Towards the top: In the long run, casted silicon is expected to lose ground to carefully grown monocrystalline silicon ingots (the top is pictured here). ITRPV expects mono to increase its share to 48% in 10 years from about 34% in 2015.
- High performance multi is expected to replace traditional material by 2026 in casted silicon segment
- Switching to Gen 7 ingot size expected to start from the beginning of 2017, while transition to G8 ingots of 1,200 kg is expected after 2020
- Continuous CZ increasingly replacing traditional CZ, attaining 44% share within monocrystalline making by 2026
- Ingot mass in monocrystalline to double in next 10 years
“Deviation” is probable an apt expression to describe technology trends in the crystal growth segment of PV manufacturing, when comparing the new 2016 International Roadmap for Photovoltaic (ITRPV) with the previous version. ITRPV came out with its 7th edition in mid March 2016 – and key crystal growth developments described in this latest version are taking place slower than anticipated. In fact, the 6th edition from 2015 put the advancements in crystal growth segment on the fast track (see article What To Expect In Crystal Growth). On the other hand, the projections made in the current roadmap are rather inline with forecasts made in the 5th edition from 2014. This may sound like back in time by 2 years – but that’s not completely true. There are certain areas where developments are at par with last year’s prediction, and there are even instances that show progress is above expectations.
Some faster, others slower
Market penetration of high performance multicrystalline silicon, transition to Gen 8 ingots and market penetration of n-type are the areas in which the developments are rather taking a slower pace. Reentry of quasi-mono and wide spread use of float zone (FZ) silicon as well as prospects for kerf-free wafering technologies are the technical branches about which the previous roadmap was quite upbeat and aggressive, while the 7th roadmap has strongly slashed theses prospects. What has not changed is the pace monocrystalline is supposed to gain acceptance over multi. Replacing traditional CZ with continuous CZ in the long run is one such technical advancement that has exceeded previous expectations. In short, when it comes to the progress in the crystal growth segment, the 7th roadmap deviates a lot from the earlier version, which also means several interesting developments to report in detail.
Multi and mono not enough
As in the past, ITRPV 2016 discusses the developments in the crystal growth segment also under two subsections – manufacturing and products. The products section of the current roadmap confirms the dominance of casted silicon material, including standard multicrystalline, high performance multicrystalline (HPMC) and quasi monocrystalline. The market share of casted silicon material was above 65% in 2015 and is expected to shrink to less than 50% in 2026, while the 6th version in 2015 predicted this level to be reached already in 2022. This is about 4 years later than what was expected than a year back, but close to the prediction of the 5th edition (see graph).
ITRPV again emphasizes that the current order of distinguishing the crystal class with monocrystalline or multicrystalline is not adequate. The crystalline silicon materials market will diversify, thus the PV community would have to get used to more specific classifications such as p-type mono, p-type mono-like, p-type high performance multi, p-type multi, n-type mono and advanced epitaxial growth technologies.
High performance multi is a better bet
In the casted silicon segment, HPMC looks quite promising. Quite similar to quasi mono, HPMC is also based on the seed-assisted casting principle, but with the difference that several seed crystals are being employed. This relatively new class of material has not only grabbed a share above 60% in the casted segment, it has taken as much as 40% absolute in the overall crystal growth segment in 2015, which closely follows the prediction of the previous roadmap. However, ITRPV lowered its estimates in the long term. HPMC is expected to replace conventional multi in about a decade – by 2026, which is 4 years later than anticipated in the previous version. And the current projections are again somewhat closer to the 5th edition released in 2014. However, the world’s leading wafer maker GCL just announced in its March released financial results 2015 to replace traditional multi with high efficiency multi technology completely this year (see article Sometimes Back In Time)
What about quasi-mono?
Regarding the third variant of the casted material class, quasi-mono (also known as mono-like), the previous roadmap expected its reentry with growing acceptance in the long term with a market share of 5% in 2019 and 8% in 2022. However, the 2016 version revised the projections considerably downwards – from about 1% in 2015 close to 2% in the next 8 years and slightly above 3% in 2026. Actually, quasi-mono was a very hot topic a few years ago and was considered to replace standard multi – with a few companies strongly focusing on the technology, such as ReneSola from China. Due to the lack of maturity of quasi-mono (mostly related to defects it could not overcome as quickly as anticipated and limited efficiency), it quickly lost momentum and its commercial followers. However, GCL also announced in its 2015 annual results to expand quasi-mono wafers from basically 0 up to 5% this year.
ITRPV is now more elaborate on its views about the progress in manufacturing of casted material. Unlike the 6th edition, which stated G6/G7 ingot sizes with masses up to 1,000 kg were already in production last year, the 2016 roadmap says that G6 ingots weighing above 700 kg represent the mainstream of today’s mass production. Then, switching to G7, which enables producing ingots of up to 1,000 kg is expected to start from the beginning of 2017. The transition to G8 ingot of 1,200 kg is expected after 2020, while the previous roadmap believed this shift would happen a year earlier.
Only Continuous CZ on the rise
However, in the long run, casted material, as a segment, is expected to lose ground to the carefully grown monocrystalline silicon ingots. The latter, from its level of about 34% in 2015, would gain 14% absolute in the next 10 years, reaching 48%, which is inline with the estimations made in 2015. Increasing acceptance of the continuous CZ process is yet another technology trend that is at par with the previous forecast (both in near and midterm). It is the only technology that is on the fast track – expected to gain market shares faster than anticipated in the 6th roadmap. The continuous CZ process, such as the one offered by GT Advanced Technologies, enables pulling several monocrystalline ingots per crucible, thereby reducing production costs.
The technology is also expected to be instrumental in increasing ingot weight to more than twice of today’s level – of about 200 kg in 2016 to 430 kg in 2026. While the previous roadmap has also talked about doubling the CZ ingot mass in 10 years, the data differs much in absolute terms – the former forecast looked at about 150 kg in 2015 and an increase to 300 kg by 2025. However, it is hard to believe that the average weight of CZ ingots changed by 50 kg, which is more than 30%, within a year. Nevertheless, continuous CZ comes with undeniable advantages and there is no doubt that technology will see a wider acceptance.
The 7th ITRPV edition expects continuous CZ to gain a market share of close to 45% in 2026 from about 10% in 2015. Even then, traditional CZ is expected to keep its leading position – with a 50% presence. However, the previous roadmap had predicted continuous CZ as well as traditional CZ to stay at somewhat lower levels – at about 40%. That’s because the 2015 version was over-optimistic about the growth prospects for the float zone (FZ) process, well known for its high efficiency potential. But the current version significantly lowered its allocations – from almost nothing to about 5% in 2026; the 6th edition expected FZ to rise from 5% in 2019 close to 20% by 2025.
N-type again delayed
Within the monocrystalline area, as in the past, ITRPV sees a clear shift from p-type to n-type. However, the roadmap team has lowered their forecast for 3 years in a row, especially for the onset of the technology. While the 6th edition estimated that n-type would account for 7% of production in 2015, the current document estimates a 4.5% share in 2015, just 5% in 2016 and become only significant, exceeding 20%, after 2020. A decade from now, by 2026, n-type is expected to reach above 30%. Casted n-type silicon has neither been mentioned in the current not the previous roadmaps, indicating its thin prospects.
Lean chances for alternate methods
Alternate approaches such as kerfless and ribbon technologies, which bypass ingot making, have entered the ITRPV’s radar last year, and also in this case, the previous roadmap was more aggressive about the prospects. These technologies were expected to make an entry in 2019, producing considerable volumes by 2025, when they would contribute 10% of global production. However, the latest ITRPV version is much more down to earth, not even allocate half of that share. Alternate approaches are expected to surface only after 2020 and gain a share of 3% by 2026.
Overall, all these developments indicate the technology trends in crystal growth segment are omni directional – at par, ahead and delayed. It is actually a lot of deviation compared to the ITRPV previous version.