Recombination Active Grain Research Articles

The electrical properties of high performance multicrystalline silicon and mono-like silicon: Material limitations and cell potential

We compare the electrical properties of p-type and n-type industrially grown High Performance (HP) multicrystalline silicon (mc-Si) and mono-like silicon. The materials are characterised in terms of their bulk lifetimes, their implied open circuit voltages, and the density and recombination velocity of recombination active grain boundaries and dislocation clusters within the materials. Quokka3 is applied to simulate the cell performance potential of the studied materials, and quantify the corresponding efficiency loss due to different recombination mechanisms occurring in the bulk. Our results show that bulk recombination causes a noticeable efficiency loss in high efficiency cast-grown silicon solar cells, varying from 0.1% to 3.1% absolute in the simulated devices, depending on the materials and also the ingot position of the wafers. The performance of p-type HP mc-Si is affected by both recombination at extended crystal defects and in the intra-grain regions. N-type HP mc-Si shows a slight advantage over p-type HP mc-Si in the intra-grain regions. The advantage of p-type mono-like Si over p-type mc-Si mainly comes from optical gains due to alkaline texturing, but the material exhibits large performance variations along the ingot of up to 2.4% absolute. N-type mono-like Si shows the most promising material properties, featuring less performance variation along the ingot than p-type mono-like Si, and the highest efficiency potential among the four studied materials.

Towards the efficiency limits of multicrystalline silicon solar cells

In this contribution, we present our recent results for high efficiency multicrystalline silicon solar cells. Based on n-type high-performance multicrystalline silicon substrates in combination with the TOPCon solar cell concept featuring a full area passivating back contact and a boron-diffused emitter as well as a plasma-etched black-silicon texture at the front side, a certified conversion efficiency of 22.3% has been achieved, which is currently the world record efficiency for multicrystalline silicon solar cells. A detailed loss analysis of the record solar cell batch discloses the nature of the remaining loss mechanisms, revealing the route for further improvements. We observe an efficiency gap between the multicrystalline and the FZ reference solar cells of ~1%abs. Compared to the FZ reference cells, the mc-Si cells also feature a significantly larger scattering in Voc and Jsc as well as a fill factor loss of ~1.5%abs. We show that the scattering in Jsc correlates with the area fraction of recombination-active structural crystal defects and the scattering in Voc additionally with lateral emitter-induced inhomogeneities. The fill factor loss is attributed to the general presence of strongly recombination-active grain boundaries. A detailed loss analysis of the record mc-Si solar cell shows that the major electrical losses are due to recombination at grain boundaries (0.7%abs) and recombination in the emitter (0.6%abs). By reducing these electrical loss channels, e.g. by an improved crystallization process together with a hydrogenation of the bulk and application of an adapted emitter, we expect to reach efficiencies for mc-Si solar cells in the range of 23%.

Analysis of grain structure evolution based on optical measurements of mc-Si wafers

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.

Improved Laplacian Photoluminescence image evaluation regarding the local diode back voltage distribution

Laplacian photoluminescence-based local diode voltage evaluation was shown recently to lead to correct mean values of the local saturation current density in mm-sized regions, but local maxima in the positions of recombination-active grain boundaries appear overestimated. It is shown here by 2-D device simulations that this effect is at least partly due to the influence of the local diode back voltage, which is caused by the voltage drop at the bulk and back contact resistances. Visually the image of this back voltage appears like a blurred copy of the local diode current density. It is shown in this work that indeed the diode back voltage can be simulated in good approximation by blurring the diode current image, which comes out of the Laplacian evaluation, multiplied with an effective vertical bulk resistance. The corresponding point spread function can be obtained e.g. by device simulation. An iterative procedure is proposed leading to self-consistent results for the diode current density and the diode back voltage. If this method is applied to simulated local cell data, the assumed distribution of the saturation current density is retrieved accurately. Applying this method to measured photoluminescence images leads to a better correspondence to non-linear Fuyuki PL evaluation results than the previously performed direct evaluation of the local diode voltage data. Remaining differences will be discussed.

Recombination sources in p-type high performance multicrystalline silicon

This paper presents a comprehensive assessment of the electronic properties of an industrially grown p-type high performance multicrystalline silicon ingot. Wafers from different positions of the ingot are analysed in terms of their material quality before and after phosphorus diffusion and hydrogenation, as well as their final cell performance. In addition to lifetime measurements, we apply a recently developed technique for imaging the recombination velocity of structural defects. Our results show that phosphorus gettering benefits the intra-grain regions but also activates the grain boundaries, resulting in a reduction in the average lifetimes. Hydrogenation can significantly improve the overall lifetimes, predominantly due to its ability to passivate grain boundaries. Dislocation clusters remain strongly recombination active after all processes. It is found that the final cell efficiency coincides with the varying material quality along the ingot. Wafers toward the ingot top are more influenced by carrier recombination at dislocation clusters, whereas wafers near the bottom are more affected by a combination of their lower intra-grain lifetimes and a greater density of recombination active grain boundaries.

Quantifying carrier recombination at grain boundaries in multicrystalline silicon wafers through photoluminescence imaging

We present a method based on steady state photoluminescence (PL) imaging and modelling of the PL intensity profile across a grain boundary (GB) using 2D finite element analysis, to quantify the recombination strength of a GB in terms of the effective surface recombination velocity (Seff). This quantity is a more meaningful and absolute measure of the recombination activity of a GB compared to the commonly used signal contrast, which can strongly depend on other sample parameters, such as the intra-grain bulk lifetime. The method also allows the injection dependence of the Seff of a given GB to be explicitly determined. The method is particularly useful for studying the responses of GBs to different cell processing steps, such as phosphorus gettering and hydrogenation. The method is demonstrated on double-side passivated multicrystalline wafers, both before and after gettering, and single-side passivated wafers with a strongly non-uniform carrier density profile depth-wise. Good agreement is found between the measured PL profile and the simulated PL profile for both cases. We demonstrate that single-side passivated wafers allow more recombination active grain boundaries to be analysed with less unwanted influence from nearby features. The sensitivity limits and other practical constraints of the method are also discussed.

Efficacy of Phosphorus Gettering and Hydrogenation in Multicrystalline Silicon

The emitter formation step (POCl3 diffusion) in p-type crystalline silicon solar cell processing includes many variables, e.g., peak temperature, gas flows, temperature ramps, which can be optimized in order to improve material quality. Diffusion parameters of an 80-Ω/L1 emitter are varied, and the resulting change in electronic quality of multicrystalline silicon is analyzed. A detailed gettering analysis of multicrystalline material, surface passivated with hydrogen-rich amorphous silicon, after POCl <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> diffusion, and an additional gettering step combined with hydrogenation from SiN <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</sub> :H is presented. The industrial-type diffusion leads to material of lower electronic quality than the extended reference diffusion. A major finding of this paper is the fact that results on different 5 × 5 cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> samples out of one 15.6 × 15.6 cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> wafer can vary significantly. Hence, conclusions about which diffusion is most efficient in gettering strongly depend on wafer position. An edge position close to crucible walls, for example, might improve less effectively than another position close to the crucible center. In fact, the opposite can also be shown, and samples originating from edge regions reach their highest lifetimes after gettering. This is explained by the different defect structure of the investigated samples. Structures exhibiting high gettering efficacy contain fewer recombination active grain boundaries and are predominantly free of extended defect clusters.

High resolution saturation current density imaging at grain boundaries by lock-in thermography

The electronic properties of multicrystalline solar silicon materials are dominated by low-lifetime defect regions containing recombination-active grain boundaries and dislocations. Besides reducing the carrier collection probability, these regions increase the dark saturation current density J01, which governs the open circuit voltage. By applying lock-in thermography with spatial deconvolution it is shown that the dominant contribution to J01 comes from recombination-active grain boundaries and to a lower degree from intra-grain defects like dislocations.

Understanding junction breakdown in multicrystalline solar cells

Extensive investigations on industrial multicrystalline silicon solar cells have shown that, for standard 1 Ω cm material, acid-etched texturization, and in absence of strong ohmic shunts, there are three different types of breakdown appearing in different reverse bias ranges. Between −4 and −9 V there is early breakdown (type 1), which is due to Al contamination of the surface. Between −9 and −13 V defect-induced breakdown (type 2) dominates, which is due to metal-containing precipitates lying within recombination-active grain boundaries. Beyond −13 V we may find in addition avalanche breakdown (type 3) at etch pits, which is characterized by a steep slope of the I-V characteristic, avalanche carrier multiplication by impact ionization, and a negative temperature coefficient of the reverse current. If instead of acid-etching alkaline-etching is used, all these breakdown classes also appear, but their onset voltage is enlarged by several volts. Also for cells made from upgraded metallurgical grade material these classes can be distinguished. However, due to the higher net doping concentration of this material, their onset voltage is considerably reduced here.

Synchrotron microscopy and spectroscopy for analysis of crystal defects in silicon

AbstractThe paper discusses the synchrotron‐based microprobe techniques XBIC (X‐ray beam induced current), μ‐XRF (X‐ray fluorescence microscopy) and μ‐XAS (X‐ray absorption microspectroscopy) and their application for studying electrical activity of defects and precipitation of transition metals in Si materials. Investigations were performed on samples of block‐cast multicrystalline Si and on model samples cut from a bonded monocrystalline wafer. To analyze the precipitation sites, Ni, Cu and Fe were introduced intentionally into the samples. The detected precipitates were found to consist of silicides. Evidence for metal precipitates was also found in virtually uncontaminated as‐grown block‐cast Si. Besides Ni precipitates detected at a recombination active grain boundary, particles containing one or several metals (Cu, Fe, Ti, V) were observed. Unexpectedly, these particles seem to exhibit low only recombination activity. Further studies are necessary to identify their nature. (© 2009 WILEY‐VCH Verlag GmbH &amp; Co. KGaA, Weinheim)