films. Moreover, it is believed that composition of various elements in the deposited film remains independent of the deposition times when the rotating working electrode is employed during co-electrodeposition [37]. It is well known that Cu-poor, Zn-rich precursor composition is necessary to obtain good kesterite CZTS film due to the narrow composition range for CZTS formation. This happens due to the volatile nature of S, SnS, and Zn during sulfurization which can vary the relative percentage of the elements to a great extent [9,10,12] and may result in poor kesterite films, characterized by the extensive formation of secondary phases. For this reason, in this study, Cu-poor and Zn-rich precursor (Cu 43.5 at. %, Zn 33 at. % and Sn 23.5 at. %; Cu/(Zn + Sn) = 0.77 and Zn/Sn = 1.4) has been employed for the fabrication of CZTS. Such CZT composition resulted in quasi-stoichiometric CZTS after annealing [12]. Figure 1 shows the surface morphology and cross-sectional view of the as-deposited CZT film observed by SEM. It is evident that the electrodeposited CZT film is quite homogenous, dense, and continuous, without visible  [38]. Moreover, CZT adhesion to the Mo substrate was found to be excellent. Average roughness of the CZT film was around 94 nm, while substrate roughness was around 100 nm, which is quite commendable for the co-electrodeposition of CZT.

Characterization of sulfurized CZTS on flexible
Mo foil Figure 2a shows the XRD diffractogram of sulfurized kesterite CZTS film. Here, the CZT precursor was sulfurized at 550 °C for 2 h with a heating rate of 20 °C/minute. All the XRD peaks of sulfurized sample correspond to the typical polycrystalline kesterite structure of CZTS (JCPDS card 26-0575) with intense (112) peak at 28.53° which is in agreement with findings of other studies [19,35,[39][40][41].
The sharpness of the diffraction peaks indicates the high crystallinity of the film. Diffraction peaks of CZTS can also be attributed to the ternary phase Cu 2 SnS 3 , since they have similar lattice parameters, or to ZnS [42]. As a result, complementary Raman spectroscopy (Fig. 2b) was employed to further characterize the sample. In order to have the better understanding of the formation of CZTS along the film thickness, the 785 nm excitation wavelength has been used. According to previous studies [43], the penetration depth of the 785 nm laser beam in CZTS film is approximately 400 nm and this allows to explore deeper the sample with respect to the 514 nm excitation line with a penetration depth of only 150 nm. Figure 2b displays the Raman spectrum of the sulfurized CZTS thin film measured at room temperature using the 785 nm excitation wavelength. All the main Raman peaks associated with CZTS were observed at 266 cm −1 , 287 cm −1 , 338 cm −1 , 366 cm −1 , and 376 cm −1 Raman shift including main intense peak at 338 cm −1 which is commonly mentioned in the literature [41,[43][44][45] and has been associated to the vibration of sulfur atoms [46]. The appearance of two distinct peaks around 367 cm −1 and 375 cm −1 is typical of Raman spectra of CZTS acquired with excitation laser wavelength of 785 nm [39,43]. At lower excitation wavelengths, only a broad peak at 367 cm −1 is evident. In the Raman spectrum, some additional weaker peaks are also observed at 303 cm −1 (cubic Cu 2 SnS 3 ) and 314 cm −1 (SnS 2 ) as reported by other authors [43,44]. Despite the presence of some secondary phases in film, crystalline CZTS is the largely dominant constituent of the film as the peak intensity of the secondary phases is very weak.
Further insight into the structural properties of the sulfurized CZTS film was gained through high-resolution TEM (HR-TEM) and EELS analysis of a focused ion beam (FIB) prepared using the cross-sectional sample. Figure 3 depicts the HR-TEM of the regions adjacent to the CZTS/ Mo interface. TEM imaging and SAED show that the sample is primarily composed of CTZS grains several hundred nm in diameter, in agreement with XRD and Raman results ( Fig. 3a and S2). Many twin boundaries and dislocations are visible inside the grains of the CZTS phase, as observed in other studies (Fig. S3 and S4) [47].
The interface between Mo and CZTS was analyzed thoroughly by using HR-TEM and electron energy loss spectroscopy (EELS). In addition to the CZTS, TEM revealed the presence of two secondary phases. As shown in Fig. 3, TEM imaging found a continuous and uniform layer separating the Mo substrate and CZTS film. EELS analysis of this layer shows only Mo and S (Fig. 4b, c), and HR-TEM shows a measured lattice spacing of 6.2 Angstroms (Fig. 3b). Given that the lattice constant of MoS 2 is 6.11 Angstroms, the EELS and HR-TEM data indicate a layer of MoS 2 formed between the Mo substrate and CZTS film. A third phase is observed in TEM, consisting of nanograin clusters. This phase has a relatively low concentration throughout the CZTS film ( Fig. S5 and S6), but the concentration increases considerably near the Mo substrate (Fig. 3a). SAED of this phase is consistent with SnO 2 (Fig. S5). Agreeing with this measurement, EELS shows that this phase only contains Sn and O (Fig. S6). We note that it is difficult to independently extract the O and Sn EELS signals since the O K edge overlaps the Sn M edge. However, Fig. 4d shows a map of the EELS signal from 470-590 eV, which includes both the O K and Sn M edges. Clearly, there is a large increase in the combined concentration and O + Sn within this third phase,  and Mo, Cu, and S are all absent. By comparing EELS spectra extracted from the CZTS phase and the SnO 2 phase, it is clear that both phases contain Sn, but O is only present within the SnO 2 phase (Fig. S6). The spectra taken from the SnO 2 region closely match those found in the literature [48]. HR-TEM of this phase shows a lattice spacing of 3.2 Angstroms, consistent with the expected value of 3.18 Angstroms for SnO 2 .
The presence of MoS 2 and SnO 2 were not detected by XRD or Raman analysis. However, it is possible that these phases were not detected with XRD because of their nanometric size and very low amount. It is also possible that they were located too deep in the CZTS layer to contribute to the Raman signal, provided that their content was above the detection limit of Raman spectroscopy. The source of the SnO 2 clusters is presently unclear; it might be guessed that their formation was caused by the oxidizing impurities of the nitrogen gas used in the sulfurization step. SnO 2 is a large bandgap oxide and a VB offset of approximately 1.5 eV is predicted to exist at the boundary between CZTS and SnO 2 ; hence this interface might act as a barrier to hole transport [49,50].

Complete CZTS solar cell characterization
After the morphological investigation of the formed CZTS layer, complete cells were fabricated. Figure 5 shows the SEM cross-sectional image of the CZTS solar cell, which demonstrates that CZTS possesses bimodal grain distribution where smaller grains are located near the interface with Mo and larger grains at the top of the CZTS.
In order to make sure that fabricated CZTS absorber layers are applicable to photonic devices (i.e., solar cells), samples were further characterized by Photoluminescence Spectroscopy (PL). A typical photoluminescence spectrum of CZTS films at 15 K is reported in Fig. S7, which shows an asymmetric broad band at 1.21 eV. Similar asymmetric emissions were observed in CZTS fabricated through various methods by different groups [45,[51][52][53]. According to the literature [51,52], this band is associated to a quasi-donor-acceptor pair (QDAP) transition, which involves potential fluctuations, as commonly observed in highly defective and compensated semiconductors. It was also demonstrated that the presence of a QDAP emission with saturated peak energy position towards 1.18 eV in the PL spectrum of CZTS layers indicates them as efficient PV absorbers [54].
In order to test the efficiency of the grown layers as PV absorbers, CZTS thin films solar cells were fabricated on Mo foil with classical configuration (Al/Al-ZnO/i-ZnO/CdS/ CZTS/Mo foil). Figure 6a shows the current-voltage (J-V) characteristics of CZTS solar cells measured under AM 1.5 illumination. As fabricated Al/Al-ZnO/i-ZnO/CdS/CZTS/ Mo foil device shows 0.55% power conversion efficiency  (PCE) with J sc = 7.60 mA/cm 2 , V oc = 203.98 mV and FF = 35.68%. The low fill factor can come from high series resistance due to back contact electrical characteristics. In addition to this, non-adherence between various layers and morphological problems because of elemental losses during sulfurization can also contribute to lowering both J sc and FF [55]. On the other hand, the small value of V oc is most probably related to the presence of recombination paths may be associated with grain boundaries or other defects. Formation of interfacial MoS 2 during sulfurization process may affect the V OC and alter the band alignment between the CZTS film and MoS 2 /Mo. Hence, it may reduce the conversion efficiency of the solar cells since it is in inverse proportion to the solar cell characteristics after the certain limit. It was reported that the formation of MoS 2 interfacial layer at the CZTS/Mo interface enhances the quasi-ohmic electrical contact when it forms with appreciable thickness [56]. On the other hand, a thick MoS 2 layer at the CZTS/Mo interface can influence the total series resistance of the device [41]. However, Zhang et al. recently demonstrated 7.2% efficient CZTS solar cells without having any MoS 2 layer [57]. By the way, HR-TEM suggest around 50 nm of MoS 2 has been formed in our solar cells (Fig. 3). Moreover, the effect of alkali metals (Na, K) on the conversion efficiency of chalcogenide thin film solar cells cannot be ignored, as it has already been reported that alkali incorporation to CIGS increases its open circuit voltage and fill factor [58]. Though the alkali effect on the conversion efficiency was not extensively investigated for CZTS/CZTSe, results similar to those reported for CIGS have been found in case of CZTS also [59,60]. Alkali incorporation in CZTS can affect the grain size of the final film, the (112) texturing, the free carrier concentration, the hole density and mobility of the film [58]. Yang et al. demonstrated that there is a directly proportional relationship between CZTS thin film solar cells efficiency and Na content in Mo layer. This is probably one of the main reasons why the efficiency of CZTS solar cells on the flexible substrate is far behind that of CZTS solar cells on Mo-coated SLG substrate. Recently Lopez-Marino et al. reported that an improvement in the conversion efficiency of CZTSe solar cells from 2.2% to 4.3% is possible using Mo-Na (Na doped Mo) layer sandwiched between regular Mo layers. After further optimizing the Cr diffusion barrier layer, they reported a record efficiency of 6.1% for a CZTSe solar cell on flexible stainless steel substrate prepared by the sputtering-annealing process [17]. Point to be noted, the device described in the present work is the first prototype of CZTS solar cells on flexible Mo substrate that is realized through electrodeposition-annealing route without any surface modification, to the best of our knowledge. Figure 6b shows the external quantum efficiency (EQE) spectrum of our CZTS solar cell between 350 and 1000 nm. The EQE demonstrates a low overall spectrum response with the maximum value of 38.5% (at 510 nm) indicating a high recombination rate, in agreement with V oc value obtained by J-V characterization. The reduction of EQE spectrum in the short wavelength portion is ascribed to the absorption losses in the window and buffer layers (Al-ZnO, i-ZnO, CdS) [41]. On the other hand, recombination processes contribute to lower the EQE values in the long wavelength region.

Conclusions
A CZTS thin film solar cell was successfully synthesized on flexible Mo foil substrate by the electrodepositionannealing method with 0.55% power conversion efficiency. A high-quality CZT precursor without any void or crevice was fabricated by the co-electrodeposition route using rotating working electrode. Formation of MoS 2 at the interface, together with SnO 2 formation near to MoS 2 , can explain the low performances of the final device. Moreover, the absence of alkali metal doping at the substrate can be also associated with such low efficiency of the cell. Nevertheless, first power conversion device based on CZTS thin film solar cell on flexible Mo foil substrate, which has been produced by electrodeposition-annealing route without any surface modification, was demonstrated. In perspective, the good adhesion obtained between the CZTS layer and the Mo substrate constitutes a suitable starting point to perform tests at controlled bending angle and defined number of bending cycles, since it will reasonably prevent active layer delamination. The CZTS cells obtained following the methodology described in the present work constitute, thus, a promising starting point for further process optimization and for additional testing under bending conditions. Funding Open access funding provided by Politecnico di Milano within the CRUI-CARE Agreement.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.