Review
Silicon solar cell production

https://doi.org/10.1016/j.compchemeng.2011.04.017Get rights and content

Abstract

A significant role can be played by the systems engineering community in the optimization of the production process for silicon solar cells. Many of the techniques utilized for cell manufacturing are of recent origin and the amount of experience in the industry as a whole is limited. Some of the individual processes and steps are poorly adapted for continuous production since they were designed for micro-electronics applications rather than photovoltaics. Only very recently has the industry grown to the point where intermediate products, such as solar grade silicon, solar silicon wafers, solar cells and solar panels are commodities having global market potential. Finally, industry consolidation has generated large commercial entities which can better take advantage of tools from process systems engineering. The chemical and process systems and engineering communities can contribute to this booming industry by providing methods for improved control, process optimization and retro-fitting of existing processes, as well as encouraging process innovation and scale-up. This paper describes the complete production process for solar cells, highlights challenges relevant to systems engineering, and overviews work in three distinct areas: the application of real time optimization in silicon production, the development of scale-up models for a fluidized bed poly-silicon process and a new process concept for silicon wafer production.

Introduction

Governments and industry invest in alternative energy sources such as solar-photovoltaics (PVs), wind, geothermal, hydroelectric, wave, tidal, nuclear, corn/cellulosic-ethanol, etc. to reduce dependence on non-renewable fossil energy sources which generate off gases that impact both the local and global environments (Jacobson, 2009). Fossil energy is also becoming more expensive to produce as new regulations are introduced due to the realization that the health and environmental costs of these technologies can be high; coal-based generation in the U.S. is set to become more expensive when mercury regulations enacted by the Bush administration in 2006 come into play (Feeley et al., 2009).

The sun supplies our planet with enormous amounts of energy (3 × 1024 J/year) and sunlight can, in principle, sustain all current and future energy needs. A significant amount is already utilized at very low conversion efficiency, about 1%, through the use of bio-fuels. Current estimates show that covering 0.1% of the earth’s surface with solar cells having an efficiency of 10% would satisfy our present energy needs (Gratzel, 2003, Gratzel, 2005). However, numerous disadvantages with current technology exist. Solar energy is very diffuse, having an intensity of 1000 W/m2 near the equator, so large areas are needed for major energy production. Also, the incident photon-to-electron conversion efficiency (IPCE) of mass produced solar cells is low, averaging 14–16% for a high quality cell (Luque & Hegedus, 2003). Production costs are too high as well, with the best being near $1.20 per peak Watt.

Since solar energy is so diffuse it is important that we find ways to integrate solar cells into buildings and the environment which are aesthetically pleasing and result in an improvement rather than an inconvenience in our daily lives. Additionally, we must improve existing technologies and develop new cell technologies to achieve higher conversion efficiencies. The theoretical maximum conversion for a silicon system is about 30% (Shockley & Queisser, 1961). Recent laboratory experiments have shown that quantum dots and/or inexpensive fluorescent dyes can be applied to further enhance this efficiency. The world record for the efficiency of a solar cell is about 50%, i.e. three times better than the typical industrial solar cell (Tisdale et al., 2010). Besides the need to increase efficiency, it is also necessary to reduce production costs.

The accumulated world solar cell capacity was 2.54 GW in 2006; 89.9% was based on mono- or multi-crystalline silicon wafer technology, 7.4% was thin film silicon, and 2.6% was direct wafering (Neuhaus & Munzer, 2007). The rapidly expanding market and high cost of silicon systems led to the development of thin-film technologies such as the cadmium telluride (CdTe), copper–indium–gallium selenide (CIGS and CIGSS) (Qijie, Ford, Hillhouse, & Agrawal, 2009), dye sensitized solar cells (Gratzel, 2005), amorphous Si on steel and many other. The market share for thin-film technology jumped to nearly 20% of the total 7.7 GW of photovoltaics production in 2009 (Cavallaro, 2010, G. van Sark et al., 2007).

There are now more than 25 types of solar cells and modules in current use (Green, Emery, Hishikawa, & Werta, 2010). Technology based on mono-crystalline and multi-crystalline silicon wafers presently dominates and will probably continue to dominate since raw material availability is not a problem given that silicon is abundant and cheap (Luque & Hegedus, 2003). Solar cells based on rare-earth metals pose a challenge since the costs of the raw materials tend to fluctuate and their availability is limited.

Fig. 1 shows the approximate distributions for the different costs in producing a silicon based solar module (Muller, Ghosh, Sonnenschein, & Woditsch, 2006). The figure illustrates areas where there are significant incentives to reduce costs. The areas of solar grade silicon (SOG) production and wafer manufacture stand out. These processes are presently not well optimized, and thus many opportunities exist to improve manufacturing technology through process innovation, retro-fitting, optimization and process control. This paper shows how process systems engineering tools can contribute in this direction in three distinct areas. Many challenges and opportunities exist that can provide further improvements.

These areas involve the use of a range of processes well suited for application of chemical engineering principles and systems engineering tools. In addition to covering the production of solar cells, we provide a summary of the fundamentals of solar cell operation in Appendix A. We have used an engineering approach based on thermodynamics and transport phenomena as opposed to the traditional physics approach. This is done to help engineers to have a better idea of the specifications for which they are designing materials, as well as to show potential areas for contributions from systems engineering in device design.

Section snippets

The production of silicon solar cells

The production of a typical silicon solar cell (Fig. 2) starts with the carbothermic reduction of silicates in an electric arc furnace. In this process large amounts of electrical energy break the silicon–oxygen bond in SiO2 via an endothermic reaction with carbon. Molten Si-metal with entrained impurities is withdrawn from the bottom of the furnace while CO2 and fine SiO2 particles escape with the flu-gas. A more detailed description of the process is given in the next section.

Metallurgical

Real time optimization of MG-silicon

Metallurgical grade silicon production is the first process step in the production of silicon solar cells. The process is very energy intensive and is performed in electrochemical reactors which have power-ratings from 5 to 40 MW. In this section we describe how real time optimization (RTO) can be used to improve the performance of a 30 MW metallurgical grade silicon reactor. RTO is a type of closed-loop process control that attempts to optimize process performance on-line and in real-time (

Production of poly-silicon

As shown in Fig. 2, the next step in solar cell production concerns the production of high purity poly-silicon. This energy intensive step removes impurities from the MG-Si so that the purity requirements for solar grade silicon can be met. MG-Si typically has 0.5% impurities by weight. This number must be reduced by another order of magnitude or better to be suitable for solar cells. It is possible to produce cells with less pure materials by compensating in the doping process, but this

Continuous production of crystalline silicon wafers

The last step in the supply chain we consider in this paper concerns how poly-silicon is converted to thin wafers suitable for cell production. The most common approaches in industry today rely on the growth of single crystalline silicon in the CZ process or the casting of multi-crystalline silicon in the Bridgman process with subsequent sawing using wiresaws. These processes have high capital and operating costs, they are carried out in small scale batch equipment and they are very wasteful,

Summary and conclusions

There has been explosive growth in the solar cell industry during the past decade. It is not likely that this technology will provide a complete answer to pressing energy and sustainability problems in the short run. However, it is expected that the industry will continue to grow at a rapid pace as new and cost effective processes come on-line in the near future. These processes will reduce cost in every stage of the supply chain. New technologies are being implemented which give higher

Acknowledgements

This work summarizes aspects of work carried out by many students at Carnegie Mellon University. Dr. Martin Ruszkowski developed the model and mass balance control system for the MG Silicon reactor. Dr. Jennifer H. Hill developed the estimation scheme and applied the real time optimization algorithm to the silicon reactor while Dr. Jin Wang worked on the implementation of the RTO to the real reactor. Dr. Christy White developed the population balance model and PDE reactor model which was used

References (84)

  • G. del Coso et al.

    Temperature homogeneity of polysilicon rods in a Siemens reactor

    Journal of Crystal Growth

    (2007)
  • T.J. Feeley et al.

    An update on doe’s phase ii and phase iii mercury control technology R&D program

    Fuel Processing Technology

    (2009)
  • T. Furusawa et al.

    Chemical vapor deposition and homogeneous nucleation in monosilane pyrolysis within interparticle spaces—Application of fines formation analysis to fluidized bed CVD

    Chemical Engineering Science

    (1988)
  • M. Gratzel

    Dye-sensitized solar cells

    Journal of Photochemistry and Photobiology C: Photochemistry Reviews

    (2003)
  • M. Guay et al.

    Adaptive extremum seeking control of continuous stirred tank bioreactors with unknown growth kinetics

    Automatica

    (2004)
  • M. Hamidi et al.

    Finite element analysis of the interface shape in the growth of rad silicon ribbons

    Journal of Crystal Growth

    (1987)
  • W.K. Ho et al.

    Optimal gain and phase margin tuning for PID controllers

    Automatica

    (1998)
  • L.Y. Huang et al.

    On the hot-zone design of Czozhralski silicon growth for photovoltaic applications

    Journal of Crystal Growth

    (2004)
  • S. Jana et al.

    A numerical simulation study for the Czochralski growth process of Si under magnetic field

    International Journal of Engineering Science

    (2006)
  • H.M. Jeong et al.

    Computational simulations of ribbon-growth on substrate for photovoltaic silicon wafer

    Journal of Crystal Growth

    (2010)
  • J.P. Kalejs

    Modeling contributions in commercialization of silicon ribbon growth from the melt

    Journal of Crystal Growth

    (2001)
  • H. Kasjanow et al.

    3D coupled electromagnetic and thermal modeling of efg silicon tube growth

    Journal of Crystal Growth

    (2007)
  • T. Kojima et al.

    Development of numerical model for reactions in fluidized bed grid zone: Application to chemical vapor deposition of polycrystalline silicon by monosilane pyrolysis

    Chemical Engineering Science

    (1990)
  • K.A. Kokh et al.

    Numerical modeling of melt flows in vertical Bridgman configuration affected by a rotating heat field

    Journal of Crystal Growth

    (2007)
  • S. Lai et al.

    Chemical vapor deposition and homogeneous nucleation in fluidized bed reactors: silicon from silane

    Chemical Engineering Science

    (1986)
  • C.W. Lan

    Recent progress of crystal growth modeling and growth control

    Chemical Engineering Science

    (2004)
  • P.T. Landsberg

    An introduction to the theory of photovoltaic cells

    Solid-State Electronics

    (1975)
  • J.H. Lee et al.

    Approximate dynamic programming based approach to process control and scheduling

    Computers & Chemical Engineering

    (2006)
  • H. Miyazawa et al.

    Numerical analysis of influence of crucible shape on interface shape in a unidirectional solidification process

    Journal of Crystal Growth

    (2008)
  • A. Muller et al.

    Silicon for photovoltaic application

    Materials Science and Engineering

    (2006)
  • M. Muller et al.

    Silicon efg process development by multi-scale modeling

    Journal of Crystal Growth

    (2010)
  • O.V. Smirnova et al.

    3D unsteady analysis of melt flow and segregation during efg Si crystal growth

    Journal of Crystal Growth

    (2008)
  • P. Sonda et al.

    Hopf bifurcation and solution multiplicity in a model for destabilized Bridgman crystal growth

    Chemical Engineering Science

    (2005)
  • W.G.J.H.M. van Sark et al.

    Analysis of silicon market: will thin-film profit?

    Energy Policy

    (2007)
  • D. Vizman et al.

    3D-time dependent numerical study of the influence of the melt flow on the interface shape in a silicon ingot casting process

    Journal of Crystal Growth

    (2007)
  • V.R. Voller et al.

    A fixed grid numerical modeling methodology for convection-diffusion mushy region phase-change problems

    International Journal of Heat and Mass Transfer

    (1987)
  • R.L. Wallace et al.

    Thin silicon string ribbon

    Solar Energy Materials and Solar Cells

    (1997)
  • C.M. White et al.

    Size distribution modeling for fluidized bed solar-grade silicon production

    Powder Technology

    (2006)
  • P. Woditsch et al.

    Solar grade silicon feedstock supply for PV industry

    Solar Energy Materials & Solar Cells

    (2002)
  • Y. Zhang et al.

    Real-time optimization under parametric uncertainty: A probability constrained approach

    Journal of Process Control

    (2002)
  • B.L. Anderson et al.

    Fundamentals of semiconductor devices

    (2005)
  • Arora, N. (2003). Nonlinear programming for data reconciliation and parameter estimation. PhD Dissertation, Pittsburgh,...
  • Cited by (95)

    • CFD simulation study of the effect of baffles on the fluidized bed for hydrogenation of silicon tetrachloride

      2022, Chinese Journal of Chemical Engineering
      Citation Excerpt :

      Unfortunately, enormous amount of silicon tetrachloride (SiCl4) is generated as a major byproduct through both two methods. And for both environmental and economic reasons, SiCl4 needs to be recycled and converted to the valuable trichlorosilane (SiHCl3) which is the starting material in a closed-loop polysilicon production process [1–3]. Nowadays, the fluidized beds have been diffusely used in the chemical industry, including catalytic cracking, methanol to olefins, polysilicon production, and biomass gasification [3,8,9], because of their excellent gas–solids mixing, heat and mass transfer performance.

    View all citing articles on Scopus

    Research supported by the National Science Foundation CBET 0932556; REC Silicon Moses, Lake, WA, Elkem ASA, Oslo Norway; Industrial Learning Systems Inc., Pittsburgh, PA.

    View full text