According to Chinese media reports, the amounts now invested across a host of perovskite start-ups has now reached “billions” – billions of yuan, so hundreds of millions of dollars given the 7:1 exchange rate.
We have been aware of some of these names – such as Microquanta, GCL Perovskite, and Utmolight – for years, as they have posted impressive and occasionally record-setting research records. Others such as Aolian Ae&Ea may be dubious, with Aolian under investigation on suspicion that its claims of a perovskite business adventure being fraudulent. Plans already being touted for multi-GW production lines to be established in just the next few years are also dubious.
Microquanta credibly claims to have begun work on a 12 MW perovskite installation to serve its own power needs last year, and to have shipped 5,000 modules which had an area of 1,245mm by 635mm. These ‘alpha’ modules have been certified under the IEC61215 and IEC61730 degradation standards as of January.
Despite this, mass adoption of perovskites may yet come from the West first. Several Western companies have announced plans for perovskite tandems in the 2025 to 2027 timeframe, with the latest such announcement coming this week from a Dutch-German consortium including QCells, pursuing a monolithic two-terminal SI tandem. One major consortium member, the Netherlands Organization for Applied Research (TNO), is also pursuing a more advanced but more difficult four-terminal technology.
The West still retains some technological edge in perovskites and even in photovoltaics in general, aside from certain process which the Chinese have mastered due to industrial scale. Not to mention Saule Technologies which already has at a small scale for the device market.
There are four main obstacles to perovskite commercialization.
- While much progress has been made, the standard set by silicon PV is a thirty-year lifespan, and this is made more relevant by many planned perovskite products being Si-perovskite tandems. This is still the biggest obstacle.
- Potential for toxic pollution from lead or other optional molecules in the perovskite compound, which might leak. While these toxic elements can be substituted, so far the substitutes make for an inferior photovoltaic product.
- Scaling up from test cells to larger cells and full module scale in the face of efficiency losses caused by deformations in the smooth coating.
- Integration of two different photovoltaic materials in a ‘cell-level’ tandem. First commercialization may dodge this with “module-level” tandems, in which the perovskite and silicon are effectively different modules placed in the same infrastructure thanks to perovskite being transparent with a different bandgap.
A fifth problem right now may be that perovskites are more expensive, but this is only because they as yet have no production scale. Perovskites are inherently cheaper.
Utmolight, one of the leading Chinese companies, has recently posted an 19.9% efficiency result over 809.8cm2, which is a record at that scale. The company states it is working from the small-scale of 10cm2 to 200cm2, then enlarging its results. At the 16.7cm2 size, the company’s perovskites rate 22.9%. Utmolight claims to have a 150 MW pilot production line commissioned in December, and several other companies in China also claim to have 100 MW production lines. The most charitable interpretation of such claims, not too unlikely, is that these production lines exist but are still being used to produce small batches of test products. There is a lot of room for testing different products in small batches because a perovskite is a molecular compound defined by its structure, not by the exact elements used. Many combinations are possible, hence the existence of lead-based and lead-free perovskites.
A few years ago, the gap between 1 cm2 test cell efficiency and larger-scale efficiency was around 10% absolute – so 20% efficiency would fall by half, to 10%, when scaled up, due to deformations in the smooth coating. In mainstream silicon PV there is an analogous efficiency gap between cells and modules, but it is only 2%, not 10%, so perovskites have a lot of work to do.
The other relevant recent perovskite efficiency records are 32.5% on a tandem cell, set by HZB in December, and 25.8% set by UNIST in early 2022 for a single-junction test cell, though the latter of those may have been slightly surpassed since. Comparing Utmolight and UNIST results, we find that the ‘small-large efficiency gap’ is now 5.6%, far below the 10% gap of a few years ago. It appears that companies no longer trumpet 1 cm2 efficiency results as much, no doubt due to their irrelevance as more research groups are capable of producing larger-scale smooth coating.