Synthesis and Characterization of Cu2ZnSnS4 Thin Films Obtained by Combined Magnetron Sputtering and Pulsed Laser Deposition

Cu2ZnSnS4 (CZTS) is a complex quaternary material, and obtaining a single-phase CZTS with no secondary phases is known to be challenging and dependent on the production technique. This work involves the synthesis and characterization of CZTS absorber layers for solar cells. Thin films were deposited on Si and glass substrates by a combined magnetron sputtering (MS) and pulsed laser deposition (PLD) hybrid system, followed by annealing without and with sulfur powder at 500 °C under argon (Ar) flow. Three different Cu2S, SnS2, and ZnS targets were used each time, employing a different target for PLD and the two others for MS. The effect of the different target arrangements and the role of annealing and/or sulfurization treatment were investigated. The characterization of the absorber films was performed by grazing incidence X-ray diffraction (GIXRD), X-ray reflectometry (XRR), Raman spectroscopy, scanning electron microscopy, and regular transmission spectroscopy. The film with ZnS deposited by PLD and SnS2 and Cu2S by MS was found to be the best for obtaining a single CZTS phase, with uniform surface morphology, a nearly stoichiometric composition, and an optimal band gap of 1.40 eV. These results show that a new method that combines the advantages of both MS and PLD techniques was successfully used to obtain single-phase Cu2ZnSnS4 films for solar cell applications.


Introduction
Cu 2 ZnSnS 4 (CZTS) is a strong material candidate for next-generation solar energy converters due to its high absorption coefficient and optimal band gap [1]. P-type chalcogenides are made of low-cost and earth-abundant elements (Cu, Zn, Sn, and S), unlike other thin-film solar cell technologies (CIS, CIGS, and CdTe) that contain expensive and/or toxic components [2]. Several physical and chemical routes can be used for the synthesis of CZTS thin films. Although vacuum techniques can be complex and require high energy, they are collectively regarded as a solution to overcome issues related to time consumption, cracks, and poor crystallinity of more used chemical routes [3,4]. They also offer the possibility to easily control the structural and morphological properties of the thin films [5].
Among all physical vapor deposition methods, magnetron sputtering (MS) and pulsed laser deposition (PLD) have largely been used in the preparation of CZTS absorber layers because of their numerous advantages [6,7]. PLD is a simple and versatile technique that is commonly used for depositing thin films of a very wide range of materials on a vast variety of substrates [8,9]. This technique is known for its high deposition rate, relatively easy transfer of species from the target to the substrate, and high-quality growth of thin

Materials and Methods
CZTS thin films were obtained using three binary sulfide targets (Cu 2 S, SnS 2 , and ZnS) with a diameter of 2 inches and a thickness of 3 mm in a hybrid system combining MS and PLD techniques. Two targets were mounted in cathodic sprayers (magnetrons), and one in the PLD carousel, and the thin films were synthesized on silicate glass and Si/SiO 2 substrates. Three CZTS films were deposited at room temperature on each type of substrate by fixing a target to be deposited by PLD and the two others sputtered by MS, all deposits taking place simultaneously. A COMPexPro KrF laser source (Coherent, California, USA) with a wavelength λ of 248 nm and a fluence of 1.5 J/cm 2 was focused on the target at an incidence angle of 45 • . The sputtering was performed using argon as the working gas with a 5 × 10 −3 Torr (Linde, Bucharest, Romania). The CZTS films were deposited using a Model AG 0313 source (RF T&C Power Conversion Inc., New York, NY, USA) and a 3G Circular Magnetron (Gencoa, Liverpool, UK). In order to define the deposition rate, calibration was performed on the three targets using Q-pod software (Inficon, Bad Ragaz, Switzerland) connected to a quartz crystal. The distance target to sample was 8 cm and 11 cm for the PLD and MS targets, respectively, and the substrate holder was rotated to achieve a uniform deposition.
The first target arrangement utilized a Cu 2 S target for PLD deposition and SnS 2 and ZnS targets for MS. The PLD parameters for the Cu 2 S layer were 10 Hz for the laser pulse frequency and 210 mJ for the power. The deposition rate was 0.50 Å/s. The SnS 2 target was deposited using an RF power supply with a power of 20 W to obtain a deposition rate of 0.49 Å/s. The ZnS target was ignited by the RF power supply, and a power of 67 W was applied, giving a deposition rate of 0.50 Å/s.
In the second arrangement, the SnS 2 target was loaded in the PLD carousel, and the Cu 2 S and ZnS targets were mounted in the magnetrons. The laser pulse frequency was 1 Hz, and the power was maintained at 90 mJ, giving a deposition rate of 0.47 Å/s. The Cu 2 S and ZnS targets were deposited using an RF power supply with a power of 27 W for Cu 2 S and 65 W for ZnS, while the deposition rate was 0.54 Å/s and 0.47 Å/s for the Cu 2 S and ZnS targets, respectively.
In the third arrangement, deposition was performed by PLD from the ZnS target, while MS was used for Cu 2 S and SnS 2 . PLD deposition was carried out using a laser frequency of 2 Hz and a power of 60 mJ. The deposition rate was about 1.30 Å/s. An RF power supply with a power of 56 W to maintain a rate of 1.30 Å/s was applied in the case of the Cu 2 S MS deposition. For the SnS 2 target, an RF power supply with a power of 40 W was used to obtain a deposition rate of 1.30 Å/s. The as-deposited samples were named depending on the target deposited by the PLD technique: Cu 2 S (PLD), SnS 2 (PLD), and ZnS (PLD). Two series of samples were prepared, and each was annealed in an oxygen-free atmosphere inside a quartz tube (2 inches in diameter) of a tubular furnace to form the CZTS phase. The first series, deposited on Si/SiO 2 substrates, was annealed without any sulfur source and denoted Cu 2 S (PLD)/annealed, SnS 2 (PLD)/annealed, and ZnS (PLD)/annealed. The second series, deposited on glass substrates, was sulfurized using 1 g of S powder. The samples were named Cu 2 S (PLD)/sulfurized, SnS 2 (PLD)/sulfurized, and ZnS (PLD)/sulfurized. The two series of thin films were annealed in Ar gas atmosphere at 500 • C for 3 h using a GSL 1600X furnace (MTI, California, USA) with an increment of 2.5 • C/min for heating and cooling.
Grazing incidence X-ray diffraction (GIXRD) with an incidence angle of 0.3 • was used to inspect the structural properties of the CZTS thin films. The measurements were performed with a Rigaku SmartLab diffractometer provided with Cu Kα radiation (λ = 1.54178 Å) and HyPix-3000 2D Hybrid Pixel Array Detector (Rigaku, Tokyo, Japan) (in 0 D mode). The X-ray reflectometry (XRR) data were recorded with the same equipment in a range of (0 ÷ 2) ( • 2θ), with a 2θ step of 0.004 • . The Raman spectrometer used was LabRAM HR Evolution from HORIBA Jobin-Yvon equipped with a confocal microscope and a He−Ne laser and employed to confirm the crystalline structure study. The measurements were performed with a red excitation wavelength of 633 nm, and the laser was focused using an Olympus 100× objective on the surface of the samples. Raman spectra were analyzed in the range of 200 to 500 cm −1 at room temperature.
The optical properties were measured using a V-VASE Woollam Spectroscopic Ellipsometer equipped with a high-pressure Xenon discharge lamp incorporated in an HS-190 monochromator.
The microstructure of the films was examined using a Zeiss EVO 50 XVP scanning electron microscope (SEM) accessorized with an energy-dispersive spectrometer (EDS) for the determination of the elemental concentration in the films.

Results and Discussion
From the XRR diagrams of the three thin films obtained by MSPLD co-deposition, in the initial state and after annealing and sulfurization at 500 • C (Figure 1a), the positions of the total reflection edge (2θ TR ) ( Figure 1b) and the average film thickness are calculated and presented in Table 1. The total reflection edge is defined as the angle for which the reflectivity decreases by 30% of the maximum. We notice that, after the heat treatment without sulfur, 2θ TR decreases slightly for all the films. The average mass density (ρ m ) of the thin films is proportional with (θ TR ) 2 , and accordingly, ρ m also decreases ( Table 1). This fact, together with the decrease in the average thickness of all annealed CZTS thin films, means that the thin films undergo a process of partial evaporation and a reorganization of atoms, which produces the formation of voids between crystallites and a decrease in the thin-film thickness. The ZnS (PLD) thin film seems to be the least influenced by annealing without sulfur (both ρ m and h m decrease with −2.1% and −2.0%, respectively), the mass loss being only −4.1%. SnS 2 (PLD) is the most affected (both ρ m and h m decrease with −14.7% and −14.3%, respectively), the mass loss being −29.0%. This might be due to the evaporation of the very volatile SnS [31] during the deposition and annealing. The Cu 2 S (PLD) thin film suffers moderate modifications (both ρ m and h m decrease with −7.0% and −10.2%, respectively), the mass loss being −17.2%. The ZnS (PLD) sample is the densest and shows the smallest modifications in mass density and thickness after annealing. This could be due to the fact that in this case the deposition rate was double in comparison to the other two samples and the deposition rate influences the growth of the film (smoother films with smaller particle sizes are obtained at higher rates). When annealing with sulfur, the XRR fringes are not visible, meaning that the surface of the film becomes rough (confirmed by SEM), which is beneficial for solar cells. However, in this case, we expect a slight increase in the film thickness [32]. Additionally, the decrease in ρ m is much diminished for Cu 2 S (PLD)/sulfurized and SnS 2 (PLD)/sulfurized, while for ZnS (PLD)/sulfurized, ρ m slightly increases, which means that the voids between crystallites are smaller than in the case of simple annealing.    Figure 2 shows the grazing incidence X-ray diffraction patterns of the CZTS thin films prepared using the MSPLD hybrid system by switching the target in the PLD carousel in each film and annealing without and with sulfur powder. The GIXRD diagrams for the as-deposited CZTS thin films are presented in Figure 2a. The Cu2S (PLD) sample has an amorphous structure, while for the SnS2 (PLD) and ZnS (PLD) samples, the amorphous phase is only 12% and 27%, respectively. The majority phase for the last two samples is polycrystalline Cu2ZnSnS4 (tetragonal, I-42m (121), as natural mineral "stannite", ICDD file 00-026-0575) with an average size of crystallites of several nanometers. In order to prove the beneficial effect of sulfurization, we also annealed a set of samples without using sulfur. In this case (Figure 2b), all the samples suffer a major transformation: the amorphous phase and the Cu2ZnSnS4 nanophase (in the case of ZnS (PLD)/annealed and SnS2 (PLD)/annealed) are decomposed, some of the Zn and Sn atoms evaporate (thus the concentration of Cu atoms increase in thin films), the sulfur atoms diffuse in the entire volume of the thin films or evaporate (as XRR revealed), and thus, a polycrystalline CuS phase (hexagonal, P63/mmc (194), ICDD file 00-006-0464) appears. The average size of the crystallites for this phase is of hundreds of nanometers. Additionally, the average size of the crystallites for the remaining Cu2ZnSnS4 phase (tetragonal, I-42m (121), ICDD file 00-026-  Figure 2 shows the grazing incidence X-ray diffraction patterns of the CZTS thin films prepared using the MSPLD hybrid system by switching the target in the PLD carousel in each film and annealing without and with sulfur powder. The GIXRD diagrams for the as-deposited CZTS thin films are presented in Figure 2a. The Cu 2 S (PLD) sample has an amorphous structure, while for the SnS 2 (PLD) and ZnS (PLD) samples, the amorphous phase is only 12% and 27%, respectively. The majority phase for the last two samples is polycrystalline Cu 2 ZnSnS 4 (tetragonal, I-42m (121), as natural mineral "stannite", ICDD file 00-026-0575) with an average size of crystallites of several nanometers. In order to prove the beneficial effect of sulfurization, we also annealed a set of samples without using sulfur. In this case (Figure 2b), all the samples suffer a major transformation: the amorphous phase and the Cu 2 ZnSnS 4 nanophase (in the case of ZnS (PLD)/annealed and SnS 2 (PLD)/annealed) are decomposed, some of the Zn and Sn atoms evaporate (thus the concentration of Cu atoms increase in thin films), the sulfur atoms diffuse in the entire volume of the thin films or evaporate (as XRR revealed), and thus, a polycrystalline CuS phase (hexagonal, P63/mmc (194), ICDD file 00-006-0464) appears. The average size of the crystallites for this phase is of hundreds of nanometers. Additionally, the average size of the crystallites for the remaining Cu 2 ZnSnS 4 phase (tetragonal, I-42m (121), ICDD file 00-026-0575) increases to hundreds of nanometers. In Figure 2c are represented the GIXRD diagrams for the samples sulfurized at 500 • C. All the samples consist of a single phase, Cu 2.13 Zn 0.84 Sn 1.04 S 4 , with the tetragonal signature, the atomic plane (101), and some disorder distribution of the cations in the lattice (space group I-4 (82), similar to natural mineral "kesterite", ICDD file 04-023-6314). The ZnS (PLD)/sulfurized sample appears to have sharper peaks than the other samples, suggesting an enhancement in crystallinity and particle growth after sulfurization [33]. Therefore, to confirm the suggested improvement, the average crystallite size D of the three sulfurized samples was calculated from the (112) peak of GIXRD data using the Scherrer formula: D = kλ/(β.cosθ), where D is the average crystallite size, k is the Scherrer constant, λ is the wavelength of Cu Kα, β is the corrected 0575) increases to hundreds of nanometers. In Figure 2c are represented the GIXRD diagrams for the samples sulfurized at 500 °C. All the samples consist of a single phase, Cu2.13Zn0.84Sn1.04S4, with the tetragonal signature, the atomic plane (101), and some disorder distribution of the cations in the lattice (space group I-4 (82), similar to natural mineral "kesterite", ICDD file 04-023-6314). The ZnS (PLD)/sulfurized sample appears to have sharper peaks than the other samples, suggesting an enhancement in crystallinity and particle growth after sulfurization [33]. Therefore, to confirm the suggested improvement, the average crystallite size D of the three sulfurized samples was calculated from the (112) peak of GIXRD data using the Scherrer formula: D = kλ/(β.cosθ), where D is the average crystallite size, k is the Scherrer constant, λ is the wavelength of Cu Kα, β is the corrected FWHM (by eliminating instrumental broadening) of the (112)   In order to confirm the results obtained by GIXRD, and to steadily identify the secondary phases in our samples, Raman spectroscopy was performed with an excitation wavelength of 633 nm. The effect of annealing without and with sulfur powder on the as-deposited thin films was studied and is presented in Figure 3. Two broad peaks at 288 cm −1 and 330 cm −1 are seen in all the as-deposited samples (Figure 3a). The main Raman peak of the kesterite CZTS phase is located at 336-338 cm −1 [34] but was reported to shift to lower values due to some disorder distribution of the cations in the lattice of the CZTS film [35]. Thus, these two peaks are identified as the characteristic peaks of the kesterite CZTS phase. Figure 3b shows the Raman spectra of the samples annealed without sulfur. CZTS peaks can be noticed in the three films at 288, 336, 365, and 374 cm −1 [36,37]; however, all the samples contain a strong peak at 474 cm −1 and a smaller one at 264 cm −1 , indicating the presence of CuS [33]. In addition, a third phase is also detected in the films at 303 cm −1 and can be associated with the CTS phase [38]. The sulfurized CZTS thin films are presented in Figure 3c. Both the Cu 2 S (PLD)/sulfurized and SnS 2 (PLD)/sulfurized samples are characterized by five peaks at 256, 288, 337, 364, and 374 cm −1 of the CZTS phase and a prominent peak at 320 cm −1 whose origin is unclear. Several researchers attributed this peak to the CTS [39,40] or SnS 2 [41,42] secondary phases, while others stated that it belongs to CZTS [43,44]. On the other hand, the ZnS (PLD)/sulfurized sample contains peaks located at 256, 288, 337, 365, and 374 cm −1 , which are uniquely characteristic of the CZTS phase [43,45], and no undesirable secondary phase was observed. As a side note, Raman measurements were also performed on all the samples using an excitation wavelength of 325 nm to detect the zinc sulfide phase [46], and no evidence of the ZnS or SnS 2 secondary phases was detected. The annealing treatment either with or without sulfur plays a considerable role in the formation and crystallization of the CZTS films. However, a simple heat treatment without any sulfur source can lead to the apparition of several secondary phases, no matter what target was used for PLD, most probably due to the lack of sulfur and the very likely evaporation of zinc and tin in the form of SnS, resulting from the decomposition of SnS 2 at high temperature [47]. As an alternative, sulfurization is primordial for synthesizing single-phase CZTS thin films, as observed in the ZnS (PLD)/sulfurized sample.
CZTS peaks can be noticed in the three films at 288, 336, 365, and 374 cm −1 [36,37]; however, all the samples contain a strong peak at 474 cm −1 and a smaller one at 264 cm −1 , indicating the presence of CuS [33]. In addition, a third phase is also detected in the films at 303 cm −1 and can be associated with the CTS phase [38]. The sulfurized CZTS thin films are presented in Figure 3c. Both the Cu2S (PLD)/sulfurized and SnS2 (PLD)/sulfurized samples are characterized by five peaks at 256, 288, 337, 364, and 374 cm −1 of the CZTS phase and a prominent peak at 320 cm −1 whose origin is unclear. Several researchers attributed this peak to the CTS [39,40] or SnS2 [41,42] secondary phases, while others stated that it belongs to CZTS [43,44]. On the other hand, the ZnS (PLD)/sulfurized sample contains peaks located at 256, 288, 337, 365, and 374 cm −1 , which are uniquely characteristic of the CZTS phase [43,45], and no undesirable secondary phase was observed. As a side note, Raman measurements were also performed on all the samples using an excitation wavelength of 325 nm to detect the zinc sulfide phase [46], and no evidence of the ZnS or SnS2 secondary phases was detected. The annealing treatment either with or without sulfur plays a considerable role in the formation and crystallization of the CZTS films. However, a simple heat treatment without any sulfur source can lead to the apparition of several secondary phases, no matter what target was used for PLD, most probably due to the lack of sulfur and the very likely evaporation of zinc and tin in the form of SnS, resulting from the decomposition of SnS2 at high temperature [47]. As an alternative, sulfurization is primordial for synthesizing single-phase CZTS thin films, as observed in the ZnS (PLD)/sulfurized sample.     (Figure 4f) shows a more homogeneous and compact surface with monodispersed particles and no voids or cracks. The difference in particles sizes and shapes in the Cu 2 S (PLD)/sulfurized and SnS 2 (PLD)/sulfurized samples can be explained by the co-existence of secondary phases [48], while a homogeneous surface with relatively monodispersed particles (ZnS (PLD)/sulfurized sample) suggests the existence of a single-phase compound, which further confirms the GIXRD and Raman results. ferent shapes and sizes along with some voids probably due to the evaporation of tin sulfide. The ZnS (PLD)/sulfurized film (Figure 4f) shows a more homogeneous and compact surface with monodispersed particles and no voids or cracks. The difference in particles sizes and shapes in the Cu2S (PLD)/sulfurized and SnS2 (PLD)/sulfurized samples can be explained by the co-existence of secondary phases [48], while a homogeneous surface with relatively monodispersed particles (ZnS (PLD)/sulfurized sample) suggests the existence of a single-phase compound, which further confirms the GIXRD and Raman results. An estimation of the average atomic ratios of the four elements (Cu, Zn, Sn, and S), obtained from the energy-dispersive spectroscopy (EDS) analysis, in the as-deposited and sulfurized samples, is presented in Figure 5. All the as-prepared samples exhibit a Sn-rich and S-poor composition compared to the stoichiometric Cu2ZnSnS4 (Cu0.25Zn0.125Sn0.125S0.5) phase. By sulfurization at 500 °C, an improvement in the elemental composition is clearly seen in all the samples (Figure 5b), since the amounts of Cu, Zn, and Sn decreased, and the S quantity increased. However, the sulfurized samples are still off-stoichiometric, rich in tin, and slightly poor in sulfur. The existence of tin in excess can cause the formation of the SnS or SnS2 secondary phases [49]. This explains the apparition of the peak at 320 cm −1 in both samples (Cu2S (PLD)/sulfurized and SnS2 (PLD)/sulfurized) in the Raman analysis, An estimation of the average atomic ratios of the four elements (Cu, Zn, Sn, and S), obtained from the energy-dispersive spectroscopy (EDS) analysis, in the as-deposited and sulfurized samples, is presented in Figure 5. All the as-prepared samples exhibit a Sn-rich and S-poor composition compared to the stoichiometric Cu 2 ZnSnS 4 (Cu 0.25 Zn 0.125 Sn 0.125 S 0.5 ) phase. By sulfurization at 500 • C, an improvement in the elemental composition is clearly seen in all the samples (Figure 5b), since the amounts of Cu, Zn, and Sn decreased, and the S quantity increased. However, the sulfurized samples are still off-stoichiometric, rich in tin, and slightly poor in sulfur. The existence of tin in excess can cause the formation of the SnS or SnS 2 secondary phases [49]. This explains the apparition of the peak at 320 cm −1 in both samples (Cu 2 S (PLD)/sulfurized and SnS 2 (PLD)/sulfurized) in the Raman analysis, which can be definitely attributed to the SnS 2 phase. One should note that the determination of elemental concentrations by EDS can be affected by several parameters such as the sample thickness and the scattering power of atoms; therefore, significant errors are expected from this characterization technique [48].
Nanomaterials 2021, 11, x FOR PEER REVIEW 9 of 13 which can be definitely attributed to the SnS2 phase. One should note that the determination of elemental concentrations by EDS can be affected by several parameters such as the sample thickness and the scattering power of atoms; therefore, significant errors are expected from this characterization technique [48]. The band gaps of the as-deposited and sulfurized CZTS thin films are shown in Figure 6. The transmittance measurements were performed using an ellipsometer, and the optical band gap was determined using the following equation [50]: where Abs(E) is the derived absorbance using the measured transmittance (T), E is the photon energy, C is the constant of proportionality, d is the film thickness, and Eg is the optical band gap. Eg was estimated using the Tauc plot by plotting (αE) 2 vs. (E) (where α is the absorption coefficient) near the absorption edge. Since the as-deposited samples in Figure 6a are mostly composed of an amorphous structure, the values of the optical band gaps are above the reported value for the CZTS thin films (≈1.5 eV). In Figure 6b are shown the band gaps of the CZTS sulfurized films. The estimated Eg is around 1.40, 1.56, and 1.70 eV for ZnS (PLD)/sulfurized, SnS2 (PLD)/sulfurized, and Cu2S (PLD)/sulfurized, respectively. The band gap is found to be correlated with the stoichiometry and the elemental composition of the CZTS material. Large Eg are caused by an increase in Sn content according to Malerba et al. [51], while Hamanaka et al. reported that the large values could be due to high Zn ratios along with a decrease in Cu content [52]. Eg also depends on the content of the three elements, more precisely on the Cu/(Zn + Sn) ratio, since higher energies are obtained with lower values of this ratio and vice versa [53]. This is confirmed in our case, as the Cu/(Zn + Sn) ratio in the Cu2S (PLD)/sulfurized sample is very small, and thus, a high value of the optical band gap (1.70 eV) is found. On the other hand, the ZnS (PLD)/sulfurized sample shows an enhancement in this ratio, leading to a lower band gap (1.40 eV). The optical properties of the CZTS thin films obtained here are in good agreement with the literature and are suitable for use in solar cells as absorber layers [54,55]. The band gaps of the as-deposited and sulfurized CZTS thin films are shown in Figure 6. The transmittance measurements were performed using an ellipsometer, and the optical band gap was determined using the following equation [50]: where Abs(E) is the derived absorbance using the measured transmittance (T), E is the photon energy, C is the constant of proportionality, d is the film thickness, and E g is the optical band gap. E g was estimated using the Tauc plot by plotting (αE) 2 vs. (E) (where α is the absorption coefficient) near the absorption edge. Since the as-deposited samples in Figure 6a are mostly composed of an amorphous structure, the values of the optical band gaps are above the reported value for the CZTS thin films (≈1.5 eV). In Figure 6b are shown the band gaps of the CZTS sulfurized films. The estimated E g is around 1.40, 1.56, and 1.70 eV for ZnS (PLD)/sulfurized, SnS 2 (PLD)/sulfurized, and Cu 2 S (PLD)/sulfurized, respectively. The band gap is found to be correlated with the stoichiometry and the elemental composition of the CZTS material. Large E g are caused by an increase in Sn content according to Malerba et al. [51], while Hamanaka et al. reported that the large values could be due to high Zn ratios along with a decrease in Cu content [52]. E g also depends on the content of the three elements, more precisely on the Cu/(Zn + Sn) ratio, since higher energies are obtained with lower values of this ratio and vice versa [53]. This is confirmed in our case, as the Cu/(Zn + Sn) ratio in the Cu 2 S (PLD)/sulfurized sample is very small, and thus, a high value of the optical band gap (1.70 eV) is found. On the other hand, the ZnS (PLD)/sulfurized sample shows an enhancement in this ratio, leading to a lower band gap (1.40 eV). The optical properties of the CZTS thin films obtained here are in good agreement with the literature and are suitable for use in solar cells as absorber layers [54,55]. From the structural, morphological, compositional, and optical results, it can be deduced that the MSPLD hybrid system can be beneficial to the formation of the CZTS kesterite phase. This approach gives the possibility to produce rough films with larger crystallites when compared to some CZTS films prepared either by PLD [11] from a single Cu 2 ZnSnS 4 target or by MS [17] from a quaternary CZTS alloy target. The XRD and Raman spectroscopy characterizations of the sulfurized films show the formation of a single CZTS crystalline phase in one set of samples (ZnS (PLD)/sulfurized) or the presence of one secondary phase (SnS 2 ) in the other samples (SnS 2 (PLD)/sulfurized and Cu 2 S (PLD)/sulfurized). On the other hand, CZTS absorber layers elaborated by MS can contain several secondary phases [21,56,57] such as Cu 2 S, CTS, and ZnS when multiple binary chalcogenide targets are used or even when a single alloyed CZTS target is sputtered, which can have a harmful impact on the performance of CZTS-based solar cells [58,59]. Additionally, one of the major drawbacks of the PLD method is the use of tiny substrates [60]. The problem can be solved and large-area substrates can be covered by implementing a hybrid MSPLD system that still benefits from PLD advantages. This can be helpful for large-scale industrial CZTS elaboration. From the structural, morphological, compositional, and optical results, it can be deduced that the MSPLD hybrid system can be beneficial to the formation of the CZTS kesterite phase. This approach gives the possibility to produce rough films with larger crystallites when compared to some CZTS films prepared either by PLD [11] from a single Cu2ZnSnS4 target or by MS [17] from a quaternary CZTS alloy target. The XRD and Raman spectroscopy characterizations of the sulfurized films show the formation of a single CZTS crystalline phase in one set of samples (ZnS (PLD)/sulfurized) or the presence of one secondary phase (SnS2) in the other samples (SnS2 (PLD)/sulfurized and Cu2S (PLD)/sulfurized). On the other hand, CZTS absorber layers elaborated by MS can contain several secondary phases [21,56,57] such as Cu2S, CTS, and ZnS when multiple binary chalcogenide targets are used or even when a single alloyed CZTS target is sputtered, which can have a harmful impact on the performance of CZTS-based solar cells [58,59]. Additionally, one of the major drawbacks of the PLD method is the use of tiny substrates [60]. The problem can be solved and large-area substrates can be covered by implementing a hybrid MSPLD system that still benefits from PLD advantages. This can be helpful for large-scale industrial CZTS elaboration.

Conclusions
CZTS thin films were co-deposited simultaneously using three targets (Cu2S, SnS2, and ZnS) in a hybrid synthesis system combining magnetron sputtering (for two targets) and pulsed laser deposition (for the third target). The effect of the heat treatment with and without a sulfur atmosphere on the material properties was studied. The results show that a simple heat treatment without any sulfur source leads to the evaporation of Zn and Sn and the apparition of CuS and CTS secondary phases in all the films, no matter the target arrangement used. On the other hand, the sulfurized samples show that annealing in a sulfur atmosphere is an essential process for the formation of the CZTS phase. It is found that the samples obtained by simultaneous deposition of ZnS by PLD, Cu2S, and SnS2 by MS and sulfurized at 500 °C, which leads to the formation of a single-phase CZTS, have the best structural, compositional, and optical properties, confirmed by GIXRD, Raman, SEM-EDX, and transmission spectroscopy. For the other two sulfurized films, the SnS2 secondary phase is observed in both samples in the Raman analysis. The Cu2S (PLD)/sulfurized sample displays a wider band gap (1.70 eV) than the two other films, which have an optimal Eg (between 1.40 and 1.56 eV) for use as an absorber layer in CZTS-based solar cell applications. In fact, ZnS being deposited by PLD limits Zn loss during deposition and annealing and helps preserve an important amount of the ZnS phase required for the CZTS phase formation. In summary, the MSPLD hybrid technique is beneficial in obtain-

Conclusions
CZTS thin films were co-deposited simultaneously using three targets (Cu 2 S, SnS 2 , and ZnS) in a hybrid synthesis system combining magnetron sputtering (for two targets) and pulsed laser deposition (for the third target). The effect of the heat treatment with and without a sulfur atmosphere on the material properties was studied. The results show that a simple heat treatment without any sulfur source leads to the evaporation of Zn and Sn and the apparition of CuS and CTS secondary phases in all the films, no matter the target arrangement used. On the other hand, the sulfurized samples show that annealing in a sulfur atmosphere is an essential process for the formation of the CZTS phase. It is found that the samples obtained by simultaneous deposition of ZnS by PLD, Cu 2 S, and SnS 2 by MS and sulfurized at 500 • C, which leads to the formation of a singlephase CZTS, have the best structural, compositional, and optical properties, confirmed by GIXRD, Raman, SEM-EDX, and transmission spectroscopy. For the other two sulfurized films, the SnS 2 secondary phase is observed in both samples in the Raman analysis. The Cu 2 S (PLD)/sulfurized sample displays a wider band gap (1.70 eV) than the two other films, which have an optimal E g (between 1.40 and 1.56 eV) for use as an absorber layer in CZTS-based solar cell applications. In fact, ZnS being deposited by PLD limits Zn loss during deposition and annealing and helps preserve an important amount of the ZnS phase required for the CZTS phase formation. In summary, the MSPLD hybrid technique is beneficial in obtaining high-quality CZTS thin films for use as absorber layers in solar cells.