The long read: Change to the cast

Share

From pv magazine, July edition

Also sometimes referred to as quasi mono or mono-like, cast mono products first began to appear on PV manufacturers’ technology roadmaps more than 10 years ago. But after an initial wave of interest in the early 2010s was cut down by a lack of suitable processes further down the production line, there was little commercial interest in the technology in the following years.

Until now, that is. Performance improvements and price reductions on the monocrystalline side have forced producers that operate a significant multicrystalline production capacity to look for new routes to a lower cost per watt in order to keep up. And black silicon and other wafer texturing processes reaching mass production means that cast mono is once again an attractive prospect for many manufacturers struggling with weak demand for multicrystalline products.

“We see a very strong shift away from multicrystalline products, and we are basically phasing these out,” says Philipp Matter, President Europe at GCL System Integration. “Now we are converting lines from standard multi to cast mono, and we have successfully shipped a lot of cast mono wafers into the market.”

Analysts also expect multicrystalline products to continue losing market share in the coming years, as players such as Longi, Tongwei and others proceed with the buildout of massive monocrystalline ingot pulling and cell-making capacities. Figures from PV InfoLink place the market share of monocrystalline products at 46% in 2018, and forecast this to increase to 62% by the end of this year, and 75% in 2020.

“New cell capacity expansions all focus on monocrystalline,” says PV InfoLink chief analyst Corrine Lin. “No new multi lines will be built anymore.”

Cast mono is particularly attractive to producers like GCL that have a significant multicrystalline capacity, as they can modify existing lines to produce cast mono, extending the lifetime of these tools and achieving higher efficiencies than would be possible with standard multicrystalline casting. “Our motivation to perfect this process is the 40 GW of furnaces we have for producing multi,” says Matter. “Either we convert them, or we produce for two more years and we shut them down.”

Matter goes on to say that GCL has already converted 6 GW of its multicrystalline furnaces to produce cast mono and plans to double this by the end of 2019, with further conversions likely, dependent on market conditions. “Based on how much we can bring into the market, and also how many other suppliers and potential clients move toward cast mono,” adds Matter, “we can increase this a lot further.”

Sowing the seeds

Cast mono ingots are produced in a directional solidification furnace (DSF) of the same type used in standard multicrystalline production, with a few modifications. The bottom of the crucible is lined with a layer of monocrystalline “seed” ingots. Polysilicon chunks are then loaded on top of these and melted in a controlled process that leaves the seed ingots partially solid at the bottom. The crucible is cooled, and the polysilicon solidifies into a largely monocrystalline ingot. Key differences between this and the standard multicrystalline method are the presence of the seed ingots and the need for even more careful control of both the heating and cooling processes.

The challenge in this process is maintaining the monocrystallinity throughout the ingot, and particularly at the edges. Essentially, the process is trying to persuade the ingot to grow “mono-like,” when its natural tendency is to become multicrystalline. Toward the edges and close to the crucible wall, maintaining monocrystallinity becomes particularly difficult, and the ingots from the edge and corners tend to have a much lower monocrystalline percentage.

One potential solution to this is to simply scrap the ingots produced close to the edge of the furnace, and either sell these as a lower performance material, or remelt them in the next batch. Simon Price, CEO of Exawatt, theorizes that this could mean in a DSF measuring 7×7 ingots, producers would likely be able to use 6×6 as high-quality cast mono, reducing the yield from 49 to 36 ingots per batch.

GCL claims that its cast mono products are more than 99% monocrystalline, and while Matter doesn’t provide numbers on the process yield, he has confirmed that with cast mono production, bricks need to be cut from the ingot “wisely,” and that the number of bricks per ingot is slightly lower compared to standard multicrystalline production.

Once the wafer is made, cast mono can be treated more or less the same as any other in cell and module production. “On the etching side and the first step of cell lines it’s a bit different. And PECVD you have to adapt a little bit, in order to have the cell surface not too shiny,” explains Matter. “But these are minor things, optimizations, not whole process changes.”

Cost and efficiency

Part of cast mono’s appeal lies in its claim to provide efficiencies comparable to monocrystalline at prices closer to multi. At the end of May this year, Canadian Solar set a new efficiency record with a cast mono cell, at 22.28%. Guoqiang Xing, Canadian Solar’s Chief Technology Officer, told pv magazine that this record was achieved on a production line rather than in a laboratory, and all the technologies combined in the cell — including selective emitter, silicon oxide passivation, multi-layer anti-reflective coating, aluminum oxide back side passivation, and advanced metallization, as well as a metal catalyzed chemical etching — or “black silicon” process, are ready for use in mass production. And the company appears bullish about the prospects for its “P5” cast mono products.

“As far as we see, the cast mono can achieve higher cell efficiencies at least for a good portion of the ingot than Cz mono, so the cell efficiency potential for cast mono is underestimated by many,” said Xing. “We are working on mono percentage improvement and dislocation density reductions. The module power output is close to mono. It also passed all reliability tests.”

Producers are also keen to point out that that cast mono wafers are fully squared, boosting the output by providing more surface area than is possible with monocrystalline — GCL claims that its cast mono cells have a 2% larger surface area than monocrystalline cells. Another stated advantage is the ability to easily change to different wafer sizes, something which is more difficult to achieve in monocrystalline ingot pulling. Currently all of GCL’s cast mono products are offered with cells measuring 158.75mm².

Both Canadian Solar and GCL report that they can achieve module efficiencies close to those of monocrystalline, the latter displayed an 18.9% efficient cast mono module at Intersolar Europe back in May, and this can be combined with other technologies to ensure maximum output. “We embed all of our cast mono cells into 12-busbar configurations. Half cut is also possible and we expect to have this available soon,” says GCL’s Matter. “With that configuration we are just one power class below standard mono.”

With efficiencies approaching, but still below, those of monocrystalline products, cast mono will need to demonstrate a further cost advantage to be attractive to manufacturers. Taking a 320 W module as an example, if cast mono is around 5 W lower in output, this works out to 1.56%. So the cast mono wafer has to be more than 1.56% cheaper, and this comes to count for even more at the system level. “The critical thing for cast mono producers is keeping their cost per watt well below that of monocrystalline,” says Exawatt’s Simon Price. “A cast mono wafer would have to be, potentially, a good bit cheaper than one cent per watt lower than mono at system level to be attractive.”

In a case study presented by GCL concerning a 110 MW PV plant in the United States, the company says that with a $0.01/W cost reduction achieved by using cast mono, it was able to reduce the project’s PPA price by 3-4%, without affecting the return on investment.

Conversion not expansion

For now, all of the cast mono that is making its way to the market comes from converted DSF ingot lines. Without a major breakthrough, this looks unlikely to change, and the technology will struggle to disrupt continuing growth in monocrystalline PV. “If you’ve already got a multi furnace that’s been around for a few years, your depreciation cost is lower. It’s a way of extending the life,” says Price. “We don’t see a reason to get into cast mono if you’re not already in multi. If you are, we think it’s plausible that you would convert to cast mono. But it’s still ‘game over’ for multicrystalline.”

The bulk of the capacity expansions happening in Asia, both in wafer production and further along in cells and modules, are focused on monocrystalline technology, though. “If you have an existing DS furnace, running it until you can’t eke any more improvements out of it is arguably a better strategy than buying a new line,” continues Price. “But if you are a greenfield ingot manufacturer, buying a Cz puller is almost definitely a better strategy.”

Price’s sentiment is reflected in the expansion plans of manufacturers. Leading monocrystalline supplier Longi Green Energy Technology plans to expand its wafer capacity to 36 GW by the end of 2019. And while it is the leading player promoting cast mono, it is telling that GCL is also in the process of a major shift toward mono. GCL expects to produce up to 7 GW of mono products this year, and to sell more monocrystalline and cast mono modules than multi for the first time. As part of a joint venture with Zhonghuan Semiconductor, the company is also in the process of expanding its mono wafer capacity to 55 GW by 2022.

“We see the Tier-1 vertically integrated players continue to upgrade their multicrystalline cell lines to produce mono PERC,” says Corrine Lin. “We think monocrystalline will dominate the market after mono wafer’s huge expansion.”

Uptake and outlook

But huge expansions like the one underway in mono wafer production take time to complete, and for now at least cast mono has the advantage of immediate availability. Canadian Solar reports that it has already shipped cast mono products to multiple customers worldwide since the middle of 2018, and GCL says it expects to sell around 1 GW of cast mono modules in 2019.

“We believe monocrystalline wafers will face a shortage in the second half of 2019, and this will leave room for cast mono to gain market share,” continues Lin. “But eventually mono wafers will turn toward oversupply, and since gross margins for this product are high at the moment, there is still room for prices to drop. When the price for mono wafers goes down a lot, we think cast mono will have a hard time surviving.”

Lin says that it is still too early to make a forecast regarding cast mono capacity, but acknowledges that it could quickly, by the end of next year for example, start to replace large portions of standard multicrystalline’s market share.

Whether the technology can last much longer than this, in the face of the rise of monocrystalline PV, remains much less certain. “If you are dependent on casting, it keeps your fleet alive for a bit longer,” concludes Price. “But these players have shown that they are switching to monocrystalline in various ways. They appear to be acknowledging that the game is up for multi. This is just a life extension.”