Search

Manufacture of photovoltaic silicon wafers through epitaxy directly from the vapor phase with the so-called sintered porous silicon method holds significant cost-saving potential compared with the traditional ingot growth and subsequent multiwire sawing because of the absence of kerf loss. In this method, the surface of a substrate wafer is porosified and sintered to provide a growth platform for the epitaxial layer and to enable its later separation. As a newly industrialized technology, kerf-free wafering requires addressing specific characterization needs. One of these requirements is a reliable quality control of as-etched porous silicon (PSi) layers. In this article, we establish a basis for mass production-scale process control of PSi layers used for epitaxial wafer fabrication by utilizing inline spectrophotometry. Specifically, we compare the reflectance spectra of the porous layers between two different inline spectrophotometers and a laboratory reference spectrophotometer. We analyze these spectra with a model-based fitting approach, relying on local optimization methods and accurate initial guess finding to maximize the speed of the fitting procedure. For model verification, we compare the fitted parameters to scanning electron microscope (SEM) images. Here, we observe a close correlation between the model-based fits and SEM in a variety of different porous structures, resulting in a mean relative deviation of 5%-11% between the two methods depending on the type of the analyzed layers. As a useful property, the inline setups inherently allow line-scan measurements across the wafers, enabling the determination of parameter profiles to evaluate the uniformity.

For measuring the current-voltage (I-V) characteristics of busbarless solar cells, there is a certain degree of freedom in the choice of the contacting configuration as none has been defined as standard yet. This leads to the question of how the energy conversion efficiency of a busbarless solar cell should be specified when there are different measured values for it and poses the risk of misinterpreting cell measurements: Designing the measurement system for highest cell efficiencies for example may be tempting but can lead to cell results disassociated from later module performance and reduced cell-to-module power factors. The loss in module efficiency can thereby significantly exceed 5 %rel. It is therefore recommended to always adapt the contacting system to the module interconnection to yield meaningful I-V results in cell measurements. As a remedy for non-adapted measurement conditions, it is shown that I-V results can be converted from various cell measurement configurations to the module configuration. An analytical approach for the conversion of I-V results between measurement systems with different numbers of current contact bars as well as different shading properties is proposed. The approach is validated experimentally on busbarless cells over a large grid resistivity range. For three measurement systems with very different contacting schemes, the differences in fill factor and efficiency between the systems before correction clearly exceeded 10 %rel for high grid resistivities. After correction, good agreement in I-V characteristics within typical measurement uncertainties were shown. Limitations to the analytical approach associated with finger interruptions are identified and discussed.

The measurement of the current-voltage (IV) characteristics is the most important step for quality control and optimization of the fabrication process in research and industrial production of crystalline silicon solar cells. We propose a methodology to determine the IV characteristics of silicon solar cells in a contactless way. We summarize the theory behind the method, describe the experimental setup and prove the validity of the concept by comparing contactless with conventionally measured IV results. We demonstrate that the differences in open-circuit voltage, short-circuit current, fill factor, and efficiency are smaller than 2.5 %rel for a test set consisting of four passivated emitter and rear cells (PERC cells), two tunnel oxide passivated contact (TOPCon) cells and a silicon heterojunction (SHJ) cell. The methodology eliminates mechanical stress on the solar cells and the time required for contacting the cells, does not require broad busbars for contacting and, hence, enables to reduce the silver consumption, simplifies the measurements of shingle cells and potentially shortens the required time, avoids shading by the contact units and consumable costs for the probes, and yields additional information (reflectance spectrum, quantum efficiency) on the cells compared to conventional IV testing. The methodology might facilitate further applications in the future, e.g., contactless outdoor testing of modules, individual diagnosis of failed cells within a module, and on-the-fly IV measurements of cells.

The measurement of the current-voltage (IV) characteristics is the most important step for quality control and optimization of the fabrication process in research and industrial production of silicon solar cells. The occurrence of transient errors and hysteresis effects in IV-measurements can hamper the direct analysis of the IV-data of high-capacitance silicon solar cells. We propose a novel procedure to reconstruct a quasi-steady-state (qss) IV-characteristics from hysteretic measurements by aligning the generalized current density of forward and backward sweeps. In the process, the base doping concentration NB and the cell thickness d are used as optimization parameters and calculated alongside the qss-IV-curve. We summarize the theory behind the method, describe the experimental implementation and verify the applicability of the approach by comparing the extracted doping concentration to multiple alternative methods. For a test set of more than 100 silicon heterojunction solar cells with doping concentrations between 3 and 7∙1015 cm-3 a RMSD

Convolutional neural networks can be trained to assess the material quality of multicrystalline silicon wafers. A successful rating model has been presented in a related work, which directly evaluates the photoluminescence (PL) image of the wafer to predict the current–voltage parameters after solar cell production. This paper presents the results of two visualization techniques to understand what has been learned in the network. First, we reveal what has been learned in the PL image by visualizing the spatial quality distribution of the wafers based on the activation maps of the network. The method is denoted as regression activation mapping. We compare regression activation maps with $j_0$ images of solar cells to show the semantically meaningful representation of the trained features. Second, we show what has been learned in the data by mapping the learned network representation of all wafers into a low-dimensional subspace. Visualizations reveal the smoothness of our representation with respect to the PL input and measured quality. This technique can be used to detect material anomalies or process faults for samples with high prediction errors.

This paper investigates deep convolutional neural networks (CNNs) for the assessment of defects in multicrystalline silicon (mc-Si) and high-performance mc-Si wafers for solar cell production based on photoluminescence (PL) images. We identify and train a CNN regression model to forecast the $I\text{--}V$ parameters of passivated emitter and rear cells from given PL images of the as-cut wafers. The presented end-to-end model directly processes the PL image and does not rely on the human-designed image feature. Domain knowledge is replaced by a model based on a huge variety of empirical data. The comprehensive dataset allows for the evaluation of the generalizability of the model with test wafers from bricks and manufacturers not presented in the training set. We achieve mean absolute prediction errors as low as $\text{0.11}\%_{\text{abs}}$ in efficiency for test wafers from “unknown” bricks, which improves handcrafted feature-based methods by $\text{35}\%_{\text{rel}}$ at simultaneously lower computational costs for prediction. Samples with high prediction errors are investigated in detail showing an increased iron point defect concentration.

We study the mitigation of light- and elevated temperature-induced degradation (LeTID) with high-intensity illumination treatments, placing special emphasis on inline feasibility. After the treatments, we investigate the stability upon degradation conditions close to the recently suggested standard, which allows estimating the LeTID behavior during the operating lifetime of solar modules in the field. Subsequently, we map the stability at different treatment intensities and temperatures achievable with an air-cooled tool. We show that, when applying short treatment times, the stability improves with increasing treatment intensity and deteriorates steeply with rising temperature above an optimum region around 250 °C. However, these intensity- and temperature-dependent differences largely vanish when increasing the treatment time sufficiently. We also investigate the significance of darkness/illumination during the cooling ramp from the treatment temperature in view of the LeTID stability. We discuss our results based on suggested defect models of LeTID, and provide hypotheses of the origin of the instabilities observed at the high treatment temperatures. After identifying optimal treatments, we demonstrate that the energy yield loss due to LeTID reduces by over 60% after an inline-feasible process consisting of only 30 s of high-intensity illumination, combined with cooling of the samples from the process temperature under a lower-intensity illumination.

Measurement images of solar cells contain information about their material- and process-related quality beyond current-voltage characteristics. This information is currently only partially used because most algorithms look for human-defined image features or defects. Herein, a purely data-driven method is proposed to derive the essential image information in terms of the electrical quality within a comprehensive and meaningful representation. This representation is denoted as the empirical digital twin of the cell. Using it, solar cells can be classified according to their defects visible in the measurement images. For this purpose, a human-in-the-loop approach to efficiently develop a classification scheme is presented. Therefore, a convolutional neural network combining various measurement data of a sample by correlating them with quality parameters is designed. The digital twin is an intermediate representation of the network capturing the quality-relevant defect signatures from the images. Human experts can analyze this representation space to identify defect clusters that relate to different process errors, such as finger interruptions and shunts. How the representations are usable to derive sorting criteria for quality inspection is shown. Finally, how the empirical digital twin and the sorting scheme can be used for segmenting the defects without additional label effort is demonstrated.

In High-Performance mc-Si [1] random grain boundaries, although being recombination active, often enhance material quality by reducing dislocations. With this work, we take a step towards statistical large-scale investigations of crystal defects via a combined analysis of different inline-measurements on as-cut wafers: photoluminescence images for the extraction of recombination-active structures and reflection and infrared transmission images for the extraction of the grain structure. The combined extraction of recombination-active structures and grain structures allows isolating dislocations from grain boundaries for all material types. To discern dislocations from other recombination-active defect structures, an image-processing-based analysis technique has been developed. By applying this separation on wafers from various bricks of our material set, typical developments of grain structure and dislocations can be identified. As a particular application, we investigate the correlation between the development of dislocations in higher parts of the brick and grain size in the lower parts. The results support theory quantitatively: Dislocation ratio in the upper brick part shows a correlation with the square root of the weighted median of the grain size in the lower brick part (R≈0.85). However, the results also show that grain size distribution, in particular grain size homogeneity, has to be considered to account for a stronger distinction between materials.

We present an approach for examining and understanding the impact of material and process variations on solar cell efficiencies using the example of an industrial feasible multicrystalline silicon (mc-Si) passivated emitter and rear cell (PERC) process. We fabricate and characterize more than 800 mc-Si PERC cells with a broad material variation and model the experimentally achieved solar cell efficiencies based on numerical 3-D device simulations, metamodeling, and Monte Carlo runs. We subject the simulated distribution of cell efficiencies to a variance-based sensitivity analysis, extracting and ranking the process- and material-related input parameters according to their share of the total variance of cell efficiencies and highlighting the parameters that need to be tuned and controlled most accurately. We are able to explain 90% of the measured total variance which divides into 68%abs. material- and 22%abs. process-related influences. Experimental indication of fill factor (FF) losses due to laterally inhomogeneous bulk lifetimes is found. The presented methodology and its findings provide a fundamental tool for a better understanding of the dependencies in a mc-Si PERC process and lays the groundwork for optimizing the quality and the yield of production lines. Furthermore, the approach is transferrable to other solar cell concepts and production lines.

Light-induced degradation (LID) of multicrystalline silicon (mc-Si) solar cell performance has been reported to be surprisingly strong for conditions relevant under field operation. In paticular, solar cells with dielectrically passivated rear sides such as passivated emitter and rear cells are affected by this LID effect that can cause a loss of more than 10%rel in cell efficiency. With the root cause for the observed degradation being unknown to date, the underlying defect, however, has also been reported to be permanently deactivated under the same conditions at even longer time scales. However, a severe power loss due to the mc-Si specific LID is observed before the cells recover due to the long time scales of the regeneration under these conditions. Hence, regeneration on short time scales similar to the fast regeneration processes being reported for the boron–oxygen defect within p-type Czochralski-grown silicon is highly desirable also for p-type mc-Si, especially since mc-Si currently dominates industrial solar cell production. Within this work, partial regeneration of the defect causing the mc-Si specific LID is shown to be possible within less than 30 s, reducing the impact of LID by up to 60%, which leads to a significantly increased performance of the regenerated mc-Si solar cells.

Novel material classes for solar cell production e.g. high performance multicrystalline silicon or epitaxially grown wafers have a huge impact on solar cell performance. A speedup of these developments calls for a rapid assessment of the material quality in the as-cut stage already.This work introduces a generic architecture for the material rating of wafers in terms of solar cell quality. Our approach allows for a simultaneous prediction of the open-circuit voltage of the solar cell and the image of the dark-saturation current density (j 0 ) from photoluminescence images of as-cut wafers. In the sense of theory-guided data-analysis, we combine a data-driven machine learning approach with known physical constraints, here given by the one-diode equation.Due to this statistical optimization, our method derives the j 0 values for the occluded regions beneath the busbar. From the derived j 0 values, we also evaluate the impact of material-related defects and grid metallization structure on the dark-saturation current density.

Bifacial solar cells and modules are a promising approach to increase the energy output of photovoltaic systems, and therefore decrease levelized cost of electricity (LCOE). This work discusses the bifacial silicon solar cell concepts PERT (passivated emitter, rear totally diffused) and BOSCO (both sides collecting and contacted) in terms of expected module cost and LCOE based on in-depth numerical device simulation and advanced cost modelling. As references, Al-BSF (aluminium back-surface field) and PERC (passivated emitter and rear) cells with local rear-side contacts are considered. In order to exploit their bifacial potential, PERT structures (representing cells with single-sided emitter) are shown to require bulk diffusion lengths of more than three times the cell thickness. For the BOSCO concept (representing cells with double-sided emitter), diffusion lengths of half the cell thickness are sufficient to leverage its bifacial potential. In terms of nominal LCOE, BOSCO cells are shown to be cost-competitive under monofacial operation compared with an 18% efficient (≙ pMPP = 18 mW/cm2) multicrystalline silicon (mc-Si) Al-BSF cell and a 19% mc-Si PERC cell for maximum output power densities of pMPP ≥ 17.3 mW/cm2 and pMPP ≥ 18.1 mW/cm2, respectively. These values assume the use of $10/kg silicon feedstock for the BOSCO and $20/kg for the Al-BSF and PERC cells. For the PERT cell, corresponding values are pMPP ≥ 21.7 mW/cm2 and pMPP ≥ 22.7 mW/cm2, respectively, assuming the current price offset (≈50%, at the time of October 2014) of n-type Czochralski-grown silicon (Cz-Si) compared with mc-Si wafers. The material price offset of n-type to p-type Cz-Si wafers (≈15%, October 2014) currently accounts for approximately 1 mW/cm2, which correlates to a conversion efficiency difference of 1%abs for monofacial illumination with 1 sun. From p-type mc-Si to p-type Cz-Si (≈30% wafer price offset, October 2014), this offset is approximately 2.5 mW/cm2 for a PERT cell. When utilizing bifacial operation, these required maximum output power densities can be transformed into required minimum rear-side illumination intensities for arbitrary front-side efficiencies ηfront by means of the performed numerical simulations. For a BOSCO cell with ηfront = 18%, minimum rear-side illumination intensities of ≤ 0.02 suns are required to match a 19% PERC cell in terms of nominal LCOE. For an n-type Cz-Si PERT cell with ηfront = 21%, corresponding values are ≤ 0.11 suns with 0.05 suns being the n-type to p-type material price offset. This work strongly motivates the use of bifacial concepts to generate lowest LCOE. Copyright © 2016 John Wiley & Sons, Ltd.

Bifacial applications are a promising way to increase the performance of photovoltaic systems. Two silicon solar cell concepts suitable for bifacial operation are the passivated emitter, rear totally diffused (PERT) and the both sides collecting and contacted (BOSCO) cell concepts. This work investigates the bifacial potential of these concepts by means of in-depth numerical device simulation and experiment with a focus on the impact of varying material quality. It is shown that the PERT cell concept (representing a structure with front-side emitter only) requires high-minority-carrier-diffusion-length substrates with Lbulk > 3 × W (with cell thickness W) to exploit its bifacial potential, while the BOSCO cell (representing a structure with double-sided emitter) can already utilise its bifacial potential on substrates with significantly lower diffusion lengths down to Lbulk ≈ 0.5 × W. Experimentally, BOSCO cells with and without activated rear-side emitter are compared. For rear-side illumination, the activated rear-side emitter is measured to increase internal quantum efficiency at wavelengths λ

Microcracks in silicon wafers reduce the strength of the wafers and can lead to critical failure within the solar-cell production. Both detection of the microcracks and their impact on fracture strength of the wafers are addressed within this study. To improve the accuracy of the crack detection in photoluminescence (PL) and infrared transmission (IR) images of as-cut wafers, we introduce a pattern recognition approach based on local descriptors and support-vector classification. The learning model requires a set of labeled data generated by an artificial insertion of cracks. Within this evaluation, the algorithm detects 81% of the cracks for PL-images and 98% for IR-images at precision rates above 98% in each case, which outperforms the quality of pure IR-intensity-based crack-detection systems with a hit-rate of 65% at a precision of 59%. The proposed algorithm may be combined with the images of the grain structure to avoid the confusion of cracks and grain boundaries. Moreover, the comprehensive set of wafers allows the impact of crack morphology on wafer strength to be investigated. Despite complex crack morphologies, the theoretically expected dependence between crack length and fracture strength is confirmed. Therefore, sorting criteria are derived to rate the cracks with respect to the expected fracture strength of the wafer based on the measured crack length only.

As the market share of bifacial cells and modules grows, solar cells measurements under bifacial illumination are increasingly becoming a focus of interest. In this work, we present an analysis of four different sorting criteria for bifacial solar cells with the aim of evaluating the impact of these criteria on the power of the modules assembled with the sorted cells. First, we develop a simulation model for bifacial modules based on the two-diode model. Secondly, we generate a representative virtual data bank made up of 50000 cells, whose parameters are determined in account with empirical parameter distributions measured on a group of 300 bifacial solar cells with passivated emitter and rear (PERC) manufactured industrially. Third, we sort and bin all cells according to the current density under front illumination at maximum power point (MPP) (jmppfront). Fourth, we assemble modules from all bins, calculate the module power under bifacial illumination for five different scenarios with varied front and rear irradiation and evaluate the average module power as well as the distribution of module power. The sorting of the cells (3rd step) and the calculation of the module power (4th step) are repeated for sorting the cells according to the current at MPP under bifacial illumination (jmppbi) and the maximum power under front only and bifacial illumination (pmppfront and pmppbi). With respect to the average module power, a very small gain of less than 2 W (in the most extreme illumination scenario) can be achieved if a bifacial instead of a monofacial sorting parameter is taken as sorting criterion. However, a strong effect on the distribution of the module power is manifested, obtaining either one homogeneous group of modules with broad power distribution or several classes/bins of modules, depending of the cell sorting criterion. Therefore, the choice of the sorting criteria of the cells depends on the interest of the solar manufacturers and has a great potential to reduce module mismatch in PV arrays.

Within this work, we present a method for fast characterization and statistical analysis of structural and electrical crystal properties. By this, we investigate the impact of grain size and grain shape distribution in the lower part of the brick on the defect development in the upper part of the brick. Our method is based on fast measurements on as-cut wafers, namely photoluminescence imaging and reflection or infrared transmission images. We combine the extracted information and isolate dislocation structures from recombination active grain boundaries via image-processing. Two large data sets (more than 40 bricks) of different materials allow us to quantify the assumed correlation between grain structure and dislocation development. The results confirm that the relative length of dark lines as a part of dislocation clusters in the upper brick part shows a locally linear correlation with the square root of a meaningful grain area measure in the lower brick part (Pearson coefficient R between 0.80 and 0.88). It also correlates with the inhomogeneity measures for grain area and shape (Pearson coefficient R about 0.88). As a result, changes in the crystallization process may be targeted according to the grain structure in the lower part of the brick. In the future, this method may serve to improve the prediction of solar cell results based on wafer-data.

Recombination-active defects e.g. dislocations in multicrystalline silicon (mc-Si) wafers impact the quality of solar cells. These defects can be quantified during the incoming control of silicon wafers with Photoluminescence (PL) imaging and used to rate the solar cell quality. In this work, we analyze the relevance of defect patterns in PL images and grain-boundary (GB) data for current-voltage (IV) prediction by means of image processing algorithms. Based on a large set of empirical data of passivated emitter and rear cells (PERC), a sparse prediction model is trained for each IV-parameter. Our results include both, the prediction of different quality parameters and the relevance of the features extracted from PL images and GB images. We achieve mean absolute prediction errors as low as 2.72 mV and 0.18 mA/cm2 for open circuit voltage (Voc) and short circuit current density (Jsc) respectively, and 0.18% for efficiency as combined parameter. In this evaluation, the wafer data set is split into training group and test group. Therefore the results show the prediction of unknown material. This makes the prediction more challenging but represents a realistic use case for production. The comparative overview of the relevant feature set shows a difference between the prediction of short-circuit current and open-circuit voltage prediction.

Bifacial solar modules are expected to increase their market share in the coming years. To date, most bifacial cells in production and under research are realized on n-type silicon substrates. However, p-type silicon has been dominating photovoltaic industry, partly due to its lower cost. In this work, we present an approach to combine bifacial applications with the industrial p-type passivated emitter and rear (PERC) structure: a bifacial PERC cell (biPERC). Simulation results are presented that showcase a high potential for bifacial gains of the biPERC cell. First experimental results of the biPERC concept on large-area multicrystalline silicon substrates yield a monofacial efficiency of 17.8%. The biPERC cell, therefore, is an easy-to-apply modification to a PERC route and enables higher output powers due to bifacial applicability.

We present new experimental data on light-induced degradation due to the boron oxygen defect in compensatedn-type silicon. We are the first to show that both defect components known fromp-type silicon are formed in compensatedn-type silicon. A parameterization of the injection dependent recombination activity of the slower formed defect component is established. The formation kinetics of both defect components are studied and modeled under different conditions. It is found that the same rate factors as inp-type can describe the degradation, if the actual hole concentration under illumination is taken into account. The regeneration process known to permanently deactivate boron oxygen defects inp-type is successfully applied ton-type material and the illumination stability of the regenerated state is tested and proven.

Light-induced degradation (LID) has been identified to be a critical issue for solar cells processed on boron-doped silicon substrates. Typically, Czochralski-grown silicon (Cz-Si) has been reported to suffer from stronger LID than block-cast multicrystalline silicon (mc-Si) due to higher oxygen concentrations. This work investigates LID under conditions practically relevant under module operation on different cell types. It is shown that aluminium oxide (AlOx) passivated mc-Si solar cells degrade more than a reference aluminium back surface field mc-Si cell and, remarkably, an AlOx passivated Cz-Si solar cell. The defect which is activated by illumination is shown to be doubtful a sole bulk effect while the AlOx passivation might play a certain role. This work may contribute to a re-evaluation of the suitability of boron-doped Cz- and mc-Si for solar cells with very high efficiencies. (© 2015 WILEY-VCH Verlag GmbH &Co. KGaA, Weinheim)

To investigate the rear side of bifacial p-type Czochraslki-grown silicon PERC solar cells, the present work combines Sentaurus Device simulation – calibrated with extensively characterized samples – and the subsequent fabrication of solar cells according to the simulation findings. The authors investigate the physical alteration of rear-side characteristics in the context of an additional rear-side illumination. The additional injection represents an further factor for the balance of carrier generation, recombination and series resistances which in turn influences the design rules for the rear side layout. Our detailed bifacial simulations include these physical aspects and we derive design solutions for different bifacial illumination scenarios for a bifacial p-doped PERC solar cell. Using an industrial PERC process, solar cells with laser contact openings (LCO) and a rear aluminum grid were produced according to the simulation results with a wide variation in rear side layout parameters. The PERC batches showed a rather constant medium (front side) efficiency of η = 20.8±0.2% and a bifaciality of 66 to 77% depending on the rear layout, allowing us to investigate the rear-side characteristics in detail and to compare them with the effects predicted by the simulations. We processed an aluminum rear contact grid with finger widths as small as 100 µm and successfully aligned it onto the LCO with 30 µm contact openings on full-area 156x156 mm 2 wafers. We reached good accordance between the monofacial measurements from front and rear side and our simulation model and could thus predict bifacial illumination results by modeling for two issues: 1. Planar rear sides have an advantage over pyramid textured rear sides for 1000 W/m² front illumination unless additional rear illumination exceeds 250 W/m². 2. As soon as any rear illumination is added to the front-illuminated PERC solar cell, 100 µm thin fingers at the rear side have an output power advantage compared to 150 µm and 200 µm wide fingers.

For bifacial applications, double-sided collecting solar cell structures can be beneficial. The recently introduced "BOth Sides COllecting and COntacted" (BOSCO) solar cell is such a structure and allows the use of standard module interconnection technology. The structure features emitter areas on both sides, which are interconnected by diffused vias. It favors the use of silicon substrates with low-to-medium diffusion length and low resistivity for the maximum benefit compared with other structures, such as Al-BSF and PERC. Within this study, we discuss the potential of the BOSCO cell structure and its applicability for certain silicon material types. Experimental results on different multicrystalline silicon (mc-Si) materials yield monofacial efficiencies (independently confirmed on a non-reflecting chuck) of 17.4% on large-area wafers from block-cast electronic-grade mc-Si and 16.9% for low-quality upgraded metallurgical-grade mc-Si. These values represent a gain of 0.6-0.7% (abs) compared with Al-BSF cells processed in parallel. The bifacial properties are investigated under outdoor testing conditions, yielding a gain in output power of 13% compared with monofacial operation.

The BOSCO solar cell represents a bifacial structure with double-sided collection. The structure allows the use of standard module interconnection technology and favours the use of material with low to medium diffusion length and low resistivity for maximum benefit towards other structures, such as Al-BSF and PERC. Within this work, we present first results on different multicrystalline silicon materials yielding a monofacial efficiency of 17.4 % on large-area wafers from block-cast mc-Si. This value represents a gain of ~0.7 %abs compared to Al-BSF cells processed in parallel. The applicability for bifacial operation is demonstrated by a significantly increased quantum efficiency for rear side illumination. These results make the BOSCO solar cell concept a promising candidate to further boost the output of utility-scale PV plants even when using low-cost wafers of low to medium diffusion length material. Sketch and photographs of the BOSCO solar cell concept.

Light-induced degradation in boron-doped p-type Czochralski silicon is caused by a boron-oxygen complex, which may be permanently deactivated by simultaneous illumination and heating leading to a permanent regeneration of carrier lifetime and solar cell performance. To date, the long-term stability of the regenerated state under thermal stress has not been fully investigated. In this work, we investigate whether the regeneration gain achieved on solar cell level is stable upon accelerated aging tests such as thermal cycling and damp-heat testing as applied during module certification according to IEC and UL standards. It is shown that regeneration is mostly conserved upon thermal cycling and cells in the degraded state may even be regenerated if current is injected during the treatment. Damp-heat testing at 85 C in the dark for 1500 h decreases the regeneration-induced gain significantly but does not annihilate it completely.

In this paper we propose a method to predict hot spot temperatures in crystalline silicon photovoltaic modules operating under critical shading conditions prior to module fabrication. We developed a unique tool to evaluate the damage risk potential for an individual solar cell. We show that shading conditions leading to the most critical hot spot temperature do not necessarily coincide with the shading conditions for the maximum total power dissipation in the shaded cell. In fact, for an adequate prediction of temperature fields and the worst case shading scenario, spatially resolved information of the power dissipation on cell level and a thermal simulation of the module system are indispensable. Our approach is divided into three steps, starting with an electric network simulation of cell operating points for different shading scenarios. For these operating points we perform spatially resolved lock-in thermography measurements of the power dissipation on cell level. On that basis we compute three-dimensional temperature fields inside the module with a finite-element analysis for different realistic module operating conditions. The model is validated with experimental data on a module of industrial cells.

The BOSCO (“ BO th S ides CO llecting and CO ntacted”) solar cell features a double-sided emitter and contact grids on both surfaces. The emitter region on the rear is connected to the front side by diffused vias. The structure allows the use of standard module interconnection technology and favours the use of silicon substrates with low to medium diffusion length and low resistivity for maximum benefit towards other structures, such as Al-BSF and PERC. Within this work, we summarize the latest results on multi-crystalline silicon (mc-Si). Monofacial efficiencies of 17.4% on large-area wafers from 9N (99.9999999% pure) block-cast mc-Si and 16.9% for low-quality 5N upgraded metallurgical-grade mc-Si have been achieved. These values represent a gain of 0.6 to 0.7% abs compared to Al-BSF cells processed in parallel. First tests of bifacial operation under outdoor conditions yield a gain in output power of 13% compared to monofacial operation. This effect makes the BOSCO solar cell concept a promising candidate to enable bifacial operation, even for low-cost wafers of low to medium diffusion length material.

Light-induced degradation (LID) in boron-doped p-type Czochralski (Cz) silicon is caused by a boron-oxygen (BO) complex, which may be permanently deactivated by simultaneous illumination and heating leading to a permanent regeneration of carrier lifetime and solar cell performance. Up to now, regeneration has only been investigated on wafer and solar cell level. In this work, we investigate whether the regeneration gain on solar cell level can be transferred to module level. For this purpose, we fabricated solar cells with passivated emitter and rear on boron-doped p-type Cz and float zone wafers. The cells are extensively characterised regarding losses in cell performance due to LID by preparing different defect states with certain temperature and illumination treatments. The effect of the injection-dependent carrier lifetime caused by the BO complex in its active state on fill factor is investigated in detail theoretically and experimentally, developing a descriptive explanation. Finally, it is shown by integrating solar cells in the degraded and regenerated state into solar modules, that the regeneration effect can be transferred to module level and is stable upon subsequent illumination of the module.

The recombination induced by soldering pads, screen printed with a commercially available paste containing primarily silver and an aluminium amount of 1-5% wt , is investigated concerning the open-circuit voltage for aluminium back surface field (Al-BSF) solar cells. The saturation current density J o,pad due to soldering pads is found to be in the range of 26.7 10 3 to 85.4 10 3 fA/cm 2 for different conditions of the rear silicon surface. The recombination under soldering pads leads to a voltage drop of Δ V oc = 6.7 mV for Al-BSF solar cells with 6.2% of the rear side being covered with soldering pads compared to a cell with full-area Al-BSF. Within a microstructural analysis of the interface between the silicon surface and the soldering pads it is found that Al-alloying from the Al share in the soldering pad paste only occurs insularly. Furthermore, Al-BSF cells with and without residual rear side emitter and varying area fraction of soldering pads, using a soldering pad paste with and without an Al-share, are investigated. It is found that the non-overcompensated emitter regions under the soldering pads of Al-BSF solar cells affect cell performance dramatically in case of full area soldering pad, especially for an Al-free pad paste.

In this paper, we introduce a predictive, physics-based model, i.e., the so-called tilted-mirror model (tm-model), for optical modeling of rough rear surfaces on silicon solar cells. An enhanced method of using transfer matrices at the rear-side interface of solar cells is developed and combined with Monte Carlo ray tracing. As a result, a physically consistent and precise simulation of the spectral reflectance is achieved, thus leading to a predictive quality of the simulations that could previously not be reached for solar cells with a remaining irregular rear-surface roughness. This advance in optical simulation enables the researcher to directly analyze the effects of varying rear-side passivation materials and thicknesses, as well as the impact of different surface morphologies on the gained charge-carrier generation rate of a solar cell. A comparison with the Phong model shows that the tm-model is able to simulate the generated photocurrent Jph more accurately, as it is shown that the Phong model tends to overestimate this value due to imprecise calculation of charge-carrier generation. In an application of the tm-model to passivated emitter and rear cells, it is shown that a strong planarization of the rear surface leads to an improvement in photogenerated current up to 0.13 mA/cm2 compared with a weak planarization.

If a solar cell is partially shaded in a photovoltaic module, it can be reverse-biased and therefore dissipate power rather than generating it. When being locally concentrated, this power dissipation can lead to hot spots damaging solar modules irreversibly. In order to assess the hot spot danger of solar cells prior to module integration, their operating points under partial shading conditions need to be known which can be determined by simulation. We show, that deviations between simulated and measured module I-V characteristics can be explained by contrary temperature- and illumination-dependent effects on the reverse characteristic of a multi-crystalline silicon solar cell. A negative temperature coefficient for voltages of V < -12.5V in the temperature range of T = 25-50°C is determined to be mainly due to type II and III breakdown while a positive illumination coefficient is found for V < -6V which predominantly originates from locally distributed currents.

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.