Bernhard Mitchell

Bulk minority carrier lifetimes and doping of silicon bricks from photoluminescence intensity ratios

Two methods for spatially resolved measurement of the bulk minority carrier lifetime on the side faces of block cast silicon bricks, both based on photoluminescence imaging, are presented. The first method uses a single photoluminescence image which is normalized for variations in the background doping density. The second method is based on the measurement of the ratio of two photoluminescence images taken with different spectral filters in front of the CCD camera. This second method is advantageous over the first method since it is not dependent on an absolute measurement of the luminescence intensity and does not require a separate resistivity measurement. It is further demonstrated, that the measurement of an intensity ratio image and comparison of the latter with an individual luminescence image also allow obtaining the doping density gradient within silicon bricks. This paper presents an in-depth analysis of the underlying models, demonstrates experimental results and discusses the calibration and un...

Temperature dependence of the band-band absorption coefficient in crystalline silicon from photoluminescence

This work has been supported by the Australian Research Council (ARC) and the Australian Renewable Energy Agency (ARENA).

The Impact of Silicon CCD Photon Spread on Quantitative Analyses of Luminescence Images

Commercial and R&D photoluminescence imaging systems commonly employ indirect bandgap silicon charge-coupled device (CCD) imaging sensors. Silicon is a weak absorber of the near-infrared band-to-band emission of silicon, and significant lateral spreading of the luminescence signal can occur within the sensor. Uncorrected, this effect reduces image contrast, introduces artificial signal gradients, and limits the minimum feature size for which accurate quantitative measurements can be derived. Empirical quantification of the spreading effect defined in terms of the point-spread function (PSF) for the imaging apparatus allows for postprocessing deconvolution, which quantitatively improves image accuracy and contrast. Assessment of the impact of a photon spread indicates that signal smear under commonly occurring high contrast ratio scenarios is sufficient to warrant the application of deconvolution to improve the accuracy of quantitative data in calibrated luminescence images. With a carefully defined PSF, corrections to within ± 10% of the true signal ratio for small-area features can be achieved. Short-pass filtering provides partial correction of the photon spread with the advantage of reduced experimental complexity but, nonetheless, with limitations on the minimum feature size for which accurate signal ratios can be measured.

On the method of photoluminescence spectral intensity ratio imaging of silicon bricks: Advances and limitations

Spectral photoluminescence imaging is able to provide quantitative bulk lifetime and doping images if applied on silicon bricks or thick silicon wafers. A comprehensive study of this new method addresses previously reported artefacts in low lifetime regions and provides a more complete understanding of the technique. Spectrally resolved photoluminescence measurements show that luminescence originating from sub band gap defects does not cause those artefacts. Rather, we find that optical light spreading within the silicon CCD is responsible for most of the distortion in image contrast and introduce a method to measure and remove this spreading via image deconvolution. Alternatively, image blur can be reduced systematically by using an InGaAs camera. Results of modelling this alternative camera type and experiments are shown and discussed in comparison. In addition to eliminating the blur effects, we find a superior accuracy for lifetimes above 100 μs with significantly shorter, but dark noise limited expos...

Contrast enhancement of luminescence images via point-spread deconvolution

We investigate the impact of point-spread in the silicon CCD sensor of a BT Imaging LIS-R1 luminescence imaging system. It is found that an experimental definition of the point-spread function allows for a significant restoration of CCD point-spread to be achieved. A comparison with short-pass filtering is performed, demonstrating that point-spread will still have a measurable influence on image quality while reducing the luminescence signal and increasing the relative noise level. An implementation of point-spread deconvolution is presented at a multi-crystalline silicon grain boundary, illustrating a practical enhancement of resolution in a typical high-contrast scenario. The characteristics of the point-spread presented here are specific to the experimental apparatus investigated, but the procedure described is universally applicable.

Photoluminescence Imaging of Silicon Bricks

Photoluminescence imaging techniques have recently been extended to silicon bricks for early production quality control and electronic characterisation in photovoltaics and microelectronics. This contribution reviews the state of the art of this new method which is fundamentally based on spectral luminescence analyses. We present highly resolved bulk lifetime images that can be rapidly extracted from the side faces of directionally solidified or Czochralski grown silicon bricks. It is discussed how detailed physical modelling and experimental verification give good confidence of the best practice measurement errors. It is also demonstrated that bulk lifetime imaging can further be used for doping and interstitial iron concentration imaging. Additionally, we show that full spectrum measurements allow verification of the luminescence modelling and are, when fitted to the theory, another accurate method of extracting the absolute bulk lifetime.

Full Spectrum Photoluminescence Lifetime Analyses on Silicon Bricks

Bulk lifetime and doping images on silicon bricks can be obtained by spectral luminescence intensity ratio analysis as established recently. Here, we report on calibrated full spectrum band-to-band luminescence measurements taken on the flat side faces of mono- and multicrystalline silicon bricks at room temperature. Our results verify the physical modeling used for the spectral intensity ratio imaging. We further investigate three fitting methods employing spectrally resolved photoluminescence data to obtain bulk lifetime information.

Imaging As-Grown Interstitial Iron Concentration on Boron-Doped Silicon Bricks via Spectral Photoluminescence

The interstitial iron concentration measured directly on the side face of a silicon brick after crystallization and brick squaring can give important early and fast feedback regarding its material quality. Interstitial iron is an important defect in crystalline silicon, particularly in directionally solidified ingots. Spectral photoluminescence intensity ratio imaging has recently been demonstrated to independently provide high-resolution bulk lifetime images and is therefore ideally suited to assess spatially variable multicrystalline silicon bricks. Here, we demonstrate this technique to enable imaging of the interstitial iron concentration on boron-doped silicon bricks and thick silicon slabs. We present iron concentration studies for two directionally solidified silicon bricks of which one is a standard multicrystalline and the other is a seeded-growth ingot. This lifetime-based measurement technique is highly sensitive to interstitial iron with detection limits down to concentrations of about 1 × 1010 cm-3. Its accuracy is enhanced, as the injection level remains below 2 × 1012 cm-3 during the measurement and, hence, avoids the influence of injection level dependences on the conversion factor, although it remains dependent on the knowledge of the electron capture cross section of interstitial iron in silicon. Access to both bulk lifetime and dissolved iron concentration provides a valuable parameter set of as-grown crystal quality and the relative recombination fraction of interstitial iron via Shockley-Read-Hall (SRH) analysis. Simulated interstitial iron concentration profiles support the presented experimental data.

Photoluminescence Imaging Using Silicon Line-Scanning Cameras

We experimentally demonstrate photoluminescence imaging using a silicon line-scanning camera. The narrow rectangular geometry of the sensor effectively lowers the image-blur, while keeping the sensitivity relatively high. Line scanning is demonstrated to provide sharp and high contrast images. To maintain fast measurement speeds, higher injections are used to compensate for the lower sensitivity of a line-scan measurement, where localized parts of the image are acquired sequentially. However, low-injection measurements remain possible at the expense of slower scan speeds if cooling is applied to the sensor. Using a detailed point-spread analysis and deconvolution, we quantify and correct for the remaining light spreading effects. Experimental results suggest an effective suppression of the cumulative nonlocal point spread commonly observed using area sensors, confirming silicon (Si) line sensors to be only susceptible to local point spreading effects.

PERC Solar Cell Performance Predictions From Multicrystalline Silicon Ingot Metrology Data

The influence of the as-grown material quality on the performance of multicrystalline silicon PERC solar cells is investigated using recently developed spectral photoluminescence imaging techniques on ingot level, i.e., on bricks, and is examined in conjunction with photoluminescence measurements on as-cut wafers. The effect of material parameters, including bulk lifetime, dislocation density, and resistivity, is studied with regard to their effect on cell output across a sample set of three directionally solidified production bricks of widely varying bulk lifetime and dislocation density. The data are analyzed statistically using a linear mixed model. Bulk lifetime is found to be a statistically significant predictor of the cell performance across the studied sample set. The prediction accuracy is found greatest for material with low dislocation density where a linear correlation between cell performance and as-grown bulk lifetime is found. The dislocation densities measured on as-cut wafers remain a more accurate predictor for medium to highly dislocated material, but predictions are improved by adding bulk lifetime as an additional predictor in the model. Dislocation density measurements taken on the side facets of silicon bricks were identified as not being significant predictors for this dataset. However, the detection may enable the classification of bricks into broad dislocation defect classes, which is expected to further improve the overall prediction accuracy for models which use brick metrology only. With increasing cell efficiencies and an ongoing trend of reducing the dislocation density in industrial multicrystalline wafers, our findings suggest that the bulk lifetime, measurable on bricks, i.e., directly after ingot growth, becomes an increasingly relevant parameter.