Novel Evaporation Process for Deposition of Kesterite Thin Films Synthesized by Solvothermal Method

Kesterite, a quaternary compound of Cu2ZnSnS4, is a promising option as a material absorber to reduce the cost of photovoltaic solar cells. -e solvothermal method is a way to synthesize nanoparticles of this material. In this work, once synthesized, particles were deposited on a substrate through evaporation, and their morphological, structural, and optical properties were studied. Results show that changes of precursor ratios during solvothermal synthesis result in a modi8cation of particle morphology but not on its size. -e deposition of already synthesized kesterite through evaporation preserves kesterite structure and permits the formation of a homogenous 8lm on a substrate. Optical re9ectance and transmittance measurements allowed estimating the band-gap energy between 1.41 and 1.46 eV for representative samples, which is near the optimum for the absorber material.


Introduction
in-lm solar cells have emerged recently as a promising option to reduce costs in comparison to current solar photovoltaic technologies.
e thin-lm solar cells are composed of material layers where each di erent material layer has a particular function in the working cell.A simpli ed classi cation of these layers is as follows: the rst is a contact layer deposited over the substrate; the second, over the contact layer, is an energy absorber material; and the third, on the absorber layer, serves as a bu er.Finally, the last is a window layer made of two di erent materials, acting as a transparent electrode [1].Among all these material layers, the absorber material is the one that transforms photons into energy.erefore, absorber materials largely de ne thin-lm solar cells.Some absorbers have been extensively developed and have reached commercial scale, including amorphous silicon, cadmium telluride (CdTe), copper indium selenide (CIS), and copper indium gallium selenide (CIGS).However, these materials have some drawbacks; for example, amorphous silicon requires a lot of energy to be manufactured.Cadmium, on the other hand, is a heavy metal, which may cause environmental and health problems if it is used massively in the production of solar cells.Finally, CIGS cells are composed of elements that are relatively scarce in the earth crust.erefore, the cost of materials may become a problem in the massive development of a ordable photovoltaic cells.Because of these problems, there is an ongoing research to develop substitute options.Among all these options, kesterite has emerged as a promising substitute for CIGS.Kesterite is a quaternary compound material formed by copper, zinc, tin, and sulphur, with the acronym CZTS.Kesterite was rst reported by Ito and Nakazawa in 1988 [2].Research on methods to prepare kesterite thin lms can be classi ed into three types: (1) Research whose purpose is to synthesize materials, for example, nanoparticles [3][4][5].(2) Research about deposition of material once it is synthesized.Some of these deposition methods already reported are roll to roll process [6] and spray coating [7].(3) Finally, research about methods used for simultaneous synthesis and deposition of several materials over the substrate [8][9][10].In the present work are reported CZTS compounds synthesized by using the solvothermal method and deposited over a substrate by thermal evaporation.Usually, thermal evaporation is applied using elemental metals in a sulphur atmosphere [11,12].To the best of our knowledge, the present paper reports for the rst time a CZTS compound applied directly through thermal evaporation without a sulfurized vapor lling the atmosphere.

Materials and Methods
2.1.CZTS Solvothermal Methods.CZTS was synthesized following the experimental procedure reported by Zhou et al. [13].e procedure consists in the use of ethylene alcohol, copper (II) chloride dihydrate (CuCl 2 •2H 2 O), zinc (II) chloride (ZnCl 2 ), tin (IV) chloride tetrahydrate (SnCl 4 •4H 2 O), and thiourea (CH 4 N 2 S) which were all of analytical grade from Sigma-Aldrich, and Polyvinylpyrrolidone (PVP) was a product of Sigma.e experimental procedure included 2 mmol copper (II) chloride dihydrate, 1 mmol zinc (II) chloride, 1 mmol tin (IV) chloride tetrahydrate, and 5 mmol thiourea.ree different samples were prepared with polyvinylpyrrolidone (PVP): (1) one of these samples with 0.640 g of PVP, (2) a second sample with 0.324 g, and (3) a third sample with 0.964 g of PVP.All three samples were prepared with 38 ml of ethylene alcohol and loaded into a 40 ml Te on-lined stainless autoclave and kept in an oven at 200 °C for 24 hours.e resultant materials were washed several times with distilled water and centrifuged, and the materials washed with water were washed again with absolute ethanol and centrifuged.e resultant materials were characterized and applied to an Mo-glass and Corning glass square substrates of 25 × 25 mm.ese materials were deposited on the substrates via a Cressington Carbon Coater.In fact, this method was modi ed to improve the deposition process on the substrates.In order to control the quantity of the deposited material onto substrates were varied both the time and the magnitude of the electric current (amperage) applied.In the method used to evaporate the synthesized materials, these materials were rst preheated by applying a current of 20 A to the electrode in the experimental setup.Next was applied an electric current of 21 A for 10 s.Finally, the current was increased by 1 A for every 10 s until it reaches 35 A.

CZTS Characterization.
Morphology and size of the synthesized materials and the deposited lms were characterized by scanning electron microscopy (SEM; JEOL JSM-7000F) equipped with high resolution energy dispersive X-ray spectroscopy (EDS, EDAX Genesis); particles were dispersed in absolute ethanol using sonication for ve minutes and deposited on a silicon wafer.X-ray di raction (XRD) was made using a PANalytical X Pert Pro X-ray di ractometer, with a scanning rate of 0.1 °/2.5 s, using CuK α radiation and a wavelength λ of 1.540 nm.In addition, selected samples were characterized by high-resolution transmission electron microscopy (HRTEM; Jeol JEM 2200FS), equipped with spherical aberration correction at an operation voltage of 200 kV.On the other hand, thin lms were also characterized by using an FIB equipment (Focus Ion Beam, FEI Quanta 3D FEG Dual Beam, operation voltage 30 kV).By using UV-Vis-NIR spectroscopy were measured the optical transmission and optical re ectance in the UV-Vis-NIR ranges employing an Agilent Cary 5000 spectrophotometer.e optical transmittance T was measured at normal incidence, and the optical re ectance R was obtained using an absolute specular re ectance accessory with an incident light beam at 7 °to the sample surface.From these measurements, the optical absorption coe cient α in the UV-Vis-NIR ranges of the electromagnetic spectrum of each one of the samples was obtained.en, the direct band gap for each one of the samples was evaluated from the absorption coe cient.e samples were characterized by Raman spectroscopy using an excitation wavelength 632.8 nm emitted by a laser power of 9 MW in a Horiba model Labram HR Vis633 spectrometer.In order to avoid structural changes due to the thermal e ect, the Raman measurements were made at lower power density.

Results and Discussion
3.1.Morphology Characterization. Figure 1 shows SEM images of the morphology and size of prepared particles.As it can be seen from Figures 1(a)-1(g), as PVP increases, the morphology of particles changes from needle-ower to spherical shape. is is in accordance with the results reported by [14].e obtained results in the present research show how particles synthesized reached nanoscale size, but they remain sintered.erefore, PVP does not control particle size and the sintered characteristic of them, but it does control its morphology.Flower-shaped material shown in Figure 1(a) was also characterized by HRTEM.Figures 2(a)-2(c) show a detail of the ower-like morphology of the particle shown in Figure 1(a).
From the EDS analysis of the synthesized materials, it was observed a 20% reduction of copper, regarding the stoichiometry values, a 16% reduction of zinc, and a 11% increase of sulphur, while tin remains close to its stoichiometry value.In order to verify the elements present in the sample synthesized with 0.964 g of PVP, shown in Figure 1(g), additionally, it was characterized by means of a particle elemental analysis.e results concerning the composition of this sample are shown in Figure 3.In this last gure, it can be appreciated that all the elements from the quaternary compound, copper, zinc, tin, and sulphur are present in the correct proportion in the particle.erefore, these results support the identi cation of this particle as kesterite.Figures 4(b)-4(e) show the spatial distribution of constituent elements Cu, Zn, Sn, and S, respectively, corresponding to sample (a) synthesized with 0.964 g of PVP. Figure 5 shows a scheme describing the e ect of PVP on particle formation during the synthesis of CZTS.
e precursors CuCl 2 , ZnCl 2 , and SnCl 2 provide copper, zinc, and tin, respectively, while thiourea is the source of sulphur.All these components were kept constant, while the quantity of PVP was varied from 0.234 g to 0.964 g. e results show changes in the particle morphology from ower like to sphere like.However, in order to verify the crystalline structure and their di erences, an XRD analysis was completed, and the 2 Advances in Materials Science and Engineering crystallite size of the samples was determined by the Scherrer equation, being required the deconvolution of peaks to determine the full width at half maximum (FWHD).e samples were also characterized by EDS analysis as powders once the materials were evaporated over the substrate.e results are shown in Table 1.

XRD Characterization.
e results of the X-ray diffraction study, shown in Figure 6, allowed identifying the crystallographic structure of the samples synthesized with di erent amounts of PVP (0.3324 g, 0.640 g, and 0.964 g).In this gure, it can be appreciated that the peaks of the three di raction patterns coincide exactly with those of the pure phase of kesterite structure (CZTS) (JCPDS 01-75-4122).
No peaks of impurities were detected.In other words, the structure of CZTS is the same for all samples, despite having di erent percentages of PVP and morphologies.In Figure 6, it can be seen how the main di raction peaks are present in all samples, supporting the assumption that despite any changes in PVP, kesterite is formed.Based on di raction data, the crystallite size was calculated according to the Scherrer formula, and the necessary information such as FWHD and the degrees was solved through deconvolution of the peaks using Gaussian curves.

Evaporation of Materials
Over a Substrate.Synthesized materials were deposited over a Corning glass and an Mocoated glass.Figure 7 shows an assembly of CZTS deposited    over a glass substrate.e sample chosen to make the assembly was that containing the particles with the smaller size, which were prepared with 0.324 g of PVP.By using the method of evaporation, it was possible to deposit thin lms with a thickness of the order of 50 nm.In a typical assembly obtained in this work, shown in Figure 7, molybdenum was deposited through a sputtering process, and CZTS were deposited through evaporation.in lms deposited over the glass substrate were prepared using an focused ion beam (FIB) in order to see the cross section using HRTEM.In Figure 7, it can be seen that the deposited material is homogeneous.In fact, the di erent layers of the same material were deposited in di erent stages varying the amperage and deposition time as it was previously discussed in Section 2.1.ree di erent powders synthesized with di erent PVP contents were deposited through physical evaporation over a glass substrate and characterized through SEM and EDS, and the results are shown in Table 1.XRD results were compared with the EDS data for each of the synthesized material as well as the same material deposited through evaporation method.Results indicate that there are small changes in crystallite size according to the PVP content, but there are major changes in the stoichiometric ratio of the elements when they undergo a physical evaporation phase.
e most important changes are the increase in the Zn content; there is also an increase in the sulphur content and a decrease in the Cu content when they undergo this physical change through evaporation.

UV-Vis-NIR Spectroscopy.
UV-Vis spectroscopy studies included measurements of optical transmittance and optical re ectance in the wavelength ranging from 200 to 2300 nm as shown in Figure 8, for the CZTS assembly synthetized with 0.324 g of PVP and deposited through thermal evaporation on a glass substrate.As can be seen in Figure 8, one of the representative assemblies of CZTS studied in this work is highly transparent above a wavelength of 350 nm.Next, these measurements of transmittance and re ectance were used to obtain the absorption coe cient α which was used to calculate the direct band gap of CZTS assembly using the relation: (αh]) 2 A (E g -h]), where ] is the frequency of the incident radiation, h is Plank's constant, E g represents the optical band gap, and A is a constant [15].Figure 9 shows the behaviour of (αh]) 2 as a function of h] for the sample of CZTS with 0.324 g of PVP deposited through thermal evaporation on glass.e E g value was evaluated by extrapolating the linear part of the plots (see solid line) to the abscissa (h]).In the case of CZTS assembly synthetized with 0.324 g of PVP, the values of E g calculated was equal to 1.46 eV.In fact, the values of E g of all the CZTS assemblies synthetized in this work were found in the range between 1.41 and 1.46 eV, which is very close to the optimum for solar cell assemblies.
e maximum experimental error found in the calculation of the band-gap energy was estimated in 0.01 eV.
ese results open the possibility that CZTS samples synthesized in this work can be used as an e cient material in solar cell assemblies.

Raman Scattering Analysis of Kesterite.
Figure 10 shows the spectrum measured on three di erent powders synthesized with di erent PVP content.From top to bottom, they appear in decreasing order of PVP content, being spectrum marked as (a), the one that was synthesized with the largest PVP content (0.964 g).Spectrum (b) shows a powder synthesized with 0.640 g of PVP, and nally spectrum (c) shows a kesterite powder synthesized with 0.324 g of PVP.
e tting of the spectra with Gaussian curves was also shown in Figure 10 and was allowed to identify two main peaks at 272 cm −1 and at 322 cm −1 ; the position of these two peaks remain constant in all the three different powders.e position of these two peaks are a good approximation of the signals at 289 cm −1 and 339 cm −1 identi ed as kesterite peaks [16].
e peak at 272 cm −1 corresponds with the experimental Raman band identi ed in kesterite with a symmetry E(LO).
e band at 322 cm −1   2 versus h] spectra for the calculation of direct band gap of CZTS assembly synthetized with 0.324 g of PVP and deposited by thermal evaporation.6 Advances in Materials Science and Engineering corresponds with a symmetry B(TO).B mode is nondegenerated, and E mode is doubly degenerated.While the B mode involves in cation displacement along the C-direction, E mode involves in plane cation vibrations parallel to the a direction.e normal displacements of the 24Ѓ point vibrational mode predict that B(TO) is produced by a tin and sulphur couple, and the E mode involves a zinc and sulphur [17].erefore, the increase in PVP quantity involves a growth in the vibration of the zinc/sulphur and a reduction in copper.

Conclusions
Kesterite materials were successfully synthesized through a solvothermal method.It was found that the content of PVP allows controlling the particle morphology, but not their size.For the rst time, CZTS materials were deposited using thermal metal evaporation controlling its thickness by means of a method based on variations of the electric current intensity and the deposition time.From the HRTEM characterization, it was observed that the deposited materials are very homogeneous.erefore, the quaternary compound (CZTS) resulting from solvothermal synthesis can be directly deposited onto substrates using the method described in this work.Raman sensibility was key to identify how materials change their structure, according to PVP content.However, these changes do not cause major modi cations in the interatomic distances, making it di cult for XRD di raction, to re ect them.For the rst time, a physical evaporation method for CZTS deposition is reported, allowing the control of Cu/Zn ratio to obtain an ideal composition for kesterite solar cells.is method will add a new tool to design better photovoltaic devices, since it allows obtaining extreme homogeneous thin lms as well as the possibility to deposit materials over exible substrates.e calculated band gap of CZTS materials deposited through thermal evaporation suggests that this material and the method described in this work can be used for e cient solar cell assemblies.Advances in Materials Science and Engineering

Figure 1 :Figure 2 :Figure 3 :
Figure 1: SEM images of morphology and sizes of prepared particles, (a) and (e) correspond to samples with 0.324 g of PVP; (b), (c), and (d) correspond to samples with 0.640 g of PVP, and (f ), (g), and (h) correspond to samples with 0.964 g of PVP.

Figure 4 :Figure 5 :
Figure 4: Particle elemental analysis of a sample synthesized with 0.964 g of PVP that shows that the principal elements of kesterite are found in the entire particle.(a) Morphology of the sample.EDS map of (b) copper, (c) tin, (d) sulphur, and (e) zinc.

Figure 6 :Figure 7 :
Figure 6: XRD patterns of CZTS samples prepared with di erent concentrations of PVP which show a good matching between the peaks in each one of the di raction patterns and those of the pure phase of kesterite structure (CZTS) (JCPDS 01-75-4122).

Figure 8 :
Figure 8: Optical transmittance spectra and optical re ectance spectra of CZTS assembly synthetized with 0.324 g of PVP and deposited through thermal evaporation on glass substrate.

Figure 9 :
Figure9:(αh])2 versus h] spectra for the calculation of direct band gap of CZTS assembly synthetized with 0.324 g of PVP and deposited by thermal evaporation.

Figure 10 :
Figure 10: Raman scattering analysis of the CZTS using di erent PVP contents in their synthesis.Spectra (a) to (c) are ordered from higher to lower content of PVP used during synthesis of the source powders, the PVP content being (a) 0.964 g, (b) 0.640 g, and (c) 0.324 g.Characteristic kesterite Raman peak positions are indicated on top, as well as experimental values resulted from Gaussian analysis.

Table 1 :
XRD di raction study and EDS analysis of the CZTS synthesized with di erent PVP contents.SampleX-ray di raction crystallite size (nm) Electron dispersion spectroscopy (atomic %)