**7. Life cycle of photovoltaic systems and ecological footprint**

As a result of the accelerated growth of the PV industry, a rigorous assessment of the environmental impacts of the systems has become necessary, conducted through a life cycle assessment (LCA) integrating all of the manufacturing, operating, collection and waste recycles. The LCA is an orderly process that analyzes the input/output impact of the PV industry from the "cradle to grave", with the inputs referring to the materials and energy consumed and the outputs illustrated by greenhouse gas (GHG) type emissions and solid and liquid waste.

A form of environmental management that is as exhaustive as possible, the LCA is a series of tools and techniques for which the ultimate objective, beyond the descriptive and quantitative aspect of the environmental profile, is to reinforce the sustained effort to limit the environmental impacts in a context of sustainable development (Fthenakis et al., 2005a; IEA-PVPS, 2011). The key factor that will determine the pertinence and the credibility of the LCA will be the voluntary and transparent cooperation of the manufacturers with respect to the accurate and full disclosure of the various inputs/outputs (Fthenakis et al., 2005a; Stoppato, 2008; IEA-PVPS, 2009; Ecoinvent , 2010; IEA-PVPS, 2011).

In addition to the energy considerations previously illustrated by the calculation of the EPBT and the ERF, the parameter most frequently estimated for the LCA assessment is the ecological footprint describing and quantifying the entire greenhouse gases (GHG) released during the lifespan of the PV system and expressed in carbon dioxide equivalents per kWh. The environmental gain expected by the reduction of GHG related to the operation of PV electricity has also to be taking into account. These two assessments are always determined in comparison with the emissions attributed to fossil energies (Fthenakis et al., 2005a; IEA-PVPS, 2011).

The estimate of the GHG attributed to PV systems is an increasingly complex exercise since it includes criteria that are as diversified as the technology used, the choice of manufacturing processes and the type of energies consumed, the techniques for assembling the cells and units, the power generated, the transportation of raw materials and the finished product, the components required for the installation of the units (Balance Of System/BOS) as well as the recycling processes. The BOS will, in turn, depend on the applications, the dimensions, the orientations and, above all, the location selected (Krauter & Rüther, 2004; Stoppato, 2008; Fthenakis & Kim, 2011; Reich et al., 2011).

A major distinction is acknowledged between indirect emissions, which concern the overall energy, electricity included, needed to manufacture the units, and the direct emissions, which concern all of the chemical compounds, raw materials included, that are involved in the manufacturing process and are a potential source of GHG (Reich et al., 2011). It is essential to point out that the GHG emission estimates for PV systems are not absolute since they are subject to a certain number of constraints, particularly the quality of the information provided by the manufacturers involved throughout the life cycle. Thus, the estimates are subject to future revisions as are the EPBT and ERF calculations (Reich et al., 2011; Held & Ilg, 2011).

With respect to the direct emissions of the silicon industry, three critical phases are identified: the development of metallurgical-grade silicon from silica, its transformation into solar-grade silicon and the development of a structure and framework in the form of panels. While the production of metallurgical-grade silicon requires the consumption of roughly 14 kWh per kg of metallurgical-grade silicon whole releasing 3 tons of CO2 equivalents for one

Photovoltaic Conversion: Outlook at the Crossroads

countries (Stoppato, 2008; EPIA/Greenpeace, 2011).

Natural Gas 360-991

(EPIA/Greenpeace, 2011).

implemented (Krauter, 2006).

**8. Potential health effects** 

Between Technological Challenges and Eco-Strategic Issues 325

Notwithstanding this disparate data, the GHG emissions of PV systems are well below those of fossil energies as summarised in table 2. The overall production of electricity, all energy sources combined, generates an average of 600 g CO2-eq/kWh, although this varies between

Energy system Average emission Reference

Coal 800 – 1280 Dones et al., 2003 ; Weisser,

Oil 519-1200 Dones et al., 2003; Weisser,

Table 2. GHG emissions (g CO2-eq/kWh) resulting from different energy systems

Moreover, the ecological footprint may be evaluated in terms of environmental gains resulting from the expected reduction in GHG caused by the use of PV systems. Based on the principle that, once installed, the systems (with the exception of the diesel generators and the transportation of maintenance services) do not emit GHG, it is possible to calculate how many CO2-eq will be saved throughout their lifespan. Scenarios have been developed to extrapolate this reduction with a forecast of 0.6 kg CO2-eq/kWh on average saved by the extension of the systems connected to the network and taking into account emissions (optimistic) of 12-25 g CO2-eq/kWh. Almost 4 billion tons of CO2-eq could be saved by 2050

As for the autonomous systems, 70% would experience a reduction of more than 200 kg of CO2-eq per year, namely 6 tons of CO2-eq and 8.9 tons of CO2-eq for 15 and 50 Wp systems respectively with a lifespan of more than 20 years (Posorski et al., 2003). The projections go even further, considering that the implementation of PV plants in developing countries, combined with a generalization of systems connected to the network in order to supplement the hybrid systems and reduce emissions related to the transfer of technology from the supplier country to the consumer country would be even more beneficial in ecological terms with more than 26 tons of CO2-eq/kWh saved per site

The photovoltaic industry, with its ambitious goal to provide clean electricity, paradoxically uses materials and/or manufacturing processes that are not free from inherent potential health and safety effects. The sector is therefore facing a dual objective: increase energy efficiency and reduce or even abandon processes that use potentially toxic compounds.

2007; Evans; 2010

2007; Evans; 2010

Dones et al., 2003; Jaramillo et al., 2007; Weisser, 2007; Evans; 2010

PV Systems 15-187 See references in Table 1

ton of metallurgical-grade silicon, the solar-grade silicon stage is by far the most energy consuming, with 150 kWh per kg obtained (Miquel, 2009) or 1190 MJ/panel (0.65 m2) (Stoppato, 2008).

Assembling the panels with an aluminum frame also consumes energy, ranging between 53 and 245 kWh with emissions varying between 15 and 19 kg CO2-eq, all per kg of aluminum consumed (Krauter & Rüther, 2004). Overall, the estimation of GHG emissions for silicon panel manufacturing is variable as shown in table 1.

The silicon technologies release also GHG directly, with the primary sources being the raw material itself, the various fluoride compounds involved in the manufacturing process as well as the incineration of the plastic used to encapsulate the solar cells, one of the common processes in the recycling of plastic materials. According to the estimates, the emission is virtually negligible, about 0.16 g CO2-eq/kWh for the raw material, whereas the incineration of plastic would be a source of 1.1 g (Reich et al., 2011).


Table 1. GHG emissions for silicon panel according to different authors

The fluoride compounds remain the Achilles heel of silicon cells since they have an even higher Global Warming Potential (GWP). CO2, methane and the nitrogen oxides have GWPs of 1, 23 and 296. There are also issues with respect to CF4 (carbon tetrafluoride), SF6 (sulphur hexafluoride), C2F6 (hexafluoroethane) and above all NF3 (nitrogen trifluoride) for which the GWPs range from 7,400 to more than 17,000 (Fthenakis et al., 2010; Miquel, 2009). Despite this fact, these fluoride compounds, excluding SF6, are not included in the Kyoto protocol whereas NF3 is considered to be the gas with a significant environmental impact (Prather & Hsu, 2008).

Concerning the thin layer technology and, more specifically, the cadmium telluride (CdTe) technology, the small amount of data available relies on a certain number of parameters such as the geographic location of the facility, the conditions at the site of the installation and, certainly, the type of databases used. The information about the recycling procedures has a particular impact on the calculation, as for all of the technologies, but the recycling process is in the experimental stage since the CdTe market is still relatively young (Held & Ilg, 2011). From 18 to 20 g CO2-eq/kWh (Fthenakis & Kim, 2005; Fthenakis, 2009) the estimates are currently being revised slightly upwards (Held & Ilg, 2011).

The autonomous PV systems include, in their calculations, the emissions generated by the storage batteries and eventually those caused by the diesel generators integrated in most of the hybrid systems. Taking into account the 1.26 kg CO2-eq released per kg of batteries produced, the cost of transportation and maintenance, and based on an operating life of more than 20 years, the individual systems, namely solar home systems (SHS), with a power of 15 Wp release an average of 160 kg CO2-eq whereas SHS with a power of 50 Wp release 650 kg (Posorski et al., 2003).

ton of metallurgical-grade silicon, the solar-grade silicon stage is by far the most energy consuming, with 150 kWh per kg obtained (Miquel, 2009) or 1190 MJ/panel (0.65 m2)

Assembling the panels with an aluminum frame also consumes energy, ranging between 53 and 245 kWh with emissions varying between 15 and 19 kg CO2-eq, all per kg of aluminum consumed (Krauter & Rüther, 2004). Overall, the estimation of GHG emissions for silicon

The silicon technologies release also GHG directly, with the primary sources being the raw material itself, the various fluoride compounds involved in the manufacturing process as well as the incineration of the plastic used to encapsulate the solar cells, one of the common processes in the recycling of plastic materials. According to the estimates, the emission is virtually negligible, about 0.16 g CO2-eq/kWh for the raw material, whereas the incineration

15-25 EPIA/Greenpeace, 2011

The fluoride compounds remain the Achilles heel of silicon cells since they have an even higher Global Warming Potential (GWP). CO2, methane and the nitrogen oxides have GWPs of 1, 23 and 296. There are also issues with respect to CF4 (carbon tetrafluoride), SF6 (sulphur hexafluoride), C2F6 (hexafluoroethane) and above all NF3 (nitrogen trifluoride) for which the GWPs range from 7,400 to more than 17,000 (Fthenakis et al., 2010; Miquel, 2009). Despite this fact, these fluoride compounds, excluding SF6, are not included in the Kyoto protocol whereas NF3 is considered to be the gas with a significant environmental impact (Prather &

Concerning the thin layer technology and, more specifically, the cadmium telluride (CdTe) technology, the small amount of data available relies on a certain number of parameters such as the geographic location of the facility, the conditions at the site of the installation and, certainly, the type of databases used. The information about the recycling procedures has a particular impact on the calculation, as for all of the technologies, but the recycling process is in the experimental stage since the CdTe market is still relatively young (Held & Ilg, 2011). From 18 to 20 g CO2-eq/kWh (Fthenakis & Kim, 2005; Fthenakis, 2009) the estimates are currently being revised slightly upwards (Held & Ilg,

The autonomous PV systems include, in their calculations, the emissions generated by the storage batteries and eventually those caused by the diesel generators integrated in most of the hybrid systems. Taking into account the 1.26 kg CO2-eq released per kg of batteries produced, the cost of transportation and maintenance, and based on an operating life of more than 20 years, the individual systems, namely solar home systems (SHS), with a power of 15 Wp release an average of 160 kg CO2-eq whereas SHS with a power of 50 Wp release

148-187 Stoppato, 2008

43-73 Weisser, 2007; Miquel, 2009

30-45 Fthenakis & Alsema, 2006; Fthenakis et al., 2008

(Stoppato, 2008).

Hsu, 2008).

2011).

650 kg (Posorski et al., 2003).

panel manufacturing is variable as shown in table 1.

of plastic would be a source of 1.1 g (Reich et al., 2011).

Emission estimates

(g CO2-eq/kWh) Reference

Table 1. GHG emissions for silicon panel according to different authors

Notwithstanding this disparate data, the GHG emissions of PV systems are well below those of fossil energies as summarised in table 2. The overall production of electricity, all energy sources combined, generates an average of 600 g CO2-eq/kWh, although this varies between countries (Stoppato, 2008; EPIA/Greenpeace, 2011).


Table 2. GHG emissions (g CO2-eq/kWh) resulting from different energy systems

Moreover, the ecological footprint may be evaluated in terms of environmental gains resulting from the expected reduction in GHG caused by the use of PV systems. Based on the principle that, once installed, the systems (with the exception of the diesel generators and the transportation of maintenance services) do not emit GHG, it is possible to calculate how many CO2-eq will be saved throughout their lifespan. Scenarios have been developed to extrapolate this reduction with a forecast of 0.6 kg CO2-eq/kWh on average saved by the extension of the systems connected to the network and taking into account emissions (optimistic) of 12-25 g CO2-eq/kWh. Almost 4 billion tons of CO2-eq could be saved by 2050 (EPIA/Greenpeace, 2011).

As for the autonomous systems, 70% would experience a reduction of more than 200 kg of CO2-eq per year, namely 6 tons of CO2-eq and 8.9 tons of CO2-eq for 15 and 50 Wp systems respectively with a lifespan of more than 20 years (Posorski et al., 2003). The projections go even further, considering that the implementation of PV plants in developing countries, combined with a generalization of systems connected to the network in order to supplement the hybrid systems and reduce emissions related to the transfer of technology from the supplier country to the consumer country would be even more beneficial in ecological terms with more than 26 tons of CO2-eq/kWh saved per site implemented (Krauter, 2006).
