In broad terms, this thesis is concerned with the measurement and interpretation of carrier lifetimes in multicrystalline silicon. An understanding of these lifetimes in turn leads to a clearer picture of the limiting mechanisms in solar cells made with this promising material, and points to possible paths for improvement. The work falls into three broad categories: gettering, trapping and recombination. A further section discusses a powerful new technique for characterising impurities in semiconductors in general, and provides an example of its application.
Gettering of recombination centres in multicrystalline silicon wafers improves the bulk lifetime, often considerably. Although not employed deliberately in most commercial cell processes, gettering nevertheless occurs to some extent during emitter formation, and so may have an important impact on cell performance. However, the response of different wafers to gettering is quite variable, and in some cases is non-existent. Work in this thesis shows that the response to gettering is best when the dislocation density is low and the density of mobile impurities is high. For Eurosolare material these conditions prevail at the bottom and to a lesser extent in the middle of an ingot. However, these conclusions can not always be applied to multicrystalline silicon produced by other manufacturers. Low resistivity multicrystalline silicon suffers from a concurrent thermally induced degradation of the lifetime. This had previously been attributed to the dissolution of precipitated metals, although we note that the crystallographic quality also appears to deteriorate. The thermal degradation effect results in an optimum gettering time for low resistivity material. Edge-defined Film-fed Growth (EFG) ribbon silicon was also found to respond to gettering, and even more so to bulk hydrogenation. Evidence for Cu contamination in the as-grown EFG wafers is presented.
Multicrystalline silicon is often plagued by trapping effects, which can make lifetime measurement in the injection-level range of interest very difficult, and sometimes impossible. An old model based on centres that trap and release minority carriers, but do not interact with majority carriers, was found to provide a good explanation for these effects. These trapping states were linked with the presence of dislocations and also with boron-impurity complexes. Their annealing behaviour and lack of impact on device parameters can be explained in terms of the models developed. The trapping model allowed a recently proposed method for correcting trap-affected data to be tested using typical values of the trapping parameters. The correction method was found to extend the range of useable data to approximately an order of magnitude lower in terms of carrier density than would be available otherwise, an improvement that could prove useful in many practical cases.
High efficiency PERL and PERC cells made on gettered multicrystalline silicon resulted in devices with open circuit voltages in the 640mV range that were almost entirely limited by bulk recombination. Furthermore, the injection-level dependence of the bulk lifetime resulted in decreased fill factors. Modelling showed that these effects are even more pronounced for cells dominated by interstitial iron recombination centres. Analysis of a commercial multicrystalline cell fabrication process revealed that recombination in the emitter created the most stringent limit on the open circuit voltage, followed by the bulk and the rear surface. The fill factors of these commercial cells were mostly affected by series resistance, although some reduction due to injection-level dependent lifetimes seems likely also. Secondary Ion Mass Spectroscopy on gettered layers of multicrystalline silicon revealed the presence of Cr and Fe in considerable quantities. Further evidence strongly implied that they resided almost exclusively as precipitates.
More generally, injection-level dependent lifetime measurements offer the prospect of powerful contamination-monitoring tools, provided that the impurities are well characterised in terms of their energy levels and capture cross-sections. Conversely, lifetime measurements can assist with this process of characterising impurities, since they are extremely sensitive to their presence. This possibility is explored in this thesis, and a new technique, dubbed Injection-level Dependent Lifetime Spectroscopy (IDLS) is developed. To test its potential, the method was applied to the well-known case of FeB pairs in boron-doped silicon. The results indicate that the technique can offer much greater accuracy than more conventional DLTS methods, and may find applications in microelectronics as well as photovoltaics.
| Identifer | oai:union.ndltd.org:ADTP/216709 |
| Date | January 2001 |
| Creators | Macdonald, Daniel Harold, daniel@faceng.anu.edu.au |
| Publisher | The Australian National University. Faculty of Engineering and Information Technology |
| Source Sets | Australiasian Digital Theses Program |
| Language | English |
| Detected Language | English |
| Rights | http://www.anu.edu.au/legal/copyright/copyrit.html), Copyright Daniel Harold Macdonald |
Page generated in 0.0793 seconds