Acceptable contamination levels in solar grade silicon: From feedstock to solar cell

https://doi.org/10.1016/j.mseb.2008.05.021Get rights and content

Abstract

Ultimately, alternative ways of silicon purification for photovoltaic applications are developed and applied. There is an ongoing debate about what are the acceptable contamination levels within the purified silicon feedstock to specify the material as solar grade silicon. Applying a simple model and making some additional assumptions, we calculate the acceptable contamination levels of different characteristic impurities for each fabrication step of a typical industrial mc-Si solar cell. The acceptable impurity concentrations within the finished solar cell are calculated for SRH recombination exclusively and under low injection conditions. It is assumed that during solar cell fabrication impurity concentrations are only altered by a gettering step. During the crystallization process, impurity segregation at the solid-liquid interface and at extended defects are taken into account. Finally, the initial contamination levels allowed within the feedstock are deduced. The acceptable concentration of iron in the finished solar cell is determined to be 9.7×103 ppma whereas the concentration in the silicon feedstock can be as high as 12.5 ppma. In comparison, the titanium concentration admitted in the solar cell is calculated to be 2.7×104 ppma and the allowed concentration of 2.2×102 ppma in the feedstock is only two orders of magnitude higher. Finally, it is shown theoretically and experimentally that slow cooling rates can lead to a decrease of the interstitial Fe concentration and thus relax the purity requirements in the feedstock.

Introduction

The requirements on the silicon quality for photovoltaics (PV) are not as high as for the electronic industry. Therefore, alternative and cheaper methods to purify silicon are being developed for the PV market. The key question is what contamination levels in silicon can be considered as acceptable, i.e. what contamination levels are allowed to maintain the performance of the solar cell? Criteria for the specification of solar grade silicon (SoG-Si) that have been published by different authors differ in orders of magnitude and the origin of these values is not always clearly justified [1], [2], [3], [4], [5], [6], [7], [8]. Some of the data together with the contamination levels found in electronic grade silicon (eg-Si) are given in Table 1. In the early eighties, the well-known Westinghouse study was carried out: the performance of solar cells that were made of intentionally contaminated Czochralski-grown single crystalline silicon (sc-Si) wafers were investigated [1]. Recently, Geerligs et al. [7] have started to repeat these extensive experiments for mc-Si.

The objective of this work is to calculate the acceptable contamination levels in silicon at different stages of the solar cell fabrication cycle (see Fig. 1) employing a simple theoretical model and some additional assumptions. An actual industrial process for mc-Si has been considered. Cañizo et al. [6] have done similar calculations for single sc-Si solar cells. In contrast to the sc-Si process, the effects of impurity precipitates in mc-Si have to be taken into account as they can act as impurity sources or as internal gettering sites. The allowed impurity concentration in the finished solar cell Cc, in the silicon wafer Cw and in the silicon feedstock Cu have been calculated for some typical impurities. Finally, several process modifications are discussed that have the potential to relax the previously established requirements on the silicon purity.

Section snippets

Theoretical model

Calculations have been done for p-type silicon devices. The starting point for our model calculations is an electron lifetime in the finished device of τe=20μs. PC1D simulations, which corresponds to a typical industrial cell structure, have shown that this lifetime value results in an efficiency of about 15%.

Model results

The acceptable contamination levels from feedstock to solar cell have been calculated for four different metal contaminants of different recombination activity and diffusivity (see Table 2): titanium (harmful and slow), chromium (harmful and intermediate velocity), iron (intermediate harmfulness and fast) and copper (benign and very fast). All these metal contaminants form interstitial point defects in silicon. In order to solve Eq. (2), the following assumptions have been made for the

Model discussion

Comparing our model results with literature, a similar trend is observed: the metallic impurity that has to be less abundant in silicon is Ti, closely followed by Cr, while quite big amounts of Fe and somewhat lower amounts of Cu are acceptable. The acceptable Fe concentration in the feedstock Cu is of the same magnitude as the experimental result for mc-Si solar cells by Geerligs et al. [7]. Their threshold value for Ti is by a factor 8 higher than our calculated one.

The main difference with

Relaxation of purity requirements

There are mainly three tools to relax the purity requirements on the silicon wafer that can be applied during the solar cell process: passivation of recombination active point defects, extraction of impurities by means of external gettering and reduction of the interstitial impurity concentration by means of internal gettering.

In a typical industrial solar cell process, a co-firing step is included at the end of the process to form the front and back metallic contacts. During this fast

Conclusions

We have calculated the acceptable contamination levels of Ti, Cr, Fe and Cu in a p-type mc-Si solar cell, in the initial mc-Si wafer and in the silicon feedstock. These four impurity types have been chosen due to their different capture cross-sections and diffusivities within silicon. It results that Fe and Cu may be present in the silicon feedstock in amounts of several ppma whereas Cr and Ti may only have concentrations of some hundredths ppma. Because of the presence of precipitates in

Acknowledgments

This work has been partially funded by the Spanish Ministerio de Educación y Ciencia, through CESIDEL (project ENE2005-9431-C03-02) and through SOLARSÍ (project PSE-120000-2007-12).

References (38)

  • D. Sarti et al.

    Sol. Energy Mater. Sol. Cell

    (2002)
  • J.P. Murray et al.

    Solar Energy

    (2006)
  • A. Rohatgi et al.

    Solid-State Electron.

    (1980)
  • R. Brown et al.

    J. Cryst. Growth

    (1991)
  • F. Duerinckx et al.

    Sol. Energy Mater. Sol. Cells

    (2002)
  • S. Martinuzzi et al.

    Sol. Energy Mater. Sol. Cells

    (2003)
  • P. Manshanden et al.

    Sol. Energy Mater. Sol. Cells

    (2006)
  • J.R. Davis et al.

    IEEE T. Electron. Dev.

    (1980)
  • S. Pizzini et al.

    J. Electrochem. Soc.

    (1986)
  • N. Yuge et al.

    Prog. Photovoltaics Res. Appl.

    (2001)
  • A.I. Nepomnyaschikh et al.
  • C. del Cañizo et al.
  • L. Geerligs et al.
  • C. del Cañizo et al.

    J. Electrochem. Soc.

    (2000)
  • T. Buonassisi et al.

    J. Appl. Phys.

    (2005)
  • T. Buonassisi et al.

    J. Appl. Phys.

    (2005)
  • J. Mohr et al.
  • J.S. Kang et al.

    J. Appl. Phys.

    (1989)
  • Z. Scheil

    Metallkd.

    (1942)
  • Cited by (70)

    • Impact of copper on light-induced degradation in Czochralski silicon PERC solar cells

      2018, Solar Energy Materials and Solar Cells
      Citation Excerpt :

      Nevertheless, the vast majority of the data on this phenomenon has been obtained at wafer level (i.e. on lifetime or semi-fabricate samples). The feedstock is indeed one of the sources for copper contamination [9,10], although other sources appearing at later phases of the solar cell fabrication, such as brass wire wafer sawing [11], may be even more detrimental. Furthermore, given the extremely high solid diffusivity of copper in silicon even at low temperatures [12], Cu impurities tend to be ubiquitous and the high-temperature steps involved in PERC cell processing (i.e. phosphorus-emitter diffusion and metal contact firing) are particularly critical for their potential to drive-in Cu and contaminate the complete cell.

    • Mechanical stirring influence on solute segregation during plane front directional solidification

      2018, International Journal of Thermal Sciences
      Citation Excerpt :

      The silicon purification is an important issue for the photovoltaic industry since impurity concentrations have a strong influence on solar cells efficiency [1,2].

    • Metallic Contaminants on Surfaces and Their Impact

      2017, Developments in Surface Contamination and Cleaning
    • Potential of producing solar grade silicon nanoparticles from selected agro-wastes: A review

      2017, Solar Energy
      Citation Excerpt :

      Gribov and Zinov’ev (2008) and Gribov and Zinov'ev (2003) present problems associated with polycrystalline silicon production for solar cells applications and highlighted new promising technologies. The criteria for the specification of SoG-Si that have been published by different authors differ in orders of magnitude and the origin of these values is not always clearly justified (Hofstetter et al., 2009). The acceptable contamination levels in solar grade silicon from feedstock to solar cell was modelled by Hofstetter et al. (2009) but the study is limited to Ti, Cr, Fe and Cu in a p-type mc-Si solar cell, in the initial mc-Si wafer and in the silicon feedstock.

    View all citing articles on Scopus
    View full text