Developments seem to have taken a fast track in the crystal growth segment, at least from the International Roadmap for Photovoltaics (ITRPV) perspective. According to the 6th edition of the roadmap, which was published in April 2015 and updated in the summer, some of the key developments in ingot making are ahead of schedule – this includes market penetration of high performance multicrystalline silicon, transition to Gen 8 ingot format, monocrystalline dominating over casted material in the long run. At the same time, the progress of n-type technology has slightly taken aback step. What has not changed is the dominating market share of casted silicon over monocrystalline. And there are several interesting developments to report in detail.
Multi and mono not enough
As in the past, the crystal growth segment is part of the ITRPV's 2015 roadmap and the developments in this part of the PV value chain are discussed under two subsections – manufacturing and products. As predicted in the 5th version in 2014, the current roadmap confirms the dominance of casted silicon material with a market share of about 65%, which it expects to shrink to less than 50% in 2022, 2 years earlier than what was expected a year back. 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 n-type multi.
High performance multi is a better bet
In the casted silicon segment, high performance multi (HPMC) looks quite promising and surpassed the expectations. Quite similar to quasi monocrystalline silicon, HPMC is also based on the seed-assisted casting principle, but with the difference that several seed crystals are being employed. The 2014 study was conservative about penetration of this new class of material, forecasting a 50% share in the casted silicon segment and 20% overall. The 2015 study, however, indicates this modified multicrystalline 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. The high performance material is expected to replace conventional multicrystalline by 2022, 2 years before than expected in the previous roadmap. A very interesting prediction of the 6th ITRPV is that not only it anticipates the re-entry of quasi-mono (also known as mono-like) in 2017, but also expects that the technology will attract an increasing number of followers, attaining 8% market share in 8 years from then and almost nothing today. Indeed, 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 so quickly and which have limited efficiency), it quickly lost momentum and its commercial followers.
ITRPV has not altered its stand about the progress in manufacturing of casted material. The 2015 roadmap continues to say that G6/G7 ingot sizes with masses up to 1,000 kg are already in production. The transition to G8 crucibles would lead to 1,200 kg ingots, which is expected to take place in the next 4 years, that´s 1 year earlier than the timeframe defined in the previous study. The 2014 roadmap voiced technical concerns associated with manufacturing Gen 8 ingot formats, especially the availability and price of graphite parts, mechanical stability, energy requirement and cycle times. It is not clear from the 2015 study if these were addressed or if ITRPV is simply ignoring them for now. However, we did see multi ingots of 1,200 kg already on display at the SNEC show in Shanghai in May 2014, though termed G7 ingots (see intro photo).
Continuous CZ gonna be heavy
However, in the long run, casted material, as a segment, is expected to lose its ground to the carefully grown monocrystalline silicon ingots. The latter, from its current level of 35%, would gain 12% absolute in the next 10 years, reaching 47%. The 2015 roadmap has given a special emphasis to monocrystalline ingot making. The study indicates that the continuous CZ process, such as the one offered by GT Advanced Technologies, is expected to increase acceptance among producers. The continuous CZ process enables pulling several monocrystalline ingots per crucible, thereby reducing production costs. The technology is also expected to be instrumental in increasing ingot weight, doubling from the current level of about 150 to 300 kg by 2025. With these sets of undeniable advantages, this technology is expected to get wider acceptance, resulting in a market share of 40% in 10 years, at par to traditional CZ, raising from its current share of about 10%.
Mid-term rise of float zone
As to standard CZ technology, the 2014 roadmap indicated that increasing the productivity at the tool level is possible through increasing pulling speed, reusing crucibles and changes in the cooling setup of the furnace. However, there are no updates available in the 6th edition. The 2015 roadmap is optimistic about the prospect of the float zone (FZ) process, especially due to its high efficiency potential. From almost negligible presence in today's market, float zone is expected to rise to well above 5% in 2019 and attain close to 20% in 10 years from now, whereas float zone was not even mentioned in the previous document.
Slightly delay for n-type
Within the monocrystalline arena, ITRPV sees a clear shift from p-type to n-type. The latter, with its share of about 5% in 2014, is expected to cross 10% in 2017, while the previous roadmap forecasted that these levels would be attained 1 year earlier. Casted n-type silicon has not even got a spot in the 2015 roadmap; the previous study was also not optimistic about its prospects, seeing it at less than 5% in 2024. The main technical hurdle of the technology is the higher variation in wafer resistivity along the ingot height originating from segregation coefficients of the dopant. Luckily, monocrystalline silicon escapes this due to its completely different growth mechanism.
Alternate methods may find a chance
Alternate approaches such as kerfless and ribbon technologies have been detected by the ITRPV's radar for the first time. These technologies are expect to make an entry in 2019 and considerable volumes will be produced by 2025, contributing 10% of global production.
All these finding indicate that the crystal growth segment is going to be very dynamic in the coming 10 years. HPMC will replace traditional casted material, while quasi-monocrystalline may get another chance. Continuous CZ is going to be as strong as standard CZ. Float zone also seems to have good prospects on account of its higher efficiency potential. On the tools front, the throughput of both crystallization technologies will increase by 40 to 50% over the next 10 years. That means developments are at least "on track," if not on the fast track.