Energy Pay-Back Time (EPBT) and CO₂ mitigation potential

Session 3
at
Environmental Aspects of PV Power Systems
IEA PVPS Task 1 Workshop
June 25-27, 1997
Utrecht, The Netherlands

Editors:
Evert Nieuwlaar
Erik Alsema

General

Although PV power systems do not require finite energy sources (fossil, nuclear) during their operation, a considerable amount of energy is needed for their production. The environmental issues associated with this energy use for PV manufacturing will also affect the environmental profile of PV Power systems. The environmental themes that are strongly related to the energy system are: exhaustion of finite energy carriers, climate change and acidification. One may consider using energy performance (e.g. Energy Pay Back Time) as an indicator for the environmental stress caused by PV power systems. Such indicators are strong regarding the exhaustion of finite energy sources, reasonably strong regarding climate change and acidification and weak or failing regarding themes like toxicity.

Unfortunately, it seems to be a popular belief that PV systems cannot ‘pay back’ their energy investment. Therefore, it is important to investigate this issue on the basis of solid data.

The Energy Pay Back Time is defined by EPBT = E_input/E_saved, where E_input is the energy input during the module life cycle (which includes the energy requirement for manufacturing, installation, energy use during operation, and energy needed for decommissioning) and E_saved the annual energy savings due to electricity generated by the PV module. For PV power systems the EPBT depends on a number of factors: cell technology, type of encapsulation, frame and array support, module size & efficiency, PV system application type (autonomous or grid-connected) and, finally, PV system performance as determined by irradiation and the performance ratio. EPBT is also affected by factors that do not directly relate to the characteristics of the PV power system itself: conversion efficiency of the electricity supply system and energy requirements of materials like glass, aluminum etc.

In his review of energy analysis studies on thin-film (a-Si and CdTe) PV modules presented at the workshop, Alsema (appendix B-6) showed that

• the EPTB of frameless thin film modules is below 2 years for present-day technology (at 1700 kWh/m²/yr irradiation)
• encapsulation materials and direct processing energy form the major part of the energy use
• a frame may add up to 1.6 years to the EPTB
• in the future EPTB of less than 1 year is feasible for frameless thin film modules

Kato (appendix B-8) presented an overview of the work done in Japan on c-Si, mc-Si and a-Si based rooftop systems. The analysis included the Balance of System (supporting structure & power conditioner). The results for (current state-of-the-art) monocrystalline-Si based systems depend on the choice made regarding allocation of energy to off-grade silicon from the semiconductor industry. If off-grade silicon is treated in the same way as silicon used in the semiconductor industry, the EPBT of the rooftop system would be 15.6 years. If no energy use is allocated to off-grade silicon, the EPBT would be 3.9 years (at 1430 kWh/m²/yr irradiation, Performance Ratio of 0.81).

Kato also considered future module production technology based on a dedicated solar-grade silicon process in combination with electromagnetic casting of mc-Si ingots. For this type of production technology the EPBT of the rooftop system was estimated at about 2 years. For systems based on a-Si modules too, an EPBT of about 2 years was found.

The analyses by Kato show that the EPBT of present-day crystalline silicon modules is affected very strongly by its dependency on silicon feedstock which was originally prepared for the electronics industry. Because the (energy) costs of silicon are virtually neglectable in the electronics industry’s products, this situation will improve only when the PV market is large enough to sustain dedicated production processes for Si feedstock (‘solar-grade Si’).

On the positive side one can remark that in the context of future energy supply scenarios the bottleneck of electronic-grade instead of solar-grade Si feedstock will most probably have been passed by the time the energy impact of PV power systems becomes really relevant.

Furthermore one should note that in other presentations at the workshop (e.g. Baumann, Frankl) as well as in another recent publication [Nijs et al., 1997] lower values for the energy requirement of present-day monocrystalline silicon modules were presented, leading to system EPBT values ranging from 5 to 10 years (under the same conditions as Kato’s systems). The reasons for these different results have been clarified to some extent during the workshop, but nonetheless we have to conclude that a clear understanding of the energy requirements of present-day crystalline silicon modules still lacks. In itself this would not be such a problem, if did not hinder a good insight in the future energy balance of c-Si modules too. Therefore our opinion is that the issue of the energy requirements of present-day and future crystalline silicon modules, should still be regarded as a ‘white spot’. In this context a further clarification of the impact from different process routes for Si feedstock production is also needed.

System aspects like Balance-of-System components, autonomous or grid-connected systems, building integration, and energy demand management options strongly influence the results of EPBT evaluations. These aspects are discussed in the session on system aspects (session 5).

It has to be remarked, however, that the energy payback times of autonomous PV systems have not been addressed specifically during this workshop. So all remarks and conclusions given here relate only to grid-connected PV systems.

The energy payback time as an indicator of energy performance has an appeal because of their similarity with economic payback times. A drawback of EPBT is that it does not account for the energy gain during the rest of the economic lifetime. The workshop expressed a desire for an indicator that combines EPBT with economic lifetime. An indicator that fulfills this requirement is the energy return factor (ERF) which expresses the total amount of energy saved per unit invested energy. The formula resembles the one for energy payback time: ERF = (E_saved * LT) / E_input, where LT represents the economic lifetime. A disadvantage of the ERF indicator is that it is not additive, i.e. ERF values of different system components cannot be added to obtain the ERF of the total system.

An approach similar to EPBT can be used to determine CO₂ pay back times as a measure for the climate change mitigation potential associated with PV power systems. Alternatively, cumulative CO₂ emissions are recorded per kWh in order to compare them with CO₂ emissions from alternative power production technologies. For a large part the CO₂ emissions originate from the use of fossil energy carriers in the life cycle of the PV power systems. In addition to these energy-related emissions, however, other CO₂ emissions occur. Examples are the CO₂ emissions caused by the silica reduction process and the CO₂ emissions from the consumption of carbon electrodes in aluminum production.

Greenhouse gas emissions other than CO₂ should also be considered since some of them have a large Global Warming Potential relative to CO₂, which make that small emissions of those gases can have a significant contribution to the total Global Warming Equivalent as expressed in equivalent CO₂ emission. Examples of such substances are SF₆ or CF₄, gases which may be used in plasma etching processes or in the cleaning of reactor chambers. Release to the atmosphere of 1 kg of these gases will cause a greenhouse effect equivalent to 24,000 respectively 6,500 kg of CO₂ [IPCC, 1996]. Alternative cleaning methods and other techniques under development within the semiconductor industry will help to avoid these emissions.

The results presented by Kato (appendix B-8) showed that CO₂ emissions for silicon-based rooftop PV power systems in Japan are less than 20 g-C/kWh, except for c-Si when CO₂ emissions from Si material production are fully included. Compared with an average of 126 g-C/kWh for the average electrical output of the Japanese utilities, a significant potential for CO₂ emission reduction exists.

The study presented by Inaba (appendix B-7; Komiyama et al., 1996; Tahara et al., 1997) showed that the choice of system boundaries is of large significance especially when the manufacturing and the installation of modules are performed in different countries.

We have seen that comparison of energy/CO₂ analysis studies is often unnecessarily difficult because of differences and unclarities with regard to the methodological approach and the reporting format (also see paper B-6, section 2 and paper B-7).

Therefore we recommend to:

• aim for more clarity on:

* system boundaries (including the way in which end-of-life disposal is treated);

* module encapsulation and framing;

* the evaluation of indirect processing energy;

* Gross Energy Requirements of input materials;

* allocation schemes used in the calculations

• Express energy requirements

* on the basis of module area;

* separately for thermal energy, electrical energy and "material energy",

* as equivalent primary energy units;

Conclusion

A final conclusion from this session is that PV technology definitively offers a significant potential for energy savings and CO₂ mitigation. Although the energy payback time and the CO₂ payback time for present-day systems is still relatively high, especially for crystalline silicon modules, it is generally lower than their expected life time.

Most important, however, is that it seems feasible to achieve a future decrease of the energy/CO₂ payback time for grid-connected PV systems to two years or less in case of c-Si modules and to one year or less for thin film modules(under 1700 kWh/m²/yr irradiation).

- enhance the energy efficiency in PV manufacturing, especially in Si feedstock production;

- avoid the use of fully fluorinated compounds such as SF6 and CF₄ in PV module production.

Alsema, E. A. (1996). Environmental Aspects of Solar Cell Modules, Summary Report, Department of Science, Technology and Society, Utrecht University.

Alsema, E. A., M. Patterson, et al. (1997). Health, Safety and Environmental Issues for Thin-Film Modules. 14th European Photovoltaic Solar Energy Conf., Barcelona.

IPCC (1996). Climate Change 1995, Second Assessment report of the Intergovernmental Panel on Climate Change, Cambridge University Press.

Nijs, J., R. Mertens, et al. (1997). Energy payback time of crystalline silicon solar modules. in: Advances in Solar Energy, Vol. 11, K. W. Boer (Ed.), American Solar Energy Society, Boulder, CO, pp. 291-327.

Steinberger, H. (1996). Umwelt- und Gesundheitsauswirkungen der Herstellung und Anwendung sowie Entsorgung von Dünnschichtsolarzellen und Modulen. München, Fraunhofer Institut für Festkörpertechnologie.

H. Komiyama, K. Yamada, A. Inaba and K. Kato, 1996. Life Cycle Analysis of Solar Cell Systems as a means to reduce Atmospheric Carbon Dioxide Emissions. Energy Convers. Mgmt Vol. 37, Nos 6-8, pp. 1247-1252.

Kiyotaka Tahara, Toshinori Kojima and Atsushi Inaba, 1997. Evaluation of CO₂ Payback Time of Power Plants by LCA. Energy Convers. Mgmt Vol. 38, Suppl., pp. S615-S620.

Chairperson: Bent Sørensen

• take away misconceptions regarding EPT values for PV
• understanding & interpretation of EPT and CO2 mitigation potential
• guidelines for calculation and use
• EPT and CO2 mitigation potential as performance criteria

Appendix B Papers delivered to the workshop