Charles Annis VP Manufacturing Research, DisplaySearch LLC

2009 proved the solar industry can be highly cyclical. After averaging 15 per cent between Q1 06 and Q4 08, profit margins for a group of 14 leading cell and module manufacturers suffered a precipitous drop to -10 per cent in Q1 09 and remained negative in Q2 09. The significant losses were caused by the capping of Spain's Feed-In-Tariff (FIT), the worldwide economic crisis and tight credit markets.

Excess manufacturing capacity and oversupply throughout the supply chain during the first half helped push average photovoltaic (PV) system prices down more than 25 per cent. These lower prices, somewhat easier financing, diversification of the demand base and positive incentives in multiple regions helped push demand and most leading producers into the black in Q3 09.

Incentives impact

Incentives are designed to increase demand by improving the economics of installing a PV system. But on the other hand, fear of incentive reduction can also push demand, as users rush to install systems before benefits expire. Expectations that the new government in Germany will further lower FIT rates is one the most important current demand drivers. Changes or new incentives in Germany, Italy, Japan, America, France, China and other regions are now forecast to cause demand to grow more than 40 per cent in 2010. Solar cell and module makers should see average profit margins increasing in Q4 09 and towards 20 per cent in 2010.

In 2010 solar cell manufacturers will continue to focus on cost reduction as they try to push PV towards grid parity while simultaneously maintaining margins. Key methods for reducing costs include using less expensive silicon, using less silicon or not using silicon at all.

Polysilicon is the primary material of wafers used to fabricate crystalline silicon (c-Si) solar cells. It alone can account for 15-20 per cent of the cost of a PV module. As the result of a large ramp up of p-Si capacity just as demand was collapsing, an oversupply caused p-Si prices to drop dramatically from the 2H 08. The rate of decline has moderated now but supply is expected to still be more than sufficient in 2010, further weighing on p-Si prices.

Besides increasing the scale of conventional Siemens-based p-Si production, new refining technologies, such as Fluid Bed Reactor (FBR) and Upgraded Metallurgical Grade (UMG) silicon, have been developed as lower cost alternatives. Interest in these technologies has waned with the collapse of p-Si pricing, but they still may offer greater savings in the future.

PV revenues, profit and margins for DisplaySearch's index of module manufacturers. Source: DisplaySearch Quarterly PV Cell Capacity Database & Trends Report Click on image to enlarge.

Thinner wafers

Adopting thinner wafers consumes less p-Si and reduces the total cost of PV modules. Cell manufacturers have transitioned from 300m wafers used in the year 2000 to 150m at leading-edge fabs today, reducing Si content by 50 per cent. Wires used to slice wafers off ingots are not much thinner than the wafer itself, so as much as half of the silicon is lost as kerf. Applying thinner wires and new slurries or diamond coated wires all can further reduce kerf loss and continue to enable thinner wafers. One hundred microns is often assumed to be the limit to conventional wafering, but alternative slicing techniques, such as Silicon Genesis' Polymax technology, can cleave mono-crystalline silicon foils down to only 20m.

There are multiple iterations of amorphous silicon (a-Si), cadmium telluride (CdTe) and copper indium gallium selenide (CIGS), the main thin film PV cell technologies, but all target lower costs by eliminating the use of crystalline silicon absorbers. However, typically there is trade-off of lower conversion efficiency.

Due to the proliferation of turn-key fab vendors, a-Si has ramped from just 296MW of capacity in 2007 to 1.6GW in 2009 and is forecast to reach more than 3GW in 2010. Much of this capacity is being run at low utilization rates now as a-Si struggles to remain competitive in an era of much lower than previously expected p-Si prices.

CIGS based solar cells have long offered the hope of both relatively high conversion efficiencies, similar to crystalline silicon levels, and potentially a very low cost structure. However, the technology has proven very hard to ramp. CIGS accounted for only three per cent of total solar cell capacity in 2009, and actual production on that capacity is quite limited. Even so the appeal of CIGS remains. An additional 415MW of new CIGS capacity is planned to be built out in 2010. And in the near future large 500MW or even 1GW CIGS fabs are being planned by companies like Solyndra and Showa Shell Solar.

CdTe, as manufactured by First Solar, is by far the cost leading PV technology. In Q3'09 the company is manufacturing modules at just $0.85/W. Not surprisingly, First Solar has a large backlog and will add 424MW of capacity in 2010 to help fulfill strong demand. And because of that, the company is forecasting 2010 earnings per share between $6.05 and $6.85! First Solar is a role model for other module manufacturers, but to date no other CdTe companies have been able to replicate its success.

Polysilicon price trends. Source: DisplaySearch Quarterly PV Cell Capacity Database & Trends Report Click on image to enlarge.

Because higher efficiency equals lower cost, whether thin film or c-Si, there will be in 2010 a renewed emphasis on improving conversion efficiencies. Centrotherm, a leading German equipment company, states that an efficiency gain of 0.5 per cent translates to a cost reduction of about three per cent.

After going through its first significant "solar cycle" in the first half of 2009, the majority of the PV supply chain has quickly recovered. With demand increasing and costs coming down, most segments of the industry are looking forward to a bright 2010 where average profit margins head towards 20 per cent. That should hold for at least the 1H 10; after that, incentive policy decisions will determine the timing of the next cycle.