Simone Bernardini

Nano-XRF Analysis of Metal Impurities Distribution at PL Active Grain Boundaries During mc-Silicon Solar Cell Processing

Metal impurities are known to hinder the performance of commercial Si-based solar cells by inducing bulk recombination, increasing leakage current, and causing direct shunting. Recently, a set of photoluminescence (PL) images of neighboring multicrystalline silicon wafers taken from a cell production line at different processing stages has been acquired. Both band-to-band PL and sub-bandgap PL (subPL) images showed various regions with different PL signal intensity. Interestingly, in several of these regions a reversal of the subPL intensity was observed right after the deposition of the antireflective coating. In this paper, we present the results of the synchrotron-based nano-X-ray fluorescence imaging performed in areas characterized by the subPL reversal to evaluate the possible role of metal decoration in this uncommon behavior. Furthermore, the acquisition of a statistically meaningful set of data for samples taken at different stages of the solar cell manufacturing allows us to shine a light on the precipitation and rediffusion mechanisms of metal impurities at these grain boundaries.

Light-Induced degradation in compensated mc-Si p-type solar cells

Light-induced degradation (LID) due to boron-oxygen complex formation seriously diminishes the efficiency of p-type solar cells. The influence of dopants concentration, net doping and oxygen on the degradation process is investigated using a large variety of B and P compensated mc-Si ingots. Our experiments indicate that the trend of LID depends on the amounts of interstitial oxygen [Oi] and total boron [B]. No clear dependence was found on the net doping.

On the recombination centers of iron-gallium pairs in Ga-doped silicon

Gallium (Ga) doped silicon (Si) is becoming a relevant player in solar cell manufacturing thanks to its demonstrated low light-induced degradation, yet little is known about Ga-related recombination centers. In this paper, we study iron (Fe)-related recombination centers in as-grown, high quality, directionally solidified, monocrystalline Ga-doped Si. While no defect states could be detected by deep level transient spectroscopy, lifetime spectroscopy analysis shows that the minority carrier lifetime in as-grown wafers is dominated by low levels of FeGa related defect complexes. FeGa pairs have earlier been shown to occur in two different structural configurations. Herein, we show that in terms of recombination strength, the orthorhombic pair-configuration is dominant over the trigonal pair-configuration for FeGa. Furthermore, the defect energy level in the band gap for the orthorhombic defect center is determined to be EV + 0.09 eV, and the capture cross-section ratio of the same defect center is determined to be 220.Gallium (Ga) doped silicon (Si) is becoming a relevant player in solar cell manufacturing thanks to its demonstrated low light-induced degradation, yet little is known about Ga-related recombination centers. In this paper, we study iron (Fe)-related recombination centers in as-grown, high quality, directionally solidified, monocrystalline Ga-doped Si. While no defect states could be detected by deep level transient spectroscopy, lifetime spectroscopy analysis shows that the minority carrier lifetime in as-grown wafers is dominated by low levels of FeGa related defect complexes. FeGa pairs have earlier been shown to occur in two different structural configurations. Herein, we show that in terms of recombination strength, the orthorhombic pair-configuration is dominant over the trigonal pair-configuration for FeGa. Furthermore, the defect energy level in the band gap for the orthorhombic defect center is determined to be EV + 0.09 eV, and the capture cross-section ratio of the same defect center is determin...

Is It Possible to Unambiguously Assess the Presence of Two Defects By Temperature- and Injection-Dependent Lifetime Spectroscopy?

This paper comprehends a systematic study of the prospects for an unambiguous assessment of the presence of two separate defects in silicon samples analyzed by temperature- and injection-dependent lifetime spectroscopy (LS). A large number of lifetime datasets are generated by simulating the presence of two defects and then fitted to a single-defect lifetime model. We have categorized the outcome in four categories: (i) a low overall fit quality and thus likely a combination of two defects, (ii) a high overall fit quality by dominance of one of the involved defect, (iii) a high overall fit quality because of symmetry effects in the model, and (iv) a high overall fit quality but no clear dominance by either involved defects nor presence of symmetry effects. We show that the presence of two defects can be ascertained through perceiving a low-quality fit to a single-defect model (category (i)), but we also show that a high-quality fit can arise from a combination of two defects (category (iv)). We show that in the case of categories (i) and (iv), it is possible to identify the two original defects through linear parameterization. In the case of (iii), however, the identification of two simultaneously occurring defects is highly ambitious and not practically feasible.

Evaluation of passivation layers via temperature-dependent lifetime measurements

The effect of temperature on the surface passivation of p-type and n-type monocrystalline silicon is evaluated by temperature dependent photoconductance decay (PCD). Wafers with different passivation layers, i.e. a-Si and SiNx are the subject of these studies. A characteristic lifetime increment is observed for p-type samples coated with a-Si(i) when compared to substrates passivated with SiNx, in agreement with previous literature reports. A different behavior is measured for the case of n-type samples, which show comparable lifetimes among samples with different passivation layers. An interesting lifetime increment is also found at high injection levels for n-type substrates coated with a-Si(i).

Nano-XRF analysis of metal impurities distribution at grain boundaries during mc-silicon solar cell processing

Photoluminescence (PL) imaging is a widely accepted tool to characterize the quality of multicrystalline and monocrystalline silicon cells. Recently a set of neighboring multicrystalline silicon wafers taken from a cell production line at different processing stages have shown an unexpected PL trend. Band-to-band PL (BPL) and sub-bandgap PL (subPL), were collected for the entire silicon wafers. Interestingly, in various regions of the wafer a reversal of the subPL intensity is observed right after the deposition of the anti-reflective coating (ARC). In this work we present the results of the nanoscale X-ray fluorescence imaging at the points of subPL reversal to evaluate the role of metal decoration on this uncommon behavior and we complement it with our previous findings on the distribution of impurities during cell processing.

Unraveling of bulk and surface behavior in high-quality c-Si material via TIDLS

The current trend in silicon photovoltaics towards high-quality thin mono-crystalline silicon substrates makes the accurate representation of surface recombination of utmost importance. It has been shown by several authors that an effective way to study detrimental defects in silicon wafers is by means of temperature and injection dependent lifetime spectroscopy (TIDLS) coupled with the Shockley-Read-Hall recombination model. Given its high sensitivity this is an excellent technique to study high lifetime substrates. However, a thorough evaluation of the surface recombination velocity (SRV) dependence on injection level and temperature is vital to the extrapolation of meaningful results regarding the defects contained in the bulk of the material. Here, we present a TIDLS study of a-Si:H(i), a-Si:H(n) and a-Si:H(p) deposited on n-type low-resistivity FZ substrates. We evaluate the impact of every dielectric layer on the total SRV temperature- and injection dependence while demonstrating its fundamental role in τeff behavior of high-quality Si substrate.

The impact of the FeGa complex on directionally solidified, mono-crystalline, Ga-doped silicon

In this work we show that the high minority carrier lifetime in as-grown Ga-Si wafers is dominated by low levels of iron contamination incorporated during silicon growth. Upon phosphorous diffusion iron is however effectively removed, increasing the bulk carrier lifetime from a few hundred microseconds to well above one milli-second. Lifetime spectroscopy in combination with Shockley Read Hall theory was used to determine the concentrations of Fei and FeGa complexes in the course of the FeGa association. Finally, we use the estimated concentrations of FeGa as a function of time of storage in the dark to validate that FeGa association follows the laws of coulombic attraction similar to FeB.

Correlating defect band luminesce to elemental distribution by X-ray fluorescence

Photoluminescence (PL) imaging is a widely accepted tool to characterize the quality of multicrystalline and monocrystalline silicon cells. Recently a set of neighboring multicrystalline silicon wafers taken from a cell production line at different stages of processing have shown an unexpected PL trend. Band-to-band PL (BPL) and sub-bandgap PL (subPL), where collected for the entire silicon wafers. Interestingly, a reversal of the subPL intensity in various regions of the wafer is observed right after the deposition of the anti-reflective coating (ARC). Regions with low subPL intensity before ARC exhibit high subPL intensity afterwards, and the opposite holds true for other regions of the wafer. Some authors have performed high-resolution cathodoluminescence spectroscopy, EBIC and dark lock-in-thermography to elucidate the origin of this phenomenon, In this work we present the results of the nanoscale X-ray fluorescence imaging at the points of subPL reversal to evaluate the role of metal decoration on this uncommon behavior and we complement it with our previous findings on the distribution of impurities during cell processing.