The objective of the workshop "Photovoltaics and the Environment" was to bring together PV manufacturers and industry analysts to define EH&S issues related to the large-scale commercialization of PV technologies.

The National Photovoltaic EH&S Assistance Center is operated at BNL, under the auspices of the DOE, to foster the safe and environmentally friendly operation of photovoltaic facilities and products, extending from R&D to manufacturing and use. The objectives of the Center are to prevent accidents, to reduce EH&S occupational and public risks, and lower the environmental- and safety-related costs.

The BNL Center undertakes the following types of activities: 1) It directly supports DOE Headquarters, the National Renewable Energy Laboratory (NREL), and Sandia National Laboratory (SNL) to ensure that their facilities and those of their contractors are operated in an environmentally responsible manner. 2) It conducts EH&S audits, safety reviews and incident investigations, as needed. 3) It assists the photovoltaic industry to identify and examine potential EH&S barriers and hazard-control strategies for new photovoltaic material, process and application options before their large-scale commercialization. To facilitate the transfer and application of knowledge derived from this work, BNL hosts workshops, tutorials and symposia, uses electronic mail and a web page, and publishes articles in the peer-reviewed journals. The current workshop is one of these "industry outreach" activities.

The Thin Film PV Partnership is the main program funded by DOE to support R&D in thin film options, such as amorphous silicon, cadmium telluride, and copper indium diselenide. The Partnership funds in subcontracts to leading companies and universities to perform this research. Subcontractors are categorized as Technology Partners if they are companies with commercial or near-commercial thin films. Others are categorized as R&D Partners, and they undertake work that will keep further advances in the pipeline for the future. The members of the Partnership work on technology-specific National R&D Teams in collaboration with NREL in-house researchers; these teams are directed to solve critical fundamental and technological problems. An additional cross-cutting team, the Thin Film Partnership ES&H Team is made up of the Technology Partners and others with commercial interests, as well as invitees, who address ES&H challenges together. The workshop in Keystone was one of their activities, planned jointly with BNL.

The workshop was opened by BNL's Paul Moskowitz who welcomed the participants and emphasized the need for open discussion and collaboration on EH&S issues among all industry players.

To safeguard the environmental friendliness of photovoltaics, the PV industry follows a pro-active, long-term environmental strategy involving a life-of-cycle approach to prevent environmental damage by its processes and products from "cradle to grave." Recycling manufacturing waste and spent PV-modules is examined as part of this strategy. Although the PV industry will not face this problem in a large scale before the year 2020, today’s choices of materials and module designs may very well set a precedent for the future.

The session was chaired by Vasilis Fthenakis, BNL who briefly summarized the previous studies and work on recycling. The first workshop on recycling was organized by BNL and NREL and was held place in March 1992 at Golden, CO. The 1992 workshop focused on Cd and Se and brought together interested parties from the PV industry, the metal-smelting industry, the utilities, the DOE, and the national laboratories. It established the state-of- current-affairs and provided a foundation for the infrastructure and technical feasibility studies, which were conducted by BNL, UNISUN, Solar Cells Inc. and DrinkCard Metalox between 1993 and 1998. In April 1994, BNL and NREL hosted a workshop that focused on understanding and managing the health and environmental risks of CIS, CGS and CdTe module production and use. The 1994 workshop covered the toxicology of these new materials, and the pertinent implications to occupational health and the environment. This workshop identified the need to examine regulatory drivers and constraints to recycling.

The first session of the current workshop expanded on all these issues. The papers presented gave industry's perspectives and analyses of the collection and recycling infrastructure, regulatory concerns, technical feasibility, economic incentives, and costs. In the open discussion forum different opinions from the industry participants were heard and discussed.

Photovoltaic modules may contain small amounts of regulated materials, which vary from one technology to another. Environmental regulations can determine the cost and complexity of dealing with end-of-life PV modules. If they were classified as "hazardous" according to Federal or State criteria, then special requirements for material handling, disposal, record-keeping and reporting would escalate the cost of decommissioning modules. Chris Eberspacher, UNISUN, discussed such issues related to recycling of CdTe and crystalline Si PV modules. He showed that several of today's CdTe modules failed the US-EPA Toxicity Characteristic Leaching Procedure (TCLP) for potential leaching of Cd in landfills. Similarly, some of today's x-Si modules failed tests for leaching of Pb. Consequently, such modules may be classified as "hazardous" waste. The TCLP is the current EPA method to characterize the toxicity potential of wastes. It assumes a particular worst-case scenario of potential leaching of compounds in a landfill. However, many parameters that affect the leachability of contaminants in the field are not addressed in the current TCLP, and the EPA is investigating a more flexible approach that can be tailored to different types of waste. Danny Cunningham, BPSolar, discussed issues related to preparing PV scrap for TCLP testing. He showed that preparation of the sample can affect the results of the test and pointed out the need for preparation guidelines.

The PV industry may choose to recycle spent modules, even if there are no regulatory issues related with their disposal, because recycling may improve market penetration. Recycling is most cost-effective if is done collectively by the industry, and in high volumes. Jef Bagby, who discussed the experience of the Rechargeable Battery Recycling Corporation (RBRC) discussed the benefits of a collective approach. Two-hundred and fifty manufacturers of rechargeable products fund RBRC, representing about 75% of the world's rechargeable producers. Bagby described the difficulties in setting up the logistics and operating the collection and transport of products for recycling. These difficulties included allocating cost among participants, non-uniformity of state laws, restrictions from antitrust laws, national laws and international agreements affecting transportation, and additional overhead due to tax laws. The RBRC program became more effective when it was expanded beyond the manufacturers and involved distributors, retailers, end-users, and the government.

The technical feasibility of recycling was proved by work funded by the DOE small-business initiative research (SBIR). John Bohland presented the results of such research at Solar Cells (SSI), which demonstrated the feasibility of recycling the basic components of spent CdTe and x-Si modules. For CdTe, their technique entails crushing the module in a hammer-mill, screening EVA flakes, and stripping metals from the crushed glass in successive steps of chemical dissolution, mechanical separation, and precipitation or electrodeposition. The estimated cost for this process for a 2 MW process (about 40,000 modules) integrated with a manufacturing facility is 0.2-0.4 $/W, excluding transportation. For x-Si cells, SSI tries to recover the functioning cells which have a high value. So far, they worked with PV coupons (not complete modules), and recovered most of the Tedlar backsheet and the functioning x-Si cells which have an electrical efficiency only slightly lower than the original ones. Their method starts by gently heating and manually peeling off the backsheet. Then, inert atmosphere pyrolysis at about 500 ^oC vaporizes the EVA lamination layer. Si-cell recovery was estimated to cost about 13 ¢/W, for an operational scale of 150,000 x-Si cells per year. For comparison, a new x-Si cell costs at least $1.50/W to produce.

Bob Goozner, DrinkCard Metalox (DMI) discussed the results of another DOE-SBIR project, dealing with recycling of CdTe and CIS modules. DMI's operations include chemical stripping of the metals and EVA, skimming off the EVA from solution, and successive steps of electrodeposition, precipitation, and evaporation to separate and recover the metals. DMI reports recovery of 95% or better of Te, and 96% or better of Pb from CdTe modules. Chemical stripping leaves the SnO₂-conducting layer intact on the glass substrate, potentially allowing the re-use of the substrates for PV deposition. They project a processing cost of 9 ¢/W or less for either CdTe or CIS modules. Goozner also showed some promising preliminary results of stripping x-Si coupons with an HNO₃ solution.

It has been shown that there is an economic incentive to design modules that will not be hazardous, or to design them in such a way that they can be recycled at a reasonable cost. This issue was discussed by Simon Tsuo, NREL, who presented alternatives in silicon solar-cell manufacturing that are friendlier to the environment than today's common practices. Tsuo started by refering to sources of information on environmentally benign semiconductor manufacturing, and continued with a step-by-step analysis of manufacturing and environmental improvements for x-Si PV. His major points were that alternatives exist that are both environmentally benign and cost-effective, and that the PV industry can benefit by coordinating efforts with the printed-circuit industries.

The following are some additional salient points from this session:

• Recycling makes sense because it conserves relatively rare minerals like tellerium, reduces the energy requirement in PV manufacturing, and preserves the environmentally friendliness of PV.
• Recycling can help market acceptance and penetration.
• Effective methods of recycling have been developed that can be used for both in-house and large-scale recycling.
• Currently, economic incentives may be inadequate to move the PV industry into voluntary recycling. However, this may change in future, as more economic incentives may be given to developing clean technologies, preventing pollution, and reducing CO₂ emissions.
• Also, in future, recycling costs are likely to decrease as the technologies mature, whereas the costs of landfill disposal are constantly increasing.
• The industry must keep an eye on regulatory developments related to TCLP, and similar waste-defining rules and exceptions as they affect the need for, and economics of, PV recycling.

Session II focused on the potential of PV to reduce emissions of carbon dioxide, and thus have an ameliorating effect on global climate change. The Chair of this session, Ken Zweibel, NREL opened the discussion with a synopsis of the fundamental terms used in analyzing PV energy pay-back and carbon dioxide mitigation. PV CO₂ emissions are zero during use because PV systems require little or no maintenance or oversight. Some CO₂ emissions can be attributed to manufacturing because it takes energy to manufacture a module. Thus, PV requires some input energy, which it pays back early in its lifetime. This energy pay back (EPB) was the focus of the paper by E. A. Alsema ("Energy Requirements and CO₂ Mitigation Potential of PV Systems") of the University of Utrecht. For PV to be useful, the energy payback must be reasonable. In the past, PV has had high EPB because technology was immature and burdened by energy-intensive processing. As Alsema points out, EPB for existing PV systems is in the 3-10 year range; and future technological options will likely allow system EPB to fall to the 1-3 year range. In this range, the amount of CO₂ displaced by PV over a thirty year lifetime outdoors is 90% to 97% as compared to the CO₂ of the energy it offsets. For example, if we assume the US mixture of energy generation causes 160 g carbon equivalent of CO₂ per kWh of electricity, then a 1 kW PV array in an average US location would produce about 1600 kWh each year, and 48,000 kWh in 30 years. That would avoid about (48,000*0.95)*160 g/kWh, or about 7 metric tons (MT) of carbon equivalent during its useful life. The "0.95" in the equation represents a reasonable guess about future energy payback at the system level. Clearly, certainty about the order-of-magnitude of this number is far more important than predicting it to three decimal places; and also more important than the expected absolute amount of carbon dioxide that PV will displace, since that is dependent on (1) the mix of energies with which PV will be manufactured and (2) the mix that it will displace (which will vary with location, application, and date).

Once it is established that substituting fossil-fuel energy generation with PV can prevent substantial CO₂ emissions, it is then necessary to establish that PV can become a large-scale source of electricity. The major barrier to PV becoming a large-scale energy source is PV electricity cost. Today’s PV systems sell for about $6-$10 per peak Watt, with an implied electricity price of about $0.4 to $0.7/kWh. This is much higher than conventional sources, which sell for about 8 cents/kWh in the US. However, PV technologies are making rapid progress toward improved output per unit cost. Existing technologies are improving, and new technologies are coming on line. Projections of future costs based on progress in PV technologies are consistent with module costs below $0.5/W (compared to $3/W today) and reduced balance of system (BOS) costs as volume and design sophistication increase. Thus the competitive economic future of PV is quite promising.

However, PV faces several other hurdles before it can be a source of energy on a global scale. Two of them are covered in two other papers delivered during the session: "Materials Availability and Waste Streams for Large Scale PV" by Bjorn Andersson of Goteborg University and "The Competition Facing PV in a Greenhouse Gas (GHG) Emissions-Constrained World" by Robert Williams of Princeton. Andersson discusses the various potential constraints on the use of several thin film PV technologies that are expected to have excellent cost potential. However, each of these technologies uses a rare raw material: germanium in amorphous silicon; tellurium in cadmium telluride; and indium and gallium in copper indium diselenide. Andersson provides some sobering parametric insights into the potential of these technologies to be used on a global scale, i.e., at volumes per year of 10-100 GW. Each of these technologies must address the materials availability issue in order to be viable at these volumes. Fortunately, research avenues such as making thinner layers, increasing device output per unit area, increasing the materials utilization of processes, recycling, and substituting other materials for the rare ones, provide excellent avenues for meeting the materials availability challenges. These challenges will not impede the near-term adoption of these technologies (prior to 2010 or 2020, depending on growth), but will need to be addressed during this grace period.

Williams addresses the issue that PV must compete for markets that will have other choices, such as wind and biomass electric power. Indeed, such competition will always be a ‘moving target’ since all energy technologies will progress substantially during the time it will take PV to mature. If they have favorable cost, favorable environmental or infrastructural qualities, they will be hard competition for PV in developed and developing nations. The special uses of PV (e.g., for rural electrification in developing countries) means that some of this future market competition will out of necessity remain uncertain while markets and technologies develop.

Finally, our last paper "MARKAL-MACRO: A Computer Tool for Integrated Energy-Environmental-Economic Analysis", presented by Vasilis Fthenakis of Brookhaven National Laboratory (BNL), discussed the predictions of a BNL model (MARKAL-MACRO) on how PV will compete in the future 30 years, against other technologies in the US energy marketplace. Fthenakis used input data from NREL about future efficiencies and cost reductions of PV, and data from DOE on the expected performance of about 200 competitive technologies and administrative options, and produced estimates on PV penetration in the US and corresponding CO₂ emissions reduction. The model predicts that by the year 2020, if the targets of 15% PV module efficiency and $0.71/W_p capital cost are attained, PV installations in the sunniest areas of the US (e.g., Arizona) will total 20 GW; if the price falls down to $0.57/W_p, then PV's sharein the US could jump to 140 GW (about 10% of the projected total 2030 US energy capacity). At that level, PV will be reducing carbon emissions by about 64 million metric tons per year, a very significant contribution by any single technology.

This workshop has addressed the effort to develop a framework for evaluating PV in relation to global CO₂ reduction. We have examined the issues of PV energy payback; PV CO₂ reduction; the potential of PV technologies to become cost competitive; competing technologies; and constraints such as materials availability and waste streams. This is only a first step. For example, we have not performed an in-depth study of PV with electric or chemical storage or other similar infrastructure issues unique to non-dispatchable sources like PV.

As PV matures and becomes a more widespread and practical energy option, we shall re-examine this question again and again. In the future, we hope to adequately characterize the value of PV in this endeavor so that the public and their representative organizations can properly weigh policy affecting PV development.

President Clinton announced in June 1997 the ultimate goal for one million solar-energy systems to be installed on U.S. rooftops by 2010. He committed the federal government to installing 20,000 solar systems on its buildings' rooftops by then. This initiative was the Administration's response to the issue of Global Climate warming. As discussed in the previous session, PV can make a significant contribution in reducing carbon-dioxide emissions in the United States, and the Million-Roof Initiative is intended to jump-start the solar market and create the momentum for necessary price reductions to achieve this goal.

Christi Herig, NREL talked about the goals of the initiative and the role the administration envisions for the different partners (e.g., industry, states, municipalities, consumers, developers and builders). She presented some heartening statistics. In 1998, there were 8,500 solar buildings in the United States, up from only 2,000 the previous year. The projected numbers of solar buildings for 2000, 2005, and 2010 are 51,000, 376,000 and 1,014,000 respectively. The DOE wants to identify potential EH&S concerns related to such a quick pace of development.

The most obvious issue relates to the hazards of electric shock and falls during the installation, connection, repair, and maintenance of PV roof systems. Ward Bower, Sandia gave an overview of codes, certification requirements, and guidelines issued by the IEEE, ANSI, ASTM, IEC and the UL. There are many guidelines and standards applicable to PV-system interconnects and current efforts at Sandia focus on ensuring that the National Electrical Code includes the PV-unique requirements for safe installation of PV building systems. Bower distributed copies of draft of NEC article 690 that covers photovoltaic systems, and discussed several considerations applicable to PV modules and arrays, (e.g., critical temperatures, voltage ratings, cable and insulation types, sizing for safe design, over-current protection, manual disconnects, grounding, anti-islanding protection, and in-surge and transient protection). An easy-to-read safety document, "IEEE Guide for Terrestrial Photovoltaic Systems Safety" will become available in early 1999. Other items that were touched upon by the audience included fire hazards to the buildings' occupants and to firefighters. Questions pending answers include: 1) How are arrays disabled so that firefighters are protected when using water on the roof? 2) Do firefighters need extra protection from potential toxic vapors emitted from a burning array? 3) Are there any environmental issues related to the disposal of roofing shingles and building-integrated modules? 4) Can the industry meet RCRA and state requirements for landfill disposal, or will treatment/recycling be necessary?

While these concerns are addressed by the industry and solutions are being worked out, it is imperative to emphasize the positive aspects of the upcoming scaling-up of building-integrated photovoltaics. Paolo Frankl, INSEAD showed that such systems could offer even greater benefits in lowering carbon dioxide emissions (on a MW basis) than large stand-alone systems. This conclusion is based on energy pay-back time, energy yield, and net CO2 balance from a life-of-cycle analysis of current silicon-based photovoltaics. The main advantage of building-integrated systems vs. stand-alone ones comes from avoiding the expense and energy intensity needed for structural supports. Frank also presented a parametric analysis of the effect of future system designs, showing that future hybrid PV/thermal roof systems are expected to further enhance the potential of PV to mitigate CO₂ emissions.