New! Sign up for our email newsletter on Substack.
Purifying silicon is brutally energy-intensive work. You start with quartzite — essentially very pure sand — and heat it in an electric arc furnace to roughly 2000°C. What comes out is metallurgical-grade silicon, still riddled with impurities. So you refine it further, converting it to a gas, distilling that, then depositing it back as polycrystalline silicon in a process that can run for days, pulling enormous quantities of electricity the whole time. Only then do you grow the ingots, slice the wafers, and begin the delicate business of turning them into solar cells.
All of which means that the technology we’re banking on to decarbonise the planet has, somewhat awkwardly, a carbon problem of its own. And as countries race to install solar at a scale measured in terawatts — the world blew past 1 terawatt of cumulative capacity at the end of 2023, and could hit 80 by mid-century — that manufacturing footprint starts to matter rather a lot.
Now a team of researchers from four UK universities has done the most comprehensive accounting yet of what it actually costs, environmentally, to build the next generation of solar panels. Their study, published in Nature Communications, finds that the solar industry’s ongoing shift to a newer cell architecture called TOPCon could, if combined with smarter manufacturing choices, prevent up to 8.2 billion tonnes of CO₂-equivalent emissions by 2035. That’s roughly 14 per cent of current global annual emissions. Not from the electricity the panels generate — from making them in the first place.
“Multi terawatt-scale photovoltaic manufacturing demands a sharper focus on its full environmental footprint,” says Nicholas Grant at the University of Warwick, one of the study’s authors.
The heart of the issue is a technology transition already underway. Until recently, the standard workhorse of the solar industry was a design known as PERC — passivated emitter rear cell. It works well enough. But manufacturers are now rapidly switching to TOPCon (tunnel oxide passivated contact), which squeezes more electricity from the same amount of sunlight through a cleverer arrangement of layers at the back of the cell. Where PERC uses aluminium oxide for rear passivation, TOPCon adds an ultra-thin tunnel oxide beneath a phosphorus-doped polysilicon layer — a tweak that improves how electrons move through the device.
The question nobody had properly answered was whether this shift is also better for the planet. Bethany Willis and Oliver Rigby at Northumbria University, along with colleagues at Warwick, Birmingham, and Oxford, set out to compare both technologies across their full manufacturing lifecycle, from quartz mining to finished module ready for shipping. They assessed 16 separate environmental impact categories using life-cycle assessment modelling, and they got real production data from an international TOPCon manufacturer rather than relying entirely on estimates from the literature.
The results were fairly emphatic. TOPCon came out ahead in 15 of those 16 categories. Its climate-change emissions were 6.5 per cent lower per unit of electricity capacity — a modest-sounding number that scales dramatically when you’re talking about terawatts of deployment. There was one exception, though: TOPCon uses more silver for its electrical contacts, pushing mineral resource consumption up by about 15 per cent. Silver is already one of the industry’s pinch points, and the team flags it as the trade-off that needs watching. Copper contacts are one potential fix; getting silver use down to 5 milligrams per watt — a target that researchers have discussed — would cut the metal-use impact by more than 40 per cent.
But the really striking finding wasn’t about which cell design wins. It was about where you build the things.
The electricity consumed during silicon purification — those furnaces, those days-long deposition runs — dominates the environmental footprint, contributing up to 62 per cent of the total module impact in some categories. So the carbon intensity of the local grid matters enormously. Manufacturing a TOPCon panel in India, where the grid still leans heavily on coal, produces 0.95 kilograms of CO₂-equivalent per watt peak. The same panel made in Europe comes in at 0.40 kg. Shift production from China to Europe and you could halve the climate impact, according to the team’s modelling.
“We are at a critical moment where solar power is rapidly scaling to become a significant portion of global electricity generation,” says Sebastian Bonilla at the University of Oxford. The study, he says, helps identify the choices of materials, technologies and manufacturing locations that will minimise harm.
The team projected these numbers forward to 2035, incorporating expected improvements in panel efficiency, reductions in polysilicon and silver consumption, and the gradual decarbonisation of national grids. Under their most optimistic scenario — where you combine the switch to TOPCon, manufacturing improvements, and cleaner grid electricity — cumulative manufacturing emissions drop by 8.2 billion tonnes of CO₂-equivalent compared with the worst case. And that’s just the supply side. Solar panels installed between 2023 and 2035 are projected to avoid more than 25 billion tonnes of carbon emissions over their lifetimes by displacing fossil-fuel electricity. Even factoring in manufacturing, solar PV will emit just 0.017 kg CO₂-equivalent per kilowatt-hour by 2035. China’s grid, for comparison, will still be at 0.608.
None of this means solar’s manufacturing footprint is trivial, mind you. The study estimates that building all the panels the world plans to deploy could generate up to 13.8 billion tonnes of CO₂-equivalent by 2035 if nothing changes. The point is that most of that is avoidable; it depends on decisions being made right now about where factories get built and what powers them.
There are uncertainties, as the researchers are careful to note. Their model for India’s electricity mix relies on data that’s over five years old. China’s grid is decarbonising rapidly, but predicting exactly how fast is tricky. And the study only covers what happens up to 2035 — beyond that, newer technologies like perovskite tandems and silicon heterojunction cells will start claiming larger market shares, and their environmental profiles remain less well understood.
Still, the core message lands with considerable force. “Even when manufacturing impacts are considered, solar photovoltaics remains one of the lowest-impact and most sustainable electricity generation technologies available over its whole life cycle,” says Neil Beattie at Northumbria University, the study’s senior author. We should, he argues, be deploying it at scale — now.
What makes this work matter, perhaps, is its timing. The solar industry isn’t debating whether to switch from PERC to TOPCon; it’s already doing it. The question is whether the other changes — cleaner factory grids, reduced silver, thinner wafers — happen quickly enough to capture those 8 billion tonnes of savings. The furnaces are running either way.
Study link: https://www.nature.com/articles/s41467-026-69165-x
If our reporting has informed or inspired you, please consider making a donation. Every contribution, no matter the size, empowers us to continue delivering accurate, engaging, and trustworthy science and medical news. Thank you for standing with us!
This site uses Akismet to reduce spam. Learn how your comment data is processed.
