Jutta Beier

Light induced changes in the electrical behavior of CdTe and Cu(In,Ga)Se2 solar cells

Abstract The electrical properties of CdTe and Cu(In x ,Ga 1− x )Se 2 (CIGS) solar cells change depending on their illumination conditions. The dark and light current–voltage (I–V) curves of both solar cells exhibit a distinct cross over. The reason for the cross over of the CIGS solar cells is located in the CdS buffer layer and its interface to the absorber. Light which is absorbed in the CdS buffer layer results in a steeper I–V curve in contrast to the dark curve. The buffer layer of the CdTe solar cell has a similar influence on the I–V curve. In this case, however, the effect described above overlaps with a second feature. Many CdTe solar cells have a Schottky contact as their back contact. This is expressed in the I–V curve as a current limitation at positive voltage bias (roll over). However, under light bias even this diode can be influenced. By means of quantum efficiency (QE) measurements under voltage bias it is possible to analyze both effects. All experimental observations can be explained by traps in CdS and/or CdTe assuming strongly different capture cross sections for holes and electrons. A model describing the influence of the illumination on the I–V and QE curves is presented.

Grain boundaries and impurities in CdTe/CdS solar cells

Near-interface grain effects and quantitative SIMS impurity profiles for close-space sublimation-deposited CdTe/CdS solar cells and test structures are described. TEM of the interfacial CdTe grains demonstrated them to be smaller than expected from extrapolation of the bulk grain size. Computer simulation showed this to be consistent with the development of grain sizes during additive growth processes, the grain size increasing approximately as the root distance from the polycrystalline hetero-interface. TEM of the near-interfacial CdTe before and after CdCl2 processing revealed grain growth from 0.1 to 0.5 μm — even though the grain size distribution in the overlying CdTe is stable to processing. Quantitative SIMS of test structures comprising co-sputtered contacts on close-space sublimation-deposited cells (Sb–Te/CdTe/CdS/TCO/glass) is reported for the structures in their fresh, light-soaked and bake-tested states. High concentrations of impurities are present in the CdTe of the fresh cells as follows: S, 7×1019; Cl, 8×1018; Cu, 7×1017; Na, 3×1017; Sb, 4×1018; and O, 7×1018 cm−3. It is considered that the unintentionally introduced impurities are unlikely to originate from the source CdTe. The effects upon the impurity distributions of light soaking and baking in air at 400°C are reported. Accumulations of impurities near to the CdS layer are discussed.

Key Aspects of CdTe/CdS Solar Cells

Recent developments in the following areas are briefly reviewed: a) the electrical structure of grain boundaries in CdTe absorbers, b) impurities and non-stoichiometry in CdTe solar cells and c) use of Sb 2 Te 3 in contacts to CdTe. Nominally identical solar cells fabricated using 99.999% pure CdTe feedstock from two different suppliers were compared. Differences in the photovoltaic response and absorber grain size were correlated with the purity of the feedstock, the purer material giving the higher V oc , FF and efficiency, and larger grain size. Quantum efficiency and C-V measurements indicated that the performance differences are most likely to result from reduced doping at the back contact surface in the less pure sample. A quantitative SIMS study of Sb-Te contacts to CdTe reveals that annealing in air at 400 °C causes an influx of Sb and O into the absorber layer. Free energy calculations indicate that this is driven by the preferential reaction of O with Sb compared to CdTe oxidation.

Effect of in situ UHV CdCl2-activation on the electronic properties of CdTe thin film solar cells

Abstract To reach reasonable conversion efficiencies of approximately 10% and above with CdTe thin film solar cells an activation step involving chlorine at elevated temperatures seems to be necessary before back contact formation. This activation process has been simulated in an ultrahigh-vacuum (UHV) system. Solar cells with a maximum efficiency of 9.1% have been prepared using this process. In addition the effect of the CdCl 2 activation process on the electronic properties of each solar cell layer, SnO 2 , CdS and CdTe has been investigated in situ using photoelectron spectroscopy. The effects of the activation on the Fermi level position of all investigated layers is presented and discussed.

Importance of electron current in p-type CdTe in CdS/CdTe thin film solar cells at forward bias

In previous work [1,2], we modeled the cross-over of the I-V curves of thin film CdS/CdTe solar cells in terms of an electron (minority carrier) current in the vicinity of the back contact. In this work, we focus on the necessary extension of this analytical model based on a series of measurement results. Especially the wavelength and voltage dependence of the current at forward bias is illustrated in these measurements. The various possible causes for this kind of behavior are discussed and modeled. The extensions to the previous model, needed to describe the voltage and wavelength dependent behavior of I-V curves of real CdTe/CdS solar cells, are proposed.

Modeling the cross-over of the I-V characteristics of thin film CdTe solar cells

The I-V curves of high efficiency CdTe thin film solar cells show a kink in the forward region (roll-over) this fact is well studied and ascribed to a contact barrier e.g. at the back contact of the CdTe layer. Also, light and dark I-V characteristics intersect (cross-over), due to an increase of the contact saturation current under illumination. We observed on many occasions that a fan of several I-V curves all intersect at one single point in the first quadrant (forward current and voltage), when some parameter is varied, like illumination intensity or minimal changes of processing conditions. We ascribe this to the contribution of electron minority current at the CdTe back contact Schottky barrier. We show how this can explain the one-point-intersection. The validity of the assumptions are discussed with theoretical considerations, numerical simulations and experiments.

STUDY OF SPATIALLY RESOLVED IMPURITY DIFFUSION IN CdTe SOLAR CELLS USING VOLTAGE DEPENDENT QUANTUM EFFICIENCY

The performance stability of CdTe/CdS solar cells is strongly determined by diffusion of impurities from the back contact into the absorber layer and hetero-junction. Impurity migration changes the effective carrier concentration and barriers in the device by compensation of donors or acceptors and by creation of defect centres. The CdS window layer is particularly affected by this phenomenon, since the impurities tend to accumulate there. This can be characterised by measuring the voltage dependent, the so called apparent quantum efficiency (AQE) in the blue wavelength region, while the back contact can be analysed by the AQE in the IR. CdTe/CdS cells with different back contact materials have been stressed in different conditions and ambiences. When thermally stressed in presence of oxygen, enhanced AQEs were observed for cells containing Cu, while cells containing Sb showed negligible changes, in the UV range as well as in the IR range. In comparison, vacuum-stressed Cu containing cells showed lower AQEs, but still higher than non-stressed cells. Results of the stressing tests for different materials and in different conditions have been analysed and interpreted using the recently developed model of a modulated barrier in the CdS bulk.

A comparative EBIC study of CdTe solar cell activation using CdCl2 and Cl2

Abstract The effect of a new Cl2 gas-based activation for CdTe/CdS thin film solar cells has been examined. Devices deposited by close space sublimation were treated using this method and compared with otherwise identical control devices activated using a conventional CdCl2 treatment. Depth-dependent carrier collection functions were obtained by measuring the current induced under electron beams of varying energy; this data was deconvoluted using carrier generation functions modelled using Monte Carlo simulations. The resultant measurements show that the performance of the chlorine treated device falls within the range of values measured for CdCl2 treated cells.