The electronic and chemical structure of Cu(In,Ga)(S,Se)(2) (CIGSSe) thin film surfaces and of relevant interfaces in CIGSSe-based thin film solar cells is investigated with a combination of X-ray emission spectroscopy (XES) and photoelectron spectroscopy. Examples of sulfur L-2,L-3 XES spectra of CdS and CIGSSe are discussed in view of resonant excitation, surface oxidation, and chemical bonding. The combination of the two techniques proves to be a powerful tool to identify spectral features correlated to certain chemical states or bonds. By monitoring these features in interface formation sequences, chemical and electronic information about buried interfaces can be obtained, which will be discussed in detail for the ZnO/CIGSSe interface. The experimental results provide valuable information on the CIGSSe surface and the ZnO/CIGSSe interface and, in general, demonstrate some of the spectroscopic advantages of X-ray emission spectroscopy.
Cu(In,Ga)Se2 based solar cells with a ZnSe buffer layer deposited by metal-organic vapor deposition are compared to reference cells with a usual CdS buffer. Current–voltage characteristics were measured at different intensities of illumination (up to 100 mW/cm2) in the temperature range between 80 and 300 K. We observe a current roll-over for cells with ZnSe buffer at much higher temperatures than for cells with CdS buffer, which seems to originate from defects either in the buffer volume, or at the buffer/absorber interface. Also wavelength-dependent I(V) measurements were performed before and after annealing at 380 K. A distortion of the I(V) curve measured under red illumination occurs for most of the cells with CdS buffer, but cannot be observed in cells with ZnSe buffer. This effect can be attributed to deep levels in the CdS volume.
Zn-compounds Zn(X,OH) (X=S,Se) buffer layers have been deposited by chemical bath (CBD) process on Cu(In,Ga)(S,Se)2 (CIGSS) with the aim of developing Cd-free CIGSS-based devices. The films are produced in alkaline aqueous solution containing ZnSO4, ammonia NH3 and XC(NH2)2. Optimum deposition conditions were established. The temperature (Tsub) of the chemical bath is found to be critical for the device quality. The thickness and good surface coverage were controlled by XPS-UPS photoemission spectroscopy. SEM study showed that the growth of ZnSe nuclei on CIGSS proceeds in lateral direction. Once the surface is covered the growth takes place in vertical direction . The ZnSe clusters grow in size and their elongated shapes cover the CIGSS surface. High efficiency of over 13% was obtained for both CIGSS/Zn(S,OH) and CIGSS/Zn(Se,OH)-based solar cells. Solar cells with CIGSS/Zn(Se,OH)x/ZnO/MgF2 structure show an active area efficiency up to 15.7%. Using Zn(Se,OH) buffer layer, efficiency of 11.7% was achieved with a 20 cm2 aperture-area monolithic minimodule.
By combining ultraviolet and x-ray photoelectron spectroscopy with inverse photoemission spectroscopy, we find that the conduction-band alignment at the CdS/CuInSe2 thin-film solar-cell heterojunction is flat (0.0±0.2 eV). Furthermore, we observe a valence-band offset of 0.8±0.2 eV. The electronic level alignment is dominated by (1) an unusually large surface band gap of the CuInSe2 thin film (1.4 eV), (2) by a reduced surface band gap of the CdS overlayer (2.2 eV) due to intermixing effects, and (3) by a general influence of the intermixing on the chemical state near the interface.
The guideline for our CIGS-thin film process development is the scalability to large areas for low cost and high throughput module fabrication. Our technology consequently applies established large area sputter coating processes for the Cu-, In-, Ga, Se-precursors as well as for the back- and front-electrode. The characteristics of our absorber formation are an advanced two-step stacked elemental layer process including Cu(Ga)–In–Se precursor deposition, rapid thermal processing (RTP) to CIGS in a sulfur-containing ambient and a controlled sodium doping technique. Within our laboratory 12-cell mini-module baseline (substrate size: 10 cm×10 cm) peak and average conversion efficiencies of 14.7% and 13.2%± 1.5%, respectively, have been achieved. By varying the sulfur content from run to run in the gas atmosphere of the absorber formation process the total S/(Se+S) ratio in the obtained CIGS films has been changed between 0% and 18%. Although TEM-EDX- and SIMS-analyses on the CIGS-absorbers reveal an increased concentration of sulfur (and also gallium) towards the molybdenum back electrode, open circuit voltage and minority carrier lifetime monotonously increase with the average S/(Se+S) ratio determined by XRF. Our scaling up efforts towards pilot module processing necessitate a transfer of our laboratory RTP-technology to a fast heating system for large area substrates (60 cm×90 cm) at a high throughput. The essential task of heating the single-side-coated glass panel homogeneously is successfully demonstrated in a prototype heating chamber proving that a temperature deviation of ±10°C is not exceeded even at high heating rates.
We demonstrate how a combination of photoelectron spectroscopy and x-ray emission spectroscopy can be utilized to derive semi-quantitative information about the localization of impurities at buried interfaces. In the case of the CdS/Cu(In,Ga)Se2 (CIGS) thin-film solar cell heterojunction, segregated Na, which stems from the soda-lime glass substrate or is deliberately added, plays an important role. We find that almost all Na atoms are located at the external CIGS surface or at the CdS/CIGS interface, and that the Na concentration in the bulk of the CIGS film is <1 ppm. Moreover, we show that the Na surface coverage at internal CIGS surfaces is significantly lower than at the external CIGS surface or CdS/CIGS interface, which demonstrates that the internal surfaces may not be regarded merely as a special case of the external surface. Copyright © 2000 John Wiley & Sons, Ltd.
The localization of Na impurities at the buried heterojunction of CdS/Cu(In,Ga)Se2 thin film solar cells has been studied by photoelectron spectroscopy and X-ray emission spectroscopy. This combination of a surface- and a bulk-sensitive technique allows to identify the localization of impurities at a buried interface in a non-destructive, semi-quantitative, and element-specific way. We compare samples with increasing CdS-overlayer thickness on (a) a CIGS film with nominal Na content and (b) a Na-rich CIGS film. The data clearly indicate a self-limitation of the Na content at this interface. The consequences are discussed in view of the possibility to tailor the electronic structure of the buried heterojunction by controlling the nominal Na content in the CIGS film.
Scanning tunneling microscopy (STM) under laser illumination and microspot X-ray photoelectron spectroscopy (micro-ESCA) have been utilized to study the local variation of the photovoltaic properties and of the stoichiometry of Cu(In,Ga)Se2 (CIGS) thin film solar cell absorbers. The STM results clearly demonstrate that the photovoltaic quantities like the local surface photovoltage (SPV) and photoinduced tunneling current (PITC) vary significantly between different grains. This observation is in accordance with the local variation of the chemical composition of CIGS absorber films derived by micro-ESCA. In addition, the (local) photoelectric parameters vary with exposure time to red light, in agreement with recently reported metastability effects.
A combination of x-ray emission spectroscopy and x-ray photoelectron spectroscopy using high brightness synchrotron radiation has been employed to investigate the electronic and chemical structure of the buried CdS/Cu(In, Ga)Se2 interface, which is the active interface in highly efficient thin film solar cells. In contrast to the conventional model of an abrupt interface, intermixing processes involving the elements S, Se, and In have been identified. The results shed light on the electronic structure and interface formation processes of semiconductor heterojunctions and demonstrate a powerful tool for investigating buried interfaces in general.
We investigate Cu(In,Ga)Se-2-based solar cells with a new ZnSe buffer layer deposited by metal-organic vapour deposition and compare their electronic properties to reference cells using a standard CdS buffer layer. The best solar cell with a ZnSe buffer layer achieves an efficiency of 11.6%. We further investigate a large series of solar cells with varied thickness of both types of buffer layers by means of quantum efficiency measurements in equilibrium and under light and voltage bias, The characterization of the devices concentrates on the analysis of the collection of photogenerated holes from the buffer layer. We introduce a nero method to determine the recombination probability of holes at the buffer/absorber ber interface, We find a similar interface recombination probability of about 40% for both devices, those with a ZnSe buffer layer and those with a CdS buffer layer. An anomalous enhancement of the quantum efficiency measured under current bias is ascribed to a barrier modulation effect which is caused by light absorbed in the buffer layer. Copyright (C) 1999 John Wiley & Sons, Ltd.
We demonstrate a general approach to identify and locate minority species at buried interfaces which are of fundamental interest in many fields of solid state research. The approach combines soft x-ray emission for bulk and photoelectron spectroscopy for surface sensitivity. In the present study, the interface between a thin CdS layer and a Cu(In, Ga)Se2 thin film solar cell absorber has been investigated, showing that Na impurities are localized at the buried CdS/Cu(In, Ga)Se2 heterojunction.
A characteristic, reversible metastability is observed for Cu(In,Ga)Se2 thin films and ZnO/CdS/Cu(In,Ga)Se2 heterojunctions. Annealing at 80 °C leads to a decrease of the dark conductivity of the thin films by up to a factor of 2 at room temperature and several orders of magnitude when measured at lower temperatures. By exposure to light, the initial state can be re-established. This reenhancement of the dark conductivity can be looked at as persistent photoconductivity. Admittance measurements at Cu(In,Ga)Se2 heterojunction solar cells display a reversible shift of the activation energy of a distinct dielectric loss peak ranging from 70 to about 160 meV upon illumination or annealing at 80 °C, respectively. We propose that both phenomena as well as commonly observed light-soaking effects of Cu(In,Ga)Se2 solar cells have a common origin.
Room-temperature recombination dynamics has been investigated in a large set of different Cu(In,Ga)Se2 absorber films and compared to the electrical device characteristics of the respective solar cell modules. For a given cell preparation process, a characteristic relation between the low-injection minority-carrier lifetime of the absorber layers and the conversion efficiency of the solar cells is observed: Long lifetimes correlate with high open circuit voltages and conversion efficiencies, while no significant influence of the lifetime on the short circuit current is found.
The controlled incorporation of trace amounts of Na as well as of S into the absorber material of Cu(In, Ga)Se2 solar cells leads to an increase of conversion efficiency. We investigate the influence of both ingredients on the electronic loss mechanisms of the cells by means of temperature dependent d.c. current-voltage measurements and admittance spectroscopy. We find that incorporation of Na gives rise to a distinct dielectric loss peak with activation energy of about 75 meV and to an increase of free carrier concentration in the absorber material. The benefit of S seems to originate from a passivation of deep trap states leading to a dramatic change in the electronic transport properties and, finally, to an enhancement of the open circuit voltage.
In order to evaluate their long term stability in the field under various climates CIS based thin film solar cells and modules were subjected to accelerated environmental stress testing. The most critical part within those test cycles involves extended high temperature and humidity exposure (damp heat conditions). Two different phenomena are generally observed, a power loss due to high temperature that is reversible under prolonged illumination and an irreversible degradation related to humidity exposure. Recent improvements in cell and module technology have reduced the extent of those degradations such that a 1000 h damp heat exposure was passed for the first time.
We report on a detailed investigation on relaxations of the open circuit voltage V-OC of ZnO/CdS/Cu(In,Ga)Se-2 solar cells. Considering that persistent photoconductivity can be observed on Cu(In,Ga)Se-2 polycrystalline absorber layers as well as on Cu(In,Ga)Se-2 single crystals, we propose a model that describes the V-OC relaxation as a property of the Cu(In,Ga)Se-2 bulk material and not as a property of a specific interface of the heterojunction.
Surface photovoltage spectroscopy (SPS) has been used for quality control of ZnO/CdS/ Cu(In,Ga)Se2 (CIGS) thin-film solar cells. The results show that SPS makes it possible to detect “hard failures” following CIGS deposition, and both “hard” and “soft” failures following CdS deposition and following ZnO deposition. In addition, a semi-quantitative screening of CdS/CIGS and ZnO/CdS/CIGS samples is possible. Hence, SPS is suggested as a useful tool for in-line monitoring of CIGS-based solar cell production lines. Moreover, SPS is shown to yield important new information regarding CIGS-based solar cells: (a) A deep gap state is found in samples of superior performance. (b) As opposed to the CdSCIGS structure, a marked decrease in the open-circuit voltage upon Na contamination in ZnOCIGS structures is found.
Recumbent chest x-ray with digital luminescence radiography is an imaging method of limited accuracy in respect of diagnosis of pleural effusions. Supplementary diagnostic methods are recommended, as the present results show, especially in such cases where the recumbent chest x-ray does not reveal an effusion or if the volume must be determined accurately. Digital recumbent chest x-ray ranks equal with conventional x-ray in the diagnosis of pleural effusions.
Purpose: The significance of the recumbent chest xray using digital luminescence radiography was to be assessed in respect of diagnosis of pleural effusions. Material and Methods: Three experienced radiologists evaluated 32 digital recumbent chest x-rays of 32 intensive-care patients. The radiologists were asked to estimate the effusion volume and to assess whether typical x-ray signs of pleural effusions were seen. These evaluations were compared with one another and with the simultaneously produced CTs. Results: Diagnostic accuracy of the digital recumbent chest x-ray is of medium quality in respect of diagnosis of pleural effusions (sensitivity: 69%, specificity: 54%, positive predictive value: 81%, negative predictive value: 34%, rate of accuracy: 65%). Diagnostic safety is the same for right-sided or left-sided pleural effusions, and increases with increasing effusion volume. The ratings by the radiologists are statistically not significantly different, but are significantly different from the CT measurements (Wilcoxon test, p <0.05). Correlations of the assessments and the measurements were weakly positive (r = 0.24, r = 0.36, r = 0.47). The pleural effusions were on the average underestimated by the radiologists. The median predictive error was 203 ml. Conclusions: Recumbent chest x-ray with digital luminescence radiography is an imaging method of limited accuracy in respect of diagnosis of pleural effusions. Supplementary diagnostic methods are recommended, as the present results show, especially in such cases where the recumbent chest x-ray does not reveal an effusion or if the volume must be determined accurately. Digital recumbent chest x-ray ranks equal with conventional x-ray in the diagnosis of pleural effusions.
AbstractTargeting large area and low cost processing of highly efficient thin film solar modules an advanced stacked elemental layer process for Cu(InGa)Se2 (CIGS) thin films is presented. Key process steps are i) barrier coating of the soda lime glass substrate combined with the addition of a sodium compound to the elemental Cu/In/Ga/Se-precursor stack and ii) rapid thermal processing (RTP) to form the CIGS compound.By this strategy exact impurity control is achieved and the advantageous influence of sodium on device performance and on CIGS film formation is demonstrated unambiguously by means of electrical characterisation, XRD, SEM, TEM and SIMS. Sodium enriched and sodium free precursor stacks were heated to intermediate states (300°C–500°C) of the RTPreaction process. The experiment clearly reveals that on the reaction pathway to the chalcopyrite semiconductor increased amounts of copper-selenide are formed, if sodium is added to the precursor films. TEM-electron diffraction unambiguously identifies the CuSe-phase which is localised at the surface of the forming CIGS-film. These experimental findings propose a sodium assisted quasi liquid growth model for the CIS formation taking into account that sodium promotes the existence of CuSe at higher temperatures and its effect as a flux agent. The model contributes to a better understanding of the observed superior crystal qualitiy for sodium enriched in contrast to sodium free CIGS films.Application of these experimental findings in the technique of the optimized and controlled sodium incorporation significantly improves process reproducibility, CIGS film homogenity over larger substrate areas and shifts the average efficiency of cells and modules to a significantly higher level. This is demonstrated by a 12-cell integrated series connected minimodule with an aperture area of 51 cm2 and a confirmed efficiency of 11.75 %.