Speeding toward growth in the fog March, 2009: The silicon industry, putting the pedal to the metal, seems to be speeding toward growth. But as the many new manufacturing hopefuls disappear into the fog, they can’t be sure if they are driving down a highway with guardrails or careening straight for a dangerous cliff. © Ronald Frommann / photon-pictures.com   Nick Sarno, LDK Solar. That, at least, is the picture that emerged when scores of silicon manufacturers presented their plans for expansion on March 3 at PHOTON’s 7th Solar Silicon Conference in Munich. All together, these roadmaps would raise the annual production capacity by a total of 150,000 tons – enough for 15 GW of crystalline modules. But, says analyst Michael Rogol, head of PHOTON Consulting, since these plans represent only about 10 percent of all silicon production worldwide, the actual growth could be a far higher figure. Nevertheless, Rogol is convinced that at least for the next five years there won’t be an excess of solar silicon. “Why are we seeing spot markets of $150 per kilogram,” he asks, “if an oversupply is forseen?” The reason is that the real production costs for silicon, and hence the absolute pain threshold for the selling price, averages at a much lower $40 per kilogram. One who sees it quite differently, however, is Richard Winegarner, the head of the independent market research company Sage Concepts and an old hand in the silicon business. If the dozens of new polysilicon producers who are currently building up their factories worldwide don’t start putting on the brakes soon, he fears, they will find themselves crashing into a wall. “The demand is clearly smaller than the supply,” Winegarner believes, a situation he expects to last for five years. “The big question is: will these companies have enough money to survive until growth returns?” he told the audience of about 800 decision makers from the silicon branch. Winegarner is forecasting a 2012 worldwide production capacity of more than 400,000 tons. A good half of this will be produced. But the demand will only be about 100,000 tons, he predicts. The comments of many conference participants was that the truth probably lies somewhere between the two extremes of these analysts. Still, the silicon producers don’t seem keen on cutting back their plans for building up production capacity. On the contrary, the companies, especially in China, are tripping over themselves to churn out polysilicon factories with capacities in the thousands of tons as fast as they can in little more than two years. The Chinese wafer manufacturer LDK, for example, announced that its 15,000-ton factory is 80 percent done, and that a 5,000-ton production line is 95 percent complete. “At the very least,” said one observer, looking at the many photos of completed and half-completed production facilities in the Middle Kingdom, “now we know that the Chinese are actually building factories.” At PHOTON’s 5th Solar Silicon Conference a year ago, most of the projects were still no more than announcements. Back then, says Rogol, developing plans for polysilicon production meant trying to discover the secret recipe for a special sauce. Everyone knew that it existed but no one knew how to make it. That has now changed, notes Rogol. “We know the recipe.” Which means that the supremacy of the large silicon producers – above all Wacker, Hemlock Semiconductor and Tokuyama – belongs to the past. The only thing that remains for them now is to differentiate in terms of pricing. With production costs of the top players of around a mere $25 per kilogram, that may certainly be their best strategy. Something else that this conference has made clear is that different market segments are moving into the solar silicon field. In addition to highly purified polysilicon, collected through gas distillation in reactors, there is also a growth in the supply of so-called upgraded metallurgical silicon (UMG). This is produced through the purification of cheap metallurgical silicon, a process so far used mostly for making solar cells with a somewhat lower efficiency. As a result, the production cost is also lower than its purer big sister: At Timminco Ltd., for example, the current UMG cost is about $30 CAD ($18 USD) per kilogram. The company plans to produce several thousand tons of it annually – along with many other firms. And that has exerted some pressure on silicon producers. Long-term contracts – up to now standard in the relationship between silicon and wafer manufacturers – are increasingly being renegotiated. At the same time, the willingness of wafer and cell producers to be locked into new contracts with such prices, and payment and acceptance obligations has been noticeably waning. The result is that the material on the spot market will almost certainly end up being as inexpensive as silicon purchased through long-term contracts, in other words costing around $60 per kilogram instead of up to $450 as last year. It would seem cell producers who once needed silicon bloodhounds to sniff out enough feedstock might now be able to discover better deals by using radar to spot the increasing number of producers lost in the fog of growth. Christoph Podewils © PHOTON International, March 2009 Duplicate only with allowance of PHOTON Europe GmbH, Aachen, Germany  back