Environment, Technology and Culture



Recycling and Product Stewardship

Recycling is an imperative for a successful, sustainable photovoltaic industry.

-Karsten Wambach, engineer for Sunicon, a former subsidiary of SolarWorld'


The photovoltaic industry's remarkable growth will eventually result in huge quantities of retired modules that have become waste in need of disposal. A report from the International Renewable Energy Associa- tion estimates 80 million metric tons by 2050.2 Assuming that each gigawatt of photovoltaic capacity represents somewhere around ten million modules, there are billions of photovoltaic modules installed worldwide today. The U.S. has installed about 50 GW of distributed and utility-scale photovoltaics as of 2018, so roughly half a billion modules. The Energy Information Agency of the Department of Energy (DOE) estimates that residential rooftop photovoltaics will continue to grow at around 7% each year through 2040.3 By 2030, cumulative glo- bal installed photovoltaics will be around 8 terawatts.4 This amounts to hundreds of billions of photovoltaic modules installed in utility-scale, distributed, and off-grid applications. Where will all this photovoltaic material go after the end of its useful life?

There are tremendous benefits coming online with these photovoltaic systems. Electricity from photovoltaics displaces the dirtiest energy sources, the fossil fuel peaker plants that provide electricity during peak hours of the day. Replacing peaker plants with solar power means we breathe cleaner air, fewer nitrogen oxide emissions that cause photo- chemical smog, and less particulate-matter pollution. There are also


Recycling and Product Stewardship I 95

unquestionable benefits for our climate. The greenhouse gas emissions related to the manufacturing of photovoltaic modules are negligible compared to the lifetime emissions they save after installation.5 Photo- voltaics are one of the few energy technologies that can generate elec- tricity close by healthy communities. They do not cause air pollution, and can be integrated with the built environment.

However, photovoltaics could present disposal challenges when large amounts start to enter the waste stream. Photovoltaic modules degrade in power output over time until they are no longer generating sufficient electricity or meeting the needs of the homeowner or power plant owner. Nearly all companies offer warranties on photovoltaic modules that guarantee a certain power output (commonly 80%) for 20 to 25 years. Because of these long warranties, companies aim to make a product with minimal degradation, and many photovoltaic modules can operate for many decades without fading appreciably. Most "solar panels" on rooftops and in power plants are not going to be decommissioned any- time soon. Still, every photovoltaic module will eventually become pho- tovoltaic waste.

Some photovoltaic module waste comes from defective or broken modules that require disposal before they are installed. Some manufac- turers may produce photovoltaic modules that fail quality control or have manufacturing extrusions, creating waste on the factory floor. Other modules may break en route to or at an installation site, or may fail at some point early in operation. These photovoltaic waste streams will represent early waste flows.

The challenge for policymakers is that the steep wave of photovoltaic waste is over the horizon, too far off to spur action, as large volumes will not show up for twenty to forty years, or even later. But as early photovoltaic installations are upgraded and replaced with new mod- ules, the waste stream will quickly rise to very high volumes. If there is no plan in place, it is not clear where they may end up. Where will they be directed? Developing countries? Landfills? Smelters? High-value recycling? Finding a way to close the loop on photovoltaics, shifting the material flows closer to "circular economy" principles, could be one way to manage future photovoltaic waste streams.

Some photovoltaics will become waste much earlier. The so-called bathtub curve shows a high rate of failure at first that slowly dimin- ishes, only to rise again as the expected lifetime approaches. Early fail- ures are often due to mechanical issues or breakage. Modules that are defective will show that right away, if they made it to installation

96 I Recycling and Product Stewardship


Photovoltaic modules (approx.) Owner/developer County

Desert Sunlight Solar Farm 8.9 million First Solar Riverside Topaz Solar Farm 8.9 million First Solar San Luis Obispo Antelope Valley Solar Ranch 2.9 million First Solar Los Angeles California Valley Solar Ranch 1.3 million SunPower San Luis Obispo Mount Signal Solar 1 million PayneCrest Imperial

without being caught by quality control. Modules can be broken during transport and installation, or damaged by random or acute events like hail, mice, heat stress at high temperatures, mechanical stress (mounts fastened too tight, for example), bullet holes, even tornadoes. These are all-low probability events, so the volume of middle-aged end-of-life modules remains small.

Another source of end-of-life photovoltaic module waste is discards from manufacturing facilities. These end-of-life modules might be clearly defective or broken, or fail quality control (even for cosmetic reasons) at the fab. A similar stream of end-of-life photovoltaics is from modules that are removed because they are defective or do not meet warranty claims, if the company directly processes warranty reserves. These arc modules that might generate electricity for a few years, but 1101 last as long as the rest of the system or power plant.

California will be a sentinel for photovoltaic waste because the state is hy for the leader in photovoltaic module installations in the U.S. Cal- ifornia also has the most utility-scale photovoltaic power plants. The means that in addition to the distributed flows of photovoltaic waste, there will be places with very high concentrations. In California, many of the rural areas where the largest solar farms are there may have dis- posal challenges due to the remoteness the of facilities or the lack of disposal infrastructure. However, utility-scale installations are likely to have some kind of waste logistics company involved in case disposal is needed. Table 8 shows the largest installations, developers, and loca- tions of some of the large photovoltaic installations in California.

Photovoltaic module waste flows can be estimated with a few assumptions about when waste will be generated and where end-of-life management opportunities arise. Dr. Vasilis Fthenakis, of Brookhaven National Laboratory and a professor at Columbia University, devel- oped a framework for addressing this question. It begins by asking how

Recycling and Product Stewardship I 97

many photovoltaic modules are produced by the industry each year. This value is readily available, because numerous industry reports estimate it regularly. Second, how much does each photovoltaic module weigh per rated power output? This is estimated in tons per MW-peak. This is more speculative, because photovoltaic modules are getting more powerful over time, but some assumptions can be made about the power output and mass of panels every year to account for increased output and lighter modules. Third, how much waste is generated in production at photovoltaic manufacturing facilities? All manufacturing results in some kind of waste. Manufacturing extrusions can damage photovoltaic modules during manufacture, and they can be removed for quality control, or downstream if the problem remains undetected before it leaves the factory gate. Fourth, what percentage of photovoltaic modules are damaged or defective en route or during operation? Installers receive broken photovoltaic modules, and others may become broken or damaged in the field. This includes damage from external factors such as extreme weather, falling debris, and so on. Finally, how long will the modules operate? When will customers start replacing them, and when will they begin to show up for disposal? Some photovoltaic modules might be sent for reuse, since many decommissioned photovoltaic modules are still capable of delivering power.


Photovoltaic modules contain a number of valuable or rare materials, and recycling will ensure that these rare materials are recovered for reuse.6 Scarce and precious metals are now even geopolitical issues. In 20n, China blocked exports of rare earth elements (the lanthanides on chemistry's periodic table) such as terbium, europium, cerium, and neo- dymium to several countries for strategic reasons. China controlled 97% of global rare earth production that year. The announcement drove prices for these commodities up over 7 50% within a year.7 Other rare inputs have similar volatility. A looming limitation for decarbonization technologies (renewable energy, energy efficiency) is the availability of a handful of rare metals, including rare earth elements, precious metals, and other metals, which are used in products such as electric vehicles, hydrogen fuel cells, photovoltaics, and next-generation LED lights. The only rare earth element associated with the photovoltaic industry has

98 I Recycling and Product Stewardship

been cerium, which is believed to have been used in a filtration process, but only by one known manufacturer over time.

Almost all photovoltaic modules require inputs that are relatively scarce, precious, or subject to price volatility: silver, tellurium, and indium. The price and availability of rare and precious metal can influ- ence costs of electronics and photovoltaics significantly. The impor- tance of ensuring adequate supplies of critical elements has long been recognized by the Department of Defense. 8 The Department of Energy's SunShot Initiative recognized this too and tracks four metals (silver, indium, gallium, and tellurium) that can influence the cost of photo- voltaics and possibly even hinder cost reductions."

Crystalline silicon photovoltaic modules contain a number of valua- ble materials that can be recovered and reused, such as copper, alumi- num, and glass. The precious metal silver is also used in these modules, in the electrical contacts (look closely at a module and notice thin the lines of metal running across the blue, grey, or black photovoltaic mate- rial). From 2008 to 2009, silver prices rose while many other costs fell for photovoltaic manufacturers as investors moved out of securities and other risky financial products to the safe refuges of precious metal com- modities. In 20u, the solar industry consumed u% of the global silver supply.'? By 2018, that number was over 20%.11 Silver is commonly recovered in mines where there are also ores associated with other met- als, such as lead, gold, copper, zinc, and cobalt. The dominant produc- ers include Mexico, Peru, Australia, the U.S., and Chile. Smelters that produce silver can be designed to recover other metals, but only about 3 2 % of that silver is recovered, and the rest disposed of as slag.12

Few materials from end-of-life photovoltaics can be recycled into high-value glass cullet (recycled glass that has been crushed and is ready to be melted) or "downcycled" into lower-value secondary glass prod- ucts. A photovoltaic module is mostly glass by weight (75-90%), but recyclers report that much of this glass cannot be made into flat glass again due to impurities. Common problematic impurities in glass cullet include plastics, lead, cadmium, and antimony; where they are present, the glass can only be downcycled. Recycled silicon wafers also have value since a significant energy investment is required to make them.13 Recovering the silicon from photovoltaic cells can reduce overall energy use because less polysilicon needs to be refined.

The tellurium, indium, and gallium used in thin-film photovoltaic technologies are among the rarest elements in the Earth's crust. Recycling is one way to recover these materials. Modules are first disassembled to

Recycling and Product Stewardship I 99

recover valuable materials like copper wiring and aluminum (if they have frames). Then they are shredded into small pieces. A process developed by Brookhaven National Laboratory and First Solar uses dilute acids to remove the Cd Te from the module. t4 Up to 9 5 % of the cadmium and tel- lurium is recovered from filter cake, and reprocessed into new semicon- ductor material by a third party. To ensure that CdTe modules are recy- cled safely and responsibly, ambient cadmium emissions are monitored in real time to ensure that occupational exposure is minimized. LS

Tellurium occurs in the earth's crust at the rate of only one part per billion. It is about a thousand times as scarce as the rare earth elements that are the subject of ongoing trade disputes among the U.S., China, Malaysia, and Japan. Annual global production is on the order of 200 to 1,000 tons. The availability of and market for tellurium have inter- esting implications for the photovoltaic industry, in part because one photovoltaic manufacturer dominates the market. First Solar, of Tempe, Arizona, currently purchases 40% of the total volume of tellurium sold.16 In z.o r o First Solar agreed to buy a significant portion of high- purity tellurium from Apollo Solar Energy, a Chinese tellurium supplier operating open-pit mines on its Dashuigou property in Sichuan Prov- ince.17 This is the only mine in the world where tellurium is the primary product. Most other tellurium supplies are secondary or tertiary prod- ucts of copper or gold. Recovery is largely driven by the price; as the price goes up mines are more willing to put in the effort to recover tel- lurium from ores. The DOE expects that by 2031 there will need to be additional main-product tellurium mines and an ample secondary sup- ply of recycled CdTe photovoltaic modules. 18 In z.o r r , financial analysts observed that First Solar had acquired a gold-tellurium mine in Mexico to secure future cadmium telluride supplies.19

First Solar's efforts to secure a tellurium supply are bolstered by the reuse and recovery of tellurium from manufacturing scrap and end-of- life photovoltaic modules. It builds recycling operations into its facto- ries, and ships the resulting filter cake to its supplier, which returns it as high-value semiconductor feedstock. It is also making its semiconductor layers thinner and more efficient over time, yielding more energy per module, with less tellurium. Its current line of photovoltaic modules contain semiconductor layers that are about three microns thick, and there is potential to reach two microns in the not-too-distant future.

Indium is a key input to some thin-film photovoltaic technologies and is also relatively rare. Competition for indium for use in flat-screen televisions could ultimately restrict its availability for commercial CIGS

mo I Recycling and Product Stewardship

photovoltaic production.'? The need to recover indium should therefore encourage the recycling of CIGS, and this is seen as an important issue for large-scale commercialization and deployment. The recycling of amorphous silicon is said to be one of the cleaner recycling processes of all photovoltaic technologies. Though most of the literature mentions this, none mentions that indium tin oxide has been linked to occupa- tional lung problems in flat-panel television recycling facilities.21 Amor- phous silicon manufacturing has been marred by several silane acci- dents, and the technology suffers from low efficiency.

The major indium producers are China, France, Canada, and Japan. Indium is mainly recovered from zinc production at smelters. Indium is also recovered from waste flat-panel displays. The price of indium has been far more volatile in recent years than that of tellurium, even though it is more widely available and more widespread in the Earth's crust (about one part per million in the Earth's crust). There is no production of indium in the U.S.

Gallium is one of the four scarce metals tracked by the DOE, and its scarcity is a potential barrier to low-cost photovoltaics.22 Gallium is used in CIGS photovoltaics, as well as multijunction and dual-junction solar cells used in satellites and space craft. The metal is widely distrib- uted, and no country dominates the supply, though the primary global suppliers are Australia, Guinea, Brazil, and Jamaica.

Some advanced photovoltaic technologies currently being explored in scientific laboratories, such as dye-sensitized photovoltaic cells, require ruthenium, a platinum-group metal of which only about 12 tons is mined annually.23 Molybdenum is used as a contact layer in some kinds of pho- tovoltaics. The presence of these valuable materials suggests that recy- cling would be both economically and environmentally beneficial in the long term. In fact, many studies show that the limited availability of some of these key inputs caps the total amount of photovoltaic module produc- tion possible without materials recovery.24 Other valuable materials used in thin-film modules include copper, aluminum, and glass (Table 9).

Today, very few of the components of a photovoltaic module are recovered for reuse as feedstock in similar products. Even where photo- voltaic modules are safely and responsibly recycled at facilities renowned for worker health and safety, not all the components of photovoltaic modules are recovered. At ECS Refining, a San Jose company locally recognized as a leader in product stewardship and recycling advocacy, photovoltaic modules result in only smelter flux, sold off in bulk with other glass and e-waste products.


Material Content by weight

Glass ~90% Steel 0--50% Aluminum 0-5% Copper ~2% Plastics 5% Silver <1% Tellurium 1% Silicon 2% Cadmium 1% Lead <1% Antimony trace Gallium arsenide 5%

Recycling and Product Stewardship I ro r


Substrate, weatherization Frame and mounting equipment Frame and mounting equipment Wires, electrical equipment Junction boxes Metallization frit paste in c-Si Semiconductor in CdTe photovoltaics Silicon wafers Semiconductor in CdTe phorovoltaics Solder; also found in glass Solar glass Semiconductor in multijunction cells

Almost all photovoltaic modules contain plastics of some kind. For example, all must have a junction box-the unit that connects the solar cell to the wiring needed to deliver electricity. Though most of the ther- moplastic materials used for junction boxes are "halogen free" (meaning they contain no bromine or chlorine), some module junction boxes may contain brominated flame retardants. The EU and California both have restrictions on selling products containing brominated flame retardants. While it is critical that photovoltaic modules use the most fire-resistant materials possible for electrical equipment and wiring, there are many such materials available that do not contain carcinogenic bromine-based materials. More recently, there have been efforts to remove halogens from backsheet, laminates, and wires and cables.


Where photovoltaic modules are disposed of, they can present hazard- ous-waste disposal issues. One concern raised with end-of-life photo- voltaics is that they may enter into the global e-waste trade in end-of- life electronics. Photovoltaics have the recipe for such e-waste: toxic materials are embedded in valuable ones. The heavy metals of concern in photovoltaics are lead and cadmium, depending on the design (a handful of modules have no toxic metals).

Regulated metals in end-of-life CIGS include cadmium, copper, and selenium, while in end-of-life CdTe, cadmium compounds are the pri- mary metal of exposure concern. California regulators debated how to

102 I Recycling and Product Stewardship

classify end-of-life photovoltaic modules, with focus on whether photo- voltaics would burden local governments and transfer stations. In 2011, CdTe thin-film manufacturers First Solar and Abound asked the Cali- fornia Department of Toxic Substances Control (DTSC) to deregulate end-of-life photovoltaic modules that otherwise would be considered hazardous waste under California law. Other thin-film producers, such as SoloPower and Solyndra, participated as well. This new rule would eliminate costly and time-consuming manifests and other paperwork. At issue is whether photovoltaic modules must pass a hazardous waste characterization test designed to estimate the amount of toxic material that might leach from landfilled material.

Because crystalline silicon is the most common photovoltaic technol- ogy, lead compounds are the most widespread toxic materials in photo- voltaics. The amounts of lead in crystalline silicon photovoltaics vary from zero to several hundred grams per module. According to an annual survey of photovoltaic manufacturers by the Silicon Valley Toxics Coa- lition (SVTC), several companies are able to make modules without lead for several or all of the module types they offer, and more plan to, though only a handful do so as of 2018.

Two thin-film CdTe manufacturers with projects in California-First Solar and Abound Solar-made modules that passed the U.S. Environ- mental Protection Agency waste determination test, but failed Califor- nia's more stringent test, meaning that they would be considered haz- ardous in the state. Both manufacturers asked for an exemption from DTSC rules for the disposal of end-of-life modules. Project developers at the time were unclear of the decommissioning costs of projects, so there was interest in confirming and clarifying the rules. The rules would treat photovoltaic waste as "universal waste" (widely produced house- hold hazardous waste, like compact fluorescent bulbs or mercury ther- mostats), which required that the state develop management programs, often paid for by companies or consumers. The agency finalized a rule in 2012, but owing to rules limiting administrative procedures, the rule required legislation, which was introduced by state senator William Monning and passed in 2014.25

Some have argued that CdTe has different physical properties than elemental Cd and could be less toxic. For example, CdTe melts at a much higher temperature (1041 °C) than cadmium (3 21 °C), and is less soluble in water. CdTe appears to be less toxic than elemental cadmium in terms of acute exposure, but the highly reactive oxidizing surface of CdTe quantum dots can damage cell membranes, mitochondria, and

Recycling and Product Stewardship I 103

cell nuclei, though this study was not able to disentangle the nano-scale effects from the effects of cadmium compounds.26 There are a few stud- ies of the toxicology of CdTe, but most do not accurately capture the primary exposure route for CdTe because they rely on ingestion, not inhalation, as the primary pathway.27

The global e-waste trade includes advanced automated facilities, but also facilities with lax labor laws and low occupational health stand- ards, and a separate informal processing sector that uses especially crude tools and equipment. Strong and enforced environmental health and worker protection standards for recycling can help minimize the toxic exposure and human rights abuses that currently plague the trade.

Possible destinations for end-of-life photovoltaics are landfills, waste transfer stations, and recycling facilities. Incinerators operate at tem- peratures high enough to volatize the CdTe into the air, although such incinerators should have pollution-control equipment to prevent sig- nificant emissions, and most of the cadmium will be dissolved into the molten glass.28 Discarding CdTe photovoltaic modules in landfills can pose risks because the cadmium could leach into groundwater.29 How- ever, one study found that the amount would not violate the drinking water standard in Germany." A report from the Norwegian Geotechni- cal Institute argued that leaching of cadmium into the environment could occur even in slightly acidic water, suggesting it could end up in landfill leachate. 31

There has been much debate among policymakers about whether photovoltaic modules that fail hazardous waste determination tests should be considered hazardous waste under the U.S. federal Resource Conservation and Recovery Act of 1976. The act sets certain national standards, but states are free to set stricter rules. This means that some modules might be hazardous waste in some states, but not others. So whether or not a photovoltaic module is considered hazardous waste largely depends on the thresholds set by regulators. As mentioned, some modules pass the EPA tests but fail the stricter California tests. This has economic implications, because hazardous waste can be more costly to transport and dispose of. Many local waste managers feared that Cali- fornia and its major cities will bear many of the costs associated with end-of-life photovoltaic module disposal. Estimates of photovoltaic module disposal costs depend on the waste classification and the dis- tance to recycling and disposal facilities. These estimates range from $0.04 to $0.13 per watt installed, somewhere in the range of $5 to $10 per module, depending on the size.32 On an analyst call with First Solar,

104 I Recycling and Product Stewardship

the only company with any recycling infrastructure at its manufacturing facilities, the company said the cost was about $0.04 per watt.33

California's hazardous waste laws are codified in the California Code of Regulations at Title 22, Environmental Health Standards for the Management of Hazardous Waste.34 This law authorizes the DTSC to implement a system of registration and permitting for the transport and disposal of hazardous waste.

Following Photovoltaic Waste

On a rainy winter day in San Jose, California, a team from SVTC visited an e-waste recycling facility, ECS Refining. SVTC has made photo- voltaic recycling the core of its Green Jobs Platform for Solar and its Solar Scorecard since 2008. SVTC's goal is to keep photovoltaics out of global e-waste streams by encouraging extended producer responsibil- ity (EPR), preventing e-waste exports, and ensuring that no prison labor is used to disassemble modules.

ECS has long offered tours to SVTC to help members understand e-waste recycling, and learned from SVTC about global e-waste prob- lems. The company earned the esteemed e-Stewards certification from the Basel Action Network for its recycling practices. The volume of e-waste generated in the Bay Area is staggering. Our guide noted that a room 150 feet across and roo feet wide, and filled to the ceiling with old cathode-ray-tube television sets, turned over several times a day. Work- ers wearing masks wielded hammers, breaking the plastic off the sets and sending the cathode ray tubes along conveyor belts toward a ham- mermill, where they would be broken into smaller bits. The facility was a simple, yet effective, means of separating the various elements of elec- tronics for secondary uses.

On the north side of the building was an outdoor area where large containers held scrap materials. One area of the facility was a room for hazardous liquids built over a containment liner. The liner prevented leakage into the soil of the toxic photo-processing chemicals that were stored at these facilities before the era of digital cameras. Several dump- sters full of broken or off-spec Solyndra photovoltaic modules were parked in the center of this room. The tubes were unmistakable, as no other company was exploring any similar form factor. Many were still clear, meaning the semiconductor layers had not yet been applied. The reason they needed to be stored in a place prepared for a chemical leak was that the tubes contained a mineral-oil-like substance that was

Recycling and Product Stewardship I 105

supposed to enhance the photovoltaic effect and long-term stability of the layers. Should the tubes break, the oily mess might be difficult to clean up. Hence, it was not the presence of cadmium but the liquid form and the inconvenience of handling spills that drove the infrastructure needs for these particular end-of-life photovoltaic modules.

Nearby were several more dumpsters of modules from a competing thin-film CIGS manufacturer, Mia Sole, Its semiconductor cocktail was an n-layer of CIGS and a p-layer of cadmium sulfide. Since the cadmium compounds were present only as a solid, they did not require the same levels of infrastructure and containment. Miabole, too, was seeking a DOE loan guarantee. It had gone so far as to seek out a life cycle assess- ment (LCA) expert to help it prepare the documentation and analysis required by the DOE background check, which requires that the project over its lifetime leads to greenhouse gas reductions. A photovoltaic module facility could claim that the modules made there would lead to returns on energy investments. The energy invested during facility con- struction would be returned or "paid back" by the electricity generated from modules made there. Energy payback time is a very common met- ric used in LCA of photovoltaics. A survey of LCA of photovoltaic modules shows extensive use of this metric in one form or another.

Like another local CIGS manufacturer, Nanosolar, Mia Sole saw interest in its technology dwindle with numerous delays, and technical and market developments that led to steep price declines in crystalline silicon photovoltaics. Mia Sole was eventually purchased by the Chinese energy giant Hanergy as that company went on a buying spree of CIGS manufacturers to complement its existing line of crystalline silicon modules. The company is owned by one of the richest men in China and began to have trouble in 2015, when the stock was the most shorted of any on the Hong Kong stock exchange.

Near Dresden, Germany, nestled among rolling hills, is a small town called Freiburg. It is famous as Germany's "solar city," and it exports to neighboring cities four times the amount of energy it uses, thanks to photovoltaics installed on the city's rooftops.35 The city is also a head- quarters for SolarWorld, a major manufacturer of crystalline silicon photovoltaics with operations in several countries, including the U.S. and Germany. It was here that SolarWorld had spun off a company called Sunicon, which was a project to investigate recycling solutions for crystalline silicon photovoltaics. The project began during the peak of the polysilicon shortage of 2008 and continued until about 2013. Its lead engineer, Karsten Wambach, was instrumental in the design and

rn6 I Recycling and Product Stewardship

implementation of the pilot facility. Yet the success of the facility hinged, not on good engineering and design, but on automation and improving the economics of recycling these high-value materials. Wambach was instrumental in the development and founding of PV Cycle in 2005, the first industry-wide EPR program for photovoltaics.

Extended Producer Responsibility

The key to effective long-term management of end-of-life photovoltaic waste is the establishment of recycling infrastructure based on EPR, a "polluter-pays" policy framework aiming to ensure that consumer prod- ucts are safely disposed or recycled. One approach to limiting the amount of end-of-life photovoltaic waste entering the environment is to employ a lifecycle management strategy based on EPR. The programs usually involve some kind of collection scheme to ensure that money is available to collect and or even recycle the product. EPR is a widely used policy instrument for products as varied as electronics, paint, carpet, batteries, and even pharmaceuticals. The rise of EPR in electronics is in part a reac- tion to the concerns raised by activists and government environmental agencies about e-waste. This e-waste is regarded as a serious environ- mental justice issue, as the development of informal sites of valuable- metal recovery operations has occurred in places with high rates of pov- erty in Ghana and China. These informal operations have few if any safeguards to protect people from the toxic materials to which these valuable materials are bound. Estimates vary, but somewhere on the order of fifty million tons of e-waste is produced annually."

There are multiple benefits from recycling end-of-life photovoltaics, including energy and resource savings, green jobs creation, and landfill diversion. Once this was recognized as an important issue in Europe, the first major site of growth for photovoltaics in the past decade, an organization named PV Cycle worked to develop a take-back and col- lection scheme in 2007. Several interview informants speculated that the industry was acting to get ahead of any proposed regulation under the directive on Waste Electrical and Electronic Equipment (WEEE). Recycling of electronics, especially photovoltaics, is recognized as a critical issue for the sustainability of photovoltaics by researchers at government labs and agencies in the U.S., such as DOE's National Renewable Energy Laboratory. Photovoltaic modules contain many toxic materials, so to protect workers and communities many safe- guards are required, akin to those warranted for other electronic

Recycling and Product Stewardship I rn7

products. Many environmental organizations and some policymakers believe that strong EPR policies can even create incentives for product designs that are safer, easier, more environmentally friendly, and cheaper to recycle.

The German Federal Institute for Materials Research and Testing, in cooperation with First Solar and Deutsche Solar-the two photovoltaic companies that operate recycling processes for end-of-life thin-films-as well as the Universities of Utrecht and Miskolc, conducted LCAs of recy- cling processes in a project called RESOLVED (Recovery of Solar Valu- able Materials, Enrichment, and Decontamination). The project found broad benefits from photovoltaic recycling, including for materials avail- ability, energy recovery, and environmental emissions. The findings from that project offered a proof-of-concept for recycling processes that a handful companies would adopt as they built recycling infrastructure.

Abound Solar, a CdTe manufacturer that also received ARRA sup- port in the form of a loan guarantee, also planned to develop a recycling program. It signed an agreement with CdTe powder refiner 5N Plus to operate a recycling facility in Wisconsin. The life of the facility was brief, however, and the company closed the plant only a few years later. After Abound Solar declared bankruptcy it worked with First Solar to recycle the end-of-life modules left behind in its shuttered factory.

Companies operating in European Union member states are required to meet the legislated requirements of the local state. These vary from one EU country to the next, because the legal basis for WEEE allows each member state to interpret WEEE's scope differently, setting slightly different objectives, benchmarks, and goals. Rules, requirements, and even procedures can be different in different markets. WEEE sets mini- mum thresholds for end-of-life e-waste recovery. Each member state has authority to go beyond these. Costs and recycling markets differ from state to state.

The first cast of WEEE included ten product categories in the scope of electrical and electronic equipment, including large household appliances, lighting equipment, information technology devices, and many more. This first cast of WEEE did not create a product category for photovoltaic modules to be included in the directive. In the scoping documentation for WEEE in 2002, photovoltaic modules are discussed in a separate article that specifically deals with WEEE's adaptation to scientific and technical progress over time. Most photovoltaic modules manufactured to date are still far from end-of-life. The EU determined that it did not yet need spe- cific WEEE regulations for photovoltaic modules. Looking forward, this

108 I Recycling and Product Stewardship

issue will rise to prominence if not planned for and anticipated properly. Text from an article in the 2002 Directive of the European Parliament briefly mentions "solar panels" and leaves open the possibility that they could be included in a future recast of WEEE:

Article r3: Adaptation to scientific and technical progress. Any amendments which are necessary in order to adapt Article 7(3), Annex IB, (in particular with a view to possibly adding luminaires in households, filament bulbs and photovoltaic products, i.e. solar panels).37

Several leading photovoltaic manufacturers, the European Photovoltaic Industry Association (EPIA), and the German Solar Business Associa- tion joined together to launch PV Cycle, "a European-wide collection, recycling and recovery system. "38 In anticipation of the possibility that WEEE could include photovoltaic modules in its recast, the photo- voltaic industry in 2007 developed an initiative to implement a take- back system to ensure that defective and used photovoltaic solar panels are properly recycled and their hazardous materials safely removed. The initiative was largely led by European industry actors and policy- makers but included global photovoltaic companies, because Europe was by far the largest market for photovoltaics. At the time, 80% of photovoltaic modules sold were installed in Europe, led by Germany, Italy, Spain, the Czech Republic, and Greece.

PV Cycle's founding photovoltaic manufacturers were Avancis, Isofo- ton, Conenergy, Schott Solar, SolarWorld, and Sulfur Cell. Membership ballooned to fifty members in two years, and over five hundred members in five short years. PV Cycle was shorthand for European Association for the Voluntary Take-Back and Recycling of Photovoltaic Modules. The association developed a framework based on "self-control and reporting." The organization would establish the broad framework for a central management system for end-of-life photovoltaics, either as a clearinghouse or a complete disposal service. WEEE mandates "insol- vency insured guarantees" to ensure that even companies that go out of business will take back end-of-life modules. Under PV Cycle's scheme, businesses must submit evidence that a financial guarantee is in place for each new module brought to market, though there were still unresolved questions about the value of the financial guarantee, and how perform- ance is overseen. PV Cycle also defined uniform quality and technical standards for collection and recycling.

The group announced that the scheme would be in operation by 2008 and cover about 90% of photovoltaic waste by 2015. They estab-

Recycling and Product Stewardship I ro9

lished benchmarks for recovery rates that would improve over time. They also laid the groundwork for a system for developing and sharing best practices for end-of-life photovoltaic waste handling, recycling and reuse projects, and for minimizing the overall waste in the design phase. Marie Latour of the EPIA said, "There is currently little need for spe- cific measures for recycling and waste control for solar panels as most of the solar panels are yet to reach the end of their first life cycles." However, she added, the "industry is already preparing for the solar panel waste issues likely to emerge over the next r 5 to 20 years. "39

Early on, the organization promoted a voluntary framework. The threat that WEEE could be extended to include solar panels in a future revision was believed to be enough to motivate the industry to act. EPIA and individual companies argued that the PV Cycle scheme would be more effective at increasing recycling rates than a mandatory scheme. "An inclusion of the photovoltaic sector in WEEE would result in 27 differently designed recycling systems, with inherent administrative pro- cedures and costs," noted a European Commission Report." This would offer an opportunity for companies to participate, but not require them to comply with rules that could differ by jurisdiction.

PV Cycle announced, "The take-back system proposed by PV Cycle will instead create a coherent EU-wide recycling system that will enable efficient and economically viable management of waste from the photo- voltaic sector. EPIA together with PV Cycle is therefore urging the Com- mission ... not to include photovoltaic products within the future scope of the revised WEEE Directive. "41 The voluntary efforts by the industry aimed to get ahead of European regulators, showing that vol- untary regulation works and that the EU should continue to exclude photovoltaics from WEEE.

PV Cycle would contract with Okopol in 2008 to research different schemes for a take-back and recovery system for photovoltaic products. One major conclusion of Okopol's report was that a mandatory system could have high regulatory costs that would be a financial burden on an immature, young industry. The Okopol study also found that a volun- tary program would have higher collection and recycling rates than a WEEE-type scheme. The voluntary framework would have high initial costs for manufacturers, but unlike in the WEEE scenario, those costs would go down over time, as a consequence of allowing companies to develop their own opportunities for innovation.

At a 2009 meeting in Brussels with SVTC, the leadership of PV Cycle and the EPIA proudly spoke of the PV Cycle effort and the "doubly

I IO I Recycling and Product Stewardship

green" benefits of the recycling policy. They enthusiastically shared a recent video of a de-installation, set to a high-quality musical score. The founding CEO of PY Cycle is Jan Clynke, an engineer from thee-waste industry, who brought intimate knowledge of recycling e-waste to the team tasked with designing an end-of-life management scheme for photo- voltaics. Clynke said that instead of regulation the industry would rely on "self-control and reporting" to ensure that it was meeting targets. How- ever, the possibility that WEEE would regulate photovoltaics remained.

By 20n, new rumblings of a recast of WEEE spurred PY Cycle to greater action. Many details remained to be decided about how the take-back, collection, and recycling system would be organized. Would it be based on direct reverse logistics, where photovoltaic installers would provide the services? Would customers have to drop them off at big box retailers or other pickup points? One key sticking point in the early debates was whether photovoltaic manufacturers making modules that were considered hazardous waste would have to pay a larger set- aside per panel, or per watt. Eventually, PV Cycle's progress on devel- oping the take-back and recycling scheme began to slow.

A key area of disagreement centered on whether the scheme would be paid for per photovoltaic module manufactured or as photovoltaic modules arrived at collection centers. How would it be financed? One scheme is a pre-paid approach, where funds are set aside in a restricted investment account. The second is a pay-as-you-go approach. The problem with the latter is that there is no guarantee that companies will be solvent-modules could arrive at recycling centers after a company is in bankruptcy. Given the turnover in the industry and the long time between installation and end of life, some modules would be "orphaned" in a pay-as-you-go model.

The solar industry set collection and recycling targets that were much higher than those for many other consumer electronics devices. Collec- tion targets refer to the amount of waste collected, out of the total sold. The recycling target is the amount of waste that is recycled. As of 2015, the target collection rate was 8 5 %, and the target recycling rate 90%.42

One might expect higher numbers if manufacturing discards were cov- ered under WEEE. Recyclers suggested that this would also enhance the business model for photovoltaic waste haulers and reserve logistics companies. Today no remnant of the word "voluntary" appears in PV Cycle's websites or brochures, or in formal WEEE documentation.

Strong EPR policies can even create incentives for product designs that are safer, easier, and less expensive to recycle. To ensure safe and

Recycling and Product Stewardship I III

responsible recycling of modules sold in the U.S., more domestic recycling networks and infrastructure will need to be developed. However, more collaboration and collective action is needed across the photovoltaic industry to encourage strong recycling policies and best practices, and to encourage design for environment and recycling. In an annual survey con- ducted since 2010, 75% of photovoltaic module manufacturers reportedly support mandatory take-backs and a responsible recycling policy.43

Private initiatives are being pursued also. In 2012, SVTC collabo- rated with the Basel Action Network to develop language to include photovoltaics in its widely used e-Stewards certification process. This is a certification standard for electronics recyclers who adopt certain best practices for material recovery and worker health and safety. The proc- ess included bringing together photovoltaic industry experts, electronics recycling industry representatives, scientists, and policymakers to develop a list of best practices for handling end-of-life photovoltaics,

The welcome rapid growth of photovoltaic installations means there is a limited window of opportunity to establish recycling policies and practices to manage end-of-life photovoltaic waste. If waste issues are not preemptively addressed now, the industry risks repeating the disas- trous environmental mistakes of the electronics industry. Toxic e-waste- made up of discarded computers and other electronics-is shipped to developing nations like India, China, and Ghana for manual disassem- bly, exposing workers and communities to highly hazardous chemicals.

A mandatory EPR and recycling law would achieve several objectives that are critical to a just and sustainable solar energy industry. First, responsible take-back and disposal can help prevent end-of-life photo- voltaic waste from adding to the already burdensome global flows of e-waste. Second, there are real material limitations on the availability of some critical inputs to photovoltaics, which cap the total production of photovoltaics containing certain materials, including tellurium, sil- ver, gallium, and indium. Third, and relatedly, a supply of these inputs from recycled sources could help stabilize the costs and price volatility of inputs that are subject to price fluctuations, such as indium and tel- lurium. Fourth, purifying materials from ores and minerals requires far more energy than recovering them from recycled materials, meaning less energy is needed, and thus less greenhouse gases are emitted, when sourcing recycled materials. Finally, the major burdens of end-of-life photovoltaics will be borne by local governments who operate and own landfills, waste transfer stations, and recycling facilities and that will have to find final disposal places.

u2 I Recycling and Product Stewardship

The long time between when a module is made and when it becomes waste suggests that mandatory recycling will be required. Many inter- views with manufacturers point to the need to eliminate free riders- companies that take advantage of recycling services, but do not pay for them. The Basel Action Network and SVTC identified several best prac- tices for end-of-life photovoltaic module management, which aid safe de-installation and handling, encourage reuse, protect vulnerable com- munities in developing countries, protect workers, and minimize emis- sions from recycling facilities. The lack of recycling in the U.S. can be partly explained by the lack of volume of manufacturing scrap, the small volume of waste arriving from the field, the lack of consideration of the value of materials recovery, and the lack of interest and funding by the government.44 Ultimately, a successful take-back, collection, and high-value-materials recycling system will depend on governance and cooperation throughout the industry.

At the urging of several photovoltaic manufacturers, including First Solar and Abound Solar, California's DTSC, part of the California EPA, engaged in several efforts from 2010 to 2014 to clarify the hazardous waste laws in California with regard to photovoltaics, as several hundred million modules were being installed. California legislator William Mon- ning, whose district includes the Santa Clara/Silicon Valley but also areas further south, in San Luis Obispo, where large utility-scale solar energy projects have been built, took an interest in photovoltaic recycling and authored legislation that would declare end-of-life photovoltaics a type of universal waste.45 Universal waste is a classification of materials that relaxes certain requirements. The state senator's staff said that this would open the door for a more comprehensive EPR law in the future. Washing- ton State became the first state in the country to require manufacturers to have (by 2020) an EPR program in order to sell in the state.

Product stewardship requires a policy framework that takes a cradle- to-grave approach and better labeling and identification schemes to ensure proper disposal and treatment of end-of-life photovoltaic mod- ules. But it is not the only way to improve end-of-life impacts. Design efforts could use the tools and principles of green chemistry to phase out toxic materials.46 Many of the concerns about environmental and worker protection standards to minimize the toxic exposure and human rights abuses that currently plague the global e-waste trade would be obviated by the absence of toxic heavy metals. While green chemistry is discussed in the solar industry in general, the latest technological innovations of interest in solar are perovskite solar cells, which can contain soluble lead

Recycling and Product Stewardship I II3

compounds.47 This suggests that some recycling and take-back scheme will be necessary to prevent heavy metals from entering the environment, even in emerging technologies. End-of-life photovoltaic management will offer the most benefits if recycling services are offered domestically and locally, with minimized transport requirements, and an added emphasis on recycling and recovery of high-value materials. Some argue that the best way to make sure that photovoltaics stay out of landfills is to make them last longer. Products made to last longer would delay the need to find management strategies and solutions. But this alone will not safe- guard the long-term sustainability of the photovoltaic industry. Some form of EPR will best ensure that there are no e-waste crises, that worker health and safety will take priority, and that a long-term supply of critical metals will be recovered, augmenting future supply chains.