Thin film chalcogenide photovoltaic materials

Resume : In the last few years, the interest and research on new thin film inorganic materials for photovoltaic (PV) applications have largely increased. There are two main motivations for this: i. the necessity to substitute scarce and/or expensive materials currently used in the thin film photovoltaic industry (such as In, Ga or Te); and ii. the necessity of materials with new properties that are required for non-conventional PV applications (flexible, transparent). In this context, kesterite type semiconductors, including Cu2ZnSn(S,Se)4 (CZTSSe), Cu2ZnSnS4 (CZTS), and Cu2ZnSnSe4 (CZTSe) are recognized as one of the most relevant and promising thin film photovoltaic technologies. In the first part of the presentation, the most relevant progresses achieved for kesterite in the last years will be reviewed, focusing on the recent strategies followed to improve their efficiency, that are mainly based on the partial cationic and anionic substitution including Cu by Ag, Zn by Cd, Mn and Mg, Sn by Ge, extended to the S, Se and Te substitutions as possible anions in the structure. In the second part of the presentation, other relevant emerging photovoltaic materials will be discussed, including oxides, pnictides and halides, giving insights about the most promising ones for different PV applications.

Resume : A non-ideal interface between MoSe2 and the kesterite absorber is often considered as a possible hindrance to device performance. Various layers on the Mo back contact (BC) have been investigated to control the formation of MoSe2, but they also altered alkali diffusion, crystallization path and absorber composition, which are parameters with limited reproducibility. In order to decouple MoSe2 growth control from other factors, we designed a structured sample allowing comparison of four back surfaces (SLG, SLG+Al2O3, Mo and Mo+Al2O3). It was fabricated by scribing Mo BC lines and then covering half of the substrate with a 10nm ALD-Al2O3 layer. The CZTS precursor was then spin-coated and annealed in Se atmosphere. MoSe2 developed on both Mo and Mo+Al2O3. Absorber composition and elemental profiles were measured by TOF-SIMS, EDX and XRF mapping whereas photoluminescence (PL) imaging was used to calculate non-radiative voltage losses (Vnon-rad). The lowest Vnon-rad of 290meV was obtained for CZTSSe on Mo+MoSe2 BC without Al2O3 on top, comparable with kesterite devices with efficiency of ≈12%. A PL peak shift of ≈15meV between SLG and Mo lines correlates with a measurable composition variation as obtained by XRF and EDX. No influence on composition or alkali distribution was caused by the Al2O3 layer, while MoSe2 layer was thinned, but not suppressed. It can be concluded that Mo+MoSe2 bi-layer has the major influence on absorber composition and provides the lowest Vnon-rad.

Resume : Aqueous precursors provide an alluring approach for low-cost and environmentally friendly production of earth-abundant Cu2ZnSn(S, Se)4 (CZTSSe) solar cells. The key is to find an appropriate molecular agent to prepare a stable solution and optimize the coordination structure to facilitate the subsequent crystallization process. Herein, we systematically investigate the coordination engineering in the CZTSSe aqueous precursor solution. The competition between metal-S and metal-COO coordination interactions has been manipulated by adjusting the deprotonation degree of TGA in solution. In the optimal structure, the metal cation is adequately coordinated with thiolate anions, and the carboxylate anions are released to become fully hydrated, thus facilitating the formation of an ultrastable metal-TGA aqueous solution and improving the nucleation and crystallization of the CZTSSe thin film. Our strategies enable the fabrication of CZTSSe solar cells with a remarkable efficiency of 12.3% and a certified efficiency of 12.0%, which is the best result for CZTSSe solar cells among the aqueous precursor systems investigated to date. This significant progress in the non-toxic solution fabrication of CZTSSe solar cells brings great promise for the future development of CZTSSe solar cells.

Resume : Kesterite material is characterized by a tunable direct band-gap energy between 1.0 eV (Cu2ZnSnSe4) and 2.25 eV (Cu2ZnGeS4). Wide band-gap absorber layers are attractive for top cells of a cost-efficient tandem device, but also for semitransparent solar cells for efficient and stable advanced BIPV concepts with enhanced efficiency in the IR and UV ranges. In this work, Cu2ZnGe(S,Se)4 (CZGSSe) thin films have been grown by sulfurization of co-evaporated CZGSe layers deposited onto Mo/soda-lime glass and FTO substrates. The influence of the alkaline element Na on CZGSSe thin films and final devices has been investigated. To do so, a NaF layer has been deposited before and/or after the co-evaporation of the CZGSe. All absorber films are systematically characterized with SEM, EDX, GIXRD, and GDOES and the processed solar cells are analyzed with J-V and EQE measurements. When no NaF layer is deposited onto the absorber film, efficiencies of 5.3 % and Eg = 1.56 eV are achieved for CZGSSe-based devices. The S distribution through the kesterite is dominated by the Na content controlled by the thickness of the NaF precursor layer. Efficiencies of 4.6 % and 1.7 % with Eg = 1.8 eV and 2.1 eV are obtained for 12 nm and 3 nm NaF layer thicknesses respectively. The device with higher Eg is characterized by a much higher S concentration at the surface and a lower Jsc, limiting the performance. The efficiency of the CZGSSe-device is enhanced to 2.5 % when a NaF layer is supplied by a post deposition treatment (PDT) and up to 3.2 % for the combination of a precursor layer and a PDT, in both cases with Eg = 2.0 eV. Up to now, first solar cells deposited onto FTO have been fabricated, obtaining performances of 5.6 % and 3.1 % with Eg = 1.3 and 1.7 eV respectively.

Resume : Solar cells based on Cu(In,Ga)Se2 (CIGSe) thin films have reached record efficiency close to 23 %; these outstanding performance result from more than three decades of academic and technological developments. However, so far the highest device efficiencies were achieved with CIGSe absorbers whose bandgap (ie. absorption threshold) ranges between 1.1 and 1.2 eV. For wider bandgap, meaning with higher gallium content, the performance of the devices drastically decrease which is clearly a bottleneck for the application such as top cell in tandem solar cells with c-Si. Wide bandgap (ie. Eg>1.5 eV) can be achived by partial or complete substitution of sulfur for selenium, leading to Cu(In,Ga)(Se,S)2 materials. The present contribution deals with sulfur based Cu(In,Ga)S2 films and related solar cells. The investigated CuIn0.7Ga0.3S2 layers were grown by coevaporation following a so-called 3-stage process. This approach offers the control of several process parameters which directly impact the characteristics of the resulting layers, thus cells performance. In particular, a possible route to control the relative distribution of group III elements (i.e. GGI) throughout the films will be presented, and the impact it has on device peroformance. Up to date, the best efficiency we have achieved with the standard structure (SLG/Mo/CIGS/CdS/ZnO/AZO/grids/ARC) is above 14 % (Voc: 930 mV; Jsc : 21.2 mA/cm² ; FF: 0.72), with an absorber bandgap of 1.6 eV. This result opens many rooms for further improvement; they will also be presented and discussed during the congress.

Resume : Identification of suitable wide band gap semiconductors for application as top cells in tandems is a topic of increasing importance. AgGaSe2 is one candidate that has not received a lot of attention. Here, Ag-Ga-Se films with a range of compositions ([Ag]/[Ga] = 0.5 - 1.5) are synthesized by co-evaporation and characterized. From photoluminescence and quantum efficiency measurements, a band gap of 1.79 eV is estimated. Unlike the CuGaSe2 system, where CuxSe is formed under Cu-rich growth conditions, Ag9GaSe6 is found to coexist with chalcopyrite AgGaSe2 in Ag-rich samples. In samples grown under Ag deficiency, a zinc-blende phase similar to Ga2Se3 but with a larger lattice constant is found alongside AgGaSe2. To identify this phase, ab initio analysis is carried out with a focus on ordered vacancy compounds (OVCs). The calculations show that AgGa5Se8 is the only stable OVC phase in Ag-Ga-Se system, in contrast to Cu-Ga-Se that has both CuGa5Se8 and CuGa3Se5. A comparison of computed convex hulls also imply that AgGaSe2 has rather narrow single-phase region. Moreover, structural models predict AgGa5Se8 to be visible in XRD as a zinc-blende phase akin to Ga2Se3, in agreement with experiment. Devices produced with Ag-rich absorbers are found to be shunted, while slightly Ag-poor conditions yields working devices with an open circuit voltage of up to 911 mV and efficiency of 5.8 %. A comparison of devices grown by 1- and 3-stage processes reveal significantly poorer performance of the latter. This is likely due to different localisation of OVCs that form as a result of the aforementioned intolerance of AgGaSe2 to off-stoichiometry. These peculiarities and the encouraging initial results highlight importance of studying Ag-based chalcopyrites in more detail.

Resume : Antimony chalcogenides have recently experienced incredible progresses as emerging photovoltaic materials with a certificated efficiency of 9.2%. This technology has increasing interest due to the quasi one-dimensional crystal structure providing interesting properties such as reduced grain boundaries recombination. In this work we propose the Se redistribution at atomic level between the Van der Waals attached ribbons and grain boundaries, showing that fine tune of the slight Se rich concentration allow obtaining a compromise between enough p-type doping avoiding recombination centres. The distribution of this Se excess is analysed and a model in which two chalcogen distributions determine the properties of the device is proposed. A first Se distribution enriches the grain and sub-grain interfaces with elemental Se, leading to a remarkable improvement of the optoelectronic parameters. A second chalcogen distribution is hypothesized for very Se-rich samples, causing the diffusion of Se toward the CdS buffer. The hypothesis is further explored by a thermal stability analysis showing that Se rich conditions under air annealing at different temperatures and times have a relevant impact in the Sb2Se3/CdS interface by the formation of a CdSSe solid solution. Under these conditions, a preferential texture around [002] and secondary phases free device is obtained and a 5.7% power conversion efficiency is reported. Huge increase over previous works based in a similar route (0.72%).

Resume : Antimony selenide (Sb2Se3) absorber material has great potential for low-cost photovoltaics due to its excellent optoelectronic properties and low processing temperatures. For controlling the optoelectronic properties of semiconductors, it is of crucial importance to understand their defect structure. Very little is known about the defects in Sb2Se3, moreover, only few experimental reports can be found. This study focuses on the modification of the defect structure of Sb2Se3 via post-growth selenization treatments. Detailed temperature and excitation power dependent photoluminescence (PL) analysis of Sb2Se3 single crystals and thin films are performed revealing the changes in the radiative recombination mechanisms when varying the selenization temperature. Detected PL bands at T=10 K originate from the band-to-band and donor-acceptor pair recombination as well as from radiative transitions involving grain boundaries. Discussion around the state of the art knowledge about the photoluminescence emission of Sb2Se3 as well as the possibility to control the defect structure of this promising absorber material for solar cells will be presented.

Resume : It is well recognized that in thermodynamic equilibrium, when defect formation and diffusion is not kinetically limited, the variation of the Fermi energy in a semiconductor is limited by self-compensation. Even though the limits can be modified by adjusting the chemical potentials of the constituents, e.g. by using Cu-poor or Cu-rich conditions in the chalcopyrites, the Fermi energy in wider gap materials can be substantially restricted. For example, self-compensation prevents n-type doping and large photovoltages in CuGaSe2. By using X-ray photoelectron spectroscopy of Cu-oxides with in-situ surface preparation and interface formation, we demonstrate an additional mechanism, which can limit the Fermi energy: The valence changes of Cu. This mechanism is independent on the chemical potentials. It limits the Fermi energy in both Cu2O and CuO to values in the lower half of the energy gap and prevents high photovoltages and conversion efficiencies close to the Shockley-Queisser limit. A comparable mechanism will occur in all Cu-containing compounds and also in other prospective thin film solar materials considered to date.

Resume : Interfaces of oxides and heavier chalcogenides layers in thin film or extremely thin absorber solar cells present defect states at the interface and often a chemical incompatibility which results in dewetting issues. Here, we establish atomic layer deposition (ALD) as a tool to address these limitations. ALD allows one to obtain highly pure Sb2S3 as light absorber layers, and we exploit this technique to generate additional interfacial layers with thicknesses between 0.2 and 2.0 nm. These ultra-thin layers modify drastically the physical-chemical properties of the interfaces enable one to eliminate dewetting, passivate defect states at the interface and slow down interfacial charge recombination. We achieve heterojunction solar devices with optimized power conversion efficiency beyond 5.0 %.

Resume : BaZrS3 is an inorganic perovskite based on abundant, non-toxic elements. Its direct band gap of 1.8 eV makes it suitable for tandem cells, and it can withstand heating in air up to 500°C without decomposing. Computational work predicts similar properties to Pb-halide perovskites, but little has been published on thin film synthesis. BaZrS3 films were grown by reactive magnetron co-sputtering with BaS and Zr targets in H2S-containing plasma and then annealed in an RTP furnace for crystallization. The temperature was varied in the range 700-1000°C. Materials and optical characterisation of sputtered precursors and annealed films were obtained using ion-beam analysis, EDS, PL, XRD, Raman, STEM imaging and STEM-EDS. XRD and Raman show the BaZrS3 distorted perovskite phase, without obvious secondary phases. STEM and SAED also indicate crystallinity. The annealed films are S-poor, while the O content increases whenever excess S is present. With increasing anneal temperature, the PL peak intensifies and narrows, converging to 1.8 eV for a 900°C anneal, at which temperature it agrees well with literature. PL emission near the band gap energy confirms the absence of extensive band gap tailing. Increasing the annealing temperature even further does not improve PL but rather reverses the trend. We discuss challenges relating to further development of thin film synthesis of BaZrS3 toward tandem cell integration and explore different sputtered-based routes to achieve better process control.

Resume : Colloidal AgBiS2 nanocrystals are gaining attention in the last years amongst heavy-metal-free ternary chalcogenides considered for photovoltaic applications. They are easy to produce from inexpensive, earth-abundant and non-toxic materials and exhibit a high absorption coefficient as well as a suitable bandgap for light harvesting. In addition, the cubic rocksalt AgBiS2 structure shows intrinsic stoichiometric stability. Here we show a modified synthetic method, based on a double-step hot-injection approach, to obtain ternary AgBiS2 nanocrystals with a significant increase in the average diameter compared with our previous process. The resulting nanocrystals preserve the crystal structure and stoichiometry. As a direct consequence of the enhanced average size, the photocurrent of solution-processed solar cells using these nanocrystals increases by 4 mA/cm2, and the power conversion efficiency jumps from 5.6% to 6.4%, owing to a higher mobility, a reduced trap-assisted recombination and a lower midgap trap state density. We also discuss on the energy losses of this system and propose strategies that may allow exploiting the full potential of this material.

Resume : Antimony selenide (Sb2Se3) has attracted increasing attention due to its low-cost, non-toxic, earth-abundant components and similar properties with CdTe. However, its one-dimensional structure makes it easier to grow along the (Sb4Se6)n chains (c-axis), so that fairly difficult to deposit a high quality film with large grain size and well orientation. In addition, its proved benign grain boundary leaves the film surface as the main defective position, but most researchers ignore this and mainly focus on the bulk defects. Here we explored these two problems respectively: (1) High growth rate and small amount seed were found to be the key factors of Sb2Se3 film morphology and orientation. Thus, higher source and substrate temperature were used in CSS system, along with more inert TiO2 substrate annealed at 550 °C. Finally, we obtained a 10-20 µm grain size Sb2Se3 film with inclined [hk1] dominant orientation. (2) A simple combined two-step surface passivation method was developed, and result in a three-fold PCE improvement, especially in FF. The reason had been proved as the formation of thin crystalline Sb2O3 layer. At last, 5.36% PCE flexible Sb2Se3 solar cell was obtained and could remain 97% initial PCE after 1000 bending cycles. Reference： [1] K. Li , C. Chao ,*, J. Tang* et. al. Adv. Mater. 2019, 31, (44), 1903914 [2] R. Kondrotas, J. Tang* et. al. Sol. Energy Mater. Sol. Cells 2019, 199, 16-23. [3] C. Wang , S. Lu , R. Kondrotas*, J. Tang* Nano Energy (submitted)

Resume : Copper-zinc-tin-sulfide (CZTS with standard stoichiometry of Cu2ZnSnS4) is a cheap, earth-abundant and non-toxic photovoltaic material for 3rd generation solar cells. Unfortunately, conventional CZTS solar cell fabrication relies on an architecture, adopted from earlier solar cell generations, which depends on an energy-intensive high temperature annealing process in a sulfur or selenium atmosphere. We want to present a more environmentally-friendly approach where CZTS is synthesized in nanoparticle (NP) form at room temperature. The method is shown to be compositionally robust and able to produce sulfide-stabilized carbon-free nanoparticle inks that are suitable for an absorber layer in solar cells. No organic residues from the process were detected. The metal-composition and the occurrence of secondary phases is here correlated with synthesis conditions: By utilizing a reactant concentration of Cu/Sn < 1.8 and Sn(II) as tin-source it is possible to avoid the formation of copper sulfide-phases, which are detrimental for the solar cell performance when present in the final absorber layer. With nanoparticle sizes approaching the Bohr radius for CZTS, the band gap can be broadened up to 1.7 eV. In addition, the conditions for forming stable, carbon-free aqueous inks of such CZTS NPs and the effect of the stabilizing ammonium/sulfide-ion concentration on the quality of the deposited absorber layer are investigated. The use of room temperature synthesis and stable aqueous ink formulations make the method suitable for roll-to-roll fabrication and upscaling.

Resume : In this research work, the performance of Cu2ZnSnS4 based solar cells was analyzed and studied by the Solar cell capacitance simulator (SCAPS-1D) software. We chosen Zn(S,O) as a solar cell buffer material in order to replace toxic Cadmium Sulphide (CdS). The photovoltaic parameters (the short-circuit current density, open-circuit voltage, fill factor, and efficiency) of the Mo/p-Cu2ZnSnS4/n-Zn(S,O)/i-ZnO/n-ZnO:Al structure have been simulated with different CZTS samples. According to the simulation results, we found that the efficiency Mo/p-Cu2ZnSnS4/n-Zn(S,O)/i-ZnO/n-ZnO:Al structure was around 10 %. We also studied the effect of the thickness of Cu2ZnSnS4 films and we found an increase of the efficiency from 10 to 17 % by increasing the thickness from 0.5 to 3 μm.

Resume : In recent years, Kesterite-based Cu-Zn-Sn-S (CZTS) solar cells are attracting attention owing to its less toxic and earth-abundant elements. However, the current efficiency of CZTS champion solar cells is stagnant around 10% that shows high disparity from its counterpart, Cu–In– Ga–Se (CIGS) solar cell. High voltage deficit in CZTS solar cells is one of the primary causes of low efficiency. Unsolicited interface photogenerated carrier recombination is the prime concern and need to be addressed to reduce voltage deficit. Conventionally used CdS buffer layer in CZTS solar cells has non-ideal energy-level alignment with CZTS which leads to severe interface recombination. In this work, a-TiO2 is proposed as a potential substitute to CdS buffer layer for CZTS solar cell. From X-ray photoelectron spectroscopy (XPS) analysis, the energy level alignment is found to be “spike”-like with conduction band offset (CBO) of 0.17 eV at CZTS-TiO2 interface while it is observed to be “cliff”-like at CZTS-CdS interface with CBO 0.16 eV. Nanoscale probing of CZTS-TiO2 and CZTS-CdS interfaces using Kelvin Probe force Microscopy (KPFM) reveals higher potential barrier at CZTS-TiO2 heterojunction, quenching the back carrier recombination between injected electrons in TiO2 and holes in CZTS layer. Finally, the fast decay response and lower persistent photoconductivity (PPC) of photogenerated carriers for CZTS-TiO2 interface based photodetector, further confirm our results. The energy level alignment and nanoscale investigation signify TiO2 as a potential alternate buffer layer for earth abundant CZTS solar cells.

Resume : Earth abundant kesterite Cu2ZnSn(S,Se)4 (CZTSSe) is known to have suitable optical and electrical properties for low-cost and high-efficient thin-film solar cells. To break through the long stagnation of the power-conversion efficiency (PCE) of kesterite solar cells at 12.6%, the open circuit voltage (VOC) deficit issues ought to be settled in all the kesterites solar cells. For making flexible and high-efficient CZTSSe solar cells, it is vital to add NaF layers during the growth or after the deposition. As they can solve the issue of inhomogeneities and suppress various defects and secondary phases formation, these effects lead the improvement of VOC, fill factor, hence PCEs. In order to optimize the Na-doping, NaF layers were deposited as one of the precursors, during the sputtering. The best efficiency of the flexible CZTSSe solar cells on Mo with NaF layers is 11.4% up to date. Depending on the deposition conditions, grain boundary (GB) properties can be changed. This is because Na segregation and defect formation affect the electrical properties on the surfaces and carrier transport within the materials. The kesterite thin films, grown at relatively low temperatures, would have deep level defects in the vicinity of the GBs. To probe the GB properties, Kelvin probe force microscopy and conductive atomic force microscopy (c-AFM) were employed to measure the local variation of surface potential barrier and surface current, respectively. As a result, we figured out the GBs cannot be appropriate current flow path anymore. Moreover, using photo-assisted c-AFM, we were able to map the photo-current on the surfaces, inducing the recombination position at local-scale. We found that Na plays a role for passivation of Cu-Zn related defects and enhance efficiency substantially. It is probable to enhance the efficiency of kesterite on flexible substrates better than kesterite on sola-lime glasses.

Resume : Environmental friendliness demands the use of non-toxic elements in all types of solar cells, and Zn(O,S) thin film as an alternative buffer layer to replace CdS layer in chacopylite and kesterite solar cells has attracted enormous attention in the past. However, Cu2ZnSnS4 (CZTS) solar cells with Zn(O,S) buffer are far inferior to those with CdS buffer despite the potentially better band alignment. Herein, by intentionally control the precursor compositions, the surface of CZTS can be modified to improve the quality of Zn(O,S)/CZTS junction for chemical-bath-deposited Zn(O,S) buffer. The CZTS surface that can jointly work well with Zn(O,S) buffer has been further investigated using X-ray photoelectron spectroscopy and ultraviolet photoelectron spectroscopy. The results indicate that an ultrathin SnS layer exisits on the CZTS surface and can effectively raise the conduction band edge of the absorber surface to form a conduction band offset barrier of ~ 0.40 eV, significantly better than without the assistance of SnS layer. Our study here may offer a key route for fabricating highly efficient and low cost Cd-free CZTS thin films solar cells.

Resume : Earth abundant Cu2ZnSnS4 (CZTS) is a promising candidate for cost effective energy harvesting devices. CZTS offers high absorption coefficient ~ 104 cm-1 and band gap 1.5 eV matching the solar spectral irradiance. However, the efficiency of CZTS solar cells is limited by the formation of co-existing secondary phases and defects. Various controlled growth mechanisms of CZTS have been proposed to remove detrimental secondary phases like ZnS, SnS, SnS2 and Cu2-xS and other defects to improve the performance of CZTS based solar cells. Several chemical etchants including KCN have been developed to remove Cu2-xS metallic phases, whereas these etchants often change the material chemistry at the surface of the film. In this work, Zn/Sn ratio was tuned to suppress the formation of Cu2-xS phase during the growth itself to eliminate the additional step of etching in device fabrication. Furthermore, Ag doping was done to partially substitute Cu by Ag to probe CuZn antisite defects at nanoscale. Our findings evidently showed significant reduction of Cu2-xS secondary phase with decreased Zn/Sn ratio that resulted in notable efficiency enhancement with efficiency 6.11% for the final composition. Additionally, nanoscale electrical properties of CZTS absorber layer clearly indicates the effective suppression of CuZn defects and increase in localized minority carrier current with Ag doping. This study pave a facile and effective way for the control of detrimental Cu2-xS secondary phase and CuZn antisite defects in CZTS absorber to enhance device properties of CZTS solar cells.

Resume : When developing a photovoltaic technology, three key metrics should be considered: low cost, high power conversion efficiency, and high stability. Although, the kesterite Cu2ZnSn(S,Se)4 compounds have attracted the interest of several research groups as a low cost and non-toxic materiel, the stability of kesterite solar cells is as yet poorly covered in literature. The reason is that most of the research activities are dedicated to solve the large VOC deficit of kesterite solar cells. A promising strategy to overcome the VOC deficit is the reduction of the recombination losses through appropriate cation substitution. Our contribution provides an investigation of the impact of partial substitution of Cu by Li on material properties, on the solar cell performance and stability of (LixCu1-x)2ZnSn(SSe)4 (LiCZTSSe) solid solution monograin. The monograin powders are synthesized from elementary precursors in molten potassium iodide at 740 oC with different Li contents. A monolayer of chemically and thermally treated monograins of homogenous size is embedded into an epoxy resin. Solar cell devices are completed using a CdS buffer layer, an i-ZnO/ZnO:Al front contact and applying a graphite based back contact. The monograin composition and quality are characterized using EDX and XRD. We proved that using slight amounts of Li reveal an increase of the average FF and VOC value of the solar cells. It has been found also that higher Li amounts correspond to higher solar cells stability.

Resume : In this study, the effect of S/Se ratio on the properties of Cu2CdGe(SxSe1-x)4 microcrystalline powders has been investigated. The structural, optical and electronic properties of Cu2CdGe(SxSe1 x)4 (x= 0-1) powders and solar cells are analyzed by X-ray diffraction (XRD), Raman spectroscopy, scanning electron microscopy, energy dispersive X-ray spectroscopy, photoluminescence (PL) as well as current-voltage and quantum efficiency measurements (EQE). XRD results showed that all compounds crystallize in orthorhombic structure and lattice parameters decrease by increasing the S-content in the Cu2CdGe(SxSe1-x)4 solid solutions. Regardless of the composition of the Cu2CdGe(SxSe1-x)4 crystals, all materials showed p-type conductivity. External quantum efficiency measurements show that the band edge of the solar cell device is shifted to shorter wavelength, which enhances the open-circuit voltages. The relative increase of the open-circuit voltage with S/(S+Se) ratio is lower than expected from the band gap energy trend. LT- PL spectra of pure Cu2CdGeSe4 consisted of two PL bands at 1.15 eV and 1.27 eV, from pure Cu2CdGeS4 the PL spectra also consist of two band at 1.69 eV and 1.92 eV. The results showed that Cu2CdGeSe4 has potential properties for single junction solar cell applications, but Cu2CdGeS4 could be used as wide band gap top cell in a tandem solar cell structure or in the photoelectrochemical devices for solar water splitting.

Resume : A way to push photovoltaic (PV) on the terawatt scale in an eco-sustainable way is to investigate and validate possible alternatives to Cu2ZnSnS4 (CZTS) among I2-II-IV-S4 Earth-abundant chalcogenides. A further alternative belonging to this class of materials is Cu2FeSnS4 (CFTS), which consists of abundant and non-toxic elements and shows high absorption coefficient and direct band gap suitable for PV applications. In this work, CZTS and CFTS thin films were grown on Mo-coated soda lime glasses by a two-step process: sputtering, that combine high-throughput with an elevated reproducibility has been used to deposit stacked metallic precursors, successively annealed in sulfur vapors. A comparative study has been carried out in order to identify the main similarities and differences among the deposition parameters for pure thin films of this two chalcogenide. In particular, the effect of Mo-Back contact, the order of sputtered precursors in the stacked structure and the steps of the thermal cycles for the sulfurization step have been investigated in order to study like this parameters affect the growth of CZTS and CFTS thin films for suitable for PV applications. The homogeneity, crystallinity, and chemical composition of the thin films have been used like references for our comparative analysis. The PV parameters of some proof of concept solar devices will be also compared and exploited to validate CFTS like a potential alternative to CZTS for PV application.

Resume : The Cu2ZnSnS4 (CZTS) semiconductor is a promising material for light absorption in photovoltaic (PV) devices because of its unique combination of intrinsic properties (high absorption coefficient, direct bandgap of 1.5-1.7 eV, abundancy, and non-toxic constituents (compared to Cu(In,Ga)Se2 and CdTe). While the classical route to integrate CZTS into solar cells has been focusing on its association with the n-CdS through a complex and multi-step processing to achieve the multilayered stack (SLG/Mo/p-CZTS/n-CdS/ZnO/ITO), we propose here an alternative, yet straightforward approach consisting of the use the pulsed laser deposition (PLD) to deposit directly highly-crystalline CZTS films onto n-silicon wafers. The optimization of the PLD process, including the in-situ adjustment of the Zn content, makes it possible to achieve CZTS films with the required structural and optoelectronic properties without resorting to any post sulfurization and/or thermal treatments. The optimally produced PLD-CZTS films (~1.2 µm-thick) were shown to exhibit the pure kestërite phase with an optical bandgap of ~1.7 eV, a smooth surface morphology (Rms ≤ 40 nm), and a relatively high work function of 4.75 eV. These PLD-CZTS films were directly integrated into a relatively simple Al/n-Si/p-CZTS/ITO structure to form heterojunction devices and found to yield a power conversion efficiency (PCE) of 2.2 % (under AM 1.5 solar illumination). The PCE performance of these p-CZTS/n-Si based PV devices was associated with the relatively high built-in potential at the heterojunction along with a significant reduction of Cu-on-Zn defects in the PLD-CZTS films.

Resume : This paper describes synthesis of earth-abundant non doped and Na-doped Cu2ZnSnS4 (CZTS :Na) ingots from high purity (99.999%): copper (Cu), zinc (Zn), tin (Sn), sulfur (S) and sodium (Na) elements easily by direct fusion. CZTS :Na thin films were prepared by thermal evaporation method on glass substrates heated at 100°C and subsequent annealing at 400°C in sulfur atmosphere for 2h. X-ray Diffraction (XRD) of ingots confirmed tetragonal CZTS. All sulfurized CZTS thin films present polycrystalline nature and exhibit a preferential orientation along (112) plane. Optical measurement analysis showed high absorption coefficients greater than 105 cm-1. The band gap of CZTS :Na is about 1,4-1,7 eV with electrical resistivity of 20-100 Ωcm and p-type conductivity. Keywords: Copper-zinc-tin-sulphide ; Na-doped ; Thermal evaporation ; Thin films ; Caracterization ; Photovoltaic material.

Resume : Quaternary semiconductor Cu2CoSnS4 (CCoTS) compound were successfully synthesized using a novel process as the thermal vacuum evaporation on low substrates temperature. The substrates temperature was changed from room temperature (RT) up to 200°C in 25°C in steps. The as-grown samples were sulfurized so as to optimize the stannite phase CCoTS. After that, the acquired thin films were annealed under vacuum at 450°C for 2 hours. In this paper, we examined the structural, morphological and optical properties by X-ray diffraction technique (XRD), Raman spectroscopy, scanning electron microscopy (SEM) and optic spectrophotometer UV-VIS-NIR. The sulfurization process promotes the formation of the desired phase without secondary phases, which was confirmed by X-ray diffraction and Raman analysis. Furthermore, the vacuum annealing improved the crystalline quality of the films deposited on the different substrates temperature. The elemental compositions of the sulfurized CCoTS thin films after annealing were close to the stoichiometry. Optical measurement analysis showed high absorption coefficients greater than 105 cm-1. The p-type conductivity of CCoTS thin films was investigated by hot probe method. Keywords: CCoTS; Thermal vacuum evaporation; Low substrates temperature; Thin films; Absorber layer; Photocatalyst.

Resume : We have successfully produced Cu2CoSnS4 (CCTS), Cu2NiSnS4 (CNTS), Cu2ZnSnS4 (CZTS) and Cu2FeSnS4 (CFTS) materials by direct melting. Constituent elements taken in stoichiometric compositions were used as starting points. After that, Cu2XSnS4 (M=Co, Ni, Zn and Fe) thin films were growth using a single source vacuum thermal evaporation method. The resulting Cu2XSnS4 films were characterized by X-ray diffraction, Raman spectroscopy, scanning electron microscopy, energy dispersive X-ray analysis and optical spectrophotometry to investigate the crystal structure, composition, morphology, absorption coefficient and optical band gap. Hot probe method shows that Cu2XSnS4 (M=Co, Ni, Zn and Fe) ingots and fabricated samples exhibit p-type semiconductor. This work demonstrates that Cu2XSnS4 (M=Co, Ni, Zn and Fe) thin films are promising absorber materials for energy-conversion applications. Keywords: Cu2XSnS4; Thin films, Thermal evaporation; Structural properties; Optical Properties.

Resume : In this work, Cu2ZnSnS4 (CZTS) and Cu2ZnGeS4 (CZGS) thin films were deposited by co-sputtering. The bi-layered CZTS/CZGS precursors were prepared by deposition of CZTS of varying thickness 600, 800 and 990 nm on top of CZGS of thickness 100 and 200 nm. TiN layers (25nm) on top of the Mo contacts were used to improve the adhesion of the CZGTS films during device processing. Cross-section scanning electron microscopic (SEM) image showed different morphology of CZTS and CZGS precursors along with the presence of TiN bottom layer on top of Mo coated glass. The different batches of bi-layered films were sulfurized for a different time (1, 6 and 8 min) in order to study the Ge-Sn gradient throughout the thickness. Cu2ZnGexSn1-xS4 (CZGTS) formed with the kesterite phase confirmed with XRD and Raman studies. Compositional profiles of bilayer precursors and sulfurized material are obtained by Glow discharge optical emission spectroscopy (GDOES). The diffusion of Ge is faster than in our previous experiments with Ge in the present study seen on the surface of CZGTS for all different sulfurization times as observed in GDOES. Slight increase of Ge amount towards the CZGTS surface is noticed by observing peak shift in grazing incidence X-ray diffraction (GIXRD) however, this shift may also occur close to the critical angle. The short circuit current (Jsc) in CZGTS solar cell is found to be lower along with similar open-circuit voltage (Voc) as compared with corresponding reference CZTS solar cells. The tuning of energy level alignment using alternative buffer layers is expected to improve performance. Further investigations are needed to understand the reasons for the faster smearing out of the Ge-Sn gradient as compared to previous experiments and annealing tunability for desired gradients.

Resume : Copper zinc tin sulphide (Cu2ZnSnS4) (kesterite) is emerging as a promising material for absorber layers in thin-film solar cells. This non-toxic semiconductor material (p-type) with direct bandgap and large absorption coefficient (104cm−1) has recently received wide attention as an absorber layer in thin-film solar cells. Among the many different growth techniques, solution-based methods are interesting for their potential low cost compared to vacuum-based techniques. In this work, CZTS thin-film solution-precursor based absorber layers were prepared by spin coating on Mo coated glass. These are subsequently annealed by different treatments, namely: 1) heating in air 2) heating in sulphur ambient 3) heating in selenium ambient 4) heating in S+Se ambient, within a temperature range from 450 °C to 550 °C for 30 min. The influence of these post-deposition annealing steps on the material properties of the absorber and on the finished devices was studied by means of different spectroscopies such as X-ray diffraction, Raman and X-ray photoelectron spectroscopy, and microscopy techniques such as atomic force microscopy, scanning electron microscopy as well as transmission microscopy. The performance of the finished cells will be discussed in sight of the information collected.

Resume : The prospect of Cu2ZnSn(S,Se)4 (CZTSSe) material applied in photovoltaic applications has attracted considerable attention owing to its elemental abundancy in Earth’s crust without any material shortage issues. In this contribution, we demonstrate that CZTSSe solar cells with a novel design of ultrathin ZnO-based transparent electrodes lead to a high efficieny. CZTSSe solar absorber layers were prepared on molybdenum-coated glass by spin-coating from an environmental-benign solar ink composed of dimethylsulfoxide (DMSO) and 0.57 M CuCl2, 0.39 M ZnCl2, 0.53 M SnCl2 and 1.85 M thiourea. This process was followed by subsequent air-annealing at 300 °C and sulfo-selenization at 620 °C under atmospheric pressure. A dense and continuous CZTSSe absorber layer with a thickness of 1.2 µm and various Se/(Se+S) compositional ratios could be prepared using this route. Analysis by grazing incidence X-ray diffraction confirmed the compositional gradient of Se/(Se+S) in each as-sulfo-selenized CZTSSe absorber layer. The fabrication of the CZTSSe cell was completed with an implementation of an ultrathin transparent electrode design comprises multilayer structure of sputtered Al-doped ZnO (AZO) and Ag stacked layers with a maximum thickness of 100 nm. This electrode design showed optical transmittance > 80 % and a sheet resistance < 12 Ohm/sq. The EQE of the cells revealed that the CZTSSe absorbers exhibit band gaps from 1.07 eV to 1.47 eV as a function of Se/(S+Se). An S-rich CZTSSe cell employing the novel transparent electrode design was produced with an efficiency above 8%. This result also opens an opportunity to reduce material budget for transparent electrode material in high-efficient CZTSSe solar cells.

Resume : Cu2ZnSnS4 (CZTS) is a benign and cheap photovoltaic absorber material that represents a valuable alternative to the more expensive chalcogenide systems, i.e. Cu(In,Ga)S2 (CIGS). In this contribution, the formation pathway of CZTS nanoparticles under scalable solvothermal conditions is systematically investigated using salts with chloride or acetate counteranions as molecular precursors. The results show that the use of chloride salts lead to the predominant formation of CTSs ternary phases (Cu2SnS3, Cu3SnS4), whereas the formation of the CZTS phase is not observed even for higher temperature and longer reaction time (250 °C, 24 h). In case of acetates, the copresence of CZTS as main product and binary and ternary phases is observed already at lower temperature and shorter reaction time (200 °C, 2 h), while by increasing the reaction time and temperature, only CZTS formed. Besides the microstructural characterization of the as-synthesized materials by Raman spectroscopy, X-ray diffraction (XRD), Energy Dispersive X-ray Spectroscopy (EDX), X-ray Photoelectron Spectroscopy (XPS) and High Resolution Transmission Electron Microscopy (HRTEM), we shed light on the reactivity between the metal precursors, the organic ligand oleylamine and the sulphur precursor carbon disulfide (CS2) by 13C-Nuclear Magnetic Resonance (13C-NMR). In this way, a rational for the lack of formation of CZTS particles from chloride precursors is provided. Ahmad et al. Inorganic Chemistry, in press.

Resume : Alloying of cations or anions is an established method for tailoring the band gap energy. In this work we discuss the solid solution series Cu2(Zn,Fe)S4, Cu2(Zn,Cd)S4 and Cu2(Zn,Mn)SnSe4 in which divalent Zn is substituted systematically by Fe, Cd and Mn. The end members Cu2ZnSnS4 and Cu2ZnSnSe4 crystallize in the kesterite structure (s. g. I4 ̅ ) [1], whereas Cu2FeSnS4 [1], Cu2CdSnS4 [2] and Cu2MnSnSe4 crystallize in the stannite structure (s. g. I4 ̅2m) [2]. Both the kesterite and the stannite structure are based on a tetrahedral coordination but different cation arrangements: perpendicular to the crystallographic c-axis in the stannite structure Zn-Sn planes change with Cu-Cu planes, whereas in the kesterite structure Cu-Sn planes and Cu-Zn planes, showing Cu/Zn disorder, are present. We studied the kesterite - stannite structural transition in these three solid solution series as well as the band gap energy variation with chemical composition. The detailed structural investigations are based on neutron powder diffraction experiments which provide the cation distribution in the crystal structure allowing a distinction between the isoelectronic cations Cu+ and Zn2+ as well as the electronic similar cations Mn2+ and Fe2+. A remarkable structural feature is the abrupt change of the tetragonal deformation c/2a (a and c are lattice parameter). In Cu2Zn1-xFexS4 and Cu2Zn1-xCdxS4 this change occurs at the same x value (x~0.3). In Cu2ZnSnS4 is c/2a~1, where as in Cu2ZnSnSe4 is c/2a<1. Thus in the Cu2(Zn,Mn)SnSe4 series c/2a increases by substituting Zn by Mn until a value of c/2a~1 is reached. The structural transition is characterized by a cation re-distribution process which will be discussed in detail. [1] Schorr et al., Europ. J. Mineral. 19 (2007) 65 [2] Schäfer et al., Z. Krist. 145 (1977) 356

Resume : The open circuit voltage deficit of Cu2ZnSn(S,Se)4 (CZTSSe) kesterite solar cells is higher than the closely related Cu(InGa)Se2 solar cells. A promising strategy to overcome this large deficit is by reducing recombination losses through appropriate cation substitution. The influence of Ag incorporation on the CZTSSe thin film solar cells has been studied by several research groups. Our contribution provides an investigation of the impact of substitution of Cu by Ag in CZTSSe solid solution monograins. We focus on its influence on the material properties and on the solar cell performance. The Ag2ZnSn(SSe)4 (AZTSSe) monograin powders are synthesized from elementary precursors in molten potassium iodide at 740oC. A monolayer of chemically and thermally treated monograins of homogenous size is embedded into an epoxy resin. Solar cell devices are completed using a CdS buffer layer, an i-ZnO/ZnO:Al front contact and applying a graphite based back contact. A photoluminescence study was executed on AZTSSe and CZTSSe absorber layers and a comparison study of the radiative recombination mechanism of both material was carried out. The monograin composition and quality is characterized using EDX and Raman spectroscopy. An investigation of the probable position of Ag atoms in the kesterite structure is carried out based on nuclear magnetic resonance (NMR). AZTSSe and CZTSSe monograin solar cell results and results of the ongoing optimization of the processes will be presented as well.

Resume : Kesterite (CZTSSe) is an attractive thin-film material for photovoltaic applications due to its excellent optical properties and the use of abundant elements. CZTSe solar cells routinely include CdS/ITO as an electron transport layer on the top surface. However, CdS is recognized as a toxic material and its use in electronics involves complex treatments of electronic waste. In this way, it can be interesting to replace the CdS based contacts by means of selective collection electrodes based on transition metal oxides (TMOs). Dopant-free materials based on TMOs such as MoOx and V2Ox, or alternatively TiO2 films are a good option to form hole or electron transport layers respectively in silicon technology. In this work, we explore applying TMOs materials in CZTSe absorbers to make the solar cell contacts. With this idea in mind, we have applied thermally evaporated MoOx and V2Ox, as well as V2Ox films deposited by Atomic Layer Deposition (ALD) as hole transport layers (HTLs). In all cases, the HTL scheme includes a Ni/Al capping metallization deposited by thermal evaporation. Specific contact resistivity was extracted from the Transfer Length Method (TLM) measurements, reaching remarkable values below 10 mΩcm-2. Furthermore, here we report a CdS-free electron selective contact based on ALD AL2O3/TiO2 stacks, which incorporate a thermally evaporated two-layer metallization scheme of Mg/Al as a capping electrode. Vertical diode test structures, using the foregoing stack as ETL in the front side and a convectional Mo-based ohmic contact at the rear side, reveal good diode-like I-V characteristics with low specific series resistance Rs (~0.35 Ω∙cm2), low reverse leakage currents, i.e. high shunt resistances, and ideality factors around 1.9.

Resume : Understanding the formation pathways of Cu2SnZnSe4 (CZTSe or kesterite) is crucial for absorbers synthesised from the selenization of metallic stacks. Strategies like annealing under high Se pressure and the use of Ge-containing or bronze-based metallic precursors have been shown to induce a formation route based on the Cu2SnSe3 (CTSe)+ZnSe reaction which, in contrast to that based on binaries, offers a better control of the composition, secondary phases and defect formation ultimately leading to better performing CZTSe devices. However, these synthesis routes have reached a dead end and innovative ideas are necessary for further improving the kesterite technology. As such, this work explores Zn-incorporation in the CTSe-CZTSe system. We present a complete structural characterisation of Cu-Sn-Zn-Se absorbers with varying Zn content synthesised from sputtered metallic precursors. Employing WDS, XPS, XRF, XRD (Rietveld) and Raman spectroscopy, we observe an effective incorporation of Zn in the samples progressively transitioning from a pure monoclinic CTSe (Cc) compound (Zn/Sn<0.1) to an unreported CTSe:Zn phase with a disordered chalcopyrite structure (I-42d) (0.10.75). No segregation of metallic Zn or ZnSe is detected at any point. These results open the way to new CZTSe synthesis strategies using Zn-poor steps followed by Zn compensation which may be a key to a better control of the kesterite synthesis process.

Resume : Kesterite Cu2ZnSnSe4 (CZTSe) structure has been studied as a promising material for a thin-film solar cell, but its solar cell has difficulties to get a high open-circuit voltage (Voc), differently from other CIGS and/or CdTe based thin-film solar cells. Anti-site defects are treated as the main problem to improve Voc in the CZTSe device, and therefore, defect engineering has been studied in an aspect of cation ratios of CZTSe film. Here in this study, the reason for the defect formation is studied related to the reaction pathway for the formation of the CZTSe film. Four types of annealing processes are performed at the same precursor with a sequence of Mo/Zn/Sn/Cu. Reaction pathways are confirmed by measurements of X-ray diffractions and Raman spectroscopy. Afterward, solar cells are fabricated at the same time, using the four CZTSe films. J-V curve and Admittance spectroscopy measurements are performed to observe the device performance and the defects in each sample. Four samples are shown different efficiencies from 8.0% to 10.9% without anti-reflection coating, and the measured defect levels are different depending on an annealing process although all CZTSe films have similar crystallinities and bandgaps of ~1.0 eV. Especially FF is significantly affected by defect energy levels. Additionally, CZTSe film has mainly shallow defects when CZTSe is formed by CuSe, ZnSe, and SnSe2 (and/or SnSe) alloys.

Resume : In recent years, efficiencies exceeding 10% for solution-processed Cu2ZnSn(SxSe1-x)4 have been demonstrated using aprotic molecular inks [1]. For tandem applications, a higher band gap of 1.5-1.7 eV is favorable, thus the sulfide Cu2ZnSnS4 and its alloys should be considered. A major component of fabricating an efficient solution-processed solar cell is transforming the liquid ink into a solid film. Today, few studies exist on the formation mechanism of the kesterite via the aprotic solvent routes. In this study, we map ink constituents, investigate different spin-coating and pre-annealing conditions, and discuss the different phenomena occurring during the drying of the films. We find that the grain size, fine-grain and MoS2 layer thickness vary with thermal exposure during pre-annealing. The best performing solar cells were produced by utilizing low thermal exposure routes, contradicting most prevalent methods used in the field. Among the processes occurring during transformation from ink to film, we touch upon solvent evaporation, complex decomposition, phase formation, hydrodynamic processes within the film and the re-dissolution of the film during spin-coating. Novel characterization techniques such as thermogravimetric analysis coupled with mass spectrometry (TGA-MS), inductively coupled plasma optical emission spectrometry (ICP/OES) and Raman spectroscopy are used to characterize liquid samples, while standard morphology, composition and phase studies are carried out on thin-film samples. [1] Y. Gong, et al., 2018 IEEE 7th WCPEC, p. 805-807.

Resume : Kesterite thin film solar cells have been widely recognised as a promising photovoltaic absorber material. It consists of earth-abundant and environmental-friendly element, and has a tunable bandgap (Eg, 1.0 – 1.5 eV). In order to investigate the electrical properties of different types of kesterite thin film solar cells under low light intensity condition (~ few hundred Lux), three different types of kesterite thin film solar cells, Cu2ZnSnSe4 (denoted as CZTSe), Cu2ZnSn(S,Se)4 (denoted as CZTSSe), Cu2ZnSnS4 (denoted as CZTS). In order to compare the indoor performance of CZTSe, CZTSSe and CZTS, we prepared the samples which have similar power conversion efficiency, 8 ~ 9 %. Also, we screened each samples which represent the electrical properties of its own bandgap (1.0, 1.13 and 1.5 eV). Since the three samples has different bandgap, it is well indicated in the J-V curves. Pure selenide CZTSe shows the highest Jsc (36.3 mA/cm2), but lowest Voc (412 mV) as its bandgap is smallest among the samples. Similar to this, pure sulfide CZTS demonstrates the highest Voc (625 mV) and lowest Jsc (19.9 mA/cm2), and those of CZTSSe are in the middle. The bandgap of CZTSe, CZTSSe and CZTS are different, so different absorption edge of the samples are shown in the EQE results. J-V curves of CZTSe, CZTSSe and CZTS samples were measured at 400 lux (halogen lamp) and their power output. Unlike 1 sun J-V measurement, as they show similar PCE, the power output of the samples are much different. At 400 lux, CZTSSe shows the highest power output (45 um/cm2) which is more than 2 times higher than that of pure sulfide CZTS (19.1 um/cm2). It is surprising that PCE of the samples at 1 sun is similar to each other, but CZTSSe shows nearly two times higher power output than other two samples. Since, there is not formula to transfer Lux to light intensity (W/cm2), we did not calculate the PCE of the samples at low light intensity, rather comparing the device performance by power output.

Resume : With c-Si solar cells approaching the Shockley Queisser limit at a competitive cost, multijunction concepts are among the few pathways to further improve the performance. In this work, we investigate on the feasibility of combining a narrow band gap (1.1 eV) Si bottom cell with a wide band gap Ge based Kesterite Cu2ZnGe(S,Se)4 (CZGeSSe) top cell. CZGSSe exhibits promising properties such as a tunable band gap value of 1.4 eV (pure Se) and 2 eV (pure S). High efficiency and excellent transmission below the bandgap are the main criteria to optimize in a top cell. Various transparent conducting oxides (TCO) have been proposed as transparent back contact, among which ITO and FTO are commonly used for Cu(In,Ga)Se2, perovskite and organic solar cells. In this work, strategies for the development of CZGSSe solar cells on such TCO’s is proposed. The suitability of transition metal oxide (TMO) ultrathin interlayers improving the contact ohmicity is evaluated; different thicknesses of TMOs’ such as MoOx, V2O5, Nb2O5 etc. were deposited by thermal evaporation, while a CZGSSe absorber layer was synthesized by selenization of a sputtered metallic precursor. Solar cells were completed a standard substrate configuration. Optoelectronic properties of the completed devices were studied through material and electrical characterizations such as Raman, UV-Vis, J-V and EQE analysis etc. along with 1D numerical modelling. The complete set of results will be presented and discussed.

Resume : In order to demonstrate back bandgap gradient profiles in kesterites (Cu2Zn(Sn,Ge)Se4-CZ(T,G)Se) a cationic substitution strategy (Sn by Ge) is implemented. During the sputtering precursor deposition, the Sn and Ge layer thicknesses were adjusted in order to produce different Sn/Ge ratios. The compositional profiles of the CZ(T,G)Se absorber layers were analyzed by Auger and glow discharge optical emission spectroscopies (GDOES). Using these techniques, the hypothesized (Ge/Ge+Sn) graded composition with an increase of Ge content towards the back side of the CZ(T,G)Se film is evidenced, further supported by XRD studies and Raman spectroscopy in the front and back interfaces of the films. Solar cells were fabricated (SLG/Mo/CZ(T,G)Se/CdS/ZnO:i/ITO) and up to 2% absolute efficiency increase was observed for devices fabricated with absorbers with Ge back graded composition (PCE ≈ 9.5%). These studies lead to conclude that a carrier collection improvement occurs due to the additional drift electric field, causing at the same time an increase in the JSC values because of the inhibition of recombination at the back contact side (back surface field reflector). This demonstrates that Ge behaves similarly than Ga in Cu(In,Ga)Se2, thus opening relevant perspectives for graded bandgap in kesterites. A complete optoelectronic characterization of the absorbers and devices will be also presented.

Resume : Tandem and multi-junction solar cells have attracted a huge attention by the researchers in the field. To better understand the effect using different absorbers in the tandem solar cell devise, we simulate the tandem solar cell using SCAPS and modelled the devise with Mathematica software. Firstly, we studied and analyzed each single absorber CZTGeSSe and Si separately; then, we checked the stack of the tandem solar cell. In the present work, the theoretic model adopted to simulate the single and tandem devise is presented, and the relevant results giving the solar cell key parameters affecting the performance of the tandem solar cell are given.

Resume : Sn based kesterite has been under study for along and has given a promising performance exceeding 12%, an efficiency that remains humble when compared to CIGS and CdTe. Many attempts have been made to optimize Sn based kesterite, and the introduction of Ge in kesterite lattice is one possibility, representing an emerging approach aiming to improve the device efficiency and widen the band gap up to 2eV. The incorporation of Ge has been explored in several ways. For example, its use as dopant has yielded to an efficiency increment from 7% to 10%. The alloy of Sn compound with Ge indicates the possible tunability of band gap without affecting the crystalline structure of kesterite. The total substitution by Ge has as purpose to enlarge the band gap up to 2eV and suppress the defects related to the multicharged character of Sn. The highest efficiency reported up to now for Cu2ZnGeSe4 (CZGSe) is 7.6% using H2Se environment. This work aims at understanding and optimizing the relevant process parameters affecting the optoelectronic properties of CZGSe such as the thermal treatment (Annealing profile and Temperature) as well as the etching agents. Noting that CZGSe is synthesized using elemental Se in Ar atmosphere at lower temperature annealing than the one used for Sn kesterite. The champion cell obtained from preliminary optimization has an efficiency of 6.5% with parameters comparable to the record performance (Voc=556mV, FF=59.6%). A complete optical, electrical, morphological and compositional characterization will be presented, and correlated with the corresponding process parameters, showing that CZGSe can be formed and crystallized at lower temperatures than the corresponding Sn-kesterite, thanks to the formation of an intermediate liquid phase.

Resume : It is of interest to improve the performance of low bandgap (Eg ≤ 1.1 eV) CuInGaSe2 (CIGS) in order for CIGS to be incorporated as a bottom cell in tandem solar cells. This work has investigated Ag alloying and alkali fluoride treatments as methods to improve low Eg CIGS devices with Ga/(Ga+In) (GGI) ≤ 0.05. CuInSe2 (CIS) and CIGS with Ag alloying (ACIS, ACIGS) have shown enhanced grain size and reduced carrier concentrations. ACIS and ACIGS with GGI = 0.05 also have improved long wavelength current collection due to an increased space charge width. Subsequently, to improve the Voc in low Eg devices, a KF-PDT was investigated. In CIS with the KF-PDT, a high efficiency of 16.0% was achieved with Voc = 526 mV. However, ACIS has responded differently to the KF-PDT, with reduced Voc due to a dominant interface recombination mechanism. Raman spectra indicated an unidentified compound on ACIS+KF, which could be related to the observed interface recombination. Furthermore, CdS coverage improves with the KF treatment, and ACIS+KF has complete coverage within less than half of the normal CdS deposition time. Additional characterization will be performed to provide insight on the role and formation of secondary phases on KF-treated ACIS as well as the enhanced CdS growth on ACIS+KF.

Resume : Thin-film photovoltaic technologies are a prevalent alternative to silicon-based systems due to their greater device efficiencies, high-throughput deposition methods, and lower material requirements. Developments in the commercial manufacturing of thin-film photovoltaic modules have been directed to optimizing a low-cost, high-rate deposition process. Notably, the success of CdTe manufacturing has much relied on the high-rate, low-temperature deposition of CdTe films followed by a post-deposition CdCl2 vapor treatment. In this treatment, CdTe films are recrystallized by a short anneal in a CdCl2 environment with some oxygen. Such treatments have resulted in the low-cost production of uniform, high performance devices with minimal chlorine contamination. The thin-film chalcogenide CuInxGa(1-x)Se2 (CIGS) is a high-efficiency photovoltaic absorber material typically deposited by an expensive and time-intensive three-stage co-evaporation process at elevated temperatures. To be more cost competitive amongst thin-film technologies, CIGS must develop a similar post-deposition recrystallization procedure to that of the CdCl2 vapor treatment of CdTe. Our previous work has investigated post-deposition heat treatment conditions necessary to recrystallize low-temperature co-evaporated CIGS films. In this work, CIGS films were co-evaporated by collaborators onto molybdenum-coated soda-lime glass at 350°C and 400°C and annealed at various times and temperatures with chloride and bromide vapors. The treatments resulted in increased grain size and improved crystallinity. Treatments resulted in varying changes to composition; it is notable that for all treatments the amount of gallium in the films decreased significantly. Additionally, varying amounts of chlorine remained in the films depending on the treatment. To better understand the recrystallization kinetics and composition changes behind these studies, this work investigated the effect of annealing films only in the presence of InCl3 vapors. For the purpose of a more controlled study, CuInSe2 (CIS) films were sputtered in-house by rf-magnetron sputtering onto silicon substrates at room temperature. Using CIS instead of CIGS removed the variable of gallium depletion noted in previous experiments. Annealing chambers were made by milling holes into graphite blocks with 1”x1” stages for a sample to sit in and create a closed environment for each hole. The holes in the graphite annealing chambers were loaded in a nitrogen-purged glovebox each with ~0.002 g of InCl3 powder and sealed with a face-down CIS/Si sample. The prepared chambers were loaded into a pre-heated tube furnace and annealed at 350°C and 450°C for 20 minutes each. The systems were removed and allowed to cool in air. Surface morphological changes were studied by a JEOL JSM-7000F Field Emission SEM. In the same system, composition was analyzed by EDAX “Octane Pro” SDD energy dispersive spectroscopy of x-rays (EDS). Crystallographic information was collected and analyzed using a Malvern Panalytical Empyrean x-ray diffractometer (XRD) and the International Center for Diffraction Data (ICDD) database. The micrographs of the films as-deposited had no obvious surface morphology, appearing to be amorphous. The structure, as determined by XRD, saw broad, low-intensity peaks for the corresponding (112), (220), and (312) CIS planes, revealing little to no crystallinity of the films. This supported the amorphous-looking surface micrograph. Sharp peaks corresponding to the silicon substrate appeared, as expected. The composition of the as-deposited film, as determined by EDS, was Cu-rich. The composition was 26.9 at% Cu, 22.0 at% In, and 51.1 at% Se. Annealing treatments conducted at 350°C and 450°C for 20 minutes each resulted in large grain size enhancement. Grains from the 350°C anneal averaged approximately 0.075 μm2 and grains from the 450°C anneal averaged approximately 0.5 μm2. The composition of the film annealed at 350°C was 26.93 at% Cu, 22.4 at% In, 48.69 at% Se, and 1.99 at% Cl. The composition of the film annealed at 450°C was 27.88 at% Cu, 23.61 at% In, 46.76 at% Se, and 1.75 at% Cl. Both treatments mostly preserved the initial composition of the film as-deposited and had little to no chlorine contamination. Both treatments resulted in increased crystallinity as noted by the sharper, higher-intensity peaks of the corresponding (112), (220), and (312) CIS planes in the XRD patterns. (112) preferential growth was seen after both treatments. Additionally, similar peak ratios were noted between the (112), (220), and (312). It is notable that more peaks from secondary phase formation appeared in the 450°C pattern compared to the 350°C pattern. This preliminary research shows promise to further investigations of post-deposition treatments of low-temperature sputtered CIS films. Additional experiments will be conducted and reported at the 2020 Spring Meeting of the European Materials Research Society. CIS films will be rf-magnetron sputtered under the same conditions onto silicon substrates at various substrate temperatures. The samples will be loaded into the same graphite annealing chambers with the same amount of InCl3 in each. The chambers will be annealed at temperatures ranging from 300°C to 500°C for times ranging from 10 to 40 minutes. Morphological, compositional, and structural characterization will be conducted to study the effects of each vapor treatment on the different films. Additionally, secondary-ion mass spectroscopy (SIMS) measurements will be collected to have a better understanding of the composition of each sample with depth resolution. These results will be used to determine a relationship between temperature, time, and grain size and a better understanding of the recrystallization kinetics of InCl3 vapor treated CIS films.

Resume : Highly efficient copper indium gallium diselenide (CIGS) solar cells are limited by different electrical and optical losses. A major cause of electrical losses is the presence of the highly recombinative CIGS/Mo interface. A GGI ([Ga]/([Ga] + [In]) gradient towards the back prevents the carrier recombination by reducing the minority carrier density at the CIGS/Mo interface. In this work, we intentionally modify the gallium and indium evaporation rates during the first coevaporation stage to obtain different back grading profiles. The height of GGI back grading (ΔGGI) resulting from different process sequences is quantified based on secondary ion mass spectrometry. In order to understand the effect from back grading, we carry out time-resolved photoluminescence measurements on CIGS absorbers with different configurations (CIGS/Mo, delaminated CIGS/air and CIGS/Al2O3). We identify the minimum ΔGGI value required to substantially reduce the recombination at the back contact interface. Carrier lifetime and open circuit voltage (VOC) deficit present a high correlation with ΔGGI. For a comprehensive understanding, the experimental data are confronted to Sentaurus TCAD numerical simulations and the effect of ΔGGI on carrier dynamics and photovoltaic properties is evaluated.

Resume : Two significant cost items for CIGS-based thin film PV modules are (1) materials cost for the absorber feedstock and (2) capex for the vacuum-based absorber deposition systems. Both of these will experience a dramatic reduction of more than 50 % if absorber thickness is reduced with the same amount, provided the deposition rate can be kept at the same level as for thick absorbers. In this work, we report on progress of thin film processes for an (Ag,Cu)(In,Ga)Se2 (ACIGS) absorber thickness of 0.7 µm, which corresponds to a thickness reduction, compared to record efficiency ACIGS devices, by a factor of 3. Several approaches to this thickness reduction are tested and compared, including Ga/(Ga+In) grading profiles as well as alkali treatment schemes and substrate temperature. Solar cells with a Voc of 750 mV and 80 % fill factor were obtained for the 0.7 µm thick absorbers. In order to achieve a high back contact reflectance and increase the current further to reach above 20 % efficiency, the robust Mo that is normally used for ACIGS will be replaced with a silver-based stack of materials, that earlier has been tested for CIGS. The integrity of this stack has been shown to depend on the CIGS process, not least the substrate temperature during absorber growth, which was the motivation for the co-optimization of temperature and process schemes for ACIGS.

Resume : Incorporation of heavy-alkali elements is indispensable in achieving highly efficient CIGS solar cells. For example, CsF post-deposition treatment (PDT) was employed in demonstrating the world record CIGS device from Solar Frontier in 2019. However, fundamental understanding of the roles of heavy alkali-PDT in improving CIGS solar cells calls for further studies. In this work, we studied the effects of CsF-PDT on the material and device properties of CIGS thin-film solar cells. The amount of Cs concentration incorporated into CIGS was optimized and resulted in the enhancement of the device performance from PCE of 15.9% to 18.4% with a significant gain in open-circuit voltage (Voc) by 50 mV. We found clear differences between the grain boundaries (GBs) and intragrain region in terms of composition and electronic potential through atomic probe tomography and Kelvin probe force microscopy. The devices were measured with electron beam-induced current to map the paths of charge transport in CIGS absorbers. Photoluminescence and capacitance-based analysis were employed to study the change of defect chemistry of the solar devices. It is revealed that both deep-level donor-like defects such as VSe and InCu and deep-level acceptor-like defects such as VIn or CuIn are passivated by CsF-PDT. Moreover, the enhanced photocarrier density and recombination lifetime after CsF-PDT were measured by the novel carrier-resolved photo-Hall system, the details of which will be discussed.

Resume : Incorporation of alkali metals into the Cu(In,Ga)Se2 (CIGSe) lattice has led to a continuous improvement of efficiency and has positioned this technology among one of the best-performing photovoltaic materials with a world record efficiency of 23.4%[1]. The incorporation of heavier alkali elements (K, Rb and Cs usually in the form of alkali fluorides) through a post-deposition treatment (PDT) results in a gain in open circuit voltage, where the role of the grain boundaries has been confirmed to be of great importance [2,3]. The complex nature of the relationship between CIGSe and alkali metals has complicated both the efforts to understand the rationale behind such improvement and to discern the effects associated with the grain interiors and the grain boundaries. Thus, in order to shed some light on this regard, we perform a set of experiments on single crystalline films. We investigate the effect of a KF PDT on the quasi-Fermi level splitting (QFLS) of a series of CIGSe single crystals grown by metalorganic vapour phase epitaxy with varying Cu content. We find that improvement of more than 60meV in the QFLS can be achieved despite the absence of grain boundaries. Preliminary results of secondary ion mass spectroscopy reveal the presence of potassium inside the bulk of the films, suggesting that transport of potassium can occur through grain interiors. In addition, we confirm that the effectiveness of a PDT depends on the Cu-content as already described in [4] for a RbF PDT, showing that the observed composition dependence is due to bulk properties rather than grain boundaries. To investigate the influence of potential changes in the doping concentration, a time-resolved photoluminescence study will also be presented. References: [1] M. Nakamura, K. Yamaguchi, Y. Kimoto, Y. Yasaki, T. Kato, and H. Sugimoto, 46th IEEE Photovoltaic Specialist Conference (IEEE, Chicago, 2019). [2] S. Siebentritt, et al., "Heavy alkali treatment of Cu(In,Ga)Se2 solar cells: surface versus bulk effects" Adv. Energy Mat. in press (2020). [3] T. Kodalle,et al., "Elucidating the Mechanism of an RbF Post Deposition Treatment in CIGS Thin Film Solar Cells" Solar RRL, 2(9), (2018). [4] T. Kodalle, T. Bertram, R. Schlatmann and C. A. Kaufmann, "Effectiveness of an RbF Post Deposition Treatment of CIGS Solar Cells in Dependence on the Cu Content of the Absorber Layer," IEEE Journal of Photovoltaics, 9(6), (2019).

Resume : One of the key methods to break the efficiency barrier in CIGS solar cells has been alkaline treatment. It has been shown that this treatment increases the open circuit voltage, carrier lifetime, hole concentration, and reduces the potential fluctuation of the CIGS layer. It has also been shown that alkaline treatment reduces recombination centers at the CIGS/CdS interface and grain boundaries. However, the effects of alkaline treatment on the recombination process are not yet understood in detail. To analyze the recombination mechanism, the open circuit voltage (Voc) at different excitation wavelengths of alkaline treated CIGS solar cells was investigated. Based on the absorption coefficient of CIGS, low energy photons penetrate deeper in the alloy than high energy one. Thus, the variations of wavelength and intensity open the possibility to study the recombination at different depths in the cells. Looking at the evolution of Voc under different light intensities, one can deduce the minority carrier collection at the CIGS absorber as a function of the depth. The results show that the untreated samples Voc is influenced by the excitation wavelength, while the alkaline treated samples does not have such a behavior. This evidence a defect level in the absorber of the untreated sample which obstruct the collection of the carriers in the bulk. Combined with the defect level analysis, we will discuss the recombination processes in the alkaline treated CIGS solar cells.

Resume : Thin film photovoltaics (PV) based on CuIn(1-x)Ga(x)(S,Se)2 (CIGS) absorber technology are marking energy conversion efficiency records and have numerous deployment capabilities, such as in the building integrated PV and mobility sectors. Their deployment is facilitated by the use of flexible, lightweight substrates, such as polyimide and metal foils (steel or aluminum). In this contribution, we will present our research on flexible CIGS solar cells on stainless steel, deposited through a roll-to-roll process. We will focus on the use of a novel dielectric/metal/dielectric front contact, fabricated by sputtering without substrate heating. The thicknesses of the contact layers are optimized by simulations to maximize the light absortption in the CIGS. As metal, copper or silver are used. The ITO-free contact has high transparency (>80%) and low sheet resistance (<10 Ohm/sq.). The front contact is less than 150 nm thick and integrates, in its most attractive configuration, the abundant zinc oxysulfide (ZnOS) buffer, giving rise to a cadmium-free, all sputtered CIGS cell, with efficiency similar to cells employing the CdS buffer.

Resume : Studies about buffer/absorber interfaces in the Cu(In, Ga)(S, Se)2 (CIGSSe)–based cells reveal that S-rich and Ga-poor feature of the absorber-surface causes favorable electronic structures such as “spike” type conduction band connection. Optimization of the backside interface between CIGSSe and Mo electrode is another key to pursue extreme performance. In this study, back-surface of CIGSSe (BS-CIGSSe) and surface of Mo electrode (S-Mo) have been studied by using in-situ direct and inverse photoemission spectroscopy. Both surfaces were prepared by lift-off in UHV from CIGSSe/Mo structure fabricated identically for cells with an efficiency above 20%. The measurements reveal that BS-CIGSSe is Ga-rich and S-poor. Owing to the inverted composition, BS-CIGSSe shows a conduction band minimum (CBM) of 1.1 and valence band maximum (VBM) of -0.5 – -0.4 eV. Considering that previously determined band gap minimum in the interior region of the CIGSSe was 1.07 eV, the high CBM indicates a presence of an adequate back-surface field. S-Mo has electronic structure of alkali-doped Mo(S, Se)2 with VBM of -0.6 – -0.55 eV. UPS measurements show that work function of S-Mo is smaller than that of BS-CIGSSe, and that a difference between them is larger than that in VBM between S-Mo and BS-CIGSSe, which indicates absence of hole-barrier at the interface. It also indicates that VBM in the CIGSSe–side is graded in the direction that hinders hole-conduction, which is to be improved.

Resume : Cu(In,Ga)Se2-based (CIGSe) thin-film solar cells have reached conversion efficiencies above 23% on a laboratory scale. To further optimize the solar cell performance, e.g., to achieve higher open-circuit voltages, the customary Ga/(Ga+In) (GGI) ratio of ~0.30 needs to be increased to create a larger absorber band gap. This is particularly relevant if the “standard” CdS buffer is to be replaced by alternative buffer materials with wider band gaps. For such devices, a comprehensive understanding of the electronic and chemical structure of the high-GGI absorbers and their interfaces is still lacking. In this work, we have investigated in-line deposited and industrially relevant CIGSe absorbers with bulk GGI ratios of ~0.30, ~0.66, and ~0.95, and have studied their interfaces with CdS, Inx(O,S)y, and Zn(O,S) buffer layers prepared by chemical bath deposition (CBD). We used a number of different electron and soft x-ray spectroscopies, including laboratory-based x-ray and UV photoelectron spectroscopy (XPS and UPS), inverse photoemission (IPES), and x-ray-excited Auger electron spectroscopy (XAES). In combination with synchrotron-based soft x-ray emission spectroscopy (XES), full pictures of the electronic and chemical surface and interface structures can be derived, including the band alignment and intermixing behavior as a function of GGI. These structures will be discussed in view of their impact on the performance of corresponding CIGSe solar cells.

Resume : Alternative buffer layers in CIGSe are deposited mainly using chemical bath deposition because of its benefits like simplicity, good film quality and surface/step coverage. All the layers in CIGSe cell stack such as back contact, absorber and window layers are deposited by vacuum–deposition methods such as coevaporation, sputtering, and sometimes thermal evaporation, except for the buffer layer. Therefore, in the present work we demonstrate the feasibility to deposit In2S3 by RF magnetron sputtering while retaining superior cell performance. In this work, we studied the extent of sputter induced damage on CIGSe absorber as well as the sputtering induced intermixing phenomenon taking place at the In2S3/Cu(In,Ga)Se2 interface at the subnanometer level using atom probe tomography. Absorber surface damage and nonuniform buffer layer thickness are the primary limitations when using sputtering, and hence need to be eliminated for reaching reasonable cell efficiencies. We have also shown that a post deposition annealing not only significantly improves the crystallinity of In2S3, but also recovers the surface damage caused by sputter-induced intermixing resulting in an improved p-n. The present work consolidates the feasibility to design efficient and industry–compatible CIGSe cells using sputter–deposited Cd-free buffer layers. Moreover, this work clearly demonstrates that this novel and fully vacuum–deposited CIGSe cell meets the standard requirements, in terms of chemistry, structure, and electrical performance of a working cell for the PV industry.

Resume : Alkali- halide post-deposition treatments (PDTs) of Cu(In,Ga)Se2 (CIGS) absorbers have repeatedly resulted in device efficiency improvements, observed mainly due to an open-circuit voltage (Voc) enhancement. Replacement of the CdS buffer layer with a higher band gap alternative can increase the short-circuit current density (Jsc) and also eliminate the use of Cd. In many alternative- buffer attempts, however, the Jsc gain was overpowered by a Voc loss, resulting in some degree of performance loss. In this work we present device analyses of CIGS solar cells in order to better understand the impact of RbF-PDT in combination with chemical-bath-deposited CdS or Zn(O,S) buffer layers based on experimental devices produced in the same in-line CIGS run. Low-temperature current-voltage curves indicate a difference in Rb impact on the CIGS/CdS and CIGS/Zn(O,S) pn junctions. For example, the diode- current barrier that creates a JV curve rollover often observed in RbF-PDT-treated CIGS/CdS samples is absent in CIGS/Zn(O,S) junction. Although the RbF-PDT had a positive impact on both junction partner combinations, the CIGS/Zn(O,S) devices’ Voc and FF benefited stronger from the RbF treatment. As a result, in our samples, the Jsc and FF gain balanced the Voc loss, thus reducing the efficiency advantage of CdS over the Zn(O,S) buffer.

Resume : Transparent photovoltaics (PV) could be a solution for integrating large amount of PV in buildings. A possible approach could be a tandem cell using the combination of IR-absorbing organic bottom cells with UV-absorbing top cells. The UV top cell would then serve as a solar cell, converting UV light into electrical energy, as well as a UV filter to protect the bottom organic PV cell from degradation from the UV light. We have developed three potential UV absorber materials: ZnS, ZnSe and Zn(S,Se). All layers were fabricated by annealing an evaporated or sputtered Zn layer in an atmosphere of H2S for sulfurization, H2Se for selenization or a mix of the two for the mixed sulfo-selenization. Typical annealing temperatures and times for this crystallization process are 450°C and 10 minutes. We will report on the optoelectronic properties of the layers and we will show that the combined Zn(S,Se) shows the ideal band gap of about 3 eV combining both decent colour and good UV filtering properties. Analysis of the conductivity of the layers will also be presented, together with some initial solar cell structures with transparent front and back contacts. Acknowledgments: This work has received funding from the European Union H2020 Framework Programme under Grant Agreement no. 826002 (Tech4Win).

Resume : Thin-film solar cells of Cu(In,Ga)Se2 (CIGS) with Zn(O,S) buffers have attracted much attention in recent years due to their non-toxic constituents and high conversion efficiency above 23% [1]. Among other post-growth treatments, the achieved performance can be significantly influenced by thermal annealing (TA). In this work, CIGS solar cells with solution-processed Zn(O,S) buffers were post-annealed at 200 ˚C. The chemical and structural properties of the cells before and after TA were studied by energy-dispersive X-ray spectroscopy (EDXS) and transmission electron microscopy (TEM). EDXS analyses showed that after TA the Cu content in the Zn(O,S) layer increases significantly. Bandgap measurements of the annealed cell by angle-resolved electroreflectance spectroscopy showed an additional resonance at 2.3 eV, indicating the existence of a secondary phase, besides the bandgap energy of 2.9 eV corresponding to Zn(O,S). Nanobeam electron diffraction (NBD) patterns were acquired in the nanocrystalline Zn(O,S) buffer to identify secondary phases formed during TA. Crystal-structure data of Cu-containing phases with bandgap energies close to 2.3 eV, e.g., CuS (2.36 eV), Cu2Se (2.23 eV), and Cu3Se2 (2.37 eV), were used for comparison with the experimental NBD patterns, among which CuS fits much better than others. The results suggest that TA at 200 ˚C leads to Cu diffusion from CIGS into Zn(O,S) to form a Cu-containing phase. [1] M. Nakamura et al., IEEE J. Photovolt. 9 (2019) 1863.

Resume : Deposition of CIGS absorbers in research laboratories with growth times comparable to conditions in production is an important topic in order to facilitate transfer of processes from research into industry. In this contribution we report on the growth of CIGS by co-evaporation in an inline-deposition system with high deposition rates, corresponding to growth times of approx. 7 minutes for the absorber layer. In order to investigate the development of the growing layer so called interrupt experiments were performed: substrates are moved at process speed till specific positions in the inline system and are then removed at maximum speed. The use of standard soda lime glass with Mo as well as zirkonium oxide with Mo, Na-doped Mo or K-doped Mo as substrates allows to differentiate effects of alkali elements. Specifically we discuss the observation of Cu-poor, Na-rich secondary phases within the CIGS layer, adjacent to the molybdenum back contact.

Resume : Thin film solar cells made of Cu(In,Ga)Se2 (CIGSe) suffer from high costs, since the rare elements indium and gallium are used for a multitude of technical applications. The concept of microconcentrator solar cells allows substantial material saving and results in an increase of solar cell efficiency. Therefore a reliable bottom up process for the preparation of the absorber islands is required. In this respect, we will introduce novel micro- and nanostructuring processes for PVD techniques such as MBE or electron beam/thermal evaporation. We have developed methods for bottom-up growth, which change the roughness and morphology of the underlying substrate. These surface and roughness modifications induce controlled self-organization and self-assembly processes. On the one hand, this provides new large-scale production capabilities, which are also being investigated in an ongoing industrial project. On the other hand, it provides new insights into the fundamental physical properties concerning chalcogenides and the underlying mechanisms of micro- and nanostructure formation. By taking into account the thermodynamically stable states of the individual compounds and their interactions with the interfaces, the morphology and local structure in micro- and nanoscale can be controlled to increase the performance of the desired micro-concentrator device.

Resume : The importance of Na for the photovoltaic performance of Cu(In,Ga)Se2 CIGS has been realized for a long time and several different Na incorporation methods have been developed. However, we found Na diffusion from soda-lime glass (SLG) substrate cannot provide an optimal amount of Na as desired and over dosage could occur. With an Al2O3 layer on SLG substrate, the amount of Na from glass can be controlled. Here, we presented the effects of Al2O3 barrier layer in precise control of the amount of Na in CIGS and improved the distribution of Na in CIGS solar cells.

Resume : In order to obtain long-term stable and low-cost photovoltaic modules, it is important to know possible degradation mechanisms. In order to identify long term behavior, unpackaged Cu(In,Ga)Se2 solar cells were simultaneously exposed to damp heat and illumination (settings: 85oC/85% relative humidity and 1000 W/m2 BAA illumination) [1]. In-situ monitoring of their electrical parameters demonstrated a rapid decrease of the efficiency, mainly driven by changes in the series and shunt resistances. Moreover, non-degraded and degraded solar cells were studied by in-depth XRD and SIMS to investigate the material changes leading to efficiency loss. SIMS revealed the migration of sodium and potassium, likely leading to a severe decrease in the shunt resistance (from 500±100Ω to 20-30Ω within 200 hours) and voltage. It also displayed the ingression of hydroxide, especially in the ZnO:Al film. Extensive XRD measurements showed that molybdenum oxide was formed, which likely affected the Ohmic contact between Mo and CIGS. Moreover, an in-plane stress increase in the ZnO:Al film from -183±29 to -450±40 MPa was observed. This stress increase is most likely due to the incorporation of species like hydroxides and carbonates in the grain boundaries of the ZnO:Al film. These phenomena could lead to the observed increased series resistance in the solar cells [2]. [1] M. Theelen et al., JoVE 140 (2018) [2] M. Theelen et al., Sol. Mat. Sol. Cells 157 (2016) 943–952

Resume : Reactive thermal annealing of CuInGa precursors in elemental Se and S is an industrially attractive route to produce CIGS thin film solar modules, due to relatively low equipment costs and high metal utilization. This route typically leads to phase separation of CuInSe2 and CuGaSe2, resulting in a steep bandgap increase at the back and limiting the minimum bandgap to 1.00 eV. Tuning this bandgap profile is still today a challenge for CIGS manufacturers. To obtain the desired bandgap grading, processing on our semi-industrial 30x30 cm2 baseline was divided into four stages: i) a CuInGa precursors premixing stage, ii) a high temperature selenisation stage, iii) a soaking stage with reduced selenium pressure and iv) a final sulfurization stage. The latter only led to mild sulfurization of the surface. By combined controlling of the thermal budget, Cu/(In+Ga) ratio, Se pressure at high temperature and Na content, the gallium profile was tuned. This approach was successful for precursors of different structures, sodium sources and concentrations. By combining this control over Ga depth profile with sulfurization of the surface area, the minimum bandgap could be varied between 1.00 and 1.20 eV with fixed gallium content and thickness of the absorber. Voc values exceeding 700 mV were obtained. A maximum efficiency of 16.4% as well as a median efficiency over 10x10 cm2 of 14.6% were reached (without anti-reflective coating).

Resume : (Ag,Cu)(In,Ga)Se2 (ACIGS) solar cells exhibit high conversion efficiencies of up to 22.9% when applying a KF post-deposition treatment (PDT) on the ACIGS surface. The present work investigates the impact of these treatments on the microscale materials properties and thus on the device performance. For this study, three solar cells with various KF PDTs (none, optimal dose, extra dose) were analyzed. Various scanning electron microscopy methods were applied on the three samples in a correlative way, combining electron backscatter diffraction (EBSD), energy-dispersive X-ray spectrometry (EDX), and cathodoluminescence (CL) in a first approach, and then also EDX and electron-beam-induced current (EBIC) measurements as a second approach. The correlative analyses were conducted on identical positions on cross-sectional specimens, which were prepared from completed ACIGS solar cells. For the extra-KF-PDT sample, the solar-cell performance was bad (conversion efficiency of about 5%). In contrast, the conversion efficiencies for the none and optimal-KF-PDT samples were similar (conversion efficiency of about 16-17%), with higher open-circuit voltage (Voc) and fill factor (FF) for the optimal-KF-PDT cell. For all ACIGS layers, low net-doping densities of about 10^14 cm^-3 (as determined by capacitance-voltage analyses) were found. EBSD and EDX maps show that the microstructure as well as the elemental distribution in the volume of the ACIGS thin films are the same for all three solar cells; this result was essential for the present comparison since the three ACIGS thin films were fabricated in different deposition runs. By absolute photoluminescence (PL) measurements, it was found that the PL intensity was much smaller for the extra-KF-PDT sample than for the other two samples, which suggests that also the electron lifetime should be smaller. This lower electron lifetime in the extra-KF-PDT sample can be attributed to secondary, K-enriched/Cu-depleted (and slightly Ag-depleted) phases detected at the ACIGS/CdS interface by EDX (as compared with the ACIGS stoichiometry in the bulk). The CL intensities of these phases were substantially lower than in the neighboring ACIGS layer, suggesting enhanced non-radiative recombination. Evaluation of the EBIC results obtained on the ACIGS solar cells shows that the drift length was largest for the optimal-KF-PDT sample (2 µm), compared with those of the none/extra-KF-PDT samples (below 0.5 µm). These analyses also show that fluctuations in the width of the space-charge region were more pronounced for the none-KF-PDT sample as compared with the optimal-KF-PDT solar cell. EDX maps reveal that these fluctuations can be linked to inhomogeneously distributed regions exhibiting strong Cu depletion at the ACIGS surface in the none-KF-PDT sample. These strong lateral fluctuations of the electrostatic potential and/or the band-gap energy can explain in part the lower open-circuit voltage in this solar cell. Overall, the results in the present work show that for ACIGS solar cells, a KF-PDT is essential to passivate the ACIGS/buffer interface and to reduce fluctuations, which can limit the Voc (and the FF) of the none-KF-PDT solar cell. KF PDTs at a too high doses lead to secondary phases at the ACIGS/buffer interface that deteriorate the solar-cell performance.

Resume : The material system of Cu(In,Ga)(S,Se)2 (CIGS) is well-established among high efficiency thin film solar cells. The properties of CIGS throughout the absorber layer depend on the depth distribution of composition profiles, especially elemental profiles of Ga and S. In this work a nontoxic and adjustable etching process was used to prepare wedge-shaped samples, both from CIGS on substrate and from exfoliated CIGS, for further direct analysis. The CIGS wedges were characterized by scanning electron microscopy (SEM) to evaluate the morphology, energy dispersive X-ray spectroscopy (EDS) at 10 kV to determine the element composition and distribution, and photoluminescence (PL) to investigate the optoelectronic properties. The directly measured band gap energy is compared to the estimated values from the composition gradients that were measured by glow discharge optical emission spectroscopy (GDOES) and by variation of acceleration voltage in area EDS.

Resume : Solution-processed chalcopyrite Cu(In,Ga)(S,Se)2 (CIGSSe) solar modules are promising alternatives to conventional crystalline silicon-based photovoltaic devices owing to their potential for lower production costs and compatibility with large-area flexible substrates. However, such modules typically exhibit a higher cell-to-module loss than expected, which, in most cases, is attributed to the emergence of critical shunt sites. Herein, we investigated the structural shunt defects induced by the absorber delamination in the alcohol-based solution-processed CIGSSe solar modules. From the lock-in thermography characterisation, we found that most of the delamination-induced shunt defects emerged locally near the patterned regions. Furthermore, the secondary ion mass spectroscopy measurements showed that the alkali elements in the CIGSSe modules were not uniformly distributed but rather, were instead concentrated near the patterned regions. The results revealed that the uniformity of the alkali element distribution is the key factor in the absorber delamination phenomena. Finally, we confirmed that delamination-induced shunt defects can be alleviated by controlling the diffusion of alkali elements. The energy conversion efficiency of the alcohol-based solution-processed CIGSSe solar modules was enhanced from 4.24 to 7.59% by using a thin-film diffusion barrier layer.

Resume : The Cu(In1-xGax)Se2 (CIGS) semiconductors have been attracting a great attention as the absorber material for the thin film solar cells. The one-step sputter process utilizing a CIGS single target can be faster and more efficient, as the post-selenization process may be omitted, but the films fabricated by this method still suffer from the poor crystallinity and the Cu-rich composition. In this work, quality improvement of the CIGS films using the Cu-Se interlayer (with melting temperature of ~580oC) between the Mo substrates and the films has been studied in detail. It was observed that the CIGS thin films sputtered from Cu-poor targets, incorporating the Cu-Se interlayer, showed dramatic increase in grain size from ~100nm to ~2um, maintaining Cu-poor film compositions without presence of any Cu-Se related phase. As a result, the CIGS solar cell fabricated using the optimized CIGS thin film (with the Cu-Se interlayer) showed 10.8% cell efficiency, demonstrating that the one-step sputtering has a good potential to be an effective fabrication method for high quality CIGS thin films. CIGS thin films were also grown from quaternary targets with various Cu concentration, employing Cu-Se interlayer films formed at various conditions. Results of detailed characterization of the films and their device performances will be discussed during the presentation, as well as the responsible mechanism(s).

Resume : To prepare for the post-Si solar cell era, alternative thin film technologies should be developed toward low cost, high throughput, and easy scale-up with high efficiency. Cu(In,Ga)Se2 (CIGS) solar cells are particularly promising because they present device efficiency up to 23.4% and sustainability, approaching those of Si solar cells. However, the current record efficiency is obtained using a vacuum process which requiring high production cost. To overcome this bottleneck, solution process has been suggested, but those processes exploit toxic and reactive solvents such as hydrazine and dimethyl sulfoxides. In this report, we have developed novel solution process based on multi-layered CuInGa precursor film and potassium (K) bulk incorporation in the middle of the layer. This process is based on environmentally benign alcohol solvent. By this approach, we could achieve efficiency of 15% which is the highest among benign solvent processed CIGS solar cells. We found out solar cell performance is enhanced by crystalline grain growth which was quantified by synchrotron based techniques for the first time. This morphology improvements were enable by K-Se affinity, promoting Se penetration deep into the film during chalcogenization. The morphology-device property relation is elucidated and we have proven that increasing grain growth is a key factor to obtain a highly efficient CIGS solar cell.

Resume : Partial shading of monolithically interconnected Cu(In,Ga)Se2 (CIGS) photovoltaic (PV) modules can lead to the formation of reverse bias induced defects. These localized defects permanently reduce the output of the PV module. In a previous study [1], we showed that the propagation of these defects strongly depends on the electric field over the CIGS absorber. For this study ten substrates (162 cells of 5×10 mm2) with varying absorber thickness and composition (CGI and GGI) were used. Exposure to reverse bias was performed in an automated setup that executed an IV sweep up to -6 V either in darkness or under illumination. From the IV, data the point at which the defects start to form was determined. These values were labeled formation current and voltage. No correlation between the formation current and absorber thickness or composition was observed. However, a strong relation between the formation voltage and the absorber thickness was found. Measurements in the dark showed that for samples with a thin absorber (216 cells, 4 substrates, ±500 nm) and thick absorbers (194 cells, 4 substrates, ±1600 nm), the formation voltage was -3.1 ±0.5 V and -4.3 ±0.9 V respectively. Our findings shows that the formation voltage of these defects is depending on the absorber thickness. Therefore, the formation of reverse bias induced defects depends on the electric field over the CIGS absorber. [1] Bakker et al., Sol. Energy Mater. Sol. Cells 205 (2020) 110249. doi:10.1016/j.solmat.2019.110249

Resume : Recently Cu(In,Ga)Se2 (CIGSe)-based solar cells have reached outstanding level of performanc. This breakthrough is associated with the development of heavy alkali fluoride Post Deposition Treatments (PDT) generally performed under Se vapour right after the absorber is completed. Nevertheless, the last record efficiency (23.35%) achieved by Solar Frontier involves sulfo-selenide CIG(Se,S) absorber. We recently reported that surface sulfurization of CIGSe is significantly enhanced when elemental S is provided together with KF, suggesting that the nature of the chalcogen (Se or S) involved during the KF-PDT plays a critical role in the resulting CIGSe surface modifications, thus related cell operation. If CIGSe modifications induced by KF-PDT in Se atmosphere (KSe) have been widely reported, the impact of elemental S evaporation, either after KF-PDT (ie. KSe+S) or during KF-PDT (ie. KS) is addressed for the very first time. Detailed pictures of those modifications are deduced mainly from X-ray photoelectron spectroscopy (XPS) analyses. For both KSe and KSe+S, an In-rich overlayer is detected; either made of In-Se-O (KSe) or K-In-Se/S-O (KSe+S). Supplying elemental sulfur to KSe absorber after air exposure therefore yields overlayer sulfurization and enhances K outdiffusion, probably from grain boundaries. In contrast for KS no such overlayer is detected and absorber surface characteristics are similar to the reference sample.

Resume : For industrial growth of Cu(In,Ga)Se2 (CIGS) thin-film solar cells high deposition rates must be achieved to become economically competitive. High deposition rates have already been achieved on glass substrates in inline production using substrate temperatures of TSub ~ 600°C to allow fast crystallization of CIGS. These high temperature processes cannot be directly transferred to flexible polyimide substrates, as polyimide starts to thermally decompose above ~ 400°C. One way to circumvent this constriction is to grow the CIGS thin film at TSub < 400°C with low growth rate, but this is not an option for industrial applications. Another route is to grow the CIGS layer on polyimide at an elevated TSub with a high growth rate in order to reduce the time of substrate decomposition. When increasing the CIGS growth rate from 38 nm/min (~ 60 min process) to 280 nm/min (7 min process) the cell efficiency, fill factor and short circuit current density decreased with increasing deposition rate, whereas the open circuit voltage remained constant on both, glass and polyimide substrates. Nevertheless, a cell efficiency of 15.6 % and 13.4 % (without antireflective coating) could be achieved on glass and polyimide substrates at Tnominal ~520°C and ~450°C, respectively, with CIGS layers grown during 7 min in a single stage batch process. We expect even higher cell efficiencies to be possible at high growth rates for an optimized CIGS composition and multistage growth process.

Resume : Three-stages process has been decisive for the improvement of CIGS-based solar cells. Indeed, the presence of CuxSe liquid exacerbates two beneficial mechanisms. On the one hand, a recrystallisation occurs leading to bigger grain sizes and on the other hand diffusivity difference between group III elements is increased leading to a double Ga gradient. This graded structure has positive effects on the shunt current density and the open circuit voltage. We have studied this graded structure by XRD. Due to the linear variation of lattice parameters with Ga concentration, the double gradient causes an enlargement and a structuration of the diffraction peak. Moreover, diffracted intensity depends on the composition, diffraction angle but mostly on the absorption. As a result, there is a strong depth dependence emphasizing different zones of Ga composition and leading to the structuration observed. We have developed a model which allows to obtain qualitative information about this gradient. In this model, the CIGS layer is discretized into several sub-layers with an assigned Ga composition. Thus we are able to reproduce the gradient and reconstruct the diffracted intensity peak. A series of solar cells ranging from x=0 to x=1 were produced using the three stages process and were characterized by high resolution XRD, I(V), SEM, and EQE. Our results show that by increasing the Ga content, the double gradient becomes less pronounced and the solar cells efficiencies decrease.

Resume : Doping of Cu(In,Ga)Se2 (CIGS) thin-film solar cells with alkali metals has a positive effect on p-type conductivity, charge carrier lifetime and defect density in the CIGS layer. Thus, open circuit voltage (Voc) and fill factor (FF) of CIGS solar cells can be strongly increased by alkali metals. To elucidate the positive effect of Na in contrast to K, alkali-free CIGS layers with varying Cu content were doped with different amounts of Na (or K) by an alkali-fluoride post deposition treatment (PDT). The resulting cells were then characterized by means of current-voltage, quantum efficiency and capacitance-voltage measurements. The cell parameters increased with increasing amount of Na or K in the CIGS layer up to a maximum efficiency of 15.8 % and 16.4 % for a Na and K doped cell, respectively, compared to 9.1 % for the undoped reference. The maximum increase in Voc and FF obtained by NaF PDT was comparable to that of KF PDT. However, a higher amount of K is needed to achieve the same cell efficiency and charge carrier density as for Na doped cells. Hence, one may conclude, that Na has a higher doping efficiency compared to K. At a PDT temperature of ~350°C we found a higher amount of K in the CIGS layer after a KF PDT compared to Na in a CIGS layer after a Na PDT. This is a hint that diffusion of K in CIGS is faster than of Na. Despite the strong chemical correlation between Na and K, both alkali metals show a distinctive behavior and physical effects in CIGS solar cells.

Resume : High-quality single-crystal CIGS layers have the potential for increasing the efficiency of CIGS solar cells since there are no grain boundaries, and the sharp interface reduces the number of carrier recombination centers. Single crystals are also an ideal structure for investigating fundamental physics of CIGS solar cells. In this study, epitaxial CIGS layers were grown on III-V compound semiconductors by molecular beam epitaxy (MBE) method. The presence of a single-crystal CIGS layer without dislocations, and sharp CdS/CIGS, CIGS/GaAs heterointerfaces were confirmed by scanning transmission electron microscopy. Sodium and potassium ion incorporations were achieved by doping during growth and post-deposition treatments, and CIGS solar cells with sodium ions have high acceptor concentrations. Ga grading structures were fabricated by two-layer deposition with different Ga contents, and the Ga grading significantly increased the fill factor and open-circuit voltage. The best efficiency of 20.9% was achieved using Cu poor CIGS layers. Heterostructures between n-type III-V compounds and CIGS layers show photovoltaic effects, and the new-type p-n junctions promise new monolithic bottom solar cells for multijunction solar cells.

Resume : In this study, the CuInSe2(CISe) absorbers are peeled off from their substrate, and the rear-surfaces are systematically investigated by means of scanning tunneling microscopy and spectroscopy (STM/STS), Kelvin probe force microscopy (KPFM) and X-ray photoelectron spectroscopy methods. The results are compared to those obtained at the top surface of the films after potassium cyanide etching (KCN) and previous results obtained on the rear surface of CuGaSe2. We focus on the backside of CISe absorbers since this provides a flat and clean surface, free of contaminations from the environment. Our STS measurements show that the top surface exhibits a highly defective surface with a reduction of density of defect states at the grain boundaries (GBs) whereas the rear surface shows an improved surface passivation with almost zero granular in-homogeneities. Moreover, compared to the highly defective KCN etched samples with a metallic surface behavior, the rear surface displays a well-defined conduction and valence band onsets. This change in electronic properties is presumably ascribed to difference in copper concentration between the two surfaces. The KCN etched sample presents a strong Cu-depleted (Cu/In~0.65) surface while the rear surface shows a stoichiometry composition (Cu/In~1). Complementary KPFM measurements show a smooth backside surface topography with a slight upward band bending (45 meV) at the GBs. A detailed analysis of these observations will be presented.

Resume : To meet demanding economic and sustainable challenges, CuInSe2 (CIS) based solar cell technology will have to reduce the thickness to sub-µm absorbers. From a device performance point of view, with no further optimization, the reduction of the absorber thickness leads to an incomplete optical absorption, which translates as a strong deterioration of short-circuit current density values. The Shockley-Queisser (SQ) model is widely used to estimate the upper limit of the conversion efficiency of a single-junction solar cell, but it assumes a complete optical absorption, which should prevent us from applying it to ultrathin devices. We extend the SQ model to make it valid in the case of sub-µm absorbers. From this, we can establish benchmark values of efficiency limits vs. the reduced absorber thickness, and compare to experimental data collected from the literature. The model shows that down to 500 nm, the value of the conversion limit can be preserved, fine-tuning the absorber bandgap to mitigate some of the losses. It also describes that below 500 nm thick absorbers, the optical losses become significant and strongly affect the efficiency theoretical limit, which may be far below the 32 % usually displayed with 2 µm thick CIS layers. From this work, we wish to provide a fair way to compare devices with dissimilar absorber thicknesses.

Resume : To reduce the usage of the rare element In in CIGS solar cells, we explore the option to reduce the CIGS layer thickness to about 600 nm. When the absorber layer is less than 1 µm, current losses are expected due to insufficient absorption of the incoming light. These losses may be recovered by applying a back reflector, allowing the unabsorbed light a second passage through the absorber layer. As back reflector a metal with high reflectivity for the long wavelengths is preferably used. In this work we applied 10 and 20 nm thin Ag layers as back reflector with 0, 1, 2 to 3 nm AlOx on top to protect the Ag. The CIGS layers were 500 or 650 thick and grown by a 3-stage co-evaporation process. We find that the best Ag/AlOx combination is 20nm/3nm and leads to an increase in current from 22 to 27 mA/cm2 for 500 nm thick layers. From quantum efficiency measurements we find that this increase is due to increased response at longer wavelengths. The efficiencies are about 10%. With EDX we measured the amount of Ag. When the CIGS layer was removed, we find that the amount of Ag was below the detection limit when the AlOx was 1 and 2 nm, while for 3 nm there was still some measurable amount of Ag left. This implies that the Ag partly diffuses into the CIGS layer. We therefore etched the CIGS layer with ammonia sulfide prior to CdS deposition, removing any potential secondary phases. This etching step increased the efficiency further to 12% due to a Voc increase of about 30 mV.

Resume : Room temperature photoluminescence (PL) of thin single and 3-stage co-evaporated Cu(In,Ga)Se2 samples is modelled with the lateral band gap fluctuation model proposed by Mattheis et. al [1]. In this model the PL peak width and position depends on the length scale of the carriers before they recombine (Lµ) compared to the characteristic length scale of the band gap fluctuations (Lg). When Lµ is of similar order or larger than Lg a peak shift can be observed. In [2] we found that the thin absorber layers have a large blue shift at room temperature with increasing light intensity. We show that this anomalous blue shift can be attributed to the decreased minority carrier lifetime which changes the diffusion length and thus Lµ. Furthermore, we found that single stage samples have an anomalous second peak below the band gap [3]. We adapted the absorptivity term in the model by adding a Gaussian peak, which allowed us to model the single-stage spectra with the band gap fluctuation model and extract the band gap fluctuations for the single stage samples. The extracted values for band gap fluctuation of the single stage absorber show a decrease from 50 to 28 meV with increasing copper content, while for the 3-stage this was about 40-50 meV. This value is independent on the copper content and likely due to the Ga gradient. [1] Mattheis et. al Journal of Applied Physics 101, 113519 (2007) [2] de Wild et. al IEEE PVSC 46, 2019 [3] de Wild et. al ACS Appl. Energy Mater. 2019, 2, 8, 6102-6111

Resume : Although the influence of sodium on CIGSe solar cells has been investigated since nearly 30 years, there is still an ongoing debate on the mechanisms leading to efficiency improvement. The relatively slow progress in understanding the role of Na is mainly due to quite sophisticated deposition methods and a complex structure of top-efficiency CIGSe devices. Therefore, in this work we propose a systematic approach in which we investigate carrier transport mechanisms in CIGSe devices fabricated in single-stage process to ensure that the absorber material was uniform in composition. The amount of sodium was controlled by changing the annealing temperature during NaF post-deposition treatment in the 100-400 C range, and quantified by SIMS measurements. In accordance with the literature, we observed a gradual gain in Voc and FF as Na concentration increased. In order to elucidate the mechanisms responsible for this improvement white- and red light current-voltage characteristics were measured as a function of temperature and irradiance. The main results are as follows: i) Suns-Voc-T analysis revealed that the transport mechanism changes from interface to space charge recombination as Na content increases from a low level of 20-80 ppm to a high level of 200-480 ppm, respectively, ii) correlation with capacitance profiles shows that Na influences not only the free hole concentration but also reduces the concentration of recombination centers, and (iii) secondary barriers (cross-over and red-kink) decrease gradually at the same pace with increasing Na content which suggests that Na-PDT influences also the window/buffer/absorber interface region.

Resume : The record conversion efficiency of more than 23% for Cu(In,Ga)Se2 (CIGS) based thin-film solar cells has been achieved by incorporating heavy alkali metals like Cs into the absorber through an alkali fluoride post-deposition treatment (PDT). As the effect of the incorporated heavy alkali metals is still under discussion, we investigated the local composition and microstructure of high efficiency CIGS solar cells in a combinatory approach. To that end, thin cross sectional lamellas were cut out of the whole devices and analyzed via synchrotron based spatially resolved X-ray fluorescence analysis and different electron microscopy techniques. An accumulation of Cs is clearly detected at the p-n junction along with variations in the local CIGS composition, showing the formation of a beneficial secondary phase with a laterally inhomogeneous distribution. Additionally, Cs accumulations were detected at grain boundaries with a random misorientation of the adjacent grains where a reduced Cu concentration and increased In and Se concentrations are detected. No accumulation was found at Σ3 twin boundaries as well as the grain interior. These experimental findings are in excellent agreement with complementary ab-initio calculations, demonstrating that the grain boundaries are passivated by the presence of Cs. Further, it is unlikely that Cs with its large ionic radius is incorporated into the CIGS bulk where it would cause detrimental defects.

Resume : Proton/electron radiation has been used to test the stability of several materials and diode-like devices for space applications. Cu(In,Ga)Se2 (CIGS) thin film solar cells have shown its high radiation resistance due to the self-healing effect via environmental temperature or light illumination. The performance was further improved by Cs-treatment, which has drawn much attention in recent years. The influence from energetic charged particles of the thin-film solar cells is usually evaluated by the remaining factors of photovoltaic parameters, such as open-circuit voltage (Voc), current density (Jsc), fill factor (FF) and efficiency and power. However, the mechanism of irradiation-induced defects enhancing recombination of proton and electron irradiation on Cs-treated CIGS solar cells is still not clear, the damage from irradiation at different region also lacks an easy characterization technique. Herein, in this work, we use impedance spectroscopy with bias flow, photoluminescence and capacitance spectroscopy to diagnose responses from different interfaces on Cs-treated devices, and compare the status before and after the irradiation to reveal the defect generation of proton/electron irradiation.

Resume : The stability of the Cu(In,Ga)Se2 (CIGS) solar cells performances is dependent on the atmosphere surrounding the cell, and more specifically, on the humidity level present in the atmosphere. During ageing, humidity is a factor influencing the integrity of the cell. Indeed, degradation mechanisms can be observed at different localizations inside the device architecture. Especially, the chemistry at the rear and front faces of the absorber is one of the critical points to control in order to prevent efficiencies losses, requiring a perfect knowledge of the corresponding interfaces evolution in time. In this study, we focus on the reactivity of CIGS absorbers when exposed to different humidity levels. The chemical evolution of CIGS surfaces is accurately described using photoelectron spectroscopy. A specific attention is paid to the chemical environments’ determination and to the quantitative aspects, on the base of the specific CIGS ratios: CGI, GGI and 2*VI/(I+3*III) ([Ga]/([Ga]+[In]) et [Cu]/([Ga]+[In]) 2*[Se]/([Cu]+3*([Ga]+[In]), respectively). The ageing performed at 0% humidity level (under ultra-high vacuum) demonstrate the intrinsic stability of CIGS surface, with no reorganization of the surface network nor chemical evolutions even after 4 months of ageing. One of our major results concerns the excellent stability of CIGS surfaces when they are exposed for several weeks to a humidity level inferior to 22%. On the other hand, at 59% humidity level, an important degradation becomes detectable after only one week ageing. Different kinetics of oxidation are observed for depending on the CIGS constitutive elements and a comparative study with immersion in water has also been performed to understand the mechanisms of oxidation.

Resume : In this contribution, we examine the near-surface region of low temperature evaporated Cu(In,Ga)Se2 (CIGSe) that underwent either NaF or combined NaF/RbF post deposition treatments (PDTs) with different RbF evaporation rates. The study is based on a depth-resolved investigation of the material by using excitation energy dependent photoelectron spectroscopy. We find that the in general Cu poor surface region is further Cu and Ga depleted due to NaF/RbF PDT. Through detailed spectral analysis we are able to identify NaF/RbF PDT induced Rb, In, and Se containing species and find evidence for the formation of a Rb-In-Se (RIS) type phase. However, in contrast to our studies of the K-In-Se type species formed upon NaF/KF PDT [1], we do not find indications for a pronounced RIS/CIGSe bi-layer formation. Furthermore, in addition to Rb in a RIS environment, the presence of a second Rb species (which is also observed at the backside of the sample), is revealed – which we presumably attribute to Rb-O. Further, we see an RbF PDT induced shift of the valence band maximum away from the Fermi level at the very surface of the absorber. In our contribution, we will relate our findings to the performance of respective solar cells. [1] Handick et al., ACS Appl. Mater. Interfaces 2017, 9, 4, 3581-3589

Resume : When the thickness of the CIGS layer decreases, interface recombination may become a limiting factor. One way to reduce the interface recombination is by implementing a passivation layer. In this work, we look at passivating the front interface. However due to surface roughness of CIGS, the creation of point openings by using conventional patterning methods from Si-based PV is difficult. Also, the size and distance of the openings in the passivation layer requires submicron scale due to the shorter diffusion length. We show that we create submicron patterns for the openings by using alkali solutions. The AlOx passivation layer is deposited on this NaCl patterned CIGS. Removal of the NaCl salts by using ultra sonic bath leave openings in AlOx layer. The size of openings is determined by SEM to be between 7um and 500nm. This part shows it is possible to implement a passivation layer with openings on the front surface of CIGS. However, before applying the passivation layer, surface has to be cleaned by reducing the dangling bonds and remove undesired oxides and/or impurities. In this study, ammonium sulfide is used and its effect was investigated for samples with different Cu/(In Ga) (CGI) ratios. We observe that the impact of the surface cleaning depends on the CGI ratio. Photoluminescence (PL) yield and the lifetime decay showed improvements for all CGI ratios, reaching 10-fold for the Cu-poor sample (CGI:0.72). The trend in PL was mirrored in the current-voltage results.

Resume : In this work, a novel rear contact architecture for ultrathin Cu(In,Ga)Se2 (CIGS) solar cells was developed in order to tackle the following drawbacks when compared with standard thick devices: i) rear interface recombination; and ii) reduced photogeneration due to the thin absorber thickness. The innovative rear contact architecture allows for: i) rear interface passivation by using an Al2O3 layer with line contacts; ii) improved rear optical reflection by introducing a highly reflective metal interlayer between the bottom Mo and the overlaying Al2O3 passivation layer; and iii) ensured ohmic contact between CIGS and Mo by using a lift-off process where a second Mo layer is deposited in the line contacts. In this manner, the effective light-path in the absorber is increased with the increased optical reflection induced by the reflective metal interlayer, while simultaneously, an effective passivation strategy is enhanced by the Al2O3. The encapsulated design prevents elemental diffusion from and to the metal interlayer during the CIGS growth. TEM analysis confirmed that the expected structure was obtained with no evidence of interdiffusion, opening the door to use more reflective metals that have previously been discarded due to their ease of diffusing. A sub-micrometer (~600 nm) CIGS solar cell with this new rear contact architecture, using Ta as interlayer, achieved an average efficiency value of 10.3 % compared to the unpassivated solar cell with an average value of 6.1 %.

Resume : Recent efficiency improvements in Cu(In,Ga)Se2 (CIGS) thin-film solar cells were achieved by the application of an alkali-fluoride post-deposition treatment (AlkF-PDT), applied after the deposition of the CIGS absorber layer by a co-evaporation process. In this work we explore the electrical characteristics in the bulk for different AlkF-PDT CIGS samples using conductive atomic force microscopy (C-AFM) and make a comparison between them. We use diamond-coated highly-doped cantilevers for the experiments and apply tip forces of several µN. This combination of high force and continuous C-AFM scanning results in a tip-induced material removal (digging) [1], [2]. In order to get C-AFM images, the sample is biased at a small voltage during this repeated scanning until a desired depth is reached. This procedure reveals C-AFM images layer-by-layer into the bulk, to assess depth-dependent electrical properties (e.g. current signals at grains and grain boundaries, I-V curves etc.). Comparison of the electrical characteristics for different AlkF-PDT CIGS samples gives a deeper insight into alkali diffusion through CIGS bulk and its effect on charge carriers. [1] Celano, U., Hsia, F., Vanhaeren, D. et al. Mesoscopic physical removal of material using sliding nano-diamond contacts. Sci Rep 8, 2994 (2018) doi: 10.1038/s41598-018-21171-w [2] J. Luria, Y. Kutes, A. Moore, L. Zhang, E. A. Stach, and B. D. Huey, ‘Charge transport in CdTe solar cells revealed by conductive tomographic atomic force microscopy’, Nat. Energy, vol. 1, no. 11, p. 16150, Nov. 2016.

Resume : Ultrathin Cu(In,Ga)Se2 (CIGS) devices are industrially relevant since when compared with standard thickness devices, they have the potential to: i) reduce costs, ii) increase machine throughput, and iii) increase device performance. However, there are two major intrinsic problems: i) increased rear interface recombination, and ii) decreased light absorption. Rear interface recombination can be mitigated by using surface passivation layer in combination with a local nanocontact structure. By having a very high percentage of the rear contact covered with a transparent dielectric layer for passivation and by having Mo being the material of contact with the CIGS in the contact openings, there is freedom to add extra functionality to these layers. In this contribution we will review some of the architectures that were developed for the H2020 project ARCIGS-M. We will review the impact of the geometry of the contacts (lines vs. points); the use of different dielectric materials and its impact on interface defect density and built-in electrical field; and how highly reflective metals and nanostructures can be incorporated in passivated architectures to increase device performance. We will showcase several examples of how different strategies can lead to increased values of open circuit voltage and short circuit current over reference devices for ultrathin CIGS solar cells. We will discuss in detail champion cells and the next steps that combine all the observed improvements.

Resume : The optical losses inherent to the reduced absorber thickness of ultra-thin Cu(In,Ga)Se2 (CIGS) solar cells are tackled in this work through the implementation of a novel light management architecture. The incorporation of nanostructures in optoelectronic devices to increase the light optical path in the absorber layer has been widely studied. However, several problems related with the processing compatibility and the low performance of the nanostructures has hindered such actions in real-life devices. In this work, a novel way of introducing gold nanoparticles in an ultrathin CIGS (~500 nm) solar cell device for light management is proposed. The nanostructures are encapsulated with a dielectric layer shielding them from the CIGS harsh growth processing conditions. Such structure is achieved through the combination of a bottom-up chemical approach of depositing the nanostructures with a top-down photolithographic approach that allows for the creation of electrical contacts. With the implementation of the proposed concept on a CIGS solar cell device, a 3.7 mA/cm2 (abs) short-circuit current increase is shown, when comparing with a reference ultra-thin device. Through optical simulations, an enhancement of the light’s effective optical path of approximately 4 times the nominal thickness of the absorber layer in the infrared region was verified.

Resume : In recent years, AVANCIS has developed a dry buffer layer for large-area in-line processing of chalcopyrite-based thin-film solar cells [1]. This process results in excellent efficiencies and avoids the often-used CdS wet-chemical bath deposition (CBD). However, the dry process lacks the surface-cleaning characteristics of the CBD and thus techniques that serve as a cleaning step (e.g., to remove surface adsorbates) at an intermediate stage of the production process can increase the flexibility of the thin-film production process. Therefore, a UV-induced ozone treatment of Cu(In,Ga)(S,Se)2 (CIGSSe) absorbers is investigated, in which UV light transforms oxygen into ozone that reacts with the surface to remove surface adsorbates. In this contribution, the chemical surface structure is investigated before and after UV treatment using a complementary set of spectroscopic methods, combining laboratory-based x-ray photoelectron spectroscopy (XPS) and x-ray-excited Auger electron spectroscopy (XAES) with synchrotron-based soft x-ray emission spectroscopy (XES). As a response to the UV-induced ozone cleaning, carbon adsorbates are removed, while the oxygen content at the surface increases, accompanied by the formation of indium-, selenium-, and sulfur-oxide species. [1] R. Lechner, P. Eraerds, M. Stölzel, T. Niesen, M. Sode, A. Weber, M. Algasinger, C. Schubbert, R. Verma, T. Dalibor and J. Palm, in presented at the 33rd EU PVSEC, Amsterdam, The Netherlands, 2017.

Resume : Reducing the thickness of CIGS solar cells is a promising way to improve its competitiveness by reducing the manufacturing costs. However, CIGS solar cells with submicrometer-thick absorbers generally exhibit lower open-circuit voltage (Voc) and fill factor (FF) than devices with standard thicknesses. These losses are related to a degradation of the CIGS material quality and a higher risk of shunt paths, as well as an increased recombination of charge carriers at the back contact that can also result in substantial Jsc losses. Ag-alloyed CIGS layers (ACIGS) has been reported to improve the Voc and so the performances of CIGS solar cells. We have fabricated ultrathin ACIGS solar cells co-evaporated at 550°C and 500°C on Mo back contacts, and we have compared them to CIGS solar cells deposited in the same conditions. We demonstrate efficiencies up to 14.9% with 500 nm-thick ACIGS solar cells at 550°C. This best efficiency is close to the current record efficiency of 15.2% for ultrathin CIGS absorbers. It is also remarkably higher than the efficiencies reached with our CIGS solar cells, in particular thanks to very high Voc=744mV and FF=81.8%. To provide a better understanding of the performances of ultrathin CIGS and ACIGS solar cells, a cathodoluminescence study was performed. We show that the spatial fluctuations of the composition are very limited (ΔEg<10meV). The best performances obtained for ACIGS layers deposited at 550°C are well correlated with a higher luminescence intensity and a narrower Urbach tail, indicating a reduced density of shallow defects.

Resume : Metal halide perovskite (MHP) materials exhibit exceptional performance characteristics for low-cost optoelectronic applications. Though widely considered defect tolerant materials, perovskites still exhibit a sizeable density of deep sub-gap non-radiative trap states, which create local variations in photoluminescence [doi: 10.1126/science.aaa5333] that fundamentally limit device performance. These trap states have also been associated with light-induced halide segregation in mixed halide perovskite compositions [doi: 10.1021/acsenergylett.8b02002] and local strain [doi:10.1039/C8EE02751J], both of which can detrimentally impact device stability. The origin and distribution of these trap states remains unknown as multiple, complimentary multi-modal techniques are required to probe their location and surrounding structure and composition and MHPs damage rapidly under the electron beam. Understanding the nature of these traps will be critical to ultimately eliminate losses and yield devices operating at their theoretical performance limits with optimal stability. In this talk, we outline a low dose, multi – technique framework to reveal the structural origins of non-radiative recombination sites in (Cs0.05FA0.78MA0.17)Pb(I0.83Br0.17)3 thin films[Accepted, Nature]. By combining scanning electron diffraction and energy dispersive X-ray spectroscopy, with photoemission electron microscopy (PEEM) measurements we reveal that nanoscale trap clusters are distributed non-homogenously across the surface of high performing perovskite films and that there are distinct structural and compositional fingerprints associated with the generation of these detrimental sites.

Resume : Solar power has a major role in the crucial transition to low carbon power. Perovskite solar cells, in particular, are a great photovoltaic alternative as they exhibit excellent energy harvesting abilities and do not require energy-intensive manufacturing techniques. While poor stability during scale-up of perovskite photovoltaics limits their widespread commercial use, the key is the relationship between the microstructure and properties. Thus, an analytical strategy is necessary for resolving the microstructure at the nanoscale to efficiently tune the perovskite properties beyond the current state of the art. We have developed a magnetic sector Secondary Ion Mass Spectrometer (SIMS) and coupled it to a Helium Ion Microscope (HIM). Owing to this, we combine the best of both worlds; the high lateral resolution in secondary electron imaging of the HIM with the high detection limit of our internally developed SIMS instrument. In SIMS mode, we are able to detect elements in the full mass range with a lateral resolution below 15 nm. The SIMS data can be correlated with sub-nm resolution secondary electron images. We will show several analytical studies completed with HIM-SIMS, among them, a cross-sectional chemical map of a multi-cation based perovskite device where different thin films at the nanoscale can be distinguished. HIM-SIMS is a versatile and powerful technique for the improved understanding of the impact of microstructure on photovoltaic performance and stability.

Resume : Lateral inhomogeneities in material properties arise in Cu(In,Ga)Se2 solar cells from fabrication processes, growth conditions and its polycrystalline nature. An essential parameter defining the quality of the material is the carrier lifetime, which directly influences the open circuit voltage (Voc) and the performance of the devices. Time resolved photoluminescence (TRPL) is typically used to characterize the carrier lifetime. The low injection regime is the preferred measurement condition but is challenging to achieve experimentally, especially for mapping measurements, due to the difficulty to obtain a sufficient signal-to-noise-ratio. Achieving such conditions allows for a better understanding of the possible Voc limitations at standard test conditions due to lateral inhomogeneities. In this contribution, we evaluate a wide-field illumination microscopy technique to achieve low injection conditions. We will present the spatial distribution of carrier lifetime at both large and small scale (cm-μm) for different Cu(In,Ga)Se2 absorbers with different materials. The intensity and lifetime maps obtained revealed significant inhomogeneities with strong non-Gaussian distributions. An assessment of the impact of such distributions and its correlation to Voc and solar cell performance will be presented.

Resume : CdTe solar cell technology has reached a levelized cost of electricity competitive with all other electricity generation sources. Yet despite its commercial maturity, there is still headroom to increase performance significantly by addressing fundamental material challenges. Like most thin film polycrystalline technologies, adjusting carrier concentration has been difficult. This presentation will describe recent work to control absorber hole density, understand buried interfaces, assess electrostatic potential fluctuations, and push the boundaries of understanding and performance.

Resume : Thin film cadmium telluride is the most commercially successful thin film photovoltaic technology especially for utility scale deployment. The efficiency of cadmium telluride (CdTe) solar cells has been improved recently to 22.1% by replacing the cadmium sulphide buffer layer with a transparent metal oxide and by alloying selenium into the CdTe absorber. Both have the effect of increasing the short circuit current density Jsc. The addition of selenium reduces the bandgap of the absorber material but this does not fully explain the improvement in efficiency. By using a combination of analysis techniques including SEM, cathodoluminescence and high spatial resolution secondary ion mass spectrometry it can be shown that the addition of selenium results in higher luminescence efficiency and longer diffusion lengths showing that selenium passivates critical defects in the bulk of the absorber layer. By combining cathodoluminescence with cross-sectional Transmission Electron Microscopy it is possible to show that selenium also passivates grain boundaries. Understanding how the recent efficiency improvements have been achieved will help guide research to make further gains in device performance.

Resume : It was already shown that the quantum efficiencies of CdTe solar cells can be improved simultaneously in short and long wavelength range by an additional alloying with selenium. The absorber layer of the thin film solar cells is composed of a Cd(Te,Se) sub layer and a pure CdTe top layer. In the frame of our work we have varied the selenium concentration of the Cd(Te,Se) sub layers between 5 and 20 at.%. The polycrystalline microstructures of the solar cells were analyzed by field emission scanning electron microscopy (FE-SEM) on ion-polished cross sections. For selenium contents below 10 at. % a block-like microstructure with 1 µm large grains and no grain boundaries in growth direction are observed. In contrast for selenium concentrations above 10 at. % grain boundaries are present between the Cd(Te,Se) sub and CdTe top layer. The chemical depth profiles of the solar cells were analyzed by glow discharge optical emission spectrometry (GD-OES) and show a chlorine enrichment at interface to the front contact and at grain boundaries between Cd(Te,Se) and CdTe layer. Measurements of the electron beam induced current (EBIC) reveal that with increasing selenium content the maximum of the EBIC average line profiles is shifted into the direction of the back contact. In contrast the EBIC signal near to front contact is continuously decreased with increasing selenium contents. The highest efficiencies of 17 % were achieved for about 14 at.% selenium in the Cd(Te,Se) sub layer.

Resume : CdTe is one of the most promising photovoltaic (PV) materials due to its near optimum band gap, high absorption coefficient and low manufacturing costs. The practical efficiencies of polycrystalline CdTe PV devices are still below the theoretical limit (~30%), due to non-radiative Shockley-Read-Hall (SRH) recombination at grain boundaries (GBs), dislocations, and point defects. Low minority carrier lifetime constrains CdTe- and CdTe based devices from reaching the theoretical efficiency limits. Reduction of non-radiative recombination at grain boundaries is believed to be the key to improve the efficiency of polycrystalline CdTe-based solar cells. CdCl2 heat treatment a well-known method to passivate non-radiative recombination at grain boundaries as well as improve minority carrier lifetime. Recently new developed device fabrication methods enable us to explore other approaches to improve minority carrier lifetime. In this contribution, we combined experimental and theoretical methods to study the co-passivation of effects of co-segregation along GBs in poly-crystalline CdSeTe devices. In particular, we use high-angle annular dark field (HAADF) imaging in combination with X-ray energy-dispersive spectroscopy (XEDS) to quantitatively analyze of the local composition and atomic structure of incoherent GBs in CdSeTe devices. First- principles DFT calculations based on the structure model extracted from STEM images indicates that co-doping has the potential to eliminates mid-gap state in band structure, thus significantly improve minority carrier lifetime.

Resume : Pulsed DC magnetron sputtering is an effective way of creating Thin Film coatings. In particular, thin film Cadmium Telluride (CdTe) is a relatively low cost material for a solar cell that can be deposited at a fast rate using this process. The current research cell efficiency of a CdTe cell is ~22.1% but with a theoretical maximum of 30% and a band gap of 1.49 eV, CdTe is an ideal candidate for a solar cell [1]. However, after sputtering, experimental analysis shows ~ 4% of the magnetron working gas, Argon, is incorporated into the film. This accumulates into large bubbles and surface blisters between 3 and 50 nanometres in size after post annealing at 400o C during a cadmium chloride treatment used to increase cell efficiency [2]. This is an obstacle which stands in the way of efficient CdTe cell production. In this presentation we use computational modelling to explain the mechanisms by which these bubbles form. Using Molecular Dynamics (MD), the threshold energy and penetration depths are determined for both Ar and Xe with CdTe in both the zinc-blende and wurtzite phases. These calculations show that more Ar than Xe can penetrate into the growing film and more across the (111) surface, which is the usual orientation in which the films are grown, than other surface orientations. The mechanisms and energy barriers for interstitial Ar and Xe to diffuse are determined showing that the barriers are reduced near existing clusters, increasing the probability of capture and thus enhancing cluster growth. MD simulations of the high temperature annealing process show that bubbles grow through diffusion and trap mutation mechanisms. Bubble size distributions are determined as a function of annealing times. Ar is shown to diffuse faster than Xe and more easily in zinc-blende CdTe compared to wurtzite, which exhibits some non-Arrhenius behaviour, providing an explanation for the difference in the sizes of voids observed before and after cell activation. Blister exfoliation was also modelled, showing the formation of shallow craters with a raised rim, in agreement with images obtained using scanning electron microscopy. Based on the simulation results, ways as to how bubble formation may be minimised will be discussed. 1. W. Shockley and H. Queisser, “Detailed Balance Limit of Efficiency of pn Junction Solar Cells”, Journal of Applied Physics, vol. 32, no. 3, pp. 510-519, 1961. 2. P. M. Kaminski, S. Yilmaz, A. Abbas, F. Bittau, J. W. Bowers, R. C. Greenhalgh, and M. Walls, “Blistering of magnetron sputtered thin film CdTe devices,” in Proceedings of IEEE 44th PVSC, 2017.

Resume : In the past few years, Cadmium Telluride thin-film photovoltaics has seen unprecedented growth. Efficiencies over 19% have been recorded by alloying Se with CdTe absorber layer forming CdSexTe1-x compounds. Doping absorber layer to increase the carrier concentration has been tried using various elements such as Copper, Arsenic etc. Published literature suggests that doping CdTe absorber layer with Copper allows it to go into the Cadmium site (CuCd). Formation of CuCd is beneficial for the CdTe absorber layer. However, it has been observed that copper may also occupy the interstitial sites (Cui) which is detrimental to the device performance. Arsenic (As) is another dopant which is of interest to the scientific community to dope CdSexTe1-x. Doping CdSexTe1-x with As can allow As to occupy either Selenium site (AsSe) or Tellurium site (AsTe). This work, therefore, uses the first-principles approach to study the behaviour of AsSe and AsTe in CdSexTe1-x bulk. Density Functional Theory (DFT) with a Linear Combination of Atomic Orbitals (LCAO) basis sets have been used to simulate the As doped CdSexTe1-x structure QuantumATK software from Synopsis. This is one of the first uses of this approach demonstrated for the study of CdSexTe1-x material. The stress behaviour has been studied with AsSe and AsTe sites. The energetics has also been calculated to understand the stability of As going into Se and Te sites. Furthermore, the Density of States has been studied to understand the presence of the defect states with As being either in Se or Te site.

Resume : Solar energy plays an important role in solving serious environmental problems and meeting the high energy demand. However, the lack of suitable materials hinders further progress of this technology. Here, we present the largest inorganic solar cell material search till date using density functional theory (DFT) and machine-learning approaches. We calculated the spectroscopic limited maximum efficiency (SLME) using the Tran–Blaha-modified Becke–Johnson potential for 5097 nonmetallic materials and identified 1997 candidates with an SLME higher than 10%, including 934 candidates with a suitable convex-hull stability and an effective carrier mass. Screening for two-dimensional-layered cases, we found 58 potential materials and performed G0W0 calculations on a subset to estimate the prediction uncertainty. As the above DFT methods are still computationally expensive, we developed a high accuracy machine-learning model to prescreen efficient materials and applied it to over a million materials. Our results provide a general framework and universal strategy for the design of high-efficiency solar cell materials. The data and tools are publicly distributed at: https://www.ctcms.nist.gov/~knc6/JVASP.html, https://www.ctcms.nist.gov/jarvisml/, https://jarvis.nist.gov/, and https://github.com/usnistgov/jarvis.

Resume : Since solar energy conversion for renewable energy resource is related with optical properties and carrier transport of semiconductor materials, it is crucial to control many factors, such as compositions, defect/dopant, and band alignements between phases. The key factors are mostly determined by processing conditions and methods. However, it is extremely challenging to find the relation between processing conditions and properties, because of the lack of thermodynamic parameters and understanding in composition(defect)-electronic structure relations. In this talk, I will introduce my recent works on the developments of alloyed chalcogenide solar cell materials. For efficient design of solar cell materials, we combined density functional theory (DFT) calculations and thermodynamic modeling to predict the equilibrium compositions and resultant band gaps of Cu2GeS3 solar cell materials when mixed with Se at S-site and CuSbS2 with Bi-mixing at Sb-site. From the computations, spinodal decompositions of Cu2GeS3-xSex and CuSb1-xBixS2 were predicted. In addition, using accurate band gap calculations of the materials, we proposed optimum thermal-processing temperature.

Resume : This contribution concerns the effect of the Ag content in wide-gap AgwCu1-wIn1-xGaxSe2 (ACIGS) absorber films and its impact on solar cell performance. First-principles calculations are conducted, predicting trends in absorber band gap energy (Eg) and band structure across the entire compositional range (w and x). It is revealed that a detrimental negative conduction band offset (CBO) with a CdS buffer can be avoided for all possible absorber band gap values (Eg = 1.0 – 1.8 eV) by adjusting the Ag alloying. This opens a new path to reduce interface recombination in wide-gap chalcopyrite solar cells. Indeed, corresponding samples show a clear increase in open-circuit voltage if a positive CBO is created by sufficient Ag addition. A further extension of the beneficial compositional range (positive CBO at buffer/ACIGS interface) is possible when exchanging CdS with Zn1-ySnyOz (ZTO), due to its lower electron affinity. Nevertheless, the experimental results strongly suggest that at present, residual interface recombination still limits the performance of solar cells with optimized CBO, which show an efficiency of up to 15.1 % (no anti-reflection coating) for an absorber band gap of Eg = 1.45 eV.

Resume : Motivated by environmental reasons, In2S3 is a promising candidate for a Cd-free buffer layer in Cu(In,Ga)(S,Se)2-based thin-film solar cells. For an impactful optimization of In2S3 as alternative buffer layer, however, a comprehensive understanding of its defect chemistry in the bulk region and electronic properties at the absorber/buffer interface is of foremost importance. In this contribution, we study—by means of hybrid density functional theory— incorporation of various impurities in In2S3 and investigate their impact on thermodynamic stability and n-type conductivity of In2S3 [1,2,3]. Furthermore, to understand how band lineups would affect open circuit voltage and the overall performance of the cells buffered with In2S3, we present band alignment of pure and chemically modified In2S3 with various chalcopyrites [4]. [1] E. Ghorbani and K. Albe, Journal of Materials Chemistry C 6, 7226 (2018) [2] E. Ghorbani and K. Albe, Physical Review B 98, 205201 (2018) [3] E. Ghorbani and K.Albe, J. Appl. Phys. 123, 103103 (2018) [4] E. Ghorbani, P. Erhart, and K. Albe, Physical Review Materials 3, 075401 (2019)

Resume : In this study, sulfur incorporation of CuInSe2 (CISe), and CuGaSe2 (CGSe) absorber layers is investigated in order to better understand the sulfurization process of Cu(In,Ga)Se2 (CIGSe). We found that for all investigated annealing temperatures from 405 to 530°C, sulfur is always incorporated into the CISe absorber, forming a thin CuInS2 (CIS2) layer at the surface. By increasing the annealing temperature, this surface layer becomes thicker and acts as an electron barrier, which deteriorates the device performance by decreasing fill factor (FF), and current density (JSc). In contrast, for the off-stoichiometric CGSe samples (Cu-poor), S could only be introduced at a temperature of 530°C or above, mainly forming a mixed crystal of CuGa(S,Se)2 (CGSSe). In Cu-rich samples, however, selenium (Se) could substitute S independent of the annealing temperature, which could be explained by a sequential phase transformation of Cu2-xS to Cu2-xSe to CuGaS2. Finally, the solar cell parameters of the CISe solar cells are improved by sulfurization with respect to their references, while the electrical characteristics deteriorated for the CGSe cells. Keywords: Sulfurization, CISe, CGSe, phase transformation, Cu2-xS, Cu2-xSe, and OVC.

Resume : As an alternative to the environmentally undesirable CdS buffer layer, Zn(O,S) materials have attracted interests for use in the Cu(In,Ga)Se2 (CIGS)-based thin-film solar cells. The Zn(O,S) thin films were prepared by conventional chemical bath deposition (CBD) and reactive sputtering process with a single ZnS target. For CBD, Zn(O,S) buffer layer was grown on CIGS layer in the bath with zinc sulfate, thiourea and ammonia in aqueous solution. For the reactive sputtering process, O2/Ar mixture gas ratio was controlled to vary the sulfur-to-oxygen composition ratio. The solar cell efficiencies were investigated for Zn(O,S) buffer layers from different processes. We also report on the color tuning of CIGS solar cells via controlling the thickness of Zn(O,S) buffer layer and the ITO transparent conducting layer. Because Zn(O,S) and ITO have different refractive indices, reflective color can be easily tuned even with a slight change in the optical path lengths of each layer. In addition, colored layers based on a nanograting structure were coated on CIGS thin-film solar cells by the direct nanoimprinting process with spin-on-glass resin. By changing the pitch and size of nano-gratings, wide range of colors in CIGS thin-film solar cells were successfully demonstrated. Photometric properties and the photovoltaic performance of the color-tuned CIGS solar cells were also characterized.

Resume : The structural, electronic and optical properties of semiconductors compounds ZnX (X=O, S and Te) have been calculated using the linearized augmented plane wave method (FP-LAPW) within density functional theory (DFT). The structural properties are investigated using the Wu–Cohen generalized gradient approximation (WC-GGA) of the exchange-correlation potential. For band structure calculations, in addition to WC-GGA), the recently modified Becke-Johnson approximation proposed by Tran and Blaha is used. Optical characteristics, such as dielectric functions, refractive indices, extinction coefficient and optical reflectivity, are calculated for photon energies up to 14 eV. The study is also extended to the ZnO1-xSx and ZnO1-yTey alloys where we studied the effect of concentration on different physical quantities.

Resume : Thin film hetero-junction solar cells are ideally positioned to become a sustainable photovoltaic technology with a reduced material consumption and a high value of recycling at the end of life. To improve their performance not only for the record efficiency solar cells (CIGS and CdTe) but also for the kesterite (CZTS) without toxic or critical raw materials an improved understanding is necessary. This understanding relies on characterization experiments. Admittance spectroscopy is widely spread and has a long tradition for characterization of junction devices and the defects therein. However, in thin film chalcogenide materials assigning the observed features to specific properties in the device stack remains very challenging. Moreover, such challenges often result in the absence of key-ideas for further improvement of the devices. Capacitance spectroscopy can help to support the modelling, not only to estimate the hole density in the absorber layer but also to sometimes estimate barrier heights and band offsets. With numerical device simulations we will evaluate which band offsets, barriers, layer properties and defects can be observed with capacitance spectroscopy. Based on a classification we will show that indeed many properties can lead to capacitance steps in admittance spectroscopy. We will evaluate their properties and show how they might help to make adequate numerical models.

Resume : In the last few years, by changing the original solar cell structure, the efficiency of CdTe solar cells has been increased from around 17% to 22%. This is a result of an improvement of all the electrical parameters but with a larger increase in the short circuit current. In fact, the standard CdS window layer has been replaced by a high band gap semiconductor. Moreover, a CdTe1-xSex layer has been introduced, which allows the lowering of the absorber band gap near the junction and by that improving the absorption in the long wavelength region. Some groups have introduced a CdSe layer at the junction, subsequently mixed with CdTe by the activation treatment, others directly deposit a CdTe1-xSex compound. In this work we present a new method: the introduction of a CdTe1-xSex compound near the junction through selenization of the CdTe layer. We fabricate CdTe solar cells in superstrate configuration by low-temperature process. On a glass/ITO/ZnO/CdS stack we deposit a thin CdTe layer (from 0.5 to 1 μm) by vacuum evaporation, subsequently the stack is annealed at 400 °C in selenium atmosphere. Both X-ray diffraction and Raman analysis confirm the formation of intermediate layers such as CdSe and CdTe1-xSex compounds in the stack. Afterwards a 5 μm thick CdTe layer is deposited, followed by the activation treatment and the deposition of the back contact. An analysis of the selenized layer properties will be presented, together with the overall performance of the completed devices.

Resume : The diode factor of a solar cell devices gives important information on the dominant recombination pathway and hence the bottleneck for higher VOC’s. It can be determined either electrically by a current-voltage (IV) curve yielding values between 1 and 2 [1] or optically by photoluminescence (PL) [2], where values of either 1 or 2 are expected based on Shockley-Read-Hall (SRH) statistics. However, experimentally, values between 1 and 2 are observed. We model the optical diode factor on a bare absorber using SRH statistics as a recombination channel. In order to model diode factors between 1 and 2 we introduce metastable defects – for instance the VSe-VCu defect complex observed in Cu(In,Ga)Se2 (CIGS) based absorbers. Upon excitation (illumination for PL or voltage for IV) these metastable defects change from donor-like to acceptor-like [3]. As a result, the majority carrier hole Fermi-level shifts towards the valence band, increases the quasi-Fermi level splitting and thus the PL yield. This effect then leads to a diode factor between 1 and 2. It is noted that these metastable defects are generally identified to be responsible for the increased carrier concentration upon light-soaking a CIGS solar cell. First experiments show a higher optical diode factor in films that are grown with lower Se pressure – in agreement with the model. It is thus mandatory to minimize the metastable defects in state-of-the-art CIGS absorbers to reduce the diode factor and thus improve VOC. 1. Scheer, R. and H.W. Schock, Chalcogenide Photovoltaics: Physics, Technologies, and Thin Film Devices. Chalcogenide Photovoltaics: Physics, Technologies, and Thin Film Devices2011, Weinheim, Germany: Wiley-VCH Verlag & Co. KGaA. 2. Babbe, F., L. Choubrac, and S. Siebentritt, The Optical Diode Ideality Factor Enables Fast Screening of Semiconductors for Solar Cells. Solar RRL, 2018. 2(12): p. 1800248. 3. Lany, S. and A. Zunger, Light- and bias-induced metastabilities in Cu(In,Ga)Se2 based solar cells caused by the (VSe-VCu) vacancy complex. Journal of Applied Physics, 2006. 100(11).

Resume : We studied structural and optical properties of high-quality In2S3 films which was synthesized either by using thermal sulfurization of InAs substrates or by using a simple and reliable physical vapor deposition on various substrates in a hot-wall tube furnace. Most of the synthesized films were high quality tetragonal β- In2S3, which was confirmed by X-ray diffraction (XRD) and Raman spectroscopy. Scanning electron microscopy (SEM) analysis showed that the film thickness and the average size of the In2S3 crystallites decreased as the growth temperature decreased. The synthesized In2S3 films with high quality tetragonal β-phase were grown in the preferred growth direction of [1 0 3]. The unusual preferential growth direction was confirmed by transmission electron microscopy and XRD data. The bandgap of the synthesized films, measured by absorption, was approximately 2.6 eV. They showed unusually strong photoluminescence (PL) at room temperature. Excitation power dependence of the strong PL intensity indicated that the strong PL signal is due to excitons related to some defect.

Resume : Cu(In, Ga)Se2 (CIGS) thin-film solar cells have been advanced considerably among thin-film technologies. The incorporation of alkali elements in the CIGS layer has been one of the most effective routes achieving a higher photo-conversion efficiency (PCE) of CIGS solar cells. In this work, Na doped Mo(600nm) layer has been introduced to fulfill the Na diffusion for STS substrate with the solar cell device structures Al-Ni/AZO/i-ZnO/CdS/CIGS/Mo/STS and Al-Ni/AZO/i-ZnO/CdS/CIGS/Mo/Mo:Na/STS fabrication. The influence of Na on interdiffusion of In & Ga (In- in diffusion, Ga-out diffusion) and V-shaped Ga grading has been discussed from the XRD, RAMAN, SIMS, and TEM analysis. Thermal admittance spectroscopy (TAS) has been analyzed to unveil the consequences of Na on signature steps N1, N2 and correlate with the PCE and current density increase from 9.06% to 12.03% and 28 to 36mA/cm2 for CIGS/Mo/Mo:Na as compared to CIGS/Mo layer respectively. From, IVT (current-voltage-temperature), CVT (capacitance-voltage-temperature) and CFT (capacitance-frequency-temperature) analysis, shallow-defects with activation energies (EA) from 12 to 22meV have been identified that can be assigned to VCu acceptor. The reduction in EA for deep-level defects from 260 to 63meV validates the core aspect directly related to the advancement in the device performance. The defect distribution has been evaluated from the trap density dependent on demarcation energy (Eω). The Eω varies from 0.26 to 0.097eV with a significant change in trap density. The energy-level shift towards the shallower position, reduced depletion barrier at the back junction, and V-shaped Ga-grading near to the front surface are the key findings for the improvement in solar cell performance.

Resume : The variable bandgap of Cu(In,Ga)(Se,S)2 chalcopyrite solid solution crystals, makes them a highly suitable material for absorber layers in thin film photovoltaic applications. Especially the possibility to tune the bandgap over the thickness of the absorber is of interest. Therefore, the crystallographic composition resolved over the depth of the thin film needs to be analyzed. Measurements in asymmetric Bragg Diffraction provide information up to various absorber depths, depending on the incidence angle of the X-rays. The resulting diffraction patterns can be analyzed with regard to the chemical composition via the lattice parameters and the shape of the observed reflections. To gain insight into the composition as a function of depth complementary measurements have been performed not only from the front side, but also from the back side of the absorber layers. Therefor the absorber layers have been separated from their substrates and back contacts on large areas, up to 3x3cm². The results show signal attributed to both some solid solution gradients, as well as significant reflection intensities from two phases with very well defined compositions, which are both present in depth dependent volume ratios in the bulk of the absorber. Exact Rietveld analysis of the measured patterns is aggravated by the necessity of various corrections on both intensity as well as reflection positions, which are caused by the small incidence angles.

Resume : The investigation and understanding of fluctuating potentials impact on the performance of solar cells based on complex absorber semiconductors is an open issue in the scientific community. Here, we present a theoretical approach and an experimental investigation of the fluctuating potentials influence on the performance limitations of Cu(In,Ga)Se2 (CIGS) solar cells. We calculated the fluctuating potentials influence in the open-circuit voltage (Voc) deficit, according to extensions to the Shockley-Queisser (SQ) model. Furthermore, we compare these simulations with the experimental results. Three extensions to the SQ model are presented according to different approaches to calculate the absorptance considering the presence of tail states described by the Urbach-rule, optimal-fluctuation-theory, and bandgap-fluctuation models. For the different approaches the expected solar cells figures of merit were estimated. From an experimental point a view, 3-CIGS solar cells were grown, to obtain such samples [Cu] ratios were changed so that the density of defects also changes. We showed both theoretically and experimentally the role played by fluctuating potentials, in particular in the Voc deficit. The theoretical calculations and experimental results showed evidences of a higher degree of correlation of electrostatic fluctuating potentials with the Voc deficit, in comparison with bandgap fluctuations. Finally, we showed the influence of fluctuating potentials at room temperature.

Resume : Thermal evaporated indium sulfide (InxSy) is already commercially used as buffer layer in chalcopyrite-based thin-film solar modules. However, sputtering of InxSy is a much faster deposition method compared to all other common methods as evaporation, chemical bath deposition (CBD) or atomic layer deposition. It can be easily implemented in an industrial environment to grow the buffer layer on top of the Cu(In,Ga)Se2 (CIGS) absorber, even in a roll-to-roll coater for flexible substrates. In this work, CIGS with a Ga/(Ga+In) ratio of 0.3 was deposited in a multi-stage co-evaporation process on Mo-coated soda-lime glass. The process was carried out in an industry-relevant in-line machine with 30x30 cm2 carriers. In the next step InxSy or Inx(O,S)y buffers were prepared by rf sputtering either from a pure In2S3 or different mixed Inx(O,S)y targets. We investigated the influence of the O/(O+S) ratio, the substrate temperature, and the sodium accumulation at the CIGS/buffer interface on the p/n-junction formation and the device performance. We present efficiency results on the different Inx(O,S)y-buffered cells and compare them to reference cells with CBD CdS buffers. We have achieved efficiencies above 15%. In addition, a detailed materials characterization of our various Inx(O,S)y layers will round up the contribution, including time-of-flight secondary ion mass spectrometry depth profiles carried out at the CIGS/buffer interface.

Resume : Photovoltaic devices based on Cu(In,Ga)Se2 (CIGSe) thin films are now among the most attractive non-Si alternatives. The key to their recent efficiency gain is a post-deposition treatment (PDT) with alkali salts. RbF was demonstrated as the most efficient alkali salt for the PDT of co-evaporated CIGSe- based solar cells. It is known that the effectiveness of the RbF-PDT strongly depends on the composition of the underlying CIGSe. However, the formation of secondary phases such as RbInSe2 during the PDT might affect the CIGSe surface and the device performance in general. To clarify this effect, RbF- and RbInSe2- covered CIGSe absorbers were investigated in-vacuo by means of surface sensitive methods. Kelvin probe and photoelectron yield spectroscopy techniques were applied to determine the work function and the valence band edge (VBM) of the RbInSe2/CIGSe interface as well as its energy band diagram as a function of the RbInSe2 deposition time. The VBM of bare CIGSe is halved after only 1 minute co-evaporation of RbInSe2 with small changes up to 4 minutes, whereas a deposition for more than 10 minutes seems to restore the VBM. These changes are correlated to the chemical composition of the surface of the RbInSe2-covered CIGSe given by X-ray photoelectron spectroscopy analysis and compared to those of bare and RbF- covered CIGSe. Finally, the chemical and electronic structures of the different CIGSe absorbers are correlated to the performance of the corresponding devices.

Resume : Imaging and luminescence mapping techniques are usually employed to study the spatial and lateral variations of material properties in Cu(In,Ga)Se2 material. For instance, steady-state photoluminescence (PL) mapping has been extensively used to evaluate potential fluctuations, chemical composition as well as structural defects. Usually, steady state and time resolved photoluminescence (TRPL) mapping are performed under high excitation conditions whereas the solar cell operates under high injection regime. Low injection regime (or 1-sun conditions) is preferable to detect Voc limitations at standard operating conditions. Achieving low injection conditions is a challenging task. Only few reports are available for TRPL mapping with micrometer resolution for Cu(In,Ga)Se2 materials at high injection conditions, and even fewer or none at low injection. In this contribution, we implement TRPL scans with wide-field illumination to achieve low injection conditions. We report on the distribution of decay parameters at the cell level as well as with high resolution of ~2 μm, for different quality grades of Cu(In,Ga)Se2 material. The excitation power is varied within one order of magnitude and its effect on the carrier dynamics and experimental lifetimes is investigated. We discuss some of the capabilities and pitfalls of the technique, together with examples illustrating what information can be extracted. A comparison with the more standard steady state PL mapping is also presented.

Resume : Most alternative buffer layers used in CIGS solar cell structures are ternary compounds. For these materials, some compositional variation during the very first nanometers of growth can be expected when grown by ALD on CIGS (or other absorber layers). The variations are often relatively small and very hard to quantify with conventional material characterization methods, but can still have a significant influence on the device performance. However, it might be possible to compensate for these effects if they are known, by modifying the ALD process. For this reason, a method based on in-situ quartz-crystal microbalance (QCM) has been developed that allows for a detailed study of the initial growth of ternary compound buffer layers (such as Sn-Ga-O or Zn-Sn-O) on a CIGS surface. Here, conventional QCM-crystals are first pre-coated with a Mo/(NaF)/CIGS stack before being mounted in an ALD reactor. The buffer layer growth on the fresh CIGS surface is then recorded using a QCM monitor. The absorber surface can also be further modified to study the influence of various post-deposition treatments. From the QCM recordings, several valuable parameters can be extracted, including the compositional depth profile. Furthermore, differences in material properties between CIGS grown on the QCM-crystals as compared to soda-lime glass are discussed.

Resume : Thin film solar cells based on chalcopyrite Cu(In,Ga)Se2 (CIGS) absorbers usually employ CdS as buffer layers, especially in high efficiency devices. The parasitic optical absorption, arising from cells with a low bandgap buffer of 2.4 eV as well as the wet deposition method pose problems in terms of a reduced short circuit current (Jsc) and waste treatment, respectively. A buffer material with higher bandgap and a vacuum compatible deposition process is needed. Among various alternative buffer materials such as SnO2, TiO2, Zn(O,S,OH), this contribution explores ZnMgO grown in sequences of atomic layer deposition (ALD) and sputtering to replace CdS as a buffer layer. ZnMgO is deposited by ALD on the substrate/Mo/CIGS stack to obtain full conformal coverage, precise control over thickness and Zn:Mg ratio, which determines the crucial band alignment between absorber and buffer. The window layers are sputtered, either by a combination of ZnMgO and ZnO:Al or i-ZnO and ZnO:Al. Solar cell devices were fabricated and characterized by (temperature-dependant) current-voltage (JV) and external quantum efficiency (EQE) measurements and time resolved photo luminescence (TRPL). It is found that the Jsc can be significantly increased, by reducing parasitic absorbtion. However, this is accompanied by a reduction in Voc and FF.

Resume : The ZnS buffer layer of the CIGS thin film solar cell has the advantage of being environment-friendly and high transmittance. Among the materials for ZnS thin film formation, thioacetamide (TAA) as an S precursor has a faster deposition rate than thiourea. However, it was confirmed that Zn (O, OH) particles were deposited on the surface of the thin films quickly and the characteristics of ZnS thin film were degraded. In this study, experiments were carried out to improve the properties of ZnS buffer layer by adding sodium citrate. The ZnS thin films was deposited on ITO glass by chemical bath deposition process and the solutions such as ZnSO4, TAA, trisodium citrate, ammonia were used. Experimental variables were fixed with ZnSO4 of 0.1 M, TAA of 0.1 M and ammonia solution of 2.5 M. And the concentration of trisodium citrate were changed from 0.0125 M to 0.2 M and the deposition time were changed from 2 min to 20 min. As a result, when sodium citrate was not used, granules deposited on the surface in the conditions of more than 3 min with low transmittance. On the contrary, as the concentration of sodium citrate was increased, granule formation was suppressed and high transmittance was shown. However, when the concentration of sodium citrate was 0.05 M or more, it was confirmed that granules were formed on the surface. As a result of the EDS comparison, the Zn (O, OH) ratio was lowered when the oxygen ratio was calculated to be about 20% lower. Also, when sodium citrate was added, it was confirmed that the light transmission in the short wavelength region was extended. As a result, in the condition of trisodium citrate concentration of 0.04 M and deposition time of 2 min, ZnS thin films having a thickness of about 40 nm and a high transmittance of 83.2% were formed.

Resume : The CdS buffer layer for CIGS thin film solar cells has the toxicity of Cd and the disadvantage of not being transmitted to the photo-absorbing layer due to low energy band gap. Therefore, the study for ZnS buffer layer as Cd-free buffer layer is actively underway. During the deposition of ZnS buffer layer, the thiourea as S precursor has the disadvantage of long process time due to slow deposition rate. In this study, ZnS buffer layer was fabricated by using thioacetamide(TAA) as S precursor in order to improve the slow deposition rate of thiourea. The ZnS buffer layer was deposited on ITO glass using chemical bath deposition. The zinc sulfate (ZnSO4) of 0.1 M and TAA of 0.025 M ~ 0.2 M were used as each precursor, ammonia of 1 M ~ 4 M was used as catalyst, and process time was changed from 2 min to 20 min. As the experiment results for TAA concentration variable, it was confirmed that Zn/S ratio was similar to 1:1 in conditions of TAA concentration of 0.1 M or more and growth of particles was made smoothly. But particles were grown excessively at concentration of 0.2 M, resulting in a non-uniform film. As the experiment results for ammonia concentration variable, the uniform growth of the buffer layer was confirmed at a concentration of 2.5 M. However, it was confirmed that the size of the particles was grown and the ZnS buffer layer did not uniform at a concentration of 3 M or more. As the experiment results for process time, granules were formed at condition of 5 min or more, and characteristics of buffer layer were deteriorated. Therefore, the composition ratio of Zn of 59.3% and S of 40.7% was confirmed under the conditions of zinc sulfate of 0.1 M, TAA of 0.1 M, and process time of 3 min. The ZnS buffer layer having a light transmittance of 82% and a high deposition rate of 13 nm/min was formed.

Resume : Inkjet printing allows fast, accurate and maskless deposition of thin films and uses virtually no solvent compared to common chemical bath deposition methods. We inkjet print ZnS buffer layers for thin film solar cells. The non-toxic and air stable ink consists of ZnCl2 and thiourea dissolved in deionized water and alcohol. At the total precursor concentration of 0.8 M, the viscosity of the ink is 3.5mPa.s at room temperature and easily jets without clogging. ZnS films were prepared on Cu(In,Ga)(S,Se)2 (CIGS) and quartz surfaces, and annealed in nitrogen and air between 200 and 300 °C. The morphology and thickness of ZnS films were investigated by SEM and profilometer showing the thickness of one printed layer is ~30 nm. XRD of ZnS films annealed at 250 oC show very broad peaks indicating a near amorphous structure. From UV-Vis spectroscopy the bandgap is estimated to be 4.2 eV similar to that of ZnS nanoparticles. CIGS solar cells using both inkjet printed ZnS annealed at different temperatures and CdS buffer thin films were fabricated. The J-V characteristics show the solar cells using ZnS film annealed at 250 oC obtained the highest efficiency with the best Voc of 536 mV while the solar cell using conventional CdS buffer with the same CIGS absorber showed a best Voc of 460 mV. Quantum efficiency measurements show superior collection in the blue region compared to that of CdS buffer device thanks to the higher band gap of the ZnS.

Resume : Polycrystalline Cu(In,Ga)(S,Se)2 (CIGS) on soda-lime glass solar cells have been reported to reach very high efficiencies (higher than 23%). The CIGS absorber in these cells is fabricated by the so called “two-step process”, consisting in the deposition of a metallic CuInGa precursor (by sputtering) followed by a selenization and a sulfurization step [1]. In this study we use, instead of sputtering, the coevaporation method typically used in the so called “three-stage process”, which allows us to control in a very versatile and precise way the composition and the microstructure of the layer by modifying the thickness and the order of the metal depositions. Our goal is to understand the links between the structure of the precursor layers and the properties of the CIGS layers obtained after the second annealing step of the process. We performed morphological (SEM), chemical (XRF, EDX, GOEDS), structural (XRD) and optical (PL) characterizations. We observe that the path used to attain the required composition of the precursor (which can be easily visualized in Cu/In/Ga phase diagram [2]) strongly influence its microstructure and so the characteristics of the final CIGS layer. The study of complete solar cells is ongoing, with a particular focus on high bandgap CuIn0,7Ga0,3S2 layers, suitable for tandem applications on silicon. This project has been supported by the French Government in the frame of the “Programme d'Investissement d'Avenir” (ANR-IEED-002-01).

Resume : One of the factors determining the formation of the CdS/CdTe interface and thus affecting the performance of the solar cells is the CdS deposition technology. Electronic properties of the CdS thin films deposited by Close-Spaced Sublimation (CSS) have been investigated by Kelvin Probe (KP) and Photoelectron Yield Spectroscopy (PYS). The CdS thin films have been deposited by CSS onto TCO (AZO, SnO2)/glass substrates, varying the substrate temperature between 280-460°C and keeping the evaporation temperature at 630°C. The substrate temperature increase reveals a saturation trend of the ionization energy for CdS/AZO heterojunctions while the same parameter for the CdS/SnO2 heterojunction depending linearly. The obtained results demonstrate an influence of the substrate temperature for CdS deposition on the Fermi level position and the charge carrier concentration, respectively. The Fermi level position changes between 180 meV and 320 meV below the conduction band edge (CBM) for the CdS/AZO heterojunction while the CdS thin films on SnO2 substrates show degeneration. The KP and PYS results are correlated with the thin-film composition, structure and optical properties.

Resume : The present contribution aims to give further insights as concerns the structural, electronic and optical properties of ZnS1-xOx in the zinc-wurtzite phase for x vary between 0 and 1, using ab-initio calculations based on the full-potential linearized augmented plane waves (FP-LAPW) method in the framework of the density functional theory (DFT). The structural properties (Lattice constant, Bulk modulus and its pressure derivative) were treated by the Generalized Gradient Approximation (GGA) parameterized by Perdew Burke Ernzerhof (PBE-GGA) formalism. While for the electronic and optical properties calculations, the modified semi-local Becke-Johnson exchange potential has been included with the PBE-GGA formalism to overcome under estimation of band-gap by (PBE). The agreement between our calculated results and available experimental and theoretical data is generally good. Keywords: Structural properties; Optical and electronic properties; DFT; Band-Gap Engineering; mBJ-GGA.

Resume : CdTe solar cells on superstrate configuration, requires a transparent conductive electrode (SnO2:F); a window layer of CdS, CdTe thin film as absorbent material and finally a back metal contact (Au/Cu) evaporated on CdTe surface. In order to improve the solar cell performance, CdTe was activated by using and specific CdCl2 thermal treatment. In this work, CdCl2 thermal treatment process was studied, due to this process produces a special coloration and it is atributed to the TeOx formation on CdTe surface promoting an additional electrical response on infrared region, it means that photons with energy lower than the band gap energy of CdTe create useful electron-hole pairs resulting in an enhance of JSC electrical response. Keywords: CdTe solar cells, TeOx, CdCl2 thermal treatment.

Resume : Actually, thin films Solar cells (TFSC) technology has achievement the records efficiencies by using CdS semiconductor material as window layer due to its ideals properties. However, CdS band gap value (2.42 eV) does not allow blue photons to contribute with the photocurrent generation. On the other hand, CdS processing at industrial scale has a significant toxicity disadvantage due to Cadmium content. Therefore, a possible solution is the implementation of new materials with similar properties. In this way (In2S3-3xO3x) has shown adequate properties to be applied in the TFSC technologies. In this work, ternary compound thermally treated under different atmospheres has been applied as window layer in CdTe Solar cells and the electric characteristics were studied. Keywords: Indium sulfide, cadmium reduced solar cells, In2S3-3xO3x

Resume : For thin film solar cells, CdS or Zn(O,S,OH) are well-known buffer material. Especially in CIGS solar cell with CdS and Zn(O,S,OH), 22.9% and 23.35% efficiency were reported, respectively. Until now, although there were many researches about changing thickness, bath temperature for reaction rate, etc., there is no research for finding right time to deposit buffer by tracking mechanism. Normally, mechanisms of forming thin layer during chemical deposition were proposed by whether metal-ammonia complex was formed or not. Therefore, heterogeneous reaction, associated with the ions attached one by one on the surface in order to form films, has more chances to form buffer layer via meeting S and metal ions instead of forming complex surrounded by ammonia. On the other hand, homogeneous reaction, with appearing agglomeration of colloidal particles pre-formed in the bulk of the solution, leads to adsorption on the surface to make particulate films. The homogeneous deposition is highly undesirable because it leads to non-adherent films resulting in a porous layer structure. However, in practice, both mechanisms may occur and/or interact at the same time during the reaction, leading to deposition of films in which colloidal materials are included. Therefore, in our study, the best time to deposit buffer layer with heterogeneous reaction was investigated and the optimum point to deposit buffer will be discussed in detail.

Resume : Printed electronics (PE) technology shows huge promise for the realisation of low-cost and flexible electronics on heat- or pressure-sensitive materials. To enable future PE markets, the ability to produce highly conductive, high-resolution patterns using low-cost additive processes, such as inkjet printing, is a critical technology requirement. Here, we demonstrate the use of Electro-Static Inkjet (ESJET) printing technology for the fabrication of non-vacuum processed CIGS solar cells to bridge the gap of additive printing between high viscosity screen printed materials and low viscosity inkjet processes. In the present work, we avoid the vacuum process of thermal evaporation by using ESJET printing to deposit silver nanoparticle (AgNP) inks for fabrication of grid electrodes for non-vacuum processed Cu(In,Ga)S2 (CIGS) solar cells. The effect of printing parameters and the sintering condition of the ESJET printed Ag grids were investigated by the measurement of photovoltaic performance parameters of CIGS solar cells. The ESJET printing process produced Ag grids with typical dimensions of 50µm width and 4µm height. It was found that sintering temperatures of 220°C caused a significant loss of performance in the CIGS cell,however, sintering of the Ag grids at temperatures up to 160°C produced a cells with good performance and efficiency comparable to the test cells using thermal evaporation processes.

Resume : Based on the fact that the traditional energy sources together with their potentials are predicted to be consumed in near future, there have been several works performed to find alternative renewable energy sources. At this point, works on the use and improvement on the photovoltaic structures have been the point of interest due to being a process of converting sunlight directly into electricity. Although Si is one of the most popular material in these applications with its earth-abundancy and well-known material characteristics, Si-based modules have approached to their theoretical maximum efficiency and they cannot offer a development of original technology. Together with this limitation, Si layer for solar cells have some disadvantages, mainly depending on its indirect optical transition characteristics. Therefore, most of the works have been directed to the alternative technology where thin film structures have gained popularity with their low cost, flexible and tunable characteristics. Among these structures, there are more efficient commercial photovoltaic applications using CdTe films, and they attract interests on technological development on photovoltaic area with their high absorption coefficient and ideal band gap energy. Although it is one of the conventional materials used as an absorber layer in photovoltaic applications, it is a member of the largest segment of commercial thin film module. Among the current technological research on alternative thin film materials, it has been at the leading position due to its long-term performance, stability, low-cost fabrication. On the other hand, CdTe-based devices are very close to the theoretical short-circuit current limit on cell performance. It is due to the band gap characteristics of this material, and therefore the current works have been directed to the strategies to shift the band gap of CdTe. In this work, this absorber layer has been engineered by substitution of Te by Se. CdSexTe1-x (CST) has been deposited as a ternary II-IV compound of equivalent binaries CdTe and CdSe under the aim of tailoring band gap that will increase photocurrent in device applications. In the fabrication step, CdSe and CdTe sources have been evaporated as sequential stacked layers. This technique is expected to be more controllable and easy way for compositional change. Instead of fabrication of evaporation sources special to each film with having different Se and Te elemental ratios (where x is between 0 and 1), it is a simple way to change only thickness and/or order of the stacked layers during the deposition. Based on the structural characteristics of the deposited films, differentiation of composition ratio of the films has been evaluated for the optical and electrical properties. The results have been discussed for the films at the limits (x=0 and x=1) and also the dependence to the x-value have been evaluated with different intermediate components. Acknowledgement: This work was partially supported by the Scientific and Technological Research Council of Turkey (TUBITAK) Project No: 118F317

Resume : Photovoltaic tandem devices will offer higher power densities per weight and per area, broadening the spectrum of PV applicability - provided the technology is robust, reliable and elegant. 2-terminal tandem photovoltaic modules promise easy and cost-effective application in the field as they can simply substitute current modules in existing photovoltaic (PV) system environments. The use of Cu(In,Ga)Se2 (CIGS) as bottom device absorber in a perovskite/chalcopyrite tandem has long been deemed difficult due to its comparatively high roughness. However, using a combination of PTAA and NiOx first and then a self-assembled monolayer (SAM) as charge selective contact components, PCEs of 21.6% and 23.26% respectively (the latter is certified by ISE, Freiburg) are achieved on as-deposited CIGS thin film bottom cells. The absorber of the top cell is a solution processed perovskite composed of a mixture of different cations (methylammonium, fomamidinium and Cs) and halides (I-, Br-). So far, the CIGS absorber of the bottom cell is thermally co-evaporated. Technology transfer to sequentially deposited CIGS absorbers is under way. This development in connection with issues relevant for the development of efficient CIGS based bottom devices - such as current matching, reduction of the Voc deficit, thermal stability and also surface roughness - will be discussed.

Resume : Recently, tandem structure of Si and thin film photovoltaics has attracted much interest as cost-competitive approach to overcome beyond the performance of Si solar cell. A wide-bandgap CuGaSe2 (CGSe) is a good candidate for its top cell in view of balanced photocurrent matching between top and bottom cell, its high stability, manufacturability, etc. However, so far the monolithic tandem device does not have the proper recombination layer between CGSe and Si solar cells, and the CGSe solar cell has considerable photovoltaic conversion loss due to open-circuit voltage loss. In this study, we suggest the successful monolithic tandem structure in which thin ITO layer is used as recombination layer and the ITO/CGS interface is manipulated to be a better ohmic contact. In addition, ALD ZnSnO buffer is applied to CGSe solar cell to reduce the junction loss of CGS/buffer. Our approach results in high efficiency ITO/CGSe single solar cell beyond 12% and Si/ITO/CGSe tandem solar cell beyond 15%. We will discuss the process and physics related to the interface engineering of ITO/CGSe and CGSe/ZnSnO.

Resume : To overcome 30% conversion efficiency, tandem solar cells based on crystalline silicon (c-Si) sub-cells are one of the most promising architectures regarding the theoretical predictions. A monolithic 2-terminal approach using wide bandgap thin films as top cell does not require significant modifications of the existing solar modules. CIGSe absorber layer is a promising candidate thanks to its high efficiency and to its tunable bandgap energy. Indium-free CuGaSe2 can be used as top cell absorber in a multi-junction device with c-Si regarding the good bandgap value (1.7eV). We report on the developments related to the elaboration of evaporated CuGaSe2 on c-Si tandem solar cells. A study on the surface morphology needed for the CuGaSe2 deposition shows that layers with good crystallographic properties could be obtained for the direct deposition on c-Si wafers and shows the interest of the chemically-polished surface for tandem integration. Two different approaches for the interface layer between both cells were investigated. The first one uses thin ITO films as recombination layers. The second is an original approach based on tunnel junction located in poly-crystalline Si films. The first functional CuGaSe2/c-Si tandem devices were recently elaborated, with for both approaches, Voc exceeding 1V, demonstrating the sub-cell voltages additivity. A study on silicon sub-cell contamination by SIMS and ICPMS will be also presented.

Resume : Despite years of research on multijunction (MJ) solar devices, single junction solar cells still dominate the photovoltaic (PV) market. On one hand, the cost of highly efficient III-V MJ devices is still too prohibitive to be economically sound, and on the other the efficiency of cheaper alternatives, such as a-Si or perovskite/c-Si MJ devices, barely exceed that of their single junction counterparts. The chalcopyrite class falls further behind, with MJ devices exhibiting efficiencies in the range of 2.7-8.0%. However, this low performance range is surprising given that chalcopyrites themselves possess highly desirable characteristics for MJ devices, the most salient of them being bandgap tunability. We believe that one reason for the poor performance of chalcopyrite MJ devices is in part due to research efforts focusing predominantly on only one wide bandgap candidate (1.6 eV CuGaSe2). Therefore in this communication, we report on our recent efforts to develop alternative wide bandgap chalcopyrites. We first present theoretical modeling of selected materials candidates, including Cu(In,Al)Se2, Cu(In,B)Se2 and Cu-poor ordered vacancy compounds (e.g., CuGa3Se5) and report on their bulk energetics and conduction band offsets with common n-type buffer materials (CdS, In2S3). Then, we present some of the strategies developed by our group to fabricate these candidates (e.g., co-evaporation, molecular ink approach) and report on their photoconversion efficiencies. As an example, we show that pairing co-evaporated 1.8 eV CuGa3Se5 with Mg(Zn,O) buffer can lead to single junction cells with open circuit voltage exceeding 925 mV. Another major reason for low performance of chalcopyrite MJ devices is due to the reliance on conventional MJ device approaches such as mechanical and monolithic stacking. To address this, we present a new strategy to integrate wide bandgap chalcopyrites into high efficiency MJ devices. In this approach, we process each individual cell independently, then exfoliate and bond the top cell onto the bottom cell using a transparent conductive binder (TCB). This way, the bottom cell is not subjected to unnecessary thermal budgets, which has been identified as the main culprit for poor efficiency in monolithic chalcopyrite-based MJ devices. Being an essential component of the semi-monolithic device, we present our development of a TCB that exhibits ~90% transmittance from 300-2000 nm as well as an out-of-plane series resistance on the order of 0.1 Ω-cm2 (comparable to that of molybdenum in Cu(In,Ga)Se2 solar cells). We then present various exfoliation methods that have been investigated by our team to lift off wide-bandgap chalcopyrite devices and the challenges and merits associated with each. We will conclude the presentation with data measured on exfoliated chalcopyrite single junction and fully integrated semi-monolithic MJ devices. We show for instance that over 80% of the baseline short-circuit photocurrent density can be preserved after the exfoliation process.

Resume : The photovoltaic market is currently dominated by monocrystalline and multicrystalline Si (mc-Si) solar modules. Recently, improved performance of mc-Si solar cells has been reported, with conversion efficiencies changing from below 21 to above 22% within only a few years. Since for the new record devices, a novel wafer fabrication process based on directional solidification of the Si ingots was used, it appears that the improvements can be attributed to the correspondingly modified microstructure of the mc-Si wafers. This situation raises the question how the microstructure of a photovoltaic material is impacts the solar-cell performance. For the present study, we provide a comprehensive comparison of how the microstructure affects the photovoltaic performance in mc-Si wafers and Cu(In,Ga)Se2 (CIGS) solar cells. This comparison is interesting since solar cells based on both material systems have currently reached conversion efficiencies of 22-23% at band-gap energies of about 1.1 eV. While the collection in mc-Si solar cells has already been optimized substantially with respect to the radiative limit, there is still room for improvement in CIGS solar cells. On the other hand, the open-circuit voltages (Voc) and fill factors (FF) for the best solar cells of both technologies exhibit considerable differences to the maximum values at the radiative limit. These Voc and FF losses can be attributed to nonradiative recombination, we may occur also at line and planar defects in the polycrystalline materials. We raise the question how the perspectives of the mc-Si and CIGS solar-cell technologies, in terms of further improvements of their performance parameters, are limited by their microstructure-property relationships. Specifically, we show that in high-efficiency devices, the recombination velocities at grain boundaries in both materials systems can be rather low (about 102 cm/s). The main difference between the microstructural properties of mc-Si and CIGS can be attributed to the recombination activities of (partial) dislocations. These line defects limit the device performance of mc-Si substantially via enhanced nonradiative recombination, mainly owing to precipitation of transition metals, exhibiting the Achilles’ heel of this photovoltaic technology. In contrast, dislocations in CIGS do not represent centers of enhanced nonradiative recombination. As found by correlative analyses employing electron channeling-contrast imaging and cathodoluminescence (CL) measurements performed on the identical positions on epitaxially grown CIGS thin films, the luminescence signals do not change at the positions of dislocations in CIGS thin films, while at grain boundaries in polycrystalline CIGS layers, the changes in the CL signal were substantial. We provide possible reasons for the benign behavior of (partial) dislocations in CIGS thin films and propose modifications of the CIGS thin-film production in order to reach conversion efficiencies of 25% and beyond.

Resume : Cu(In,Ga)Se2 solar cells yield one of the highest efficiencies among all thin film photovoltaics. Compositional variations of the absorber elements as well as incorporated alkali elements significantly affect the conversion efficiency. Hence, there is a strong need to determine the composition and functionality spatially resolved. High resolution synchrotron based methods like X-ray fluorescence analysis (XRF) and X-ray beam induced current (XBIC) enable insight into such compositional and functional properties. Simultaneous XRF and XBIC measurements of complete solar cells were conducted in plan-view geometry: The highly focused X-ray nanobeam at the ID16B-NA station of the European Synchrotron Radiation Facility scanned the solar cell and spatially resolved maps were obtained by analyzing the emitted X-ray fluorescence radiation together with the corresponding induced current. As the spatial resolution is about 50nm, we can show how different elemental compositions, grains and grain boundaries influence the measured current, which is indicative for the functionality of the solar cell.

Resume : In this work we study the impact of sodium on persistent photoconductivity effect (PPC). After being subjected to illumination, conductivity of Cu(In,Ga)Se2 thin films increases, and the effect can persist for up to the order of hours. The increase of conductivity was linked through capacitance profiling to an increased hole density in the metastable state and is known to enhance efficiency of CIGS solar cells. Incorporation of sodium was likewise linked to the increased efficiency through the increased doping density (hence the higher VOC) and passivation of defect states. While there are some studies correlating the presence of sodium in the absorber and the magnitude of metastable electronic effects, the research so far is only qualitative and limited to comparison of samples grown on sodium-free substrates and on soda-lime glass. Here, we present a systematic study of the magnitude of persistent photoconductivity in samples made with varied and controlled sodium concentration. We observe a close relationship between sodium concentration and the magnitude of PPC, which is in opposition to the Lany-Zunger model linking PPC to a change of charge state of (VSe-VCu) amphoteric vacancy complexes. Our results indicate that the defect (defect complex) responsible for the PPC is directly related to the presence of sodium.

Resume : Sb2Se3 is a promising material candidate for thin film solar cells, which received increased research attention in the last years. The device structure is adapted from CdTe (superstrate) or from Cu(In,Ga)Se2 (substrate). In the substrate configuration generally a CdS buffer layer is employed, which however leads to elemental interdiffusion at the Sb2Se3 absorber layer interface. Using an ALD deposited TiO2 buffer layer between Sb2Se3 and CdS the efficiency could be enhanced from approx. 4.5 % to the current world record of 9.2 % [1]. In this contribution, we grow Sb2Se3 based solar cells in substrate configuration. We investigate the recombination properties of TiO2 and Sb2S3 layers deposited by ALD on the Sb2Se3 front surface. Subsequently, a CdS layer is deposited by chemical bath deposition. The reference stack does not include an ALD layer. All absorbers originate from the same process run and a record efficiency of 4.6 % is achieved using the Sb2S3 intermediate ALD layer. Measurements of the photoluminescence yield demonstrate a reduced recombination rate upon the ALD layers. Using temperature dependent current-voltage measurements on full solar cell devices, we show that the dominant recombination pathway shifts to higher activation energies with the use of the ALD layers. This indicates that the devices without ALD layers are limited by interface recombination and that the ALD layers (at least partially) passivate the interfaces. 1. Li, Z., X. Liang, G. Li, H. Liu, H. Zhang, J. Guo, J. Chen, K. Shen, X. San, W. Yu, R.E.I. Schropp, and Y. Mai, 9.2%-efficient core-shell structured antimony selenide nanorod array solar cells. Nature Communications, 2019. 10(1).

Resume : Thin-film solar cells with polycrystalline chalcopyrite absorbers have shown conversion efficiencies of up to 23.4% [1]. Grain orientations in these materials are known to be crucial in determining the interfacial wetting and buffer layer growth at the surface [2]. As a single grain can facet in different crystallographic directions at the surface, it is likely that also the electronic properties of the chalcopyrite surfaces are linked to the surface faceting. In the present work, we use a combined analytical approach including scanning electron microscopy (SEM), current imaging tunneling spectroscopy (CITS) and electron back scatter diffraction (EBSD) to correlate the surface electronic properties of CuInSe2 with the morphology and crystallographic orientations of surface facets. SEM images of chalcopyrite absorbers indicate that the surface is dominated by two main morphological features: micro-faceted, long, basalt-like (BL) columns and their short nano-faceted (NF) terminations. BL columns exhibit spectral signatures in the CITS data that are typical for the presence of pronounced surface dipoles in conjunction with an increased density of electronic defect levels. In contrast, CITS maps show that the NF terminations are largely passivated in terms of electronic defect levels within the band gap region. Corresponding EBSD maps from the same areas as the CITS maps show that the surface of BL columns form polar crystal facets, while the passivated NF terminations exhibit non-polar {110}/{102} planes. Our results emphasize the correlation between morphology and surface facet orientations and help in gaining fundamental understanding of the interfacial properties in chalcopyrite solar cells. [1] Solar Frontier GmbH, Press Release, January 17th, 2019: Solar Frontier Achieves World Record Thin-Film Solar Cell Efficiency of 23.35% [2] W. Witte, D. Abou-Ras, and D. Hariskos, Appl. Phys. Lett. 102, (2013).

Resume : In the solar industry, Cu(In,Ga)Se2 (CIGS) has to guarantee a standard lifetime of minimum 20 years and high efficiency. Ultra-thin CIGS solar cells were produced for accelerated lifetime testing to study the evolution of the performance and the impact of alkali atoms on their expected lifetime in a damp heat environment (85ºC and 85% room humidity for up to 1000h). As they aged, we measured the solar cells at regular intervals using a novel frequency dependant capacitance-voltage (CVf) mapping. In addition, the composition of select cells was studied using atom probe tomography (APT). The combination of both methods allows us to make a link between the evolution of the absorber layer and the results of the electrical characterization. This contribution lays out how both techniques were combined to identify some characteristics of the working and ageing of the solar cell material. By using CVf mapping, we were able to observe a response that can be identified as an interface defect. This response is also accompanied by a tail-like feature that trails into the bulk of the absorber material. We were able to show that this feature disappears after KF treatment and reappears as the solar cell ages. APT results suggest that the disappearance of the tail is due to the passivating property of K in the grain boundary after KF treatment. The reappearance of the tail, in turn, can be related to the seeping of water into the grain boundary.

Resume : Solar cells based on thin film Cu(In,Ga)Se2 (CIGS) present efficiency of up to 23.4%1 and is highest under thin film category. Even with its small grain size of 2 μm, its maximum efficiency is achieved in its polycrystalline form, the reasons of which are debated over the years. Previous studies have shown an important role of grain boundaries (GBs) 2–4 towards its output efficiency. Here we have used electron beam induced current (EBIC) technique and have found that most of the GBs are indeed beneficial/benign for the device, however a small percentage (15%) of GBs are found detrimental. Hence, in this work we have isolated GBs with different traits and have performed correlative EBIC-EBSD and atom probe tomography (APT) on them to identify its type, structure and composition. Results show that detrimental GBs are mainly enriched in Cu and O and depleted in Ga. Whereas characteristics of a beneficial/benign GB are Cu depletion, no O and no change in Ga. With sufficient statistics, we have obtained characteristics of ‘detrimental’ and ‘beneficial/benign’ GBs5 and found that these characteristics are governed by the composition of the GB. This information will be helpful towards better understanding of thin film solar cells like CIGS. Also, this helps in future developments of CIGS solar cells which will assist in optimization of deposition parameters and post deposition treatments. References: 1 Solar Frontier, Solar Frontier Achieves World Record Thin-Film Solar Cell Efficiency of 23.35% http://www.solar-frontier.com/eng/news/2019/0117_press.html, January 2019 2 C. Persson and A. Zunger, Phys. Rev. Lett. 91, 266401 (2003). 3 O. Cojocaru-Mirédin, T. Schwarz, and D. Abou-Ras, Scripta Materialia (2017). 4 M. Raghuwanshi, E. Cadel, P. Pareige, S. Duguay, F. Couzinie-Devy, L. Arzel, and N. Barreau, Applied Physics Letters 105, 13902 (2014). 5 M. Raghuwanshi, B. Thöner, P. Soni, M. Wuttig, R. Wuerz, and O. Cojocaru-Mirédin, ACS Appl. Mater. Interfaces 10, 14759 (2018).