Summary 'Environmental life-cycle assessment of multicrystalline silicon solar cell modules' (Report number 95057)

Phylipsen, G.J.M., E.A. Alsema, Environmental life-cycle assessment of multicrystalline silicon solar cell modules, Department of Science, Technology and Society, Utrecht University, Utrecht (no. 95057).

Abstract

In this report the environmental aspects of solar cell modules based on multicrystalline silicon are investigated by means of the Environmental Life Cycle Assessment method. Three technology cases are distinguished, namely present-day module production technology, future probable technology and future optimistic technology.

For these three cases the production technology is described, the material requirements and environmental emissions are inventarised and the energy requirements and energy pay-back times are discussed.

Finally recommendations with respect to Dutch photovoltaic R&D policy are given.

Summary and conclusions

Introduction

In this report we have described an environmental life-cycle assessment of multicrystalline silicon solar cell modules. The assessment was mainly focused on energy and material flows during the production of the PV modules. In order to be able to identify opportunities and possible bottlenecks in future multicrystalline silicon PV development we discerned three cases. The worst case represents an estimate of the present state of production technology. The base case is based on technologies which are most probably to be commercially available within 10 years. The best case represents an optimistic view on production technology available in 10-15 years.

In this chapter we summarize the conclusions of the energy and material flow analyses presented in the previous chapters. Based on these conclusions we give recommendations for the R&D policy in the Netherlands with regard to m-Si solar cells.

Case descriptions

In chapter 2 a schematic representation of the multicrystalline silicon solar cell studied in our report is given. We assume an improving cell efficiency from worst to best case of 13%, 16% and 18% respectively, that is to be achieved by introducing new technologies and solar cell features. In table 5.1 an overview is given of the most important differences between the cases. A more elaborate description of the discerned cases can be found in chapter 2.

The life cycle of a multicrystalline silicon PV module starts with the mining and refining of silica (quartz). Silica is reduced with carbon and the reduction step is either followed (in worst and base case) or preceded (in the best case) by a purification step. The resulting high purity silicon is melted and cast into blocks of multicrystalline silicon. The blocks are portioned into ingots, which are subsequently sliced into wafers. The wafers are processed into solar cells by etching, texturing, formation of the emitter layer, application of back surface layer and contacts and passivation and antireflective coating. The solar cells are tested, interconnected and subsequently encapsulated and framed into modules. The application of a back surface layer and the passivation step are omitted in the best case.

The general trend in the expected future developments is towards improved energy and material efficiency. This can be seen in higher process yields for high purity silicon production, casting, portioning and material production, in the usage of thinner wafers, in lowering of the metal coverage factor in contact formation, in the reduction of contouring and wafering losses and in the reduction of process energy requirements.

The most influential differences regarding energy and material requirements are the usage of thinner and larger wafers, reducing portioning and wafering losses in base and best case, and the development of a production process for solar grade silicon in the best case.

Table 1: The most important differences in used technologies and process parameters between the discerned cases.

Material analysis

Material requirements and resource depletion

If we compare the three cases to each other with regard to material requirements, we can see the influence of a decreasing wafer thickness (especially on quartz and carbon requirements), increasing wafer size (on SiC and Cu requirements) and increasing module size (on polyester and aluminum requirements).

Dominating material requirements are glass, EVA and aluminum. Other materials that are used in relatively large amounts are the input materials for mg-silicon production (i.c. quartz and carbon sources) and for wafering (mineral oil and silicon carbide). A point of attention in the best case is the large amount of HCl used in producing high purity carbon and high purity silicon. The used HCl is neutralized with Ca(OH)₂, and CaCl₂ is discharged as solid waste.

Quartz sand is a widely available material, so no problems with raw material supply are expected, even if solar cell technology is implemented on a large scale. Also with other materials no immediate problems with regard to resource depletion are expected, although the reduction of silver consumption for the contacts deserves attention.

Silver is considered as a scarce resource and 30% of the current annual silver production would be required to achieve a 5% contribution to the world electricity supply by base case modules. A second reason for reduction of silver use arises from possible waste management problems when modules containing almost 5 g of silver per kg of waste are decommissioned. Moreover, it seems probable the reduction of silver in the contacts will also be necessary from a cost perspective before large-scale introduction of m-Si modules becomes viable.

For other, more abundant materials in the module recycling can be of interest from an energy point of view. Recycling technologies for reusing silicon from solar cells (from production waste or after module decommissioning) are not (yet) commercially available. Also for the glass and EVA recycling is difficult. The EVA foil is per definition not recyclable and the glass waste from modules may contain too much plastic to be accepted by glass recyclers. Possibly a pretreatment of the glass could suffice in removing EVA remains. The aluminum module frame seems to be the only module component which is easily recyclable.

In any case it seems necessary to reconsider the overall module design to enhance recyclability while maintaining its durability in outdoor conditions.

Process emissions

Emissions in the PV module's life cycle are at this moment largely limited to the production phase. Environmentally relevant substances which may be released in multicrystalline silicon PV module production are fluorine, chlorine, nitrate, isopropanol, SO₂, CO₂, respirable silica particles and solvents.

Fluorine and chlorine may be emitted to the air as a component of dust particles by the best case silicon purification technology. The estimated air emission is maximally 0.16 kg F and 430 kg Cl per TWhe of electricity supplied by PV modules, which is orders of magnitude smaller than the corresponding emissions of a coal plant.

Fluorine and chlorine are also emitted to the water in all three cases (1,800 kg F and 89,000 kg Cl per TWhe in the base case), resulting from neutralization of etching and texturing solutions and flue gases. Fluorine and chlorine contribute to the human toxicity, as does nitrate, which stems from neutralizing acids used in etching and texturing. Water-borne F- and Cl-emissions of base case PV technology are significant but still 3-5 times smaller than for a coal plant.

The non-energy-related 1 The emissions of SO₂, NO_x and CO₂ can be distinguished into energy-related, i.e. resulting from energy use, and non-energy-related emissions, i.e. resulting from the production process itself. emissions of SO₂ (in worst and base case) are caused by using sulphur-containing carbon sources in the reduction of silica. These carbon sources are also responsible for the non-energy-related emissions of CO₂. However, the non-energy-related SO₂ and CO₂ emissions are small compared to the energy-related emissions of these gases.

Silica particles can be released in the mining and refining stage. If they are small enough to be inhaled they may cause the lung disease silicosis.

Emissions of solvents and alcohols contribute to photochemical ozone formation and both direct (the solvents itself) and indirect (ozone) respiratory problems.

Comparing the three cases to each other with regard to process emissions the influence of decreasing wafer thickness, increasing wafer size and use of a different process for producing high purity silicon is clear (especially for CO₂, SO₂ and Si powder emissions, SiC, mineral oil). Also the use of different etching, texturing and purification methods has a noticeable effect on emissions (of CaCl₂, NaCl and KCl).

Occupational health

A number of substances are considered to pose acute and/or chronic hazards on the work force in PV industry, e.g. etchants, acids, solvents etc. Most of these materials, however, are not specific for the multicrystalline silicon PV industry. We expect the hazards of these substances to be controllable by the safety measures usually employed in chemical or semiconductor industries.

Incidental releases could result in the presence of more or less hazardous substances (e.g. silane, carbon monoxide, ammonia and silica particles) on the work floor. Safety management will have to be sufficient to take care of these risks.

Energy analysis

Energy requirement

The gross energy requirement for a multicrystalline silicon PV module (without a frame) is estimated at 400 kWht per m² cell area in the base case (for worst and best case this is 970 and 180 kWht/m² cell area respectively). It should be noted that this rather low energy requirement is partly due to our choice for the UCC process for silicon purification, which is not a commonly used method at present. Use of the Siemens process, for example, may increase the GER of a multicrystalline PV module with about 20% in the worst and base case (the best case incorporates an entirely different approach for Si purification).

Under Dutch insolation conditions the energy pay-back time of a frameless module as described in the base case is estimated at 1.3 years (3.8 resp. 0.5 yr for worst and best case) With the Siemens route the energy pay back time would have been 4.6 yr in the worst case.. Addition of an aluminum frame increases the energy pay-back time with 0.4 yr in the base case.

Table 2: Gross energy requirement for multicrystalline silicon PV modules in kWht/m² cell area.

The direct process energy requirement is about half of the total required energy. The direct process energy of the cases is heavily influenced by our choice of the production route for high purity silicon. From worst to best case direct process energy for this step is expected to decrease from 250 to 23 kWht per m² cell area. Another important factor is the increasing scale in cell processing, which causes direct process energy to drop from 162 kWht/m² in the worst case to 19 kWht/m² in the best case. However the decrease in both steps is also partly due to the implementation of thinner and larger wafers. Using thinner wafers results in a lower Si requirement per m², so less high purity silicon is required for producing 1 m² of PV module (cell area). Larger wafers result in less wafers per m² cell area, so less process energy will be needed for producing 1 m² of cell area.

The amount of energy required for material production is relatively low (excluding the aluminum frame), but its relative importance is increasing from worst to best case (from 9 to 28%). Using an aluminum frame adds substantially to the gross energy requirement (18%, 30% and 44% in worst, base and best case respectively). If, however, secondary aluminum is used for framing the gross energy requirement of the frame can be reduced substantially (with up to 80%). The energy requirement for material production in the best case is about halved compared to the worst case. This is mainly due to a lower carbon requirement for silica reduction and using thinner EVA foil in the best case.

The indirect energy requirement and the energy required for investments have a significant contribution to the l energy requirement (about 20% each). These last two contributions however have a large uncertainty and depend heavily on plant lay-out, etc.

The indirect process energy requirement and the energy requirement for investments is expected to decrease from worst to best case due to increasing batch size, a more continuous way of plant operation and lower energy requirement in equipment manufacturing.

Primary energy requirement and energy related emissions

The amount of primary energy carriers necessary for producing a certain amount of electricity with base case multicrystalline silicon PV modules is estimated to be about 20 times lower than for the present Dutch electricity supply system (see table 5.3). Also for CO₂ the base case emissions are about 20 x lower than for conventional electricity generation. The non-energy-related CO₂ emissions are about 2% of the energy-related emissions.

Table 3: Energy-related emissions and primary energy requirements for electricity production with m-Si PV modules compared to the average Dutch electricity supply system in kg/GWhe. Also the accumulated acid emission (Acidification Potential) is given in SO₂-equivalent kg's.

The modules are installed in the Netherlands, insolation = 1000 kWh/m².yr).

Energy-related acid emissions for the base case are 16x less with PV electricity generation than with the average Dutch electricity supply system. The non-energy-related releases of SO₂ are about 5% of the energy-related emissions for base case. Although sulphur emissions are substantially lower than for conventional electricity generation, for an energy source considered to be sustainable they are rather high, especially in the worst case.

The energy requirement and emissions per amount of electricity generated calculated in this study are in the same order of magnitude as those found for amorphous silicon modules in a previous study [5]. This is caused by a 2 times higher solar cell conversion efficiency for m-Si modules which compensates the 2 times higher energy requirement.

Policy recommendations

In view of the points of attention that were identified in the environmental assessment of multicrystalline solar cell modules, the following recommendations can be formulated:

• The possibilities for recycling silicon wafers, both from production waste (rejected cells) and from decommissioned modules, should be investigated;

• Possibilities for using secondary aluminum for producing module frames should be assessed;

• Methods should be investigated to reduce the emission of fluorides in water, but also the emission of fluorides as solid waste. Alternative methods to etch the emitter without use of HF etchant should be considered;

• The CF₄ gas which is used in the plasma etching process has a large Global Warming Potential. Therefore routine or incidental emissions of this gas should be avoided;

• Emissions of volatile solvents and alcohols should be avoided;

• The possibilities for reuse of silicon carbides (sawing slurry) and argon should be given attention;

• The use of silver for contacts should be reduced to avoid problems with resource availability and problems with the management of waste from decommissioned modules.

Important options for reduction of the energy requirement of the modules are:

• the use of thinner wafers;

• the use of wafers with a larger area;

• the use of secondary aluminum for framing;

• the introduction of processes for solar grade silicon;

• monitoring and optimization of cell and module production processes with regard to energy consumption.

The first two options are largely in line with current R&D activities, so no special efforts are necessary.

Optimization of energy consumption in cell and module production may be possible by increasing batch sizes and/or plant shifts, introduction of new equipment (like IR ovens) and energy-efficient operation of equipment.

Further points of attention, which however lie outside the scope of Dutch R&D policy with regard to photovoltaic energy, are:

• A reduction in SO₂ emissions should be achieved by using low sulphur fuels or desulphurization of flue gases, especially in the fuel-intensive processes, such as material and feedstock production and purification.

• The introduction of a process for solar grade silicon is considered as a good option for reducing the energy requirements. However, alternatives should be considered for the use of HCl for purification of the silica and carbon input materials for the solar-grade Si process considered in this study.

This publication is available as PDF file. Click here to download.

(Tip: you can save the pdf to directly your hard disk by clicking the link with the right-hand mouse button and choosing Save target as... / Save link as... from the context menu)

Last update:
For any question or suggestion, please feel free to contact the Webmaster.