Thin film chalcogenide photovoltaic materials

Resume : Due to its high scalability and low production cost, CdTe solar cells have shown a very strong potential for massive energy production, however tellurium scarcity is a limitation for this technology. We have already demonstrated efficiencies above 11% for devices with 1.5 micron thick CdTe. Moreover a different electrical operation has been shown, considering that the depletion region takes a very large part of the device. In this work, many CdTe solar cells, made with different absorber thickness, were prepared for different accelerated lifetime tests, showing different reactions to the aging and in particular a remarkable stability as CdTe thickness reduces with a standard Cu/Au back contact. Among solar cells with absorbers from 0.7 to 6 microns thickness, with efficiencies ranging from 8 to 14%, the thin absorber devices show negligible reduction of efficiency under dark and open circuit (Voc) conditions for more than 9000 hours, whereas in the same conditions, for standard devices, the efficiency reduces up to 20% in less than 1000 hours. The solar cells have been also tested in accelerated lifetime conditions, one sun illumination and 85°C, in both short circuit (Jsc) and open circuit, showi ng different behaviors in stability as the CdTe thickness changes: Statistics of several tens of cells also with different amount of copper in the back contact and analysis of the electrical parameters will address the key points to this phenomenon, important for module production.

Resume : CdTe has enjoyed the strongest commercial success of all thin-film solar cell technologies despite its historical moderate conversion efficiency. CdTe's excellent manufacturability attributes, including extremely high deposition rates, ease of uniform large area processing, and forgiving process tolerances have been instrumental in establishing capable manufacturing at a strong cost-performance ratio. The highest reported CdTe research cell efficiency has increased by more than absolute 3% over the last 2.5 years and at the time of this writing stands at 19.6%, as compared to the S-Q limit of nearly 33%. Device performance improvements over the past 20 years have been dominated by Jsc and FF increases. Voc has been largely stagnant around 850mV and only recently been taken above 880mV in very high efficiency devices (>19%) and above 915mV in a polycrystalline CdTe high-voltage device. Recent research indicates that Voc at these levels is strongly correlated to minority carrier lifetime, which it is believed can be improved through techniques currently in development. If successful, Voc of 925mV with modest improvements in Jsc and FF should enable a 22% CdTe cell in the near-term without assuming any radical change to materials or device architecture. Based on this visible trajectory for research cell performance and CdTe's relatively narrow gap between research cell and large area efficiency, module performance can be expected to reach 17% in the coming few years. Such conversion efficiencies will put CdTe field performance significantly ahead of commodity crystalline Si module technologies at real operating temperatures which can average 50C to 60C on a power-weighted basis in typical dessert conditions.

Resume : CdS layers were deposited by close space sublimation (CSS) and chemical bath deposition (CBD) methods. CdTe layer had been grown by CSS on the two CdS layer types. Further, CdS/CdTe interfaces were investigated by HRTEM and XPS. TEM observations of cross sections showed large grain sizes in the range of 50-70 nm for CSS CdS layer. The interface between CSS CdS and CdTe appeared clear and sharp, like an abrupt hetero-junction. Besides, CBD CdS layer had very small grain sizes in the range of 5-10 nm. The interface between CBD CdS and CdTe was not as clear as CSS CdS. XPS analysis on Cd 3d and S 2p core levels were performed in both cases while growing the CdTe layer. In the case of CSS CdS layer, a sudden shift to lower binding energies was observed in relation with the CdTe layer coverage. Whereas, Cd 3d and S 2p core levels binding energies were gradually shifted with CdTe coverage for CBD CdS layer. In addition, XPS depth profile analyses indicated that a higher diffusion might occur at the CBD CdS/CdTe interface. Finally, CdTe solar cells based on CSS CdS layers exhibited higher efficiencies than those based on CBD CdS layers. The relationships between the solar cell performances and properties of CdS/CdTe interfaces will be discussed on the basis of HRTEM and XPS analyses.

Resume : Cu(In,Ga)(S,Se)2 (CIGS) has held during the last decades the title of “promising technology” to achieve high efficiency photovoltaic modules at low cost. CIGS has demonstrated its capabilities at the laboratory scale with a record efficiency of 20.8% [1] in small cells, overtaking the highest efficiency of multicrystalline Si cells. However, the industry has not reached maturity mainly due to the large gap between record small cell and commercial module efficiencies. One of the main advantages of electrodeposited stack metal layers for precursor formation to form CIGS is the high homogeneity and tight control of thickness and morphology of the precursors. Moreover, as the deposition is done under non-vacuum conditions the quality control process monitoring of the electrodeposited layers can be done in-line and in-situ under real-time conditions. This is achieved with non-contact techniques which have been developed for this application. This fact provides a clear advantage to perfectly asses the quality of precursors arriving at the step of thermal treatment under the chalcogen species. The most important challenges in terms of material, deposition process, characterization, and final fabrication of photovoltaic modules are strongly dependent on the precursor formation kinetics and its final quality, as well as the thermal conditions the stack layers face before absorber formation.

Resume : In this work, hydrazine-hydrate solution-based methods have been successfully employed to fabricate Cu2ZnSnS4 (CZTS) nano-particle. CZTS precursor thin films on Mo-coated glass substrates were grown by the doctor-blade coating technique. The CZTS thin films were selenized at temperatures of 550oC with Se vapor in quartz tube furnace. Microstructure investigation and A-modes speculation of Cu2ZnSnSe4 (CZTSe) thin films have been systematically studied with the increase of selenization ramping rate and temperature. Selenization was performed at temperatures of 550 oC with the ramping rate in the range 5 °C/min to 30 °C/min. The Raman peaks at 170 cm-1 and 192−194 cm-1 are found to have asymmetry signals. The Raman signal at 170 cm-1 is found to be composed of two gaussian fitting peaks at 168 cm-1 and 172 cm-1 are related with MoSe2 and CZTSe signals, respectively. The deconvolution of A-mode at 191-195 cm-1 Raman signal into two Gaussian peaks. The A-modes are decomposed of two symmetry gaussian peaks at A’-mode 191 cm-1 (FWHM 10 cm -1) and A-mode 194 cm-1 (FWHM 5 cm-1). The area ratio of gaussian peaks A/A’ increases with decreasing ramping rate. We attribute that broaden A’ Raman peak is probably originated from disordered-phase in matrix as well as A sharpen 194 cm-1 peak is kesterite ordered-phase in polycrystalline Cu2ZnSnSe4 grains.

Resume : The I2-II-IV-VI4 (I=Cu,Ag; II=Zn,Cd; IV=Si,Ge,Sn; VI=S,Se) series of quaternary chalcogenide semiconductors have drawn wide interest for their application as solar-cell absorbers. In this study we investigate the optoelectronic and structural properties of Cu2Zn(Sn1-xGex)Se4 (CZTGeSe) alloy compounds with x varying from 0 to 1 with a step of 0.1. Radiative recombination processes in CZTGeSe polycrystals were studied by using low-temperature photoluminescence (PL) spectroscopy. A continuous shift from 0.96 eV to 1.33 eV of the PL band position with increasing Ge concentration was detected. Recombination mechanisms responsible for the PL emission are discussed. The crystal structure and the lattice parameters of the CZTGeSe polycrystals were determined by using X-ray Diffraction (XRD) analysis. A linear decrease of the lattice parameter from a = 0.569 nm to a = 0.561 nm with increasing Ge concentration was detected. Raman spectroscopy analysis revealed unimodal behavior of the dominating A1 Raman mode that showed linear shift from 196 cm-1 to 205 cm-1 with increasing Ge concentration.

Resume : Cu2ZnSnSe4 (CZTSe), containing cheap and earth abundant elements, has a great potential for wide scale PV due to its direct band gap, possibility of p-type doping and high absorption coefficient. However, very little experimental evidence has been reported on the nature of its intrinsic defects. Improving the structural quality of CZTSe provides an opportunity to clarify the defect nature and electronic properties using optical spectroscopy. Thin films of CZTSe were synthesised by selenisation of magnetron sputtered metal precursor layers deposited on soda-lime glass. The X-ray diffraction and Raman spectra demonstrate single phase material. Photoluminescence (PL) spectroscopy is used to study the defect nature in a number of films of different structural quality. In lower quality films PL spectra show a broad low intensity band at 0.98 eV, which doesn’t shift with changing excitation power. In high quality material the band becomes narrower and two phonon replicae are resolved. Increasing excitation power shifts the band towards higher energies, suggesting that the mechanism is donor-acceptor pair (DAP) recombination. We propose a recombination model involving two pairs of donors and acceptors, which is supported by the Arrhenius quenching analysis and evolution of the DAP line spectral position with temperature. The band gap is determined from absorption spectra and observed to shift from 1.05 eV at 4.2 K to 1.01 eV at room temperature.

Resume : Cu2ZnSiSe4 (CZSiSe) is a promising potential alternative material to obtain a high band gap absorber for multi-junction thin film solar cell devices. Theoretically, CZSiSe can reach bandgap of up to 1.7eV, however not many work have been reported on the fabrication and characterization of this absorber material. In this work, CZSiSe absorber is fabricated using a multistep process sequence. In the first step, Cu and Si are sequentially deposited using evaporation techniques on molybdenum coated soda lime glass substrate. The sample is then annealed to form copper silicide. This is followed by sputter deposition of Zn. Finally, the sample is selenized using H2Se gas at high temperature to form Cu2ZnSiSe4. The composition of the CZSiSe absorber was controlled through the optimization of the thicknesses of Cu, Si and Zn stacked layers with the annealing temperature. X-ray Diffraction (XRD) was used to show the formation of copper silicide peaks at annealing temperature as low as 200 degrees. Scanning Electron Microscopy (SEM), Energy Dispersive X-ray spectroscopy (EDX), Photoluminescence Spectroscopy (PL) and optical measurement analysis were also used to characterize CZSiSe absorber material. An analysis of the electronic and physical properties of the studied CZSiSe absorber will be reported and discussed. Finally, we will show tests on n-type CdS/p-type CZSiSe based solar cells in order to assess the potential of this material as an absorber in thin film photovoltaic device

Resume : One of the most critical steps in the synthesis of solution-deposited Cu2ZnSn(S,Se)4 absorber layers is the inevitable annealing at high temperatures in order to obtain the polycrystalline semiconductor material. At this point the loss of Sn as a result of decomposition and evaporation of SnS has to be prevented by controlling the overpressure of these volatile phases. In this work the annealing process was performed in evacuated glass ampules thus offering the possibility to tailor exactly the Se partial pressure. Two series of annealings under different conditions were investigated: with an increasing amount of Se and with different precursor compositions. For the first series, XRD and XRF measurements show increased incorporation of Se with increasing amount of Se put into the ampules. The second series uses precursors with different metal ratios starting from copper poor to copper rich compositions (0.75 ≤ Cu/(Zn Sn) ≤ 1.11 ). SEM and EDX measurements reveal segregation of copper rich phases on top of the initial layer in the Cu-rich regime. All samples were processed to complete devices and IV-measurements were performed for electrical characterization. Device efficiencies up to 5.2% were obtained for copper poor and zinc rich precursor compositions and a comparably high amount of Selenium provided during the annealing process.

Resume : New solar cell absorber materials like Cu2ZnSnS4 (CZTS), are nowadays object of intense study in order to improve the power conversion efficiency of the devices. For that same purpose, in this work we prepared CZTS thin films and evaluated the influence of the sulfurization conditions (temperature, duration and evaporated sulphur mass) on their physical properties. The precursor structures were prepared with of eight periods each consisting on layers of the binaries (CuS, SnS2) and of elemental Zn. This approach allows a better control of the film’s composition in depth. Scanning electron microscopy revealed higher grain size for samples sulfurized at higher temperature and with more sulphur mass. The structural analysis, performed by X-ray diffraction and Raman spectroscopy, has shown that the dominant phase in all samples was the CZTS. The photoluminescence analysis of the samples with higher grain size and higher cell conversion efficiency showed that the dominant band corresponds to radiative recombination of electrons localized in tails of the conduction band and holes localized in acceptor energy levels. This type of transition is observed in CZTS for the first time. It was estimated an ionization energy for the acceptor levels of 284±8meV and a depth of 22±1meV for the electrons energy levels in the tail of the conduction band. The non-radiative de-excitation mechanisms were assigned to two channels, one involving a discrete energy level and another involving a band.

Resume : Cu2ZnSn(S,Se)4 (CZTSSe) absorbers are produced by two-step process by means of selenization of stacked precursors. The stacked precursors are prepared from Cu,ZnS,SnS2 targets by Ion Beam Sputtering deposition(IBSD) onto Mo-coated glass substrates with the stacking order :SLG/Mo/ZnS/SnS2/Cu .The stacked precursors are annealed at 550°C in a graphite box with Se and SnS powder. The effects of the amount of tin sulfide during selenization on the structure and morphology, as well as the electrical properties of the CZTSSe absorbers have been investigated. SEM studies reveal that the grain sizes decreased as the amount of tin sulfide increased. Raman spectroscopy studies show CZTSe peak intensity decreased and CZTS peak intensity increased as the amount of tin sulfide increased. EDX and ICP studies demonstrate that S/Se+S ratio which ranging from 0 to 0.5 can also be preciously controlled by tuning selenization condition. Device characteristics show that open circuit voltage increased obviously and short circuit current density almost kept the same when the amount of tin sulfide increased. The best cell performance can be achieved under the certain amount of SnS powder and the conversion efficiency is 7.2% with Voc=410mV , Jsc=30.61mA/cm2 and FF=0.58.

Resume : Thin films of Cu2ZnSnS4 (CZTS) have been produced with Pulsed Laser Deposition (PLD) using a KrF excimer laser operating at a wavelength of 248 nm. Pulsed laser deposition is one of the most promising techniques for achieving high quality films of complex stoichiometry. By this process all elements from the CZTS target are arriving simultaneously at the substrate, but if the target material includes volatile species, e.g. S or Sn atoms, the films may lose a considerable amount of these elements during the deposition. In this study we will investigate the loss of volatile elements during the transfer process as well as the crystalline quality of the thin CZTS films as a function of the laser energy (fluence range 1-3 J/cm2) and substrate temperature (up to 250 Co). Further loss of S and Sn during the post-deposition annealing will also be investigated. Compositional analysis is done with energy dispersive x-ray spectroscopy (EDX), the crystalline quality is investigated with Scanning Electron Microscopy (SEM) and X-Ray Diffraction (XRD) analysis. The band gap is evaluated with UV-visible spectroscopy. By this technique the use of the toxic and explosive hydrogen sulfide gas during deposition is avoided.

Resume : This work reports the analysis of structural and optical properties of Cu1.85CdxZn1-xSnS4 (0

Resume : Cu2ZnSnS4 (CZTS) is a quaternary semiconductor which attracted much attention in the recent years as environmental friendly and very promising absorbing material for photovoltaic applications. The best efficiencies of CZTS-based solar cells are obtained using a Zn-rich and Cu-poor composition, but the influence of stoichiometry on CZTS optical and microstructural properties has not yet thoroughly investigated. In this study, CZTS (kesterite) thin films with different compositions were prepared by sulfurization of precursors co-sputtered from CuS, SnS and ZnS targets. XRD, Raman Spectroscopy, EDX and SEM were used for microstructural, compositional and morphological characterization. XRD data were analyzed using Rietveld approach in order to estimate the amount of possible spurious phases as well as the Sn-site occupancy in the CZTS phase. The optical properties were investigated by spectrophotometric measurements and Photoluminescence Spectroscopy. Preliminary results show a dependence of the optical and microstructural properties on the tin content pointing to an improvement of the morphology and to an increase of the energy gap as the tin content increases. All the materials were also used as absorber layer in solar cells. Up to now the best material gave a device with a conversion efficiency of 5.7%.

Resume : Cu2ZnSnS4 coatings are widely studied for their use as absorber layer into photovoltaic devices. A small deviation from this stoichiometry may induce the formation of binary and/or ternary phases. Since the occurrence of such phase lowers the conversion efficiency of the cells, a detailed study of the properties of these binary and ternary phases is required. This communication aims to bring relevant information about the effect of the chemical composition of copper sulfide films on their structural, optical and electrical properties. The films were deposited on silicon and glass substrates by magnetron sputtering of a CuS target. The total pressure in the deposition chamber was adjusted from 0.3 to 2 Pa. As-deposited films are poorly crystallized as confirmed by XRD. An annealing treatment in air at 200°C allows improving the crystallization. An oxidation of the samples is evidenced when the annealing temperature reaches 300°C. However, no significant change is observed by Raman spectrometry after such annealing treatments. The increase of the deposition pressure induces a small variation of the film composition that does not influence the structure. The film morphology is strongly influenced by the increase of the total pressure. Optical properties of the films have been studied by UV-visible spectrophotometry. A maximum of transmittance is observed in the visible range. Finally using Hall effect measurements, the metallic character of copper sulfide films is evidenced.

Resume : Stacked Cu-Zn-Sn layers on Mo-coated glass substrates were prepared by galvanostatic electrodeposition as precursors for crystallizing kesterite Cu2ZnSnS4 films. Due to the nature of the MoOx passivating layer on the Mo surface, initial trials were particularly addressed to deposit strong adhesive precursors by firstly depositing Cu layers on Mo-coated substrates from acidic as well as alkaline solutions. It was found that the acidic Cu electrolytes was less ideal for Cu deposition on Mo-coated glass substrates since it led to the low Cu deposit coverage inhomogeneity in addition to extremely poor adhesive Cu deposits. In contrast, utilizing an alkaline Cu solution yielded homogeneous coverage Cu deposits having strong adhesive to Mo-coated glass substrates which was essential for subsequent acidic Zn or Sn galvanostatic electrodepositions. The crystallization of kesterite Cu2ZnSnS4 from metallic stacked Cu/Zn/Sn and Cu/Sn/Zn precursors was investigated as well. The crystallization was completed using a two-stage process namely sulfurization of precursors in a tube furnace under sulfur vapor at 550 ?C. On a basis of complementary structural and chemical compositional analysis, the interdiffusion of elemental phases occurred in the as-deposited stacked precursors as revealed by the presence of binary Cu-Zn and Cu-Sn intermetallic phases. Sulfurisation of precursors seems to crystallise Cu-Zn-Sn precursor films into films exhibiting predominant kesterite phase. This contribution demonstrates that the established metals gavanostatic electrodeposition technique has a technological potential for preparing kesterite precursors.

Resume : Although kesterite-based thin-film solar cells efficiencies gradually increase, many material issues are still under investigation. In particular, a full understanding of the cationic ordering in Cu2ZnSnS4 and its derivatives (so called CZTS) is probably needed to improve the performances of such devices. The present study deals with the solid state chemistry investigations of compounds with compositions close to Cu2ZnSnQ4 (Q=S and/or Se) on powder samples. From EDX and microprobe chemical analyses the stability domains of these compounds in the pseudo ternary Cu2Q-ZnQ-SnQ2 diagrams were determined. It is worth noticing that the selenide compounds appear to be much more flexible than the sulfide ones since larger shifts from the 2:1:1:4 stoichiometry are observed (towards Cu-poor Zn-rich compositions). For sulfide compounds the influence of the synthesis conditions (especially the cooling rate at the end of the thermal sequence) on the Cu/Zn disorder in the kesterite structure was investigated by NMR spectroscopy. In the case of the quenched sample, the broadening of the NMR peaks (119Sn and 67Zn spectra) is consistent with a high level of Cu/Zn disorder. Such a disorder has been confirmed by resonant X-ray diffraction on the corresponding single crystal. The same NMR investigation was done on Cu-poor samples and demonstrates that the level of disorder depends not only on the cooling rate but also on the nature of the substitution process which occurs.

Resume : In recent years the question about the polymorphic structure of CZTSe absorber films has become a matter of debate. Still there is no method available to clearly identify the present polymorph in a thin absorber film, e.g. kesterite and stannite. This is considered a critical issue as CZTSe is expected to show significant differences in its characteristics, e.g. optical band gap, depending on its polymorphic structure. In this work we investigate the crystal structure of epitaxial CZTSe films by polarized Raman measurements. The results allow for a distinction between the possible polymorphs kesterite and stannite which are expected to show substantially different Raman results depending on the polarization configuration in the Raman system. Our results indicate the presence of the kesterite as the main structure in the analyzed films. The epitaxial films are furthermore analyzed by ellipsometry, allowing for the extraction of the dielectric functions of the films. Hence, the resulting optical functions and constants can be assigned to the polymorph of the CZTSe material class, which was found by the Raman analysis. Further, we show ellipsometry results for polycrystalline CZTSe films with different composition. The results are discussed in terms of optical constants and band structure and are compared to the epitaxial reference results.

Resume : The bandgap of Cu2ZnSn(S,Se)4 can be tuned between 1.0 eV (pure selenide compound) and 1.5 eV (pure sulfide compound) by varying the S/Se ratio. In this study, we compare the variation of crystallization behavior of Cu2ZnSn(SxSe1-x)4 as a function of annealing conditions. First Cu-Zn-Sn-S precursors are deposited on glass/Mo substrate by a cosputtering of Cu, ZnS, and SnS. The crystallization behavior of these precursors during selenization is then studied at different temperatures between 400 and 600?C in order to have a better understanding of annealing on the mechanisms of formation of CZTSSe and secondary phases. The structures of these layers are investigated by XRD and Raman spectroscopy. The results indicate k?sterite phases containing both S and Se. The impact of the annealing conditions on solar cells properties will be presented.

Resume : Cu2ZnSnX4, where X = S or Se have the suitable optical band-gap energy of 1.0 –1.5eV, which match the preferred range of solar irradiation and large optical absorption coefficient of 104 cm-1, contain only abundant elements, which makes these materials promising candidates for engineering on their base of different high-efficient and low-cost photovoltaic devices Cu/Sn/Cu/Zn metallic multi-stacks were deposited onto Mo coated soda lime glass by DC magnetron sputtering and annealed under S+Se+Sn atmosphere. The absence of secondary phases at the surface was confirmed by Raman spectroscopy. The structural characterization of the thin films was carried out by grazing incidence X-ray diffraction (GIXRD) and a subsequent Rietveld analysis of the diffraction data using the FullProf suite [2]. The kesterite structure was used as starting model for the refinement procedure. The dependence of lattice parameters a and c on S/S+Se ratio is in good agreement with Vergard’s law. Microstructural analysis consisted in separation of strain and size effects for CZTSSe thin films, based on Williamson-Hall method [3]. The dependance of domain size and strain on the ratio of S/S+Se will be presented. [1] M.T. Winkler, W. Wang, O. Gunawan , Energy and Envir. Science. (2014) [2] Juan Rodriguez-Carvajal and Thierry Roisnel, www.ill.eu/sites/fullprof/ [3] G. K. Williamson, W. H. Hall, Acta Metall. 1, (1953), 22-31.

Resume : Epitaxial Cu2ZnSnSe4 (CZTSe) thin films have been grown via high temperature coevaporation on GaAs(001). Electron backscattering diffraction confirms that epitaxy is achieved in a wide compositional range. However, as we will show, different secondary phases are always present in the epitaxial layer. We present a detailed compositional study and the occurrence of secondary phases as a function of composition via Raman spectroscopy and photoluminescence. The main secondary phases are Cu2SnSe3 and ZnSe which grow epitaxially on top of the CZTSe. This is a direct consequence of the similar lattice constants of CZTSe, ZnSe and Cu2SnSe3. In addition, conventional scanning transmission electron microscopy measurements are used to study the occurrence of secondary phases in the epitaxial layers, the quality of the epitaxy and the GaAs/CZTSe interface. The photoluminescence and Raman measurements on the epitaxial samples will be compared to polycrystalline absorbers grown on molybdenum glass substrates. Finally, we present the first epitaxial CZTSe solar cells with a maximum power conversion efficiency of 2.2% an open-circuit voltage of 223 mV and a current density of 16mA/cm2.

Resume : In recent years, the Cu2ZnSnS4 (CZTS) and Cu2ZnSnSe4 (CZTSe) compound semiconductors are potential alternative materials to CuInGaSe2 (CIGS) for thin film photovoltaic absorber layers. This is because all of the elements in CZTS and CZTSe are abundant in the earth, and the bandgap energy is controlled by the ratio of S and Se. Therefore, environmental concerns are eliminated with these materials, paving the way for gigawatt scale mass production of solar cells. Despite their brief history in solar cell technology, CZTS, CZTSe, and Cu2ZnSn(S,Se)4 (CZTSSe) solar cell devices are rapidly advancing in the solar cell markets. For example, CZTSSe solar cells fabricated at IBM have achieved an efficiency of 11.1% using a hydrazine solution process [1]. Shin et al. have reported CZTS thin-film solar cell with an efficiency of 8.4% using a vacuum based thermal evaporation process [2]. Repins et al. have reported CZTSe solar cells with an efficiency of 9.2% using co-evaporation process [3]. In this work, CZTS thin film on glass substrate is grown by dipping-coat from Cu-, Zn- and Sn-xanthate solution as precursor materials. The samples are annealed under nitrogen atmosphere. X-ray diffraction (XRD), electron probe microanalysis (EPMA), scanning electron microanalysis (SEM), thermoprobe analysis and the four-point probe method are carried out. The XRD spectra indicate that a peak of CZTS (112) starts to observe at 150 °C. This temperature is lowest in non-vacuum process of CZTS film. The all samples indicate chalcopyrite structure and polycrystalline as evidenced by the XRD spectra. A value of full width at half maximum of (112) peak increases with increasing temperature. This indicates that an grain size increases with increasing annealing temperature. The samples are non-uniform composition such as Cu-poor and Zn-rich at low temperature and become S-poor with increasing temperature. Sulfur evaporates because vapor pressure of sulfur is high. The all samples are p-type conductivity by thermo prove analysis because Cu vacancy (VCu) defects is dominant in the samples from EPMA results. The resistivity increases with the increasing annealing temperature. It is assumed that this reason is due to decreasing carrier concentration. Donor type defects such as (Vs) increases with increasing the annealing temperature. [1] T. K. Todorov, J. Tang, S. Bag, O. Gunawan, T. Gokmen, Y. Zhu, and D. B. Mitzi, Adv. Energy Mater. 3, 34 (2013). [2] B. Shin, O. Gunawan, Y. Zhu, N. A. Bojarczuk, S. J. Chey, and S. Guha, Prog. Photovol: Res. Appl. 21, 72 (2013). [3] I. Repins, C. Beall, N. Vora, C. DeHart, D. Kuciauskas, P. Dippo, B. To, J. Mann, W. C. Hsu, A. Goodrich, and R. Noufi, Sol. Energy Mater. Sol. Cells 101, 154 (2012).

Resume : One major hurdle to production of Cu2ZnSnS4 (CZTS) thin films for photovoltaic applications is the narrow phase diagram region in which CZTS is expected as a single phase at most temperatures of interest. Unwanted secondary phases such as ZnS, CuxSnSx+1 and SnxSy are thus likely to be included in CZTS films independently of the chosen deposition technique. Identification by standard X-ray diffraction (XRD) of some of those phases is challenging since their diffraction peaks overlap with CZTS peaks. In this study we employ Raman spectroscopy to determine which secondary phases are incorporated in CZTS films grown by pulsed laser deposition (PLD) for a range of laser energies and substrate temperatures. Film properties, such as absorption coefficient, refraction index and thickness are extracted from ellipsometry measurements. The same set of properties is evaluated for chemical-bath-deposited CdS due to its important use as a buffer layer in chalcogenide solar cells. The validity of the optical model used to derive optical constants by ellipsometry is discussed in relation to results from direct measurement methods such as UV-visible spectroscopy, Scanning Electron Microscopy (SEM) and profiling. Identification of secondary phases in CZTS films under different PLD process parameters and their effect on optical constants is an important factor in optimizing the deposition process for production of high-efficiency CZTS solar cells.

Resume : Cu2ZnSnSe4 (CZTSe) and Cu2SnSe3 (CTSe) thin films have been produced using the two-stage process via the conversion of Cu-Zn-Sn multilayer precursors capped with thermally evaporated Selenium. Precursors with Cu/Sn ratios from 1.9 to 2.3 were converted using the same process conditions. In addition, Cu-Zn-Sn precursors with Cu/(Zn+Sn) ratio of 0.9 and a Cu/Sn ratio of 2, and binary Cu-Sn precursors, were converted at temperatures between 380 °C and 550 °C. X-ray diffraction (XRD), Energy Dispersive X-ray Spectroscopy (EDS) and Scanning Electron Microscopy (SEM) analyses indicated self-regulation of the Cu/Sn ratio and the formation of CTSe. Formation of this phase is consistent with reports on the formation path and the reaction with ZnSe to form CZTSe. The XRD analysis of the CZTSe layers was performed using the PowderCell 2.4 software. A Pseudo-Voigt function was used for fitting the peaks profile. The analysis yielded values for lattice parameters, a and c, a trend that increased to a maximum for Cu/Sn ratio of 1.9.The Williamson and Hall method was used to evaluate the strain and domain size in the CZTSe and CTSe. The dependence of the domain size on the conversion temperature and Cu/Sn will be presented.

Resume : Due to the instability of CZTS,Se in terms of decomposition and Sn loss at high temperatures, most of the existing approaches to grow high-quality CZTS,Se absorbers consist of a subsequent annealing, sulfurization or selenization step after the deposition of a CZTS precursor. To prevent this decomposition, a high chalcogen partial pressure has to be applied during the crystallization process. Herein we report about fundamental growth mechanisms and properties of CZTS,Se absorber layers, produced by selenization of various different precursors at high selenium partial pressures. We compare growth mechanisms of PVD deposited precursors, nanoparticles and nanorods by applying different temperature and selenium partial pressure profiles. Additionally the role of Na during grain growth is investigated in detail. Grain growth mechanisms are found to be substantially different for different types of precursors: The formation of CZTS,Se grains from nanoparticle precursors is dominated by the interdiffusion of cations and the existence of an unsintered layer, while grain growth of PVD precursors is dominated by nucleation of grains and stress and thus is much more depending on the elemental composition of the precursor. Cu-rich precursors tend to form large grains (~10 µm) under various selenization conditions while the grain growth in case of Cu-poor precursors is much more specifically depending on the parameters of selenization and the supply of Na.

Resume : GaN crystals of high structural quality are very much required for expanding applications in full color light sources and high power -high frequency electronics. However due to its extreme melting conditions GaN cannot be grown from stoichiometric GaN liquid. New melting data coming from very high pressure (up to 10.0GPa) and temperature (up to 3400K) experiments will be discussed in the context of theoretical simulations of GaN melting. GaN bulk crystals of high quality are therefore grown at pressures much lower than the one expected for melting. Sophisticated approaches (A-DEEP, VAS) based on HVPE on foreign substrates have been developed for obtaining free standing GaN wafers with quality and size sufficient for laser applications. The HVPE is at present, the only method supplying GaN substrates for industry. Its main advantage is high growth rate exceeding 100 m/h. Real bulk GaN crystals of very high quality are grown by ammonothermal method at moderate pressures of 0.1-0.3GPa and low temperatures of about 400-600oC. Development of this technology is limited by discouragingly low grow rate of about 1 m/h. This can be improved by increasing both pressure and temperature of the process. Higher rate of about 20 m/h can be also achieved in Na-flux system where pressures lower than 100MPa and temperature of about 850oC are used. The existing methods and their current state will be discussed. A new approach to GaN bulk crystallization based on growth by HVPE on Ammono-GaN seeds will be also presented. It will be shown that thick (d>2mm) GaN crystals with quality as good as the quality of the seeds can be grown with a rate exceeding 200 m/h. These studies are crucial for establishing physical limitations of real bulk GaN crystallization process by HVPE.

Resume : Adsorption of several molecular species, pertinent for crystal growth of the semiconductors at the polar surfaces of GaN, SiC and ZnO were investigated by DFT calculations. The investigated cases include adsorption of ammonia and hydrogen at polar GaN(0001) surface, hydrogen, silicon and carbon at SiC(0001) surface and zinc and oxygen at ZnO(0001) surface. The results of DFT calculations confirm recently obtained predictions that charge transfer between surface and the bulk of semiconductor may affect the adsorption energy. The process may change the adsorption energy, depending on the availability of empty states at the surface and pinning of Fermi level at the surface. In case of the nonpinned Fermi level, the adsorption energy depends on the doping in the bulk. Crystal growth from the vapor is reviewed showing that the adsorption of growing species leads to the increase of the adsorbate to the point where Fermi level is unpinned. Thus majority of the growth processes occurs at this condition, so that the adsorption depends on the doping in the bulk. This mechanism explains the dependence of the growth and doping on the Fermi level in the bulk. These predictions is verified by thermodynamic analysis of the growth of GaN, SiC and ZnO with application of DFT data.

Resume : Wafers cut from aluminium nitride (AlN) bulk crystals are most promising substrates for devices based on high Al content AlGaN epitaxial layers, due to their chemical stability, low thermal and lattice mismatch, and compressive strain to AlGaN layers. While AlN substrates are now available commercially, serious technological challenges still prevent mass production of AlN substrates having an industrially relevant diameter and defect density at reasonable cost. In this talk, we will present our status and progress in homoepitaxial AlN bulk crystal growth. AlN crystals are grown on N-polar basal plane AlN seeds prepared from spontaneously nucleated freestanding AlN crystals. The excellent crystal structural quality of the spontaneously nucleated AlN crystals is inherited in subsequent homoepitaxial bulk growth. In order to preserve the seed quality during bulk growth and to provide for single-crystalline diameter enlargement, seed backside evaporation, crystal cracking, and parasitic nucleation adjacent to the seed have to be prevented. These and other technological challenges are addressed in the presentation. Optical properties with corresponding impurity issues of the substrates are discussed. Finally, we show that proper surface preparation results in a smooth morphology of AlN layers grown by MOVPE on substrates sliced from the AlN crystals. Lasing of optically pumped AlGaN/AlN laser structures demonstrate the quality of the obtained substrates.

Resume : We present a high temperature coevaporation process to make efficient Cu2ZnSnSe4 (CZTSe) thin film solar cells. CZTSe decomposition at high temperatures is suppressed through a Sn and SnSe evaporation source. Selenium is supplied through a valved cracker source to maximize the Se reactivity. The process consists of three stages. A short first stage at 420°C ensures a high SnSe supersaturation and minimizes large amounts of ZnSe at the back. It is followed by a second stage at 470°C where the final Cu/Zn ratio is fixed. In the third stage Sn, SnSe and Se are supplied in excess to improve the crystal quality. The composition of the absorber can be controlled by the Cu and Zn flux solely while Sn and Se are incorporated in the film in a self-limiting way. Solar cell efficiencies as high as 7.2% have been achieved so far. We present a detailed compositional study and the influence of each stage on solar cell performance is described. The optical properties of the absorbers are analyzed via photoluminescence (PL) and Transmission/Reflection measurements. Both methods indicate that several materials with different bandgaps are present in the absorber. We observe a broad absorption spectrum and multiple transitions in PL. We therefore attribute the broadening of the low energy slope in the quantum efficiency to secondary phases in the absorber. The results achieved via this coevaporation process are compared with a precursor and annealing process with maximum efficiencies of 7.5%.

Resume : Secondary phase formation in a Cu2ZnSn(S, Se)4 (CZTSSe) based p-type layer for photovoltaic applications is one of the major problems which must be overcome to improve solar cell efficiency. To better understand the crystallization mechanism of secondary phases in the material system Cu-Zn-Sn-Se, we investigated the selenization of binary metallic layers as a function of temperature. The sputtered thin film precursors comprise Cu-Zn, Cu-Sn, and Zn-Sn with different sequence. A total six precursors were studied by time-resolved X-ray diffraction while temperature increased from 30 ˚C to 550 ˚C. Selenium had been always deposited on top of the metallic precursors in a separate thermal evaporation step. The observed reaction sequences were shown different results depending on the metals which are in first contact with selenium. From these experimental results, the mutual affinities of metals are determined, and a way of reducing the ZnSe crystallization is presented. Preferable precursor composed of ternary metallic layers with Se is also suggested in conclusion with respect to a stacking order of the metals.

Resume : Cu2ZnSnSe4 (CZTSe) is a potential replacement for the already commercialized Cu(InGa)Se2 (CIGS) thin film photovoltaic absorber. Nevertheless, it exhibits much lower photoelectric conversion efficiency (9.7%) than CIGS (20.9%). To further improve the record efficiency of CZTSe, optimization of the CZTSe/buffer hetero-junction is also of paramount importance to achieve a good device performance. Up to now few reports can be found about the CdS growth conditions and their effects on the optoelectronic properties of the CZTSe based devices. This work reports on the effects of annealing treatment under air of CZTSe/CdS hetero-junction after the CdS chemical bath deposition and the impact on the electro-optical properties of the devices. Parameters such as annealing temperature and time were investigated. A complete characterization of CdS layers grown onto glass substrate will be presented using X-ray diffraction, Raman spectroscopy and scanning electron microscopy. An increase from 2.2% for the non-thermally treated junction, to 6.1% efficiency for the annealed one was achieved. Further optimization of the processes allows us to obtain a champion cell with 7.3% efficiency. Additionally, the CZTSe/CdS hetero-junction properties were investigated combining external quantum efficiency measured under different conditions (bias voltage, white bias light, monochromatic bias light) and X-ray photoelectron spectroscopy, showing remarkable differences before and after annealing.

Resume : Silicon carbide has been thoroughly studied in various crystal growth processes. The common method for SiC epitaxial growth is based on chemical vapor deposition that is typically carried out at 1500-1600 degrees for research in transistor applications. There are benefits in high temperature growth while also challenges appear. Interestingly, we have explored a high temperature sublimation epitaxy approach at 1800-1900 degrees for growth of fluorescent silicon for white LED layers with potential in general lighting. SiC doped with certain elements act as a monolithic rare earth metal free light converter and produces a pure white light. We have also initiated growth of cubic SiC, which doped with boron fits nicely in to the model of intermediate bandgap solar cell theory. The quality of cubic SiC has been an obstacle in use of this polytype. We have shown that the cubic SiC can reach similar quality as produced in commercial hexagonal polytypes. This may open a new research area in SiC, in particular for studies of the photovoltaic properties. In addition, graphene is an emerging material. Our high temperature epitaxial graphene process carried out at 2000 degrees, which is more than 300 degrees higher than in other methods, has shown an outstanding quality. Potentially, it could be a contact material on photovoltaic cubic SiC. We describe the challenges with high temperature crystal growth of fluorescent and photovoltaic SiC, as well as graphene on SiC.

Resume : Second generation of the solar cells is based on thin films of e.g. cadmium telluride (CdTe), material with a direct optical band gap of 1.45 eV and a high optical absorption coefficient. This generation of photovoltaic (PV) devices is commercialized and has chances to compete with PV devices based on silicon. In this work we describe the properties of single-crystalline CdTe-based solar cells that contain zinc oxide (ZnO) or cadmium sulphide (CdS) films as an n-type partner to p-type CdTe. In first type of deposited by us structures we verify if zinc oxide can replace commonly used cadmium sulphide. In the second we tested use possibility of deposition CdS films by Atomic Layer Deposition (ALD) method. In both cases structures were covered with TCO films based on ZnO layer doped with Al. In addition to structures mentioned above, we also investigated another possibility of using ZnO-based films in PV devices. We deposited low resistivity ZnO films by the ALD as ohmic contacts to p-type CdTe, CdZnTe and ZnTe. As a metal contact gold, palladium and copper prepared by various deposition method were applied. Present study confirms that high conductivity ZnO films can be used as a low resistivity ohmic contact and transparent electrode (TE) to p-type tellurides, especially to ZnTe. This work was partially supported by the Innovative Economy grant (POIG.01.01.02-00-108/09, the National Centre for Research and Development grant (PBS1/A5/27/2012).

Resume : Recent achievement of over 20 % cell efficiency (AM1.5, 0.5 cm x cm) in high-performance Cu(InGa)Se2 (CIGS)-based thin film solar cells is very promising. Typically, CIGS cell has a structure of SLG/Mo/CIGS/CdS/i-ZnO/ZnO:Al, and its current can be limited by several optical and collection losses, i.e., shading from front grid, reflection from ZnO, absorption in ZnO and CdS layers, and incomplete generation and collection in CIGS absorber. In this paper, rare-earth doped ZnO (ZnO:RE) layer was employed to replace i-ZnO in the conventional CIGS cell structure for the potential use as a down-shifting or down-converting layer. Firstly, pure and RE (Yb, Nd and Pr)-doped ZnO films were deposited onto Si wafer, quartz and glass substrate by magnetron sputtering system starting from a Zn target covered with small pieces of RE, and then characterized by several techniques including XRD, Raman, SEM, TEM, RBS, UV-VIS, PL etc. For examples, the PL measurements revealed that the main emission peaks of Yb and Nd were detected in the near infrared region, while the primary emission peaks of Pr appeared in the visible light region. Then, ZnO:RE layer was inserted into CIGS cell to form SLG/Mo/CIGS/CdS/ZnO:RE/ZnO:Al structure, where CIGS was prepared by 2-step sputter-selenization process and CdS was deposited by standard chemical bath deposition. Device performance of CIGS cell with RE-doped ZnO layer was characterized by QE and I-V measurements.

Resume : Recently developed CIGS thin film solar cell has ZnO/CdS/CIGS/Mo/SLG structure and the CdS buffer layer plays an important role for device performance. However, effect of such buffer layer on PV cell characteristics is not still clear. Especially, absorption spectrum of the cell is very hard to observe due to the presence of opaque Mo back contact. We have developed the photothermal (PPT) methodology for solving this problem by using a transparent transducer LiNbO3 for a detector and could observe the absorption spectra of the CIGS film samples by measuring the nonradiative transitions of photo excited carriers. We, then, investigate the effect of buffer layer on the absorption edge comparing with the photoluminescence (PL) measurements. Different buffer layer of CdS and ZnS were grown by chemical bath deposition technique. The Ga/(Ga+In) ratio was determined to be 0.3 by an XPS measurement. PL spectra showed dominant peak around 1.05 eV at 4 K and no difference was found for three samples. However, we found the absorption edge of 1.12 eV estimated from the square root plot for the absorption spectra shifted to the lower energy side about 0.04 eV by buffer layer deposition at room temperature. The observed red-shift indicates the band tailing conflicting to our knowledge that the presence of the buffer layer improves the interface conditions and reduces the nonradiative transition. Temperature variations of the PL and PPT spectra are carrying out for making clear the results.

Resume : In this work, we show that in order to optimize the efficiency of Cu(In1-x,Gax)Se2 (CIGS) solar cells with Cd-free ZnS-based buffer layers, the Ga concentration in the CIGS absorber layer towards the heterointerface has to be adapted. We varied the In and Ga deposition rates in the last stage of our 3-stage co-evaporation process leading to different compositional ratios of 0.15 ≤ x ≤ 0.6 in the top 400 nm of the absorber layer. All absorber layers were then completed with both CdS and ZnS buffer layers by chemical bath deposition. While cells with our standard grading of x ≈ 0.4 in the front region result in a best performance of 15 % with a CdS buffer, similar efficiencies with a ZnS buffer layer are only obtained when the Ga content near the heterointerface is reduced down to x ≈ 0.25. Interestingly, the maximum efficiency for the CdS buffer layer coincides with the maximum Voc and FF. For the ZnS buffer layer this is not the case: the Voc increases steadily for higher Ga ratios while the FF is fairly constant for 0.25 < x <0.5 and decreases drastically for more extreme values. Material and diode properties are extracted from chemical, structural and opto-electronic characterization and the results are explained on the basis of the electronic band structure and its alignment at the CIGS/buffer interface. The results illustrate the importance of the absorber layer adaptation for different buffer layers and are an important step on the way to Cd-free buffer layers.

Resume : Both the Ga gradient and the Cu content strongly influence the electrical properties and thus the conversion efficiency of Cu(In,Ga)Se2 thin film solar cells. However, the material composition also affects the atomic arrangements on the nanometer scale which deviate significantly from the average long-range crystallographic structure [1,2]. The anion displacement, in particular, depends sensitively on the kind of neighboring cations [1] and has a significant influence on the material bandgap [2,3]. We have therefore used extended X-ray absorption fine structure spectroscopy to compare the atomic-scale structural parameters of Cu(In,Ga)Se2 powders and polycrystalline thin films with varying composition. No difference in the two sample types is observed for the average Cu-Se, Ga-Se and In-Se bond lengths which are nearly independent of both the Ga and the Cu content. This demonstrates the strong tendency towards bond length conservation typical for tetrahedrally coordinated semiconductors. In contrast, the bond length variation is significantly smaller in the thin films than in the powders, particularly for Cu-poor material. This difference in the nanostructure is proposed to originate from differences in the preparation conditions, most prominently from the different history of Cu composition.
[1] C.S. Schnohr et al., Phys. Rev. B. 85, 245204 (2012).
[2] S. Eckner et al., Appl. Phys. Lett. 103, 081905 (2013).
[3] J. Vidal et al., Phys. Rev. Lett. 104, 056401 (2010).

Resume : In order to meet the increasing demand for energy, earth abundant and low-cost materials fabrication are utmost criteria for photovoltaic (PV) technologies. Very recently, I-V-VI ternary system (I= Cu: V = Sb, Bi; VI = S, Se), which contain inexpensive and earth abundant elements, has been suggested as an alternative material for PV thin-film technologies [1]. Therefore in this study, employing first-principles modeling within the density function theory, we analyze the electronic and optical properties of I-V-VI system. As per different stoichiometries, there are three sets of compounds, namely Cu(Sb,Bi)(S,Se)2, Cu3(Sb,Bi)(S,Se)3, and Cu3(Sb,Bi)(S,Se)4 in I-V-VI ternary system. Cu(Sb,Bi)(S,Se)2 and Cu3(Sb,Bi)(S,Se)3 compounds crystallize in orthorhombic structure, whereas Cu3(Sb,Bi)(S,Se)4 has a base-centred tetragonal structure. Our calculations show that the fundamental band gap energies of these compounds are in the region of 0.4–2.1 eV, which is suitable for PV solar cells applications. Furthermore, the flat energy dispersion of conduction band implies a large optical absorption, and the calculations reveal that the absorption coefficient of these compounds are stronger than other Cu-S based materials like CuInSe2 and Cu2ZnSnS4 [2]. Thereby, I-V-VI ternary system has the potential to be a suitable absorber material in thin-film PV technologies. [1] L. Yu et al., Adv. Energy Mater. 3, 43 (2013). [2] M. Kumar et al., Appl. Phys. Lett. 102, 062109 (2013).

Resume : Of the II-VI type binary compounds, zinc sulfide (ZnS) having a wide energy band gap is one of the most attractive semiconductor materials. ZnS is an important semiconductor material with a wide energy band gap (Eg > 3.7eV). It can be used for the fabrication of optoelectronic devices. In special, ZnS became an important material for a buffer layer in thin film hetero-junction solar cells because it does not have any environmental issue unlike CdS, which is used for a buffer layer of the thin film solar cells. It was reported that some amount of ZnO and Zn(OH)2 as the impurities, which are denoted in ZnS(O,OH), are required to be present in the ZnS thin films for fabricating the efficient photovoltaic devices. In this study, ZnS(O,OH) thin films were synthesized by a continuous flow microreactor (CFM) process designed by a modification of CBD process, and then the influences of the pH and annealing temperature on the properties of ZnS(O,OH) thin films were investigated. The ZnS(O,OH) thin films were prepared on several different pHs of zinc precursors. Control of pH was carried out using ammonium hydroxide. The mixing ratios of zinc sulfate heptahydrate and ammonium hydroxide were varied in terms of the pH of the prepared solution. The deposited ZnS(O,OH) thin films were annealed at 300℃~600℃. The films were characterized by XRD, SEM, EDX, XPS, and UV-vis spectroscopy. Their transmittance and the estimated optical band gaps are over 70% in the visible region and 2.9eV~3.7eV, respectively. Based on the XRD and EDX, it was confirmed that the CFM process enables us to carry out the band gap engineering of ZnS(O,OH) by varying the S/O ratio.

Resume : Sputtered intrinsic ZnO (i-ZnO) layer is essential to high-efficiency CIGS solar cells. However, its role in enhancing the efficiency has not been fully revealed, especially with respect of sputtering process conditions. In this work, its sputtering conditions were varied to have different O2 concentrations in sputtering gas (0%, 2%) and different sputtering powers (50, 100, 150 W). An incorporation of O2 increased substantially the efficiency from 15% to 17%. Open-circuit voltage, short-circuit current density, and FF were all improved in companion with the reduction of series resistance. Increase of the sputtering power also tends to improve the efficiency with a similar trend to the case of oxygen incorporation, but its effect is limited. CV profiling and admittance spectroscopy revealed that the oxygen incorporation increased the carrier density and reduced the defect energy from 120 meV to 35 meV, suggesting that it could modify the defect quality of CIGS surface. Electroluminescence (EL) spectroscopy proved that such changes occurred within CIGS layer, by showing the incorporated oxygen reduced the red-shift degree in EL peak (CIGS bandgap) with decreasing temperature. EBIC analysis showed that the carrier diffusion length in the quasi-neutral region of CIGS layer was significantly increased. In summary, the oxygen incorporation influenced through CdS directly CIGS bulk as well as surface. Such an oxygen effect seems related to higher Na doping level in O2 2% sample which is much higher near the interface of CdS/CIGS.

Resume : Recently, Ikeda et al. reported a CuSbS2 (CAS) solar cell with a conversion efficiency of 3.1% [1]. CAS is expected to be a promising photovoltaic compound because it has an optimal band gap of 1.52 eV and high absorption coefficient of >10⁴ cm^-1. CAS and CuSbSe2 (CASe) have a wurtzite-based chalcostibite structure, which is different from zinc-blend-based chalcopyrite and kesterite structures. In order to clarify the features of CAS and CASe, it is necessary to study their electronic structures. In this study, CAS and CASe’s band structures and density of states were calculated with the HSE06 hybrid functional and compared with that of chalcopyrite CuInSe2.
In the band structure, the valence band maximum of CAS is located at the Γ-point and the conduction band minimum is located at the R-point. However, the second lowest conduction band is located at the Γ-point and the energy level at the Γ-point is nearly equal to the R-point. Therefore, CAS has an indirect band gap (Γ-R) of 1.16 eV and direct band gap (Γ-Γ) of 1.25 eV. The partial density of states of CAS (CASe) shows the character of Cu 3d and S 3p (Se 4p) at the top of the valence bands and Sb 5p and S 3p (Se 4p) at the bottom of the conduction bands. In the conduction band, CAS and CASe have a p-orbital character (Sb 5p) that differs from the s-orbital character (In 5s) of CuInSe2.
[1] W. Septina, S. Ikeda, Y. Iga, T. Harada, and M. Matsumura, Thin Solid Films 550 (2014) 700.

Resume : As severe Schottky barrier formation occurs when CdTe is contacted directly to a metal, an interlayer is needed to provide good contact. Due to its strong p-type character, narrowing the Schottky barrier and therefore allowing holes to tunnel through, Sb2Te3 is a promising material. Furthermore, the use of such a Cu-free back contact material might enhance the stability of the solar cell. However, the typical fabrication of Sb2Te3 by co-deposition on a hot substrate can be challenging due to Te re-evaporation, significantly complicating stoichiometry and property control. In contrast to this commonly used method, our process presented in this work is based on a newly developed deposition technique, circumventing the mentioned problems and obtaining high quality back contacts. We discuss this deposition process and the obtained properties of single Sb2Te3 films in detail. We also present performance studies on superstrate-based CdTe solar cells with such Sb2Te3 back contacts. High quality single-phase Sb2Te3 with large grain sizes, carrier mobilities > 400 cm²/Vs and large Seebeck coefficients was obtained. Changing deposition parameters allows to precisely adjust carrier concentration, demonstrating the flexibility of the method. An efficiency of 11.7% was achieved with an Sb2Te3/Cu/Mo back contact and with an Sb2Te3/Cu/Au contact in a first experimental series with significant potential for further optimization.

Resume : Efficiency of CIGS-based electronic devices depends significantly on point defects. Among them, IIICu (group III element on copper) antisites are one of the most important because of its very low formation energy and strong coupling with lattice. Recent theoretical calculations predict that they are compensating donors and exhibit metastable character. In particular, defects related to IIICu are thought to be important close to the junction. However, these papers do not give an estimation of the corresponding energy barriers which are the most important parameters from the experimental point of view. In this contribution we focus on GaCu defects (in complexes with VCu) performing ab-initio calculations in LDA and LDA+U approximation. We calculate configuration coordinate diagrams for GaCu and GaCu in complex with one or two VCu. We found that in the case of point defect GaCu, the energy barrier for the relaxation of the DX state by a hole capture is low (~0,15 eV) while the relaxation to the DX state by an electron capture can be spontaneous. In case of complexes combining GaCu and VCu, energies of transition to DX state shift up. In effect, the barriers for a hole capture are smaller (comparing to isolated GaCu) whereas there appears the barrier for electron capture. These findings are very important as the energy barriers can be measured directly and can be used for an experimental verification of theoretical model of IIICu.

Resume : Compositional and structural properties of the absorber significantly affect the efficiency of Cu(In,Ga)Se₂ solar cells. The integral composition is commonly determined using X-ray fluorescence (XRF). However, the absorber is typically inhomogeneous and thus there is a particular interest in collecting information from defined spatial regions in the nanometer range. Hence, XRF mapping with an X-ray beam diameter of approximately 200 to 300 nm was used to investigate cross section of Cu(In,Ga)Se₂ solar cells. In order to implement this improved spatial resolution in the measurement, thin lamellas were prepared using a focused ion beam. A thin lamella offers the opportunity to collect information even from a single grain. Best results were obtained using a lamella thickness of about 250 nm. This enabled us to determine quantitatively the depth dependent composition of the absorber, particularly the Ga gradient, and its spatial distribution in high resolution.

Resume : In this paper, we report on the use of pulsed dc magnetron sputtering and compare the stoichiometry and optical and structural properties of the thin film CdS layers with those obtained using conventional Chemical Bath Deposition (CBD). This study allows us to compare the advantages and disadvantages of the new pulsed DC magnetron sputtering process. In both cases we have deposited the CdS on flourine doped tin oxide (FTO) coated glass substrates. The FTO was TEC10 and TEC 15 supplied by NSG Pilkington. The CdS thin film microstructure, composition and morphology was examined using Scanning Electron Microscopy (SEM), Transmission Electron Microscopy (TEM), Electron Back Scattering Diffraction (EBSD), Energy Dispersive X-ray Spectroscopy (EDS), X-ray Diffraction (XRD), X-ray Photoelectron Spectroscopy (XPS) and Spectrosopic Ellipsometry (SE). The deposition rate for CdS obtained using pulsed DC magnetron sputtering was 2.7nm/sec using a power of only 500W. The rate increases linearly with power. The deposition rate using CBD is comparatively slow. For example, a 50nm thick film of CdS required 30 minutes and a thickness of 150nm needs about one hour. The new pulsed dc sputtering process produced stochiometric CdS thin films which are crystalline with hexagonal columnar grains. The films are also highly uniform with an optical band gap of 2.3 eV. The use of an argon/oxygen plasma pre-treatment on the substrate results in pinhole and void free thin films

Resume : Low cost processes are very important problem for solar cell devices. The spray method is one of non-vacuum processes and is respected as low cost method. In our previous works, high quality transparency ZnO films were successfully grown at 100 °C by a conventional atmospheric spray pyrolysis using diethylzinc (DEZ) based solution [1]. The DEZ was diluted with diisopropyl ether to control its reactivity to air and water (supplied by Tosoh Finechem Corporation, JAPAN). Moreover, the growth of Ga-doped ZnO (GZO) /glass films was carried out by spray pyrolysis at 150 °C [2]. The samples had an average optical transmittance of more than 80% and were strongly a-axis orientated. The sheet resistivity of 30 Ω/sq. could be obtained. In this work, the GZO films were successfully grown on ZnO/CdS/CIGS/Mo/glass by spray pyrolysis using DEZ based solution. After covering clean SLG substrates with back electrodes of Mo films by sputtering, CIGS films were deposited using physical vapor deposition. Buffer layers of CdS films were prepared by chemical vapor deposition. Buffer layers of ZnO films (≈ 20 nm) were also prepared by RF sputtering method [3]. The efficiency of the obtained device was 10.3%. The short-circuit current density (Jsc) of 34.3 mA/cm2, open circuit voltage (Voc) of 0.50 V and fill factor (FF) of 0.60 are obtained. [1] K. Yoshino, Y. Takemoto, M. Oshima, K. Toyota, K. Inaba, K. Haga, and K. Tokudome, Jpn. J. Appl. Phys. 50, 040207 (2011). [2] Y. Takemoto, M. Oshima, K. Yoshino, K. Toyota, K. Inaba, K. Haga, and K. Tokudome, 50, 08801 (2011). [3] T. Minemoto, Y. Hashimoto, W. S.-Kolahi, T. Satoh, T. Negami, H. Takakura, Y. Hamakawa Solar Energy and Solar Cell Materials, 75, 121 (2003).

Resume : To correlate the device functionality with the respective material properties, both the absolute composition and the in-depth gradient in Cu(In,Ga)Se2 thin films need to be reliably determined. Common depth profiling methods rely on reference materials for comparison and are mostly destructive. Appropriate reference materials for graded Cu(In,Ga)Se2 solar cell absorbers are not available. To address this analytical challenge, synchrotron-radiation based Grazing Incidence X-Ray Fluorescence analysis (GIXRF) with calibrated instrumentation in conjunction with the fundamental parameter approach for quantification is used. The variation of the angle of incidence over a broad range provides quantitative access to the elemental in-depth distribution in the nano- and micrometer range. This GIXRF-methodology can be used for the quantification of elemental depth gradients in Cu(In,Ga)Se2 thin film solar cell absorber layers, in particular for the determination of different In to Ga gradients [1]. This methodology can substantially contribute to the qualification of upcoming solar cell production processes, e.g. for calibration samples for in-line monitoring methods. [1] Streeck et al., Appl. Phys. Lett. 103, 113904 (2013)

Resume : We report on the microstructure analysis of kesterite (CZTSe) layers from RTP-SEL (rapid thermal processing of sequential elemental layers) by spatially resolved CL (cathodoluminescence) in a SEM (scanning electron microscopy). In addition EDX (energy dispersive x-ray fluorescence), XRD (x-ray diffraction) and Raman spectroscopy were carried out for validation of the findings. Kesterite thin film samples were prepared on glass by DC sputtering of the Mo back side electrode and the absorber layer components Cu, Zn, Sn. The chalcogen Se was thermally evaporated on top. To study secondary phase formation, the layer sequence of Cu, Zn, Sn was varied in four ways. Rapid thermal annealing was carried out for all samples at ca. 550°C for 5 min in a quasi-closed graphite box. CL spectra exhibit broad emission lines between 0.8 eV and 2 eV with a superposition of distinct narrow lines, e.g. at ca. 1 eV corresponding to the energy gap of CZTSe. The broad emission is attributed to secondary phases, defects and disorder in the kesterite lattice. Strong indication for secondary phases was found due to emission around 1.35 eV and 2.0 eV. XRD and Raman spectroscopy clearly identified secondary phases, however, it was not yet possible to precisely attribute and relate this finding to the CL data. Nevertheless, the spatially resolved CL data allow the precise determination of the layer homogeneity which is believed to be a key issue for the optimization of the layer formation technology.

Resume : In this study, a comparison between Al:ZnO TCO for CIGS-based solar cells grown by Pulsed Electron Deposition (PED) and RF-magnetron sputtering (RFMS) has been performed. By means of PED technique, polycrystalline [002] mono-oriented thin films with low electrical resistivity and high optical transparency can be obtained from room temperature to 300 °C. The electrical resistivity of these films can be widely varied depending on the growth temperature with a minimum of 3.5x10-4 Ωcm at 150°C and an average transmittance over 90% in the VIS-NIR range. The optimized AZO film were successfully applied on CIGS-based solar cell obtaining an efficiency value of 15.2% This result clearly shows that the PED is a suitable technique for the growth of ZnO-based thin film in particular for devices/applications where low deposition temperature is required. On the other hand, an optimized AZO thin film front contact for CIGS-based solar cells was fabricated via RFMS. The parameters of the technique were tweaked to obtain highly conductive and transparent AZO thin films. The optimized, 890 nm thick AZO film was found to reach a resistivity value of 6.7x10-4 Ωcm and an average transmittance of 86% in the 400-1100 nm wavelength range.

Resume : Reactive sputtering deposition technique is suitable for large area, thus offering the possibility for more competitive industrial scale-up. In this work, we develop a hybrid one-step co-sputtering/evaporation Cu(InGa)Se2 deposition process, where Cu, In and Ga are sputtered simultaneously with the thermal evaporation of selenium, thus avoiding H2Se use. A parametric study has been performed by varying pressures and power values. With in-situ optical emission spectroscopy (OES), we analyzed the different species present in the reaction chamber and controlled the selenium evaporation temperature. A precise regulation of the Se flux allowed an accurate control of the film composition. We also pointed out the influence of the Se flux on CIGS films roughness and growth rates. OES has allowed us to determine the Se flux threshold so as to avoid targets poisoning. We measured film composition and thickness of the samples with X-ray fluorescence. Different phases formed during the process were identified by Raman spectroscopy and X-ray diffraction. Morphology and nucleation were studied by scanning electron microscopy (SEM). The optoelectronic cell properties showed promising efficiency of 10.3% for an absorber with composition ratios of [Cu/In+Ga] = 0.81 and [Ga/In+Ga] = 0.54.

Resume : CdS with its large band gap and chemical stability is an n-type semiconductor used as a window layer in many types of solar cells. CdS bi-layers were deposited on SnO2:F by using chemical bath deposition technique (CBD). SnO2:F substrates were treated with HCl (0.1M), Ar and O2 at 500°C and CdS thin films were thermal annealed on CdCl2, Ar and air atmospheres at 400°C. Optical and morphological measurements reveal that HCl treatment is good for CdS growth due to the defects levels are reduced and the grain size is increased respectively. CdS thin films obtained were applied on CdTe solar cells; the photovoltaic efficiency was increased from 4% to 11%.

Resume : Synchrotron-radiation based Grazing Incidence X-ray Fluorescence analysis (GIXRF) with varying excitation angles provides non-destructive access to the compositional depth profile of thin film matrix elements in the nano- and micrometer range. Reference-free GIXRF in conjunction with fundamental parameter based quantification allows for an analysis without the need for any calibration standards. This XRF-methodology can be used e.g. for the non-preparative determination of elemental depth gradients with an In to Ga gradient in Cu(In,Ga)Se2 thin film solar cell absorber layers. As a key metrological aspect, the uncertainty of the components of the effective solid angle of detection and its impact on the uncertainty of the detected count rate will be presented: with varying angle of incidence the irradiated area on the sample changes over two order of magnitude, the Gaussian shape of the beam leads to an intensity distribution and the field of view of the detector is dependent on the distance from the sample. The uncertainty of all components shows a different angle dependency. Therefore, a detailed uncertainty analysis and their implication is prerequisite for a reliable calibration procedure.

Resume : For pursuing higher performances of CZTSSe-based solar cells, it has been pointed out that the clarifications of intrinsic features of this multinary absorber, and of junctions in the cells are crucial. In this study, dependence of electronic structure of CZTSSe layers on S/(S+Se) ratio and band alignment at interfaces between CdS buffer and CZTSSe have been studied by UPS, XPS and IPES. CZTSSe layers with S/(Se+S)= 0 ~ 1.0 were synthesized by sulfurization and/or selenization of metal precursors. Surfaces of them were cleaned by NH3 etching. Buffer/ absorber junctions were fabricated by step-deposition of CdS layer on the cleaned surface by MBE. In each step, electronic structure was examined in-situ. UPS/IPES results of the CZTSSe layers have revealed almost linear expansion of band gap energy Eg with an increase of S/(S+Se) ratio; Eg(CZTSe) = 0.9 ~ 1.0 eV, Eg(CZTS)=1.4 ~ 1.5 eV. This expansion mainly originates in the rise of conduction band minimum CBM; CBM(CZTSe) ~ 0.5 eV, CBM(CZTS) = 0.9 ~ 1.0 eV. The in-situ measurements of the interfaces have revealed that CdS/CZTSSe; S/(S+Se) ~ 0.3 has so-called “type I” band alignment with conduction band offset CBO and interface induced band bending of ~ +0.2 and ~ 0.4 eV, respectively. Sign-inversion of CBO is confirmed for the CdS/CZTS; S/(S+Se) = 1.0. The observed changes of band alignment are consistent that the variation of cell-performances; e.g. the cells using the absorber with S/(S+Se) ~ 0.3 show a high efficiency > 10 %.

Resume : The efficiency of Cu(In,Ga)Se2 (CIGS) solar cells is strongly affected by Cu and Ga concentration variations in the absorber layer, which locally influence the electronic properties such as the band gap energy (Eg) and the mean inner potential (MIP). Therefore, the measurement of nanoscale variations of these properties is of high interest to improve the understanding of local mechanisms related to compositional inhomogeneities and structural defects. Valence electron energy loss spectroscopy (VEELS) provides a promising tool to measure Eg variations on the nanometer scale. However, VEEL spectra can be influenced by various artifacts which complicate the data interpretation. Based on a systematic study, we discuss the feasibility and reliability of local Eg measurement in CIGS by VEELS. Further, we assess the precision and accuracy of the results in consideration of error estimations and compare the results with simulations. The obtained measurement precision allows detecting relative Eg variations in CIGS and the measured variation corresponds well to the expectations based on compositional gradients measured by energy dispersive x-ray (EDX) spectroscopy. As a second method, local variations of the MIP can be accessed by in-line electron holography in transmission electron microscopy (TEM) using the transport of intensity equation. First results of local MIP measurements are presented and the impact of fluctuations in Eg and MIP on the solar cell efficiency is discussed.

Resume : Cu(In,Ga)Se2 (CIGSe) based chalcopyrite-type solar cells fabricated with multi-stage co-evaporation show high power-conversion efficiencies of more than 20%. However, in many cases efficiencies obtained with CIGSe fall behind this value. This is in particular true in low-temperatures processes without a Cu-rich stage. Reasons for efficiency loss as well as limitations for further efficiency increase are not fully understood. In this work, we analyse structural defects in CIGSe absorber layers with sub-nanometer resolution to gain a better understanding of the defects’ chemical characteristics. For that purpose we use electron energy loss spectroscopy (EELS) combined with high-resolution scanning transmission electron microscopy (HR-STEM). HR-STEM and EELS analyses show striking chemical characteristics for a number of observed defects in the Cu-poor CIGSe thin films. The chemistry of twin boundaries along the {112} planes of the chalcopyrite structure, which are very frequent in these samples, is entirely identical to the rest of the grains, with a homogeneous distribution of all the elements. By contrast, Cu enrichment in combination with In and Se depletion were systematically observed within complex defects closely related to stacking faults, as well as at random grain boundaries. Finally, CuxSe-rich channels seem to form immediately outside (not within) dislocation cores suggesting these defects play a crucial role for the photovoltaic properties of the material.

Resume : CdS is widely used as a buffer layer in CIGS based solar cells. Appreciable efficiencies are obtained using this buffer layer. In view of the detrimental environment impact of Cd, Cd-free buffer layers are attractive. ZnO1-ySy is considered as the best material that can replace the CdS but the conversion efficiencies obtained with this alternative buffer layer are generally lower than those of the conventional CdS layer. In this paper, device modeling and simulation were conducted to investigate the performance of CuIn1-xGaxSe2/ZnO1-xSy solar cells by varying the sulfur content, doping concentration, as well as Zn(O,S) layer bulk defect densities. Furthermore, an inverted surface layer, n-type CIGS, is inserted between the buffer layer and the absorber layer. Donor interface defects were placed 0.2 eV below the conduction band to pin the Fermi-level at the absorber/Zn(O,S) interface. The resulting performance parameters of open-circuit voltage (Voc), short-circuit density (Jsc), fill factor (FF) and efficiency (ƞ) are determined using current density-voltage (J-V) characteristics. The obtained results show that the best solar cells with the Zn(O,S) buffer layer can be achieved when the S content in the buffer layer is approximately 0.2. In comparison to the conventional CdS buffer layer, the best solar cells with the Zn(O,S) buffer layer has slightly lower Voc, FF, and higher Jsc which result in slightly lower conversion efficiency. The simulation results suggest that the high defect density in the Zn(O,S) buffer layer may be the cause of poor performances of Zn(O,S)/CIGS based solar cells. A comparison of the simulation results with published data for the CIGS cells with the Zn(O,S) buffer layer shows an excellent agreement.

Resume : In this work, one-dimensional device simulator AMPS-1D (Analysis of Microelectronic and Photonic Structures) was employed to study the performances of superstrate SLG/TCO/Cu(In,Ga)Se2 (CIGS)/ODC/In2Se3/Metal thin film solar cells. The impact of TCO/p-CIGS and n-In2Se3/Metal interfaces has been investigated. The combination of optical transparency and electrical conductivity for TCO front contact layer is capable of yielding high efficiency. Several transparent conducting oxides (TCOs) materials and metals have been tested respectively as a front and buck contact layers for superstrate CIGS solar cells. The presence of barrie rs in the front and back contact in the structure can significantly affect the cell performance by limit the carriers current flow. The influence of various parameters for the front and the back structures was studied and the corresponding design optimization was provided. The depletion region overlapping between the TCO/CIGS and In2Se3/Metal junctions will result in the decrease of the solar cell performance. The In2Se3/Metal Schottky contact can be utilized as the back reflector in the buffer layer. The best energy conversion efficiencies have been obtained with ZnSnO3 front contact layer. An efficiency of 20.17% (with Voc ≈ 0.71 V, Jsc ≈ 35.35 mA/cm2 and FF ≈ 0.80) has been achieved with ZnSnO3-based as TCO front contact layer and Zn-based back contact layer. All these simulation results give some important indication to lead to higher efficiency of superstrate CIGS solar cells for feasible fabrication. Key words: Superstrate solar cells, Cu(In,Ga)Se2, Thin films, AMPS-1D.

Resume : Indium monochalcogenide (InSe) with a band gap of 1.25 eV is a promising material for photovoltaic applications. In this work, photosensitive anisotype n-ZnO/p-InSe heterojunctions were fabricated for the first time by means of radio-frequency magnetron sputtering of the zinc oxide onto freshly cleaved (0001) van der Waals surface of p-InSe single-crystal. Structural properties of the obtained heterostructures were investigated by means of X-ray diffraction. Surface morphology of the grown ZnO thin films was studied by means of atomic force microscopy. The electrical and photoelectrical properties of the heterojunctions were investigated using the I-V characteristics measured at different temperatures, C-V characteristics and photoresponse spectra. The dominating current transport mechanisms through the heterojunctions under investigation were determined at forward and reverse bias. It was found that the developed heterojunctions n-ZnO/p-InSe show photosensitivity in the photon energy range (1.25 - 3.20) eV at room temperature. The energy diagram of the n-ZnO/p-InSe heterojunction was built using the obtained experimental data. In addition, we analyzed the influence of vacuum annealing of the heterojunctions at different temperatures on their photoelectric properties.

Resume : The incorporation of sodium (Na) and potassium (K) into CIGS leads to enhanced cell performance, but the influence of these alkali on degradation is unclear. Therefore, three types of CIGS cells were prepared with different K and Na contents (high, low and alkali poor) by changing the molybdenum microstructure [1] or adding a SiNx barrier for the alkali poor CIGS. The cells have been degraded in a hybrid damp heat – illumination setup, which allowed continuous in-situ monitoring of the degradation [2]. It was observed that the alkali poor samples had a relatively low initial efficiency, mainly caused by a lower Voc (average 11.4% & 553 mV for alkali free vs 13.2/11.9% & 646/607 mV for low/high alkali content). After 500 hours of degradation, it was observed that the alkali poor cells’ efficiency had decreased to 8.4%, while the other cells had average efficiencies as low as 1.1% (low) and 3.0% (high). In order to better understand the degradation behavior, SIMS measurements were executed on similar samples before and after degradation. It was observed that the Na had migrated from the CIGS bulk to the CdS/CIGS interface and the ZnO:Al layer. This effect was of course very limited for the alkali poor samples, which is likely linked to the stability of these samples. Alkali poor CIGS cells are thus much more stable than CIGS containing K and Na. [1] Bommersbach, Prog. Photovolt: Res. Appl. 21 (2013) 332–343 [2] Theelen, proc. 39th IEEE PVSC (2013)

Resume : Thin-film Cu(In,Ga)Se2 (CIGS) solar cells are multiple-layer structures. In such complicated structures the presence of capacitance spectroscopy signals originating from non-ideal contacts hamper the detection of defect related signals [1]. Still, a numerous types of intrinsic defects are possible in the chalcopyrite structure, that very probably influence the efficiency of the solar cell. Therefore, we made Mo/CIGS/Metal structures from (the original) cells by etching away the buffer and window layer and evaporating gold or aluminum directly on the p-type CIGS absorber. The N1 signal, observed at low temperature (T<120K) is still present in these Schottky diodes, confirming its assignment to the non-ohmic Mo/CIGS back contact [2,3]. Besides N1 no other DLTS signal that exhibits the properties of a non-ohmic contact is observed in these simplified devices. The signals observed at near room temperature where the N2 signal is expected, do no longer obey the model for the back contact. Hence, these Mo/CIGS/Metal diodes appear favorable for studying defects that induce deep levels which can be detrimental for the solar cell efficiency, and that were masked by the presence of contact related signals near room temperature, when measuring on complete solar cells. [1] J. Lauwaert et al. Sol. Energy Mater. Sol. Cell. 112 (2013) 79-83 [2] T. Eisenbarth et al. J. Appl. Phys. 107 (2010) 034509 [3] J. Lauwaert et al. Prog. Photovoltaics 20 (2012) 588

Resume : An understanding of carrier dynamics is of vital interest for improving the efficiency of the solar cell. To realize behavior of e-h pair excited by photon, femtosecond optical technique using THz spectroscopy is very useful but, not applied in the studies of dynamics of nonequilibrium carrier relaxation in chalcogenide-based solar cell. Interestingly, Na incorporation is considered as an important variable to enhance the cell performance in Cu(In,Ga)Se2 (CIGS) thin-film solar cell. Na is typically supplied from the soda lime glass (SLG) widely used as a substrate of solar cell. Although Na is known to have positive influences on CIGS-based solar cell, a deep understanding of their response to optical excitation on a sub-picosecond time scale has not yet been studied. In this study, we performed the comparative experiment by depositing the same CIGS on different substrates of Na-containing SLG and Na-free borosilicate (BS), respectively, and then forming CdS on CIGS using chemical method. Optical pump-THz probe spectroscopy (OPTP) and photoluminescence (PL) were used to investigate the Na effect on ultrafast carrier dynamics at interface between CIGS and CdS. We found that different defect states were formed in the band gap of CIGS depending on Na presence from the PL spectra. The relaxation time of the carriers excited by optical pump beam (400 nm) was significantly increased at CdS/CIGS on SLG as compared with that on BS from the OPTP results.

Resume : in monosulfide (SnS) is considered to be a promising candidate for a cost-effective and earth-abundant inorganic material to fabricate next-generation solar cells. Although the theoretical conversion efficiency of SnS-based solar cells is high, the demonstrated efficiencies of such cells are still low. One of the reasons for low efficiencies is the poor crystal quality of a SnS layer. In fact, the morphology and crystal structure of SnS thin films strongly depend on the heating profile and the time of each step during the sulfurization process. However, few experimental results using the sulfurization technique have been reported compared to other growth techniques such as co-evaporation or chemical vapor deposition. In this presentation, we reveal the growth mechanism of SnS thin films by a simple and commercial-friendly sulfurization technique. As well, defect-related PL peaks in the as-grown SnS thin films will be investigated. In this way, we found that sulfurization conditions, especially the S flow rate, strongly affected orientation and electrical properties such as carrier density or mobility of SnS thin films. These results indicated that Sn vacancy (VSn) and S vacancy (VS) act as a shallow acceptor and deep donor, respectively. These investigations will greatly assist our ability to optimize the conditions required for the production of SnS-based solar cells.

Resume : Forbes and Peter [1] have predicted that the market share of CdTe and CIGSe PV will peak and then fall as supply limitations begin to impact as the global installation rate of PV increases. It follows that materials substitution needs to be addressed now if thin film PV is to take on the rapidly emerging challenge of TW renewable energy production. PVTEAM will use semiconductor materials based on abundant sustainable low-cost elements such as copper, tin, zinc and bismuth as substitute layers in thin-film devices. PVTEAM will accelerate the adoption of new earth-abundant PV by developing processes technologies for materials and systems to a level where they can be taken up by manufacturing industries. The programme covers materials specifications and performance, integration into cells and mini-modules and scale up. Using a multi-level screening approach, PVTEAM will choose substitute materials and incorporate the best performing candidates into a generic, industrially-oriented process coupled with appropriate in-line metrology, calibration and stability testing. The PVTEAM work programme consists of three themes: material specification, integration at cell and module level and scale-up towards manufacture. The overall objectives of the programme are: ? To specify and assess high quality binary, ternary and quaternary PV layers produced by scalable solution based processing. ? To develop sustainable non-vacuum routes to the fabrication of viable thin film architectures (t

Resume : The formation of interface Cu2ZnSnS4/CdS in monograin layer solar cells is not well understood but seems to be the key to further improvements of their performances. Cu2ZnSnS4 monograin powders were synthesized from elemental precursors in molten KI as flux materials in sealed quartz ampoules at 740 C. In the present work, chemistry of Cu2ZnSnS4 monograin powder surfaces submitted to various chemical treatments was investigated by X-ray photoelectron spectroscopy. The surface analysis allowed us to compare the surface composition with the bulk one as a function of the treatments. The chemical nature of the etchant had a dramatic influence on both, surface composition and interface chemistry. The CdS film formed by chemical bath deposition is reproducible and yields good photovoltaic performance. Due to the aqueous environment, however, the as-deposited CdS film may contain significant amounts of oxygen and hydrogen which degrade the quality of the film. The effect of heat treatment on Cu2ZnSnS4/CdS interfaces in air and vacuum has been also studied. We have developed a combined process based on chemical etching and Cu2ZnSnS4/CdS interface annealing, which improved the active area efficiency to more than 8 %.

Resume : Cu2ZnSnS4 (CZTS) is an interesting photovoltaic material owing to its earth abundant and environmentally friendly constituents and an optimal direct band gap of about 1.5 eV. We have grown CZTS thin films using a two-steps process, which consists in the precursor deposition, followed by a thermal treatment. The precursors are deposited by a co-sputtering process from three sulfides sources (CuS, ZnS, SnS) giving films with a homogeneous elements distribution and nearly stoichiometric sulfur content. This is not a widely explored route and nevertheless it gives, after a suitable thermal treatment, CZTS thin films with easily controllable stoichiometry, good homogeneity and large grain size. CZTS films and the respective precursors are investigated by XRD, Raman, SEM-EDX and Spectrophotometric measurements. Compositional depth profiles are obtained by XPS. The films have also been employed to fabricate solar cells, with a standard structure: Mo/CZTS/CdS/iZnO/Al:ZnO/Al-grid. The device performances are evaluated by measuring the J-V characteristic curves, under both AM1.5G illumination and dark condition, and External Quantum Efficiency. CZTS films composition and thickness have been optimized in terms of cell efficiency. Up to now our best CZTS-based photovoltaic device has shown an efficiency of 5.7%. The optimization of the material stoichiometry is still ongoing and its relation with the optical properties of the material and with the device performances will be discussed.

Resume : A graded bandgap structure proved to be an important factor for increasing an overall efficiency of the chalcopyrite-based thin film solar cells. This contribution will be focused on the effects of sulphur incorporation into the surface region of industrial sequentially grown Cu(In,Ga)(Se,S)2 absorbers. A front grading due to such a sulphurization step enhances the bandgap in the space charge region, whereas the bulk of the absorber exhibits a lower bandgap which determines absorption and photocurrent. The question which will be addressed in this contribution is whether such graded bandgap structures allow to separate the absorption and recombination processes, therefore resulting in highly efficient solar cells with improved open circuit voltages without compromising short circuit currents. In order to assess the effective bandgap for nonradiative recombination low temperature measurements have been performed on the Cu(In,Ga)(Se,S)2 and Cu(In,Ga)Se2 – based solar cells. The extrapolated open circuit voltages at 0K demonstrate a shift of the activation energies of about 50 meV for the device with a S-gradient at the interface. However, the effective bandgaps for current collection of the same devices determined from external quantum efficiency measurements showed a negligible difference and roughly corresponded to the pure CuInSe2 material. Furthermore, a high Ga content at the back contact will be discussed in terms of a back contact passivation preventing injection of electrons to the back contact.

Resume : Control of defects in CuInGaSe2 (CIGS) solar cells is of prime importance to improve their electrical properties and they have been extensively investigated using different analytical methods. Despite these efforts, the properties of traps in devices have not been fully understood yet, partly because of the complex structure of the active layer as well as that of the cells. In this work, we have investigated traps in Metal-CIGS-Metal (M1-S-M2) diodes (M1 = Mo, M2 = Al or Mo) by using the charge based Deep Level Transient spectroscopy (Q-DLTS), a variation of the conventional DLTS technique. Two sets of distributed traps A and B have been determined in diodes with different metal electrodes. Type A defects contain several levels of shallow traps with a mean activation energy of 25mV while type B contains deep traps with a mean activation energy of 370 mV. The capture cross sections of both types are determined to be in the range of 10-21-10-22 cm2 and their density in the range of 1015-10-16 cm-3. The trap filling process of the two trap sets was found to be strongly dependent of the temperature conditions. Replacing Mo top electrode by Al one did not modify significantly the trap distributions but decreased the trap density, suggesting that interfacial interactions between the active layer and electrode may have occurred in the devices. Q-DLTS spectra of a Mo-CIGS-Mo device measured at T = 300 K and by using a charging voltage of ΔV = 3 V and different charging times tC in the range 500 µs (○) – 1 s (▼).

Resume : Kelvin probe force microscopy (KPFM) has contributed significantly to the understanding of Cu(In,Ga)Se2 (CIGSe) materials on the nanometer length scale throughout the last 15 years. In combination with sample illumination, some important fundamental properties like charge carrier generation and separation processes can be investigated. While previously KPFM with illumination has been applied as a static technique, here we present the investigation of the dynamics of the surface photovoltage in the range between nanoseconds and milliseconds. We present results on different sets of CIGSe samples from different growth processes and with stacks of CIGSe/CdS and CIGSe/CdS/ZnO. For lower efficiency CIGSe we measure relaxation times of excited and separated charge carriers on the order of a few tens of microseconds. For higher efficiency CIGSe we obtain 10 times faster relaxation times. Analysis of spatially resolved transient surface photovoltage shows a homogeneous distribution of relaxation times, without significant differences between grains and grain boundaries. We discuss the differences of the physical origin of the measured relaxation times with respect to relaxation times obtained from other techniques, such as transient photoluminescence and deep level transient spectroscopy.

Resume : If the need of Se supply during CIGSe co-evaporation has been widely investigated and discussed, its need after the deposition, while the substrate is still at high temperature, remains unclear. The present contribution investigates the impact of chalcogen supply after CIGSe growth on cells properties. With this aim, absorber layers have been deposited by the 3-stage process and once the growth completed, the SLG/Mo/CIGSe structures have been kept at high temperature for one hour with or without Se supply. Both the resulting CIGSe layer and related device properties have been compared to those obtained when the substrate is cooled down right after the deposition (i.e. standard process). Moreover, such experiments have been performed using substrates yielding different CIGSe alkali content (i.e. Na-free, high and low Na). The results show that keeping the absorber at high temperature impacts Ga/III gradients differently depending on whether Se is supplied. More surprising is the impact on the cells, this work indeed shows chalcogen supply can be detrimental for cells performance when the CIGSe contains Na. This latter observation suggests an intimate relationship between Na and Se, which will be discussed with the help of advanced material and device characterization.

Resume : Determination of free hole and defect distributions in CIGS devices is problematic due to the amphoteric character of VSe - related defects. Recently we have proposed a method, based on measurements of capacitance profiles in different metastable states, which allows for a quantitative evaluation of densities of metastable defects (Nt), and net shallow acceptors (Na). A standard analysis of C-V measurements assumes high asymmetry in the doping level between both sides of the junction. However, after illumination, when the free hole concentration in CIGS is significantly increased, the r-ZnO/CdS/CIGS structure cannot be treated as a one-sided n+p junction. Hence, in order to accurately estimate Nt and Na, we take into account in this contribution the influence of the n-side on capacitance space charge profiles. We analyze the junction assuming both, fully and partially depleted buffer layer. We show that the explanation of the shape of the space charge profiles close to the interface is possible only by taking into account free carriers diffusing into the depletion layer. Analysis of profiles obtained for positive voltages can provide information about concentration of donors in CdS. We develop an analytical model and juxtapose it to numerical simulations. Basing on this we indicate in which conditions capacitance profiling gives reliable results, in particular free hole and metastable VSe defect densities.

Resume : SnS is a promising semiconductor for photovoltaic (PV) applications. Owing to its bangap value, SnS is well suited to be used as efficient absorber in thin film photovoltaic solar cells. Besides SnS is nontoxic and earth-abundant material. The purpose of this work is to check the effectiveness of SnS-based PV devices. Different configurations have been analyzed by using dedicated software, such as Solar Cell Capacitance Simulator (SCAPS). The proposed solar cell consist of an SnS absorber layer with a bandgap of 1.4 eV, a thin buffer layer of about 50 nm based on a n-type semiconductor without containing heavy metals and a window layer of about 100 nm. Two structures were studied. One containing the standard and well-known CdS buffer (SnS/CdS/SnOx) and a second one made of layers containing Sn in all layers in order to avoid diffusion (SnS/SnS2/SnOx). After optimization of different parameters a conversion efficiency of 10.50%, VOC of 0.91-volts, JSC 0.0134 A/cm2 and Fill-Factor (FF) of 86.48% was obtained for the SnS/CdS/SnOx structure. These results will be compared with that obtained for the greener SnS/SnS2/SnOx structure. It is worth noting that simulation studies are based of experimental data available in the literature. As SnS material has just being produced as thin film, its characteristics will be still improved for its use in PV devices and therefore the performance of related devices would be further enhanced.

Resume : In the past decade the CIGS-based solar cells have proven that they are one of the promising cost-effective thin-film solar cells by demonstrating significant technological advances from manufacturing as well as laboratory scale. In this work, with Ag alloying we attempted to improve the microstructure and device performance of low-temperature grown CIGS solar cells. Ag precursors with various thicknesses were deposited onto Mo prior to CIGS growth. CIGSS absorbers were then formed onto Mo/Ag with a single-step co-evaporation at the substrate temperature of 450 ?C. The surface morphologies and microstructures of the CIGS films were changed by the Ag alloying. A CIGS film without the Ag-alloying was found to have quite a fine microstructure, while Ag-alloyed films exhibited significant recrystallization. Devices were also fabricated with the CIGS films to elucidate the effects of the Ag-alloying to device performances. A CIGS cell without the Ag-alloying turned out to have a poor device performance as an efficiency of about 3.0% and a VOC of 310 mV. In contrast, a CIGS cell with Ag/(Ag Cu) = 0.2 exhibited a significantly improved device performance as an efficiency of about 8.6% and a VOC of 530 mV. This improvement by the Ag alloying appeared to be associated with the improved crystallinity and defect passivation. The more details of the Ag-alloying effects to a CIGSS film will be discussed in the conference.

Resume : A comprehensive study of the thermodynamic and electronic properties of intrinsic point defects in the solar absorber materials CuInSe2 and CuGaSe2 based on screened-exchange hybrid density functional theory is presented. GaCu is found to be the most detrimental intrinsic point defect in CuGaSe2 and high-gallium Cu(In,Ga)Se2 above approximately 50% gallium, since it constitutes a minority carrier trap. In contrast with results in the literature, InCu is a very shallow donor, which explains the good tolerance of CuInSe2 to off-stoichiometry rather than complex formation with vacancies. Complex formation with copper vacancies cannot occur under thermodynamic equilibrium conditions, because the formation energies are higher than that of the individual point defects. CuIn and CuGa hole traps and Cui may be contained in high quantities under certain preparation conditions such that they can significantly alter the properties of the material. Based on these results optimal preparation conditions in terms of the point defect physics of CuInSe2 and CuGaSe2 are discussed.

Resume : We have performed a DFT study of point defects in Cu(In,Ga)Se2 (CIGS), using the HSE06 hybrid functional. Our calculations support the claim that Cu vacancies have low formation energies and act as shallow acceptors, accounting for p-type conductivity in Cu poor conditions. In these conditions, In_Cu and Ga_Cu are also defects with low formation energies and thus prevalent. These defects act as compensating shallow donors. Ga_Cu is likely to be the shallow donor level responsible for the broadened PL spectra in Ref. [1]. When the amount of Cu increases, their formation energy increases, consistent with the peaked PL spectra for increased Cu richness in Ref. [1]. Our study also indicates that p-type behavior is not restricted to Cu poor conditions. Cu_In and Cu_Ga, shallow acceptors as well, occur more frequently in Cu rich conditions. At the same time, less acceptor states are compensated. It may explain the experimental observation of the increasing hole concentration with Cu/Ga ratio in Ref. [2]. Finally, we have studied C impurities, motivated by growth methods based on nanoparticle inks. We find that C_Cu is a shallow donor, while C_In and C_Ga form deep donor levels. Interstitial C is an amphoteric defect. All C impurities can in principle harm the p-type conductivity. However, their formation energies are high leading to the prediction that C is expelled. [1] A. Bauknecht et al, J. of Appl. Phys. 89, 4391 (2001). [2] A. Gerhard et al, Thin Solid Films 387, 67 (2001).

Resume : CdS is a frequently used buffer layer in chalcogenide thin film solar cells. It can be prepared by a variety of deposition techniques as thermal evaporation (TE), close-spaced sublimation (CSS), chemical bath deposition (CBD), magnetron sputtering (MS), and others. This contribution summarizes extended studies of the Fermi level position at the surfaces of CdS layers using photoelectron spectroscopy. Different substrates and different deposition conditions are compared. It will be shown that the Fermi level depends noticeably on the deposition technique. In particular, as-deposited films grown on fluorine doped SnO2 show Fermi levels of EF-EVB = 1.8-2.2 eV (TE), 2.1-2.5 eV (CSS), 1.7-2.4 eV (MS), and ≈2.0 eV (CBD), respectively. The Fermi level of TE-CdS shows no apparent dependence on substrate material. Considerably lower Fermi level positions of EF-EVB ≈ 1.5 eV are reproducibly found for CBD-CdS grown on Cu(In,Ga)Se2. The results show that the Fermi level in CdS can be strongly pinned by intrinsic defects, which can even lead to a modification of energy band alignment at interfaces. It will furthermore be demonstrated that annealing in oxygen modifies the pinning position and changes the energy band alignment to the underlying substrate. A discussion of the observations in terms of intrinsic defects in CdS calculated using density functional theory will be provided.

Resume : Zn(O,S) grown by chemical bath deposition (CBD) is well established as an alternative buffer to CdS in Cu(In,Ga)Se2 (CIGS) solar cells. Nevertheless, these devices often have a lower open-circuit voltage Voc than CdS-buffered references. This contribution promotes the understanding of Voc loss in CIGS cells buffered with Zn(O,S). A series of CIGS cells with different Zn(O,S) buffer thicknesses between 0 and 40 nm was generated by varying the time of the CBD process. All other layers of the cell stack were processed similarly: sputtered Mo, co-evaporated CIGS, as well as (Zn,Mg)O and ZnO:Al by sputtering. Reference cells from the same CIGS run were fabricated with CdS/i-ZnO/ZnO:Al. The Voc of the CIGS cells with increasing Zn(O,S) buffer thickness increases significantly from values below 500 mV for the 0 nm to 660-670 mV for the 40 nm samples. The latter values are close to the 690-700 mV found for the CdS-buffered references. Voltage-dependent external quantum efficiency (EQE(V)) measurements reveal almost no bias dependence for the cells with thick Zn(O,S), very similar to the CdS-buffered cells. The EQE(V) curves of CIGS cells with thin Zn(O,S) buffers, however, exhibit a distinct dependency on the applied bias voltage. Collection losses at positive bias voltages are possibly a result of interface recombination. This topic will be discussed with the focus on Voc of the Zn(O,S) series, complemented by the influence of a post-annealing step on the electrical cell properties.

Resume : Time resolved photoluminescence (TRPL) is a promising method for the investigation of carrier dynamics and recombination kinetics in semiconductor devices. To characterize Cu(In,Ga)Se2 (short: CIGS) solar cells, we measured the TRPL for different applied external forward voltages and excitation intensities. We show that the TRPL decay time increases with increasing voltage in case of a high excitation intensity. This result is valid for a wide range of excitation frequencies of the laser. However, for low excitation intensity the decay time is unchanged by an external voltage. By simulation of the measured transients with Synopsys TCAD we determined a set of third level parameters which allow to fit the experimental photoluminescence transients for different voltages. The calculated quantities were Schockley-Read-Hall lifetime, doping density and minority carrier mobility of the solar cell's absorber layer with values of 10ns, 10E15/cm3 and 5cm2/Vs, respectively for a standard CIGS solar cell. We note that the minority carrier mobility is difficult to obtain by other methods. We further studied the appearance of a photovoltage in TRPL experiments with single-photon-counting-methods. By experimental and theoretical results we show, that the voltage externally applied gives the same transient than the internal photovoltage at open-circuit conditions. Moreover the open-circuit voltage shows a dependence of the laser repetition rate.