Shota Furuya

Pieces日本語版

Doesn't manufacturing solar panels emit CO₂ and consume rare metals?

“Making solar panels emits so much CO₂ that they are bad for the environment in the end.” “They use so many rare metals that they cannot be sustainable.” These doubts come up again and again. On social media you even find the claim that a panel never generates as much energy as it took to build. How much of this is true?

The short answer. (1) A modern panel repays the energy of its manufacture in one to three years and then runs for 25 to 30, so life-cycle CO₂ emissions are a twentieth to a fortieth of coal’s. (2) The main materials are silicon, glass, aluminium, copper and silver, and silicon is among the most abundant elements on Earth; the image of heavy rare metal consumption is far from the reality. (3) Real problems do exist, in silver use, the concentration of polysilicon manufacturing, and immature recycling arrangements. What follows separates the numbers from the arguments.

Life-cycle emissions and energy payback time

The standard method for judging how clean solar is comes from life-cycle assessment, which counts everything from mining raw materials through manufacturing, transport, installation, operation and disposal. Per kWh generated, solar emits about 20–50 gCO₂/kWh, against 820–1,050 gCO₂/kWh for coal and about 490 gCO₂/kWh for gas — more than an order of magnitude apart (IPCC, 2014; UNECE, 2022). It is higher than nuclear (5–12 gCO₂/kWh) or onshore wind (11–12 gCO₂/kWh), but against fossil fuels it is overwhelmingly cleaner.

The other important measure is energy payback time (EPBT): how many years of generation it takes to recover the energy put into manufacturing. Fraunhofer ISE’s 2024 analysis puts EPBT for crystalline silicon panels made and installed in Europe at about 1.0–1.3 years, and at about 1.3–1.8 years for Chinese modules operating in Europe (Fraunhofer ISE, 2024). In sunnier regions the estimates fall below a year. Since panels run for 25 to 30 years, some 95% of their life is net positive generation. The claim that manufacturing uses more energy than the panel produces rests on 1990s data and does not describe today’s technology.

What a panel is made of

By weight, a crystalline silicon module is typically about 70% glass, 10% aluminium frame, 6–7% polymer encapsulant, 5% silicon cells, 1% copper and around 0.05% silver (IEA-PVPS, 2024). About 97% of the world market is crystalline silicon, so nearly everything we call a solar panel is this type.

Silicon is among the most abundant elements in the Earth’s crust, second only to oxygen. Its feedstock, silica, occurs all over the world and is not a scarce resource. The same goes for glass, aluminium and copper: common materials with established recycling chains. The idea of a solar panel as a lump of rare metal is wrong at the level of the bill of materials.

One material does deserve attention: silver, used in the electrodes on the cell surface. Estimates put the solar industry at about 13% of world silver consumption (Silver Institute, 2024). Heraeus and Japan’s AIST among others have cut silver per panel by about two-thirds over the past decade, and work on copper electrodes continues. Major manufacturers — Longi, JinkoSolar, Trina Solar — have set out plans to move to silver-free cells in stages by the early 2030s.

Examining the “rare metals” claim

“Rare metals” is often confused with rare earths. To be precise: rare earths are a subset (the 17 elements including neodymium and dysprosium) of the 31 mineral species designated as rare metals — industrially important and supply-constrained. Crystalline silicon cells use no rare earths at all. The IEA’s critical minerals report states plainly that rare earth demand from solar power is close to zero (IEA, 2023).

The rare-earth-intensive renewable technology is rather wind turbines with permanent magnets, which use neodymium — a significant difference between wind and solar.

Thin-film cells (about 3% of the market) do use tellurium in CdTe and indium, gallium and selenium in CIGS, and tellurium and indium are genuinely scarce. But that leaves the 97% of the market held by crystalline silicon untouched, and thin-film uses these materials in grams, which is not heavy consumption. Perovskite cells, the next-generation technology attracting attention, contain a little lead, which is a real issue, but their raw materials are not scarce — and this is a field where Japan leads in research.

In short, the picture of solar exhausting the world’s scarce metals is contrary to fact. Where scarce-resource constraints really bite is in battery materials for EVs — lithium, cobalt, nickel — and lumping solar in with them is inaccurate.

Geographic concentration and the supply chain

The real risk in solar is not scarce metals but the geographic concentration of manufacturing. On the IEA’s 2024 figures, China accounts for about 93% of polysilicon production, 97% of wafers, 85% of cells and 80% of modules (IEA, 2024a). Xinjiang alone is estimated at about 40% of world polysilicon, and concern over forced labour led the United States to enact the Uyghur Forced Labor Prevention Act in 2022 and to block imports under it.

Concentration is a problem for the carbon footprint too. Many Chinese polysilicon plants run on coal-fired electricity, which puts the life-cycle emissions of their modules 1.5 to 2 times above European modules made with renewable power. How clean solar is depends on where, and on what electricity, it was made.

Policies responding to this — the US Inflation Reduction Act, the EU’s Net-Zero Industry Act (2024) and India’s production-linked incentive scheme — are diversifying manufacturing geographically. Japan’s METI has set a target of raising the share of domestically or near-domestically produced panels to 50% by 2030, with mass production of perovskite cells (Sekisui Chemical, Panasonic and others) as a pillar of that.

Disposal and recycling in practice

Worry about what happens to panels at the end of their lives is persistent. Large volumes of retired panels are expected in Japan from the 2030s: environment ministry and METI estimates put disposal at 500,000–800,000 tonnes a year in the late 2030s (Ministry of the Environment, 2023).

But 85–95% of a panel by weight is recyclable, and the glass and aluminium go into existing recycling lines. The EU classes panels as electrical and electronic waste under the WEEE Directive, putting collection and recycling responsibility on manufacturers. France’s Soren (formerly PV CYCLE France) has expanded capacity and achieved 92% recovery of cell materials with new equipment in 2024.

In Japan, the scheme requiring funds to be set aside for decommissioning renewable installations entered full operation in 2024, obliging FIT and FIP operators to accumulate disposal costs through withholding. A NEDO project is demonstrating horizontal recycling, returning glass to the low-iron glass used in new panels. A nationwide processing network is still being built, though, and traceability to prevent illegal dumping and improper handling is the major task ahead.

Objections, limits and common misreadings

Several caveats belong here.

First, clean is not the same as no impact. Mining, refining and transport do not fall to zero. Silica quarrying and the local air pollution associated with carbon electrodes are real. “Far less than fossil fuels” and “zero impact” are different claims.

Second, LCA figures come in ranges. Irradiance where the panel is installed, the power mix where it was made, transport distance and recycling assumptions all move the number within that 20–50 gCO₂/kWh band. It pays to check the assumptions rather than treat a single figure as absolute.

Third, land use and ecosystems at the installation stage remain a separate question from manufacturing emissions — the discussion around nature-positive energy and agrivoltaics. A large solar farm that clears forest provokes conflict at the siting stage however clean its manufacturing was.

Fourth, social acceptance. Human rights and labour at the manufacturing stage, illegal dumping at the disposal stage: these are problems of governance, not technology. Solvable in principle, they breed the belief that solar cannot be trusted when the rules lag behind.

What matters next in Japan

Three points stand out. First, information literacy. Old LCA data from the 1990s, and claims about thin-film cells or batteries applied to crystalline silicon panels, still circulate. Today’s EPBT is one to two years, and rare earth use is zero — sharing those basic facts is where the discussion should start.

Second, rebuilding domestic manufacturing. Japan’s solar industry, once the world leader, fell back badly in the 2010s, but perovskites, building-integrated PV and plug-in solar leave room for a return. A domestic manufacturing base running on renewable electricity also has strategic value in lowering the emissions of manufacturing itself.

Third, making the disposal and recycling system work. The funding scheme has started, but collection, transport and material recovery are largely left to the private sector, and differences between municipalities are emerging. Completing the circular economy is the foundation of the industry’s credibility — and the best answer to the claim that renewables are bad for the environment.

Summary

  • A modern panel’s life-cycle emissions are 20–50 gCO₂/kWh, a twentieth to a fortieth of coal’s.
  • Energy payback time is down to one to two years, so most of a 25–30 year life is net positive.
  • Crystalline silicon, 97% of the market, uses no rare earths. Its materials are glass, aluminium, copper, silver and silicon, and silicon is among the most abundant elements on Earth.
  • Silver use keeps falling, and silver-free cells are in view for the early 2030s.
  • The real risk is not scarce metals but the concentration of manufacturing (China at 80–97%) and human rights and supply chain transparency.
  • 85–95% of a retired panel by weight is recyclable. Japan’s funding scheme began in 2024; traceability is the work ahead.

The instinctive worries — heavy CO₂ in manufacturing, heavy consumption of rare metals — both miss the reality by a wide margin on current data. The questions that matter have moved beyond technology, to where manufacturing happens, on what electricity and under what labour conditions, and how disposal is brought into a loop — that is, to governance and supply chain policy. Those questions apply to every source of electricity, nuclear and fossil included. Solar is not a special case.

References and data sources

Life-cycle analysis and EPBT

Critical minerals and supply chains

Disposal and recycling

Policy and industry

← Back to list