Present address: Arizona State University, School of Electrical, Computer and Energy Engineering, 551 E. Tyler Mall, Tempe, AZ 85287, USA.Reducing wafer thickness while increasing power conversion efficiency is the most effective way to reduce cost per Watt of a silicon photovoltaic module. Within the European project 20 percent efficiency on less than 100-mm-thick, industrially feasible crystalline silicon solar cells ("20plms"), we study the whole process chain for thin wafers, from wafering to module integration and life-cycle analysis. We investigate three different solar cell fabrication routes, categorized according to the temperature of the junction formation process and the wafer doping type: p-type silicon high temperature, n-type silicon high temperature and n-type silicon low temperature. For each route, an efficiency of 19.5% or greater is achieved on wafers less than 100 mm thick, with a maximum efficiency of 21.1% on an 80-mm-thick wafer. The n-type high temperature route is then transferred to a pilot production line, and a median solar cell efficiency of 20.0% is demonstrated on 100-mm-thick wafers.

Lock-in thermography (LIT) is a well-established tool for defect analysis of solar cells, but so far has been restricted to the measurement of metallized samples. The new light-modulated lock-in thermography (LimoLIT) described in this paper overcomes this restriction by generating the voltage modulation needed for detection from photovoltaic conversion of modulated light. Thus wafers can be measured during all stages of fabrication, a pn-junction provided. The contactless LimoLIT method shows a stronger measurement signal and invokes a current flow close to illuminated operating conditions of solar cells, whereas conventional LIT is only comparable to a dark I-V measurement.

The hydrogenation of crystalline Si by methods used to passivate defects in Si solar cells has been studied by infrared spectroscopy. For these experiments, floating-zone Si that contained Pt impurities that act as traps for H was used as a model system in which H could be directly detected. In this model system, the concentration and indiffusion depth of H were determined for different hydrogenation treatments so that their effectiveness could be compared. The postdeposition annealing of a hydrogen-rich SiN x surface layer was found to introduce H into the Si bulk with a concentration of ϳ10 15 cm −3 under the best conditions investigated here.

Defect engineering of the oxygen-vacancy clusters formation in electron irradiated silicon by isovalent doping: An infrared perspective J. Appl. Phys. 112, 123517 (2012) Band bending and determination of band offsets in amorphous/crystalline silicon heterostructures from planar conductance measurements J. Appl. Phys. 112, 123717 (2012) A transition of three to two dimensional Si growth on Ge (100) substrate J. Appl. Phys. 112, 126101 (2012) Fabrication of large-grained thin polycrystalline silicon films on foreign substrates by titanium-assisted metalinduced layer exchange J. Appl. Phys. 112, 123509 (2012) Additional information on J. Appl. Phys. Micro-Raman spectroscopy has been used to investigate the acceptor distribution in highly p-doped silicon. As an example, the dopant distribution in crystalline thin-film layers, as developed for solar cells, was mapped. The method is based on the analysis of the Fano-type Raman peak shape which is caused by free charge carriers. For calibration of the Raman acceptor measurements (excitation at a wavelength of 532 nm), we used mono-crystalline reference samples whose acceptor concentration was determined by electrochemical capacitance voltage. We find a significant influence of light induced free charge carriers on the peak shape which results from typical Raman excitation. Thus, the selection of a suitable intensity is important to avoid a too low signal-to-noise ratio on the one hand and systematic errors due to light induced carriers on the other hand. Different evaluation methods, i.e., peak asymmetry versus peak width analysis, are compared in respect to interference caused by random noise of the spectra or else by internal stress in the sample. While the width analysis method is more robust to a low signal-to-noise ratio, the symmetry analysis is more reliable in case of high intrinsic stress. V C 2013 American Institute of Physics. [http://dx

We present an advanced analytical model which applies to light beam induced current contrast profiles to determine reliably the effective surface recombination velocities ͑S eff ͒ of grain boundaries ͑GBs͒ and diffusion lengths ͑L diff ͒ in the grains, in cases where a GB is close to the studied one or when L diff of the neighboring grain differs. We introduce additionally a new method for a very accurate determination of the plateau value of the investigated linescan and make use of simultaneously fitting GB profiles measured at various laser wavelengths both aiming at increasing the accuracy of the L diff determination. Through several special case investigations, the various applications and limitations of the model are demonstrated. We discuss the influence of the electrical parameters of the semiconductor on the various zones of the profile as well as the influence of measurement technique parameters on the experimental profile and point out the need of an accurately determined small laser beam radius to ensure a reliable extraction of S eff . We discuss the occurring discrepancy between fit and measured data and show that it gives hints about particular material features and the reliability of the extracted parameters. We finally point out the possibility of determining L diff in small grains. This model thus allows more realistic GB situations to be investigated.