Search

In this paper, first generation back-contact back-junction (BC-BJ) silicon solar cells with cell efficiencies well above eta = 20% were fabricated. The process sequence is industrially feasible, requires only one high-temperature step (co-diffusion), and relies only on industrially available pattering technologies. The silicon-doping is performed from pre-patterned solid diffusion sources, which allow for the precise adjustment of phosphorus-and boron-doping levels. Based on the investigated process technologies, BC-BJ solar cells with gap and without gap between adjacent n+-and p+-doped areas were processed. On the one hand, a strong reduction of the process effort is possible by omitting the gap regions. On the other hand, parasitic tunneling currents through the narrow space charge region may occur. Hence, deep doped areas were realized to avoid tunneling currents in gap-free BC-BJ cells. This paper finishes with a detailed characterization of the manufactured cells including important cell measurements like I-V, SunsVOC, quantum efficiency, and an analysis of the cell specific fill factor losses.

In this study, we investigate the influence of different diffusion temperature profiles on the interstitial iron concentration before and after contact firing. It is shown that the positive influence of a low-temperature anneal after phosphorus diffusion is strongly reduced after contact firing as the formed precipitates dissolve again. On the other hand, a dissolution peak before the diffusion aims on a decrease in precipitate density and, therefore, leads to a reduced interstitial iron concentration especially after firing. The physical mechanisms during diffusion and firing are investigated with Sentaurus Process simulations. Furthermore, lifetime improvements in standard multicrystalline material could be achieved applying the dissolution peak before diffusion.

We simulate the precipitation of iron during the multicrystalline ingot crystallization process and the redistribution of iron during subsequent phosphorus diffusion gettering with a 2-D model. We compare the simulated size distribution of the precipitates with the X-ray fluorescence microscopy measurements of iron precipitates along a grain boundary. We find that the simulated and measured densities of precipitates larger than the experimental detection limit are in good agreement after the crystallization process. Additionally, we demonstrate that the measured decrease of the line density and the increase of the mean size of the iron precipitates after phosphorus diffusion gettering can be reproduced with the simulations. The size and spatial distribution of iron precipitates affect the kinetics of iron redistribution during the solar cell process and, ultimately, the recombination activity of the precipitated iron. Variations of the cooling rate after solidification and short temperature peaks before phosphorus diffusion strongly influence the precipitate size distribution. The lowest overall density of iron precipitates after phosphorus diffusion is obtained in the simulations with a temperature peak before phosphorus diffusion, followed by moderate cooling rates.

The present work deals with the attack of different chemicals used in nickel and copper plating electrolytes, on screen printed silver front side contacts, which lowers the adhesion and hinders industrial implementation. Dipping tests with four different self-made pastes in isolated electrolyte compounds show the dissolution of silver for pastes with glass content and acidic exposure conditions in chemical trace analysis (ICP-OES) of the dip solution. These results can be correlated to the macroscopic contact adhesion, measured with quantitative peel force tests. The results show that there is a link between the amount of silver that is found in the dipping solutions after the exposure of the contacts, the measured contact adhesion, and the presence of glass in the contact. The reaction between glass and silver during the firing process forms an oxidation layer between silver and glass which provides good adhesion. This interface seems to be dissolved by acidic plating chemicals. An additional SEM investigation shows the failure mechanism to take place exactly between the glass that covers the wafer after the firing process and the silver bulk, indicating a chemical attack at this interface.

Photoluminescence (PL) imaging is a promising characterization technique for rating and sorting of multicrystalline silicon (mc-Si) as-cut wafers concerning to their material quality. It is inline applicable and yields high resolution images showing recombination active defects from crystallization which influence solar cell performance. In this contribution the basic concepts in ongoing work concerning relevant defects, rating results, statistics, algorithms and general approaches for the rating are summarized. Since 2009 wafers are sorted into five quality classes at Fraunhofer ISE. Details on the rating criteria are given and three examples for application are presented: (i) the impact of edge contaminations on the solar cell results, (ii) a comparison of wafer suppliers based on random samples from a statistical basis of 10,000 wafers and (iii) the results of an advanced rating and sorting of wafers for the manufacturing of highly efficient MWT-PERC solar cells. The results confirm that PL imaging can be used for a very precise rating of material quality.

Carrier mobility in silicon plays a crucial role for photovoltaic applications. While the influence of doping on mobility in standard monocrystalline silicon is well understood, recent research has been focused on the effects of crystal defects in multicrystalline (mc) silicon and of the presence of both acceptors and donors in compensated silicon, both introducing additional scattering centers influencing carrier mobility. In this work measurements of the majority carrier mobility have been carried out in two blocks of compensated multicrystalline silicon. Confirming existing results we come to the conclusion that with increasing compensation level mobilities may be significantly lower than predicted by Klaassen's mobility model, which is basically suited for the description of mobilities in compensated silicon as it accounts for scattering on both acceptors and donors. However, as this model is based on mobility data from uncompensated silicon, a compensation-related reduction of screening is not taken into account sufficiently. To describe mobilities in compensated silicon, a modification of Klaassen's mobility model based on published mobility data in compensated silicon is suggested.

In this paper we measure, spatially resolved, the efficiency potential of multicrystalline material. The excess carrier lifetime in the bulk affects the open-circuit voltage, the short-circuit current and the fill factor of a solar cell differently as different injection conditions apply. Therefore, we measure the bulk lifetime injection-dependent and analyze each IV -parameter separately. Furthermore, multicrystalline wafers with laterally varying bulk recombination need a spatially resolved analysis to explore the impact of large grains or dislocation clusters on the efficiency. The crucial information we need is the spatially resolved injection-dependent bulk lifetime. For this purpose, we use lifetime calibrated photoluminescence imaging with a varying generation rate. Measuring a surface passivated wafer after the important high temperature steps of a solar cell process ensures that we attain the bulk recombination that is relevant for the solar cell. Combining this information with a cell simulation, we deduce the maximum efficiency for every image point. We also discuss how to obtain global values from these images and investigate the effect of charge currents flowing laterally through the emitter using Sentaurus Device and Spice network simulations. A comparison of these simulated global values to measured IV -parameters on finished solar cells shows good agreement between the two. A detailed analysis of two mc-wafers from different feedstock shows limiting IV -parameters in different regions, efficiency reductions due to dislocations and explains differences in the fill factor of two equally processed solar cells. The analysis shows that evaluating the material related efficiency limitation of silicon wafers has to be done at the efficiency level, as the lifetime at one generation level alone is not sufficient.

This paper is revisiting the problem of recombination on locally processed area (contacts, local doping …), based on the concept of point recombination rate (pLPA). The newly introduced effective point recombination rate (peff) can be easily determined from the effective surface recombination velocity. It also allows predictions and comparisons in most of the practical cases. Further analysis allows the separation of the influence of pLPA from the one of the transport of the carrier in a transparent way. Due to its simplicity, transparency and accuracy, the model proposed here is believed to be more suitable to recent local processing technology than the existing analytical models.

To enable a concentration control in the acidic texturing process we have closed the feedback loop from analytical data to the dosing mechanism of the used process tool. In order to analyze the process bath we used near-infrared spectroscopy in an online setup as well as ion chromatography as an inline method in a second approach. Using the developed dosing algorithm allows a concentration optimization of HF and HNO3 in dependence of the Si concentrations. This allows a further optimization of the acidic texturing process and leads to more homogeneous process results.

Microscopic laser-doped regions in advanced solar cell concepts are analyzed to determine the doping density and to identify the damage caused by the laser process. For these investigations, microphotoluminescence spectroscopy and micro-Raman spectroscopy are utilized to measure doping density, internal stress, and carrier lifetime with micrometer resolution. This analysis proves the high applicability of the microspectroscopic techniques for the characterization of laser-doped regions by analyzing the profile of the advanced local doping process and the laser-induced damage particularly at the edges of the highly doped regions.

Well suited and reliable values of the optical and electrical properties of thin indium tin oxide (ITO) films are needed in order to choose the optimal deposition parameters and to perform reliable modeling for solar cells design. In this work, a new method will be presented to evaluate the ITO transparency directly on silicon substrates. The effects of each deposition parameter that influences the ITO transparency and conductivity force a trade-off in the frame of values useful for SHJ solar cells. The deposition of our optimized ITO film on a textured wafer yields a weighted average reflectance as low as 4.4±0.2%. The deposition of an MgF2/ITO double-layer anti-reflection coating (DL-ARC) on textured cells increases the efficiency from 17.9%, measured immediately after contacts have been added to the ITO, to 18.4% after the MgF2 deposition. An annealing step at 200 °C for 10 min proved to further increase the efficiency up to 18.9%, for a total gain of 1%.

Reliable process control or predictions of solar cell efficiencies from minority carrier lifetimes on silicon wafers require precise lifetime measurements. For inhomogeneous material quality this implies the necessity of adequately averaged spatially resolved lifetime measurements. Materials such as, e.g. multicrystalline upgraded metallurgical grade silicon frequently feature relatively low lifetimes, high trap densities, and several material parameters (charge carrier mobilities and net dopant concentration) that are not straightforwardly predictable or measurable. As this may substantially compromise conventional lifetime measurements, we present a solely luminescence based lifetime imaging technique, which requires virtually no a priori information about material parameters. Our approach is based on a calibration of a wafer's photoluminescence image through a precise lifetime determination of a part of this wafer via quasi-steady-state photoluminescence. Carrier mobilities, net dopant concentration, and surface morphology leave the determination of lifetime virtually unaffected, the injection dependence of lifetime is properly taken into account, and lifetimes down to the timescale of a microsecond can be reliably measured.

In back-contact solar cells, both external polarities are located at the back surface of the device, which allows for higher photocurrent generation on cell level and reduced series resistance on module level, leading to higher energy conversion efficiencies compared to conventional solar cells and modules. However, the majority charge carriers, which are generated near the back emitter, have to flow laterally e.g. through the base in order to reach the external majority carrier contact. In the present work, we analyse the lateral series resistance by means of measurement and simulation for high-performance metal wrap through (HIP-MWT) solar cells. We compare theoretical models and experimental methods to extract the effective series resistance from simulated and measured current–voltage characteristics and show that lateral voltage variations significantly increase the local recombination current. If the width of the gap between the external majority carrier contacts is reduced from the typical value of 3.5 mm to ideally 0 mm, we expect an increase of the energy conversion efficiency of approximately 0.1%abs. for cells with three continuous rear emitter contacts on 125 mm×125 mm large silicon wafers. In a simulation study, the bulk doping concentration N A and the bulk lifetime are varied yielding an optimal base resistivity of 0.6 Ω cm–1.5 Ω cm for HIP-MWT solar cells based on Czochralski-grown silicon in the degraded state of the boron–oxygen defect and an optimal resistivity of less than 1.0 Ω cm for the case of bulk lifetimes larger than ~300 µs.

Current losses at surfaces play a crucial role in the optimization of high-efficiency silicon solar cells. We present a new approach to characterize the surface recombination activity of interdigitated-back-contact (IBC) silicon solar cells by comparing experimental and simulated photoluminescence images (PL). The different recombination properties of the p- and n-doped regions of IBC cells in combination with the operating condition lead to contrast profiles in the PL image that vary with bias voltage. We achieve a good matching of experimental and simulated data for the investigated cells enabling the analysis of how sensitive the simulated contrast patterns are to changes in the surface recombination at the emitter, back- and front-surface-field and if a better matching of the experimental PL images is possible. Using these PL images in combination with simulations around the open circuit voltage Voc the determination of surface recombination velocities S and emitter saturation currents J0 on finished cells is made possible for the first time. This approach also opens a new path towards loss analysis of finished PERL, PERC, MWT and other silicon solar cells.

In the context of nickel silicide formation from plated nickel layers for solar cell metallization, there are several open questions regarding contact adhesion and electrical properties. Nanoscale characterization by transmission electron microscopy has been employed to support these investigations. Interfacial oxides and silicide phases were investigated on differently prepared samples by different analytical methods associated with transmission electron microscopy analysis. Processing variations included the pre-treatment of samples before nickel plating, the used plating solution and the thermal budget for the nickel–silicon solid-state reaction. It was shown that interface oxides of only few nm thickness on both silicon and nickel silicide are present on the samples, depending on the chosen process sequence, which have been shown to play an important role in adhesion of nickel on silicide in an earlier publication. From sample pretreatment variations, conclusions about the role of an interfacial oxide in silicide formation and its influence on phase formation were drawn. Such an oxide layer hinders silicide formation except for pinhole sites. This reduces the availability of Ni and causes a silicide with low Ni content to form. Without an interfacial oxide a continuous nickel silicide of greater depth, polycrystalline modification and expected phase according to thermal budget is formed. Information about the nature of silicide growth on typical solar cell surfaces could be obtained from silicide phase and geometric observations, which were supported by FIB tomography. The theory of isotropic NiSi growth and orientation dependent NiSi2 growth was derived. By this, a very well performing low-cost metallization for silicon solar cells has been brought an important step closer to industrial introduction.

The effective minority carrier lifetimes on epitaxial silicon thin-film material have been measured successfully using two independent microwave-detected photoconductivity decay setups. Both measurement setups are found to be equally suited to determine the minority carrier lifetime of crystalline silicon thin-film (cSiTF) material. The different measurement conditions to which the sample under investigation is exposed are critically analyzed by both simulations and measurements on a large number of lifetime samples. No systematic deviation between the lifetime results from different measurement setups could be observed, underlining the accuracy of the determined lifetime value. Subsequently, a method to separate the epitaxial bulk lifetime and the total recombination velocity, consisting of front surface and interface recombination between the epitaxial layer and the substrate, is presented. The method, based on different thicknesses of the epitaxial layer, is applied to all batches of this investigation. Each batch consists of samples with the same material quality but different epitaxial layer thicknesses whereas different batches differ in their material quality. In addition, the same method is also successfully applied on individual cSiTF samples. From the results, it can be concluded that the limiting factor of the effective minority carrier lifetime for the investigated solar-grade cSiTF material is the elevated recombination velocity at the interface between epitaxial layer and the substrate compared with microelectronic-grade material. In contrast, the samples cannot be classified into different material qualities by their epitaxial bulk lifetimes. Even on multicrystalline substrate, solar-grade material can exhibit high epitaxial bulk lifetimes comparable to microelectronic-grade material.

A map of local short-circuit current density (JSC) of a solar cell at standard irradiance spectra is a desirable source of information for current-loss analysis and device optimization. In this work, we present a new and easily implementable method to obtain such JSC maps as well as local external quantum efficiency graphs with a spatial resolution of about 100 µm, exemplified for silicon solar cells. The method is based on a pixel by pixel interpolation of external quantum efficiency maps obtained by spectrally resolved light-beam induced current measurements (SR-LBIC). The local data is integrated with an AM1.5G spectrum to obtain a JSC map with no additional measurement time beyond SR-LBIC image acquisition. The averaged currents over a map are in good agreement with global current–voltage measurements. Some of the possible applications of this method are presented and exemplified, including local quantum efficiency of a cast-mono solar cell, separation of JSC contributions from front side and bulk, quantitative determination of the electrical shading effect in IBC cells, the impact of the irradiance spectrum on local JSC for indoor photovoltaic applications and also the separation of optical and electrical current losses with newly introduced Pseudo-JSC maps based on internal quantum efficiency maps. Finally, the resulting JSC maps also present valuable information for several existing local cell-parameter imaging methods. This basic approach leads to an easily accessible, spatially resolved short-circuit current loss-analysis, which is in principle applicable to any solar cell structure.

An evolutionary approach to increase the cell efficiency of the industrially widely used Al-BSF Si solar cell design is the implementation of a lightly doped emitter and plated Ni-Ag front side contacts. This work shows that this approach can give a significant advantage to fully screen printed solar cells with cell efficiencies up to 19%. However, the adhesion of the plated contacts strongly depends on the tempering induced silicidation of the Ni layer. The application of macroscopic and microscopic characterization techniques indicates that the tempering induces local non-ohmic shunts due to a silicide spiking in the space charge region. The application of an additional selective emitter doping in this experiment could decrease the shunting probability but could not prevent it completely due to inhomogeneous doping depths on the textured surface.

The electrical material quality of multicrystalline (mc) silicon for photovoltaic applications suffers from crystal defects as well as from impurities that originate from the feedstock, the quartz crucible, and its coating. In this study, we investigate the influence of impurities from the crucible on efficiency losses in mc silicon solar cells, focusing on the limitation due to iron. The applicability of p-type mc silicon, crystallized in G1 sized crucibles of industrial material quality and very pure electrically fused silica, for a high-efficiency solar cell process is examined by measuring lifetime and interstitial iron concentration in the wafers after different processing steps and by estimating the cell efficiency potential from injection-dependent bulk lifetime measurements. Interstitial iron concentrations extracted from 2-D simulations of iron precipitation at crystal defects and gettering during processing agree well with Fei measurements at different process stages and explain the observations. Efficiency losses are quantified to losses due to segregated impurities diffused into the silicon melt, losses due to decorated crystal defects and losses due to solid-state diffusion into the crystal. By using a high-purity crucible, losses are reduced significantly and an efficiency gain of 0.5% absolute is estimated to be attainable on wafers with edge region.

The nanostructured surface – also called black silicon (b-Si) – is a promising texture for solar cells because of its extremely low reflectance combined with low surface recombination obtained with atomic layer deposited (ALD) thin films. However, the challenges in keeping the excellent optical properties and passivation in further processing have not been addressed before. Here we study especially the applicability of the ALD passivation on highly boron doped emitters that is present in crystalline silicon solar cells. The results show that the nanostructured boron emitters can be passivated efficiently using ALD Al2O3 reaching emitter saturation current densities as low as 51 fA/cm2. Furthermore, reflectance values less than 0.5% after processing show that the different process steps are not detrimental for the low reflectance of b-Si.

Differential light-biased dynamic measurements of charge carrier recombination properties in semiconductors have long been known to yield only differential rather than actual recombination properties. Therefore, the determination of injection-dependent recombination properties from such measurements was previously found to require integration over the entire injection range. Recent investigations of the phase shift between a time-modulated irradiation of silicon samples and excess carrier density reveal a striking analogy to the above findings: the phase shift is greater than the actual effective carrier lifetime in the case of a positive derivative of lifetime with respect to excess carrier density, and vice versa. This work attempts to rearrange the mentioned previous findings in a quantitative theoretical description of light-biased dynamic measurements of effective carrier lifetime. Both light-biased differential lifetime measurements as well as harmonically time-modulated methods without additional bias light are shown to represent a limiting case in a general treatment of light-biased dynamic lifetime measurements derived here. Finally, we sketch a way to obtain actual recombination properties from differential measurements-referred to as a differential-to-actual (d2a) lifetime analysis, which does not require integration over the entire injection range.

The aim of this paper is to analyze the limiting role of crucible and coating impurities on material quality of multicrystalline silicon. Both solid body diffusion and diffusion into the silicon melt are considered in this study. Two ingots of size G1 have been analyzed. One of them was crystallized in a standard crucible, whereas the other was crystallized in a quartz crucible of very high purity. Focus is put on iron and cobalt as examples of typical impurity species. Iron was found in large concentrations in standard crucibles, and cobalt was proven to be a suitable marker impurity that is mainly found in the coating. Inductively coupled plasma mass spectroscopy data are exploited for the determination of impurity concentrations in crucible, coating, and within the crystal. With higher sensitivity for low concentration, PL imaging is applied for carrier lifetime and interstitial iron concentration measurements. The different findings are compared with modeling results of iron and cobalt in-diffusion by Sentaurus Process. The analysis of silicon wafers before and after gettering steps enable a quantification of impurity-limiting cell efficiency potential. Conclusions about the role of impurities from coated crucibles in large-scale crystallization are deduced.

We present improvements and calibrations for boron and phosphorus doping models, which allow highly predictive simulations of various doping processes used for solar cells. Besides boron doping from a boron tribromide (BBr3) source and phosphorus doping from a phosphorus oxychloride (POCl3) source, oxidation and implantation processes for solar cell applications are simulated. The use of one set of model parameters for different doping techniques improves the calibration quality, ensures the generality of the model and allows a separation of the physical mechanisms influencing the dopant diffusion process. The simulations are compared to measured depth profiles of a large number of processes, which vary in time and temperature. In addition to excellent agreement to measured doping profiles, the sheet resistances and emitter depths of almost hundred diffusion processes are in good consistency with the simulations. Multiple experiments to optimize doping profiles regarding the shape, surface concentration or sheet resistance can thus be avoided.

To date, gallium-doped Czochralski (Cz) silicon has constituted a solar cell bulk material free of light-induced degradation. However, we measure light-induced degradation in gallium-doped Cz silicon in the presence of copper impurities. The measured degradation depends on the copper concentration and the material resistivity. Gallium-doped Cz silicon is found to be less sensitive to copper impurities than borondoped Cz silicon, emphasizing the role of boron in the formation of copper-related light-induced degradation.

The light-biased dynamic analysis of excess carrier decay has been known to yield differential rather than actual recombination properties since 1995. This finding implies a mismatch between actual lifetime and the measured differential lifetime in the case of injection-dependent carrier lifetime. Therefore, a time-consuming integration of differential recombination properties over the entire injection range was deemed inevitable in order to obtain actual recombination properties from differential measurements. We recently observed a striking analogy to the above finding: Time-modulated luminescence measurements - notably without any additional constant bias light - feature a systematic mismatch between actual lifetime and apparent lifetime determined from the phase shift of carrier density with respect to carrier generation if lifetime is injection-dependent. This gave rise to rearrange previous findings in a quantitative theory of light-biased carrier decay, which also comprises time-modulated lifetime techniques without constant bias light. This theory directly relates measurable differential lifetime to actual carrier lifetime. It is applied in a new differential-to-actual (d2a) lifetime analysis, which allows the determination of actual carrier lifetime from differential measurements without integration over the entire injection range. In terms of practical relevance, the d2a approach brings about a drastic experimental simplification. Combined with the general advantages of dynamic lifetime techniques, it could upgrade differential carrier decay techniques and time-modulated lock-in techniques intopowerful quantitative characterization options in silicon photovoltaics. This paper provides an experimental proof ofconcept of the d2a technique based on harmonically time-modulated photoluminescence.

In this paper, the injection-level-dependent measurement of the effective minority carrier lifetime of solar cells from dynamic time-modulated electroluminescence is introduced. The analogy between optical and electrical injection of excess carriers is elaborated, the relevant specifications of the experimental setup are addressed, and an experimental proof of concept is given. Lifetime measurements over a very broad injection range are presented and found to agree well with dynamic photoluminescence measurements.

In this Letter, we introduce a method of calculating the optimal wafer thickness for silicon solar cells with multicrystalline bulk material. The optimal thickness depends on the relation of bulk recombination to surface recombination and the light trapping. For multicrystalline silicon bulk recombination strongly varies laterally and with injection level, which complicates the calculations. A thickness optimization using the "Efficiency Limiting Bulk Recombination Analysis" (ELBA) takes all these effects correctly into account.

Photoluminescence-based imaging is most commonly used to measure the excess minority carrier density and its corresponding lifetime. By using appropriate surface treatments, this high-resolution imaging technique can also be used for majority carrier concentration determination. The mechanism involves effectively pinning the minority excess carrier density, resulting in a dependence of the photoluminescence intensity on only the majority carrier density. Three suitable surface preparation methods are introduced in this paper: aluminum sputtering, deionized water etching, and mechanical abrasion. Spatially resolved dopant density images determined using this technique are consistent with the images obtained by a well-established technique based on free carrier infrared emission. Three applications of the technique are also presented in this paper, which include imaging of oxygen-related thermal donors, radial dopant density analysis, and the study of donor-related recomb ination active defects. These applications demonstrate the usefulness of the technique in characterizing silicon materials for photovoltaics.

A quantitative doping density mapping technique for silicon samples with micrometer spatial resolution is presented. Being based on confocal microphotoluminescence spectroscopy, the technique allows for detailed quantitative analyses on the doping concentration of microscopic technological structures in silicon solar cells. The confocal microscope setup enables laser illumination with micrometer-sized focus and fast low-noise detection of the emitted luminescent radiation which depends on doping and free excess carrier density. The doping density is determined by calibrating the depth-dependent luminescent radiation with results from 2-D simulations of the carrier density. The high-resolution method is demonstrated on a state-of-the-art doping structure for back contact solar cells, where small deviations in the doping homogeneity can be detected. This proves the potential of the proposed technique to enhance the processing of doping microstructures and to strengthen the understanding of their physical properties.

In this paper we investigate the material quality of n- and p-type multicrystalline silicon wafers after different high- temperature steps, as applied during cell processing. Both materials start with a high initial bulk diffusion length of around 440 mm (harmonic mean of the whole wafer) which is further improved by the solar cell process. A diffusion length of 510 mm was measured after phosphorus and boron diffusion and firing in the n-type material. The p-type wafers showed diffusion lengths of 540 mm after phosphorus diffusion and firing. These diffusion lengths were measured at a generation rate of 1/20 sun close to maximum power point injection conditions of a solar cell. At higher injection levels both materials reach 600 mm diffusion length. The high material quality of n-type material maintained after the high temperature boron diffusion is remarkable. An efficiency analysis shows that these excellent diffusion lengths allow for high efficiency devices exceeding 20% efficiency.

During the production of silicon solar cells crack detection systems can help to sort out damaged wafers and reduce wafer breakage before they enter the production line. In order to be cost effective, the crack detection system needs to minimize false detections as much as possible. False detections in crack detection systems occur when bad wafers are not detected or when good wafers are falsely detected as bad. The first error leads to an increase in cell breakage, the second error raises cell costs because non-damaged wafers are sorted out prior to cell processing. In this work a model has been developed to calculate the maximum allowable error rates of crack detection systems in order to achieve a cost per wafer benefit. Therefore a breakage rate dependent throughput calculation, based on manufacturing data, has been implemented. A sensitivity analysis shows that avoiding a high sorting out rate is crucial to favor the implementation of a crack detection system.

Most high-efficiency solar cells are fabricated from monocrystalline Czochralski silicon (Cz-Si) wafers because the material quality is higher than multicrystalline silicon (mc-Si) wafers. However, the material study presented in this paper reveals strong variations in the material quality of commercially available Cz-Si wafers, leading to a loss in solar cell efficiency of 4% absolute. The reason for this is the presence of defects, which appear as dark rings in photoluminescence (PL) images of the finished solar cells. It is shown that these efficiency-limiting defects originate from oxygen precipitation during emitter diffusion. It is demonstrated that an incoming inspection in the as-cut state is difficult, as strong ring structures in as-cut wafers turn out to originate most often from thermal donors. These are dissolved during high-temperature treatments and are therefore harmless, whereas moderate ring structures in the as-cut state may become severe. That is why critical wafers can be identified and sorted out reliably only after emitter diffusion, by using QSSPC-based lifetime measurements or PL imaging. The two-year statistics gathered from the research line at Fraunhofer ISE on the occurrence of ring defects in Cz-Si wafers indicate that ring defects are highly relevant in terms of material yield.

Broad range injection-dependent carrier lifetime measurements in crystalline silicon are most relevant for both defect level spectroscopy, and for the investigation of recombination properties of novel solar cell technologies. The approach presented in this paper combines a determination of the effective carrier lifetime via time-dependent (quasi-steady-state) photoluminescence with a steady-state photoluminescence lifetime scan over a broad range of excess carrier densities. Thereby, the power of a virtually artifact-free time-domain approach is combined with the high sensitivity of steady-state photoluminescence at extremely low injection conditions. Time-dependent fluctuations of dark photocurrent measurements are identified as the essential source of uncertainty under such low injection. A measurement design which eliminates this source of uncertainty is presented and tested. Lifetime measurements at excess carrier densities as low as 10 8 cm −3 are shown in detail.

We are studying the thermal stability of thick hydrogenated amorphous aluminum oxide (Al2O3) layers (20-50 nm) prepared by a high-throughput plasma-enhanced chemical-vapor-deposition (PECVD) technique for the electrical passivation of crystalline silicon surfaces. These passivation layers can be applied alone or covered by a capping layer like amorphous hydrogenated silicon nitride (SiNx) or amorphous hydrogenated silicon oxide (SiOx), also prepared by PECVD. After firing at 870 degrees C for approximately 3 s, the layers show blistering for Al2O3 of 30 nm or higher, independently from the capping layer. For thinner Al2O3, no blistering can be observed even using scanning electron microscope (SEM). Very long carrier lifetimes up to 900 mu s was obtained in passivated p-Si (1 Omega cm) wafer after annealing and firing, without observing a strong influence of the layer thickness and the capping layer. All the layer stacks, including the stacks with SiNx capping layer, show high negative charge densities in the layer (1-4 x 10(12) cm(-2)). Additionally, low interface defect densities (similar to 10(11) cm(-2) eV(-1)), which could be achieved with and without a hydrogenated capping layer, were measured even after firing. To explain these phenomena, hydrogen concentration depth profiles were measured by nuclear reaction analysis. These measurements have shown that, at the Al2O3-Si interface, hydrogen atomic concentration ranging 5-7% after annealing and 4% after firing are obtained independently from the capping hydrogen concentration. We conclude that PECVD Al2O3 layers of 20 nm or thicker can provide enough hydrogen to passivate the interface defects, even after a high te! mperature step. However, the layer thickness should be limited to 30 nm in order to avoid the blistering.

The passivation mechanisms and qualities of Al2O3, SiNx, SiO2 and a-Si:H( i ) on p - and n -type silicon are investigated by quasi-steady-state photoluminescence measurements. This technique allows effective lifetime measurements in an extremely large injection range between 1010 cm-3 and 1017 cm-3. The measurements are discussed focusing on injections below 1012 cm-3 in order to determine the most effective passivation layer for solar cells arranged for indoor applications. Fixed negative charges in the passivation layer cause field-effect passivation due to band bending leading to either accumulation or inversion at the passivation layer/silicon interface. Accumulation causes a stable passivation quality at low level injection. Inversion leads to effective lifetime losses similar to the losses in the space charge region. On p -type silicon the most effective surface passivation at low injections is provided by Al2O3 or a-Si:H( i ). The n -type silicon samples passivated with a-Si:H( i ) show the best effective lifetimes. SiNx and SiO2 show lifetimes one order of magnitude below a-Si:H( i ). Al2O3 on n -type is the most effective passivation at high injections around 1015 cm-3. Due to inversion losses at low level injections the passivation quality decreases more than two orders of magnitude for injections around 1010 cm-3.

An accurate quantitative description of the Auger recombination rate in silicon as a function of the dopant density and the carrier injection level is important to understand the physics of this fundamental mechanism and to predict the physical limits to the performance of silicon based devices. Technological progress has permitted a near suppression of competing recombination mechanisms, both in the bulk of the silicon crystal and at the surfaces. This, coupled with advanced characterization techniques, has led to an improved determination of the Auger recombination rate, which is lower than previously thought. In this contribution we present a systematic study of the injection-dependent carrier recombination for a broad range of dopant concentrations of high-purity $n$-type and $p$-type silicon wafers passivated with state-of-the-art dielectric layers of aluminum oxide or silicon nitride. Based on these measurements, we develop a general parametrization for intrinsic recombination in crystalline silicon at 300 K consistent with the theory of Coulomb-enhanced Auger and radiative recombination. Based on this improved description we are able to analyze physical aspects of the Auger recombination mechanism such as the Coulomb enhancement.

Quasi-steady-state photoluminescence is a versatile technique to determine carrier lifetime in silicon. A recent approach extracts carrier lifetime without a priori information about dopant concentration [Appl. Phys. Lett. 97, 092109 (2010)]. It utilizes the phase shift between a time-modulated optical irradiation and the radiative recombination of a sample, while requiring a minimum of two measurements. The present paper is aimed at a generalization thereof that requires only one measurement. It brings about a substantial experimental simplification, significantly improved accuracy and precision, and it opens up paths to access material properties other than lifetime.

Modulated quasi-steady-state photoluminescence is used in photovoltaics in order to measure the injection-dependent effective minority carrier lifetime of silicon wafers. In spite of the very advantageous properties of this measurement technique, its wide dissemination in photovoltaics has been hampered so far because of a relatively poor sensitivity limit in terms of carrier lifetime. A systematic analysis of the sensitivity-limiting mechanisms led to a substantial upgrade of sensitivity, tackling the range of effective lifetimes of 1s. This paper discusses the requirements to reach this level of sensitivity. Most prominently, the dependence of a silicon photodiode's response time with respect to the wavelength of the detected light is addressed. As an alternative, InGaAs photodiodes are implemented. The sensitivity of this method with respect to carrier lifetime is evaluated.

This paper presents a class of methods to determine net dopant concentration of silicon wafers by means of dynamic or quasi-steady-state photoluminescence. In contrast to resistivity measurements, this approach is independent of assumptions about dopant type and majority carrier mobility. The latter fact makes it particularly interesting for the determination of net dopant concentration in compensated silicon, where conventional mobility models involving only one dopant species have been reported to fail. Our approach also allows access to majority carrier mobility via combination of net dopant concentration as inferred from photoluminescence with a resistivity measurement.

In the last years photoluminescence (PL) imaging has become a standard characterization method for conventional silicon solar cells and wafers providing spatially resolved information for material characterization and process control. In this work, the method of the c oupled determination of d ark saturation c urrent and series r esistance (C-DCR) is evaluated on conventional silicon solar cells in terms of its accuracy, reliability and informative value based on a large number of conventional multicrystalline silicon solar cells. The statistical evaluation is based on a comparison of the series resistance mean values obtained from the C-DCR method with the global values obtained from the IV-characteristics within the standard cell testing. Furthermore, examples for its application are given and demonstrate that the minimum exposure time needed for the C-DCR method is short enough to allow an inline application of the method.

We studied the precipitation of chromium in multicrystalline silicon during the crystallization process and temperature treatments typical for solar cell processing. A model which was already successfully used for simulating heterogeneous precipitation of iron is transferred to chromium, allowing two-dimensional simulations of dissolved chromium and precipitate density. The observed accordance with spatially resolved measurements demonstrates the similarity of chromium to iron precipitation and the ability of our model to predict the dissolved chromium concentration in multicrystalline silicon. After the crystallization process, a high impact of chromium on the carrier lifetime of wafers originating from an ingot intentionally contaminated with 20 ppma chromium in the melt was observed. The concentration of dissolved chromium was significantly reduced by phosphorus diffusion gettering or oxidation at 815°C. Copyright © 2012 John Wiley & Sons, Ltd.

We present a qualitative method for series resistance imaging that generates lateral current flow in solar cells without electrical contacting. The applicability of the method is demonstrated on a multicrystalline silicon solar cell, and the results are briefly discussed with respect to lifetime artifacts and injection dependence of the series resistance distribution. The method overcomes one of the major drawbacks of state-of-the-art series resistance imaging methods, which require electrical contacting for current injection or extraction. It allows for all-contactless measurements, reduces the cell breakage rate, and simplifies the measurement setup and sample handling.

The preparation of ultra-thin oxide layers on mono-crystalline silicon substrate surfaces with ozone dissolved in ultra pure water at ambient temperature was investigated as a low cost alternative to current wet-chemical cleaning and passivation processes in solar cell manufacturing. Surface photovoltage technique was applied as fast, nondestructive, and surface sensitive method, to provide detailed information about the influence of oxidation rate and substrate surface morphology on electronic properties of the oxidised silicon interfaces and subsequently prepared hydrogen terminated surfaces. Sequences of wet-chemical oxidation in ozone containing ultra pure water and subsequent oxide removal in diluted hydrofluoric acid solution could be utilised to prepare hydrophobic substrates, which are predominantly required as starting point for layer deposition and contact formation. On so prepared hydrogen-terminated substrates values of interface state densities D it,min ≈ 5 × 10 11 eV −1 cm −2 could be achieved, comparable to values obtained on the same substrates by the standard RCA process followed by HF dip.

Luminescence imaging is a very rapid technique for the characterisation of silicon samples. Megapixel luminescence images on silicon bricks, unprocessed wafers, partially processed wafers and fully processed cells can be captured with acquisition times of typically a few seconds or less than one second. A number of specific luminescence imaging applications have been developed over the last three years, allowing quantitative spatial information to be gained about a variety of crucial material and device parameters. This paper reviews some of the intriguing possibilities for inline monitoring in PV production at an early stage of production that result from the above mentioned short measurement times.

Consolidated tables showing an extensive listing of the highest independently confirmed efficiencies for solar cells and modules are presented. Guidelines for inclusion of results into these tables are outlined and new entries since June 2010 are reviewed.

Micro-Raman (μRS) and micro-photoluminescence spectroscopy (μPLS) are demonstrated as valuable characterization techniques for fundamental research on silicon as well as for technological issues in the photovoltaic production. We measure the quantitative carrier recombination lifetime and the doping density with submicron resolution by μPLS and μRS. μPLS utilizes the carrier diffusion from a point excitation source and μRS the hole density-dependent Fano resonances of the first order Raman peak. This is demonstrated on micro defects in multicrystalline silicon. In comparison with the stress measurement by μRS, these measurements reveal the influence of stress on the recombination activity of metal precipitates. This can be attributed to the strong stress dependence of the carrier mobility (piezoresistance) of silicon. With the aim of evaluating technological process steps, Fano resonances in μRS measurements are analyzed for the determination of the doping density and the carrier lifetime in selective emitters, laser fired doping structures, and back surface fields, while μPLS can show the micron-sized damage induced by the respective processes.

Summary The impact on solar cell performance of transition metals like iron, chromium, nickel, titanium and co pper is the topic of this extended abstract. Each impurity has been intentionally added to silicon feedstock used to gr ow p-type directionally solidified multicrystalline silicon i ngots. A state of the art screen print solar cell process ha s been applied to wafers cut from the bottom to the top of these ingots. Adding 50 ppmwt of iron or 40 ppmwt of nickel or chromium to silicon feedstock, results in comparabl e solar cell performances to reference uncontaminated material in the range 40% to 70% of the ingot height. Addition of 10 ppmwt of titanium dramatically reduces the efficien cy along the entire ingot. Impurities like iron, chromium an d titanium cause a reduction in the diffusion length. Nickel d oes not reduce the diffusion length. On the other hand affe cts strongly the emitter recombination reducing the sol ar cell performance significantly. Copper has the peculiari ty to impact both bulk recombination as well as emitter recombination. A model based on Scheil distribution of impurity ha s been derived to fit the degradation along the ingot. So lar cell performance has been modeled as function of base bulk recombination and emitter recombination. The model fits very well the experimental data and has been also successfully validated. Unexpectedly, the Scheil di stribution of impurity along the ingot leaves its finger-print also at the end of the solar cell process. A measure of impurit y impact has been defined as the level of impurity which cau ses a degradation of less than 2% up to 90% of the ingot height. The advantage of this parameter is that comprises t he different impurities physical characters in one sin gle parameter, easy to compare.

Minority carrier lifetime is the most crucial material parameter for the performance of a silicon solar cell. While numerous methods exist to determine carrier lifetime on solar cell precursors prior to metallization, only very few techniques are capable of implicitly extracting effective minority carrier lifetimes from metallized silicon samples. In this paper, a measurement technique for effective minority carrier lifetime on silicon solar cells and metallized cell precursors via quasi-steady-state photoluminescence is presented. The setup requirements for this measurement technique are elaborated, experimental evidence of the reliability of such measurements down to carrier lifetimes in the range of a microsecond is provided, and a lifetime calibration of spatially resolved photoluminescence images of solar cells via the presented measurement technique is sketched. Finally, the very good agreement between the obtained effective carrier lifetime and the corresponding open circuit voltage of a solar cell is demonstrated.

We studied the behavior of iron in multicrystalline silicon during phosphorus diffusion by spatially resolved measurements and physical modeling. We present improvements to the previously used models for internal gettering in multicrystalline silicon and phosphorus diffusion gettering. 2-dimensional simulations are used for optimization of the phosphorus diffusion processes for intentionally contaminated wafers regarding the iron distribution, without changing the emitter characteristics. Simulations and experimental results show a reduced interstitial iron concentration after an additional low temperature step at the end of the phosphorus diffusion. The concentration of iron precipitates was reduced by a short annealing at 900 degrees C before the phosphorus diffusion, leading to a carrier lifetime three times higher than compared to the standard process.