Polysilicon from Xinjiang: a balanced view

Most of the operating cost of polysilicon is electricity, with labor cost only a small component, so using unskilled and unhappy workers is likely to cost more in outages and waste than is saved. However, polysilicon producers based in China’s Xinjiang region are likely to be cooperating closely with the local government to gain access to low prices for coal-fired power. Practically, the wider issue is unlikely to go away, as it is a lever for Western countries to exert pressure on China’s government.

Trade wars in solar have seldom helped the sector (or even most Western manufacturing companies), but they have also not slowed it much in the long term, either. BloombergNEF’s view on possible sanctions on polysilicon from Xinjiang is that they might increase the profitability of companies like First Solar, which does not use polysilicon, and of polysilicon producers with capacity outside Xinjiang, but they would be unlikely to have a much wider impact on solar new build. Solar manufacturers have over and over again proven themselves flexible and willing to adapt supply chains to trade legislation, for a small price.

Xinjiang will make half the polysilicon the PV sector takes in this year, BNEF projects, while the United States and Europe will combine to account for a quarter of global demand. There are large Chinese companies – Tongwei, Asia Silicon, GCL-Poly – making silicon in other parts of China. Xinte Energy, which thus far only manufactures in Xinjiang, will set up its next 100,000-ton per year plant in Inner Mongolia, which would by itself meet roughly one-fifth of current demand.

Complex story

All of this suggests that Western nations should be able to procure PV equipment without a trace of Xinjiang labor involved in its production – if they need to do so. The story is a good deal more complex, however. Few ingot makers are fully committed to any one supplier of polysilicon. While long-term contracts for polysilicon supply are once again being signed (BNEF tracks contracts for delivery of 1.6 million tons of polysilicon – enough to make about 597 GW of solar modules – between 2021 and 2026), these will inevitably be augmented with product bought on the spot market.

No ingot maker wants to risk being fully dependent on a single supplier, or even a handful of suppliers, to keep their capex-intensive factories running and to fulfil their own contracts to supply wafers. Consequently, almost any silicon-based module may well contain some Xinjiang polysilicon. Even ingot makers that have signed contracts for polysilicon made outside Xinjiang would need to be offered premium prices to commit to using these products only, because a silicon supplier factory outage or unexpected demand could be extremely expensive.

All of this is entirely consistent with how global commodity markets typically work. It is what allows consumers generally to buy the products we want whenever we want them. It is also why those of us who, for example, buy clothes from brands marketed as ethical can experience higher prices and long delivery delays. (In the case of my household, it’s why my rapidly growing daughter has trundled off to nursery in pyjamas the last few weeks as we await trousers that fit to arrive in the post.)

In reality, few consumers have the wherewithal (or patience) to allow concerns about sourcing to dictate their everyday purchases. The same certainly applies to the solar supply chain. The solar industry has been defined by brutal competition for a decade and a half now. Manufacturers in a constant fight for survival have little time to analyze the sourcing trail of the equipment they receive, especially if they are several links in the value chain removed from polysilicon production.

Any downstream player that decided not to buy equipment which might contain Xinjiang silicon would likely find there is essentially no “pure” module choice beyond supply-constrained First Solar. This would inevitably inflate the final prices consumers pay – and be a boon to fossil fuel alternatives.

Rigorous requirements

All of that said, it is possible that Western governments may apply sanctions on polysilicon made in Xinjiang, backed with rigorous supply-chain tracing requirements. If so, solar manufacturers will need to respond – and should increase transparency in any case, where it is feasible.

Improving supply chain tracing would have the added benefit of facilitating carbon footprint monitoring, which to date has been used to erect trade barriers without breaking World Trade Organization rules. South Korea, for example, already has a carbon footprint specification, which essentially excludes modules with a calculated footprint above 830kg of carbon dioxide per kW. This favors six Korean manufacturers, which have obtained the certification. France also has a program to do this, based on calculations at the country level.

I am not convinced that carbon footprint certifications for solar are a net help in the urgent business of global decarbonization, but they are a tool which helps some governments feel comfortable subsidizing, or increasingly just allowing, solar build.

One thing is certain – the solar industry will continue to adapt and adjust, driving for efficiency and lower resource intensity as it sails past 150GW/year new build for the first time in 2021.

Jenny Chase, BloombergNEF