PHOTON's 10th Solar Silicon Conference

CONFERENCE REPORT

FOUR MORE YEARS 

March, 2012: The silicon sector’s woes are indeed as dire as you’ve been hearing. And there’s little relief in sight. That was the theme at PHOTON’s 10th Solar Silicon Conference, in Berlin, where the dour industry outlook was reflected in the mood of the nearly 200 attendees.

Leading industry and company representatives agreed that a massive supply overhang will only get worse this year, forcing prices down toward the cash cost to produce. Even after the likely bankruptcy of dozens of non-competitive players, the six or so largest producers will still have more than enough capacity to meet expected demand – for at least four more years.

 
“Only a few solar silicon (producers) are likely to survive,” said Martin Meyers, of PHOTON Consulting. He thinks just 11 manufacturers will make it to 2013, with any company unable to produce for less than $30 per kg in danger of irrelevancy. There is 200,000 kg of supply that can be produced for less than that cost already, easily exceeding anticipated 2012 demand. “The days of $40 (per kg) contracts are gone,” says Goran Bye, CEO of LDK Silicon. Henning Wicht, a consultant with IHS iSuppli, predicts spot prices will fall to $22 per kg in 2012 – and stay there in 2013. “The overcapacity is not going away,” he says. Meyers says prices will range from the all-in cash cost, roughly $20 to $25 at industry leaders, to a ceiling of $35 for entrenched suppliers with established customers. A year ago, silicon sold for 2.5 times that upper figure. “The willingness to pay for silicon will never approach $80 to $100 again,” he says.

 
That will be doubly true in 2012, as the solar sector gets set for its worst year of installation growth. Meyers predicts global installations will actually contract in 2012; Wicht says they will remain essentially flat at 27.5 GW or so. Either way, the 50 percent compound annual growth rates the industry has become accustomed to will not occur. That will lead to capacity reductions in silicon, with Meyers predicting roughly 25 percent of 2012 capacity not making it to the following year. “Plants that idle for 12 months have nearly zero chance of starting again,” Meyers said. He and Wicht both expect massive expansion plans announced by the largest companies – including Hemlock Semiconductor, Wacker Chemie and OCI – to be delayed, and in some instances, possibly scrapped.

 
If any of the smaller companies present were offended by Meyer’s gloomy outlook, none said so. Instead, they put forth enthusiastic cases for why their businesses will persist, either as niche producers or with differentiated technologies. Activ Solar, based in Vienna but producing in Ukraine, says its systems development business will provide an outlet for silicon output. The company, which joined essentially all presenters in refusing to disclose costs, has very low capital costs after refurbishing a decades-old plant. Silicor, formerly CaliSolar, says its upgraded metallurgical silicon will provide no performance hindrances at the cell level for less cost than traditional silicon. Elkem, likewise a uMg Si producer, makes similar claims.

 
“You have to be innovative when the market is bad,” said Louis Parous, head of advanced silicon materials at equipment maker centrotherm photovoltaics. He predicts silicon companies will increasingly integrate upstream into metallurgical silicon production, making, instead of buying, the main raw ingredient in solar cells. Companies that do so can shave up to $1 per kg from the cost. Centrotherm Photovoltaics offers integrated systems back to metallurgical production, which CTO Peter Fath says greatly influences silicon quality. He warned that too much focus on cost could compromise quality, especially with increasing demand for higher-efficiency cells. “You need a stable and constant supply of silicon to give a steady ingot yield,” he said. Fath thinks 50 percent of the current silicon capacity is not economical. “We cannot lose sight of quality as we reduce costs.”

 
But costs will ultimately determine the winners, the conference presenters agreed. Meyers says companies need to secure feedstocks and reduce electricity consumption to drive costs lower. Controlling the capital costs of any new plant also matter, says Peter Keck of GT Advanced Technology, the equipment supplier. Size and location can greatly influence capital spending per kg, which turns into depreciation costs, a key component of the all-in cost. A 15,000 ton facility costs roughly $80 per kg in capex, for example. But a 3,000 ton plant will need $140 per kg. Location matters as well. Cheap labor in China makes capital spending on a new plant much less there than in the US or Korea. There are also companies working on ancillary processes, such as cutting the cost to dry silicon and improving kerf recovery.

 
While the industry hunkers down to focus on cutting costs during what look to be a rough few years, there remains hope due to the cyclical nature of the sector. “We will have in 2017 and 2018 an undersupply situation again as in 2008,” Wicht predicts, mostly based on estimates that demand will push global installations toward 60 GW a year. “We will see a constrained situation that will lead to higher prices,” he said. LDK’s Bye, who expects 15 to 25 percent installation growth in the coming years, does expect an upturn. But he warned that the days of plenty, of 50 percent profit margins, will not return. “We are making a commodity. Keck echoed that, saying silicon has been and will be a commodity, and a cyclical one at that, with roughly four years from trough to peak. If 2012 is the trough, a peak will occur in 2016. The question remains, who will be around to enjoy such a scenario?

 
Source: PHOTON

Jeremy Herron
© PHOTON International, March 2012
Duplicate only with allowance of PHOTON Europe GmbH, Aachen, Germany