Continuing the crusade to lower solar manufacturing costs, this week 1366 Technologies officially opened its new manufacturing site in Bedford, Mass., a 42,000 square foot facility with 25-megawatt capacity. The site just across the road from fellow solar company Spire, is part of the company's ambitious plans to remove a major chunk of costs and processes out of the solar manufacturing chain, beginning at the very start of the process: the wafer itself.

In a speech kicking off the open house event, Massachusetts Lieutenant Governor Tim Murray reviewed the state's renewable energy trajectory: 11% growth in clean energy jobs in 2012, five thousand companies operating, more than 70,000 700,000 workers, over 190 MW of total solar power capacity (well ahead of the goal of 250 MW by 2017), and solar project development in all but 9 of its 350 towns and cities. 1366 Technologies represents the nexus of the state's desire to work with and support a clean energy industry that will bring both jobs and help power its citizens and businesses.

The new facility's 25-MW capacity (1366 execs said it can actually support up to 20 MW) translates to about 6 million wafers/year. Right now 1366 expects 1-MW of output this year (in the "low 10s of thousands" of wafers) as it debugs and perfects its machines, according to company CTO Ely Sachs who led a tour inside the facility. Next year's target is 10 MW, and then 100 MW the following year — at which point it hopes to be moving into a new 1-GW facility.

The Technology

Today's multicrystalline silicon solar wafers are made by melting chunks of the material in a large quartz crucible (~2 foot long and a foot deep); once the material is molded and cooled it is chopped into a rectangular block and sawed into individual wafers, a process that uses a lot of slurry and wires and results in much wasted material. 1366 uses a much shallower container for its silicon melt, according to Sachs, from which a single wafer is produced on the top; it is subsequently laser-trimmed to a standard 156 × 156 mm size. Sachs wouldn't divulge the company's secret to creating a wafer from the shallow melt, but alluded to an "a-ha" moment about overcoming molten silicon's tendency to bead up vs. making it stick to the mold.

1366's value proposition is to "sell the world's best wafer" with added value including texturing, but made at a fraction of the cost, with better uniformity and performance specifications, summed up CEO Frank van Mierlo in an interview following the open-house. Bottom line: the company says it can reduce silicon costs across the board by to just a third of today's costs: 1/3 off of standard processing, 1/3 the labor, and 1/3 the consumables. Even if silicon prices are a fraction what they were a few years ago, any way to squeeze out more costs — while not altering processes or quality elsewhere in the value chain — still resonates.

Much of 1366's processes, and the equipment to perform them, were invented and created in-house: patterning the wafers with a low-cost polymer, and a wet etch chemistry for texturization are both proprietary, for example. (Diffusion processing is standard, as is metallization/screen printing.) There's an in-house machine shop with a fulltime operator, plus the company outsources some equipment work.

The company also has invested in a lot of characterization capabilities, to quickly get information about the cells made on its wafers; qualification information is obtainable within minutes from when a wafer is made, as opposed to sending out product and getting back results in weeks or even months. "That's a big part of what we do," noted Sachs, because it can indicate whether to go slightly heavier or lighter on a particular gas in the process, or raise/lower the temperature by five degrees, to come up with a different electrical characteristic in the final product. Analyzing the wafer's resistivity can lead to adjusting the dopant that goes into the silicon, for example.

Besides helping 1366 learn how to build a better end product, quick in-house characterization feedback also provides "feed-forward" information to the company's solar-cell customers, who typically tune their processes anyway to accommodate even minor variances in wafers; they can "spin less of their wheels and get to a sweetspot" in process efficiency, Sachs said.

1366 says it has created solar cells with 17 percent efficiency in customer trials, which it deems "industry average." During the facility tour Sachs pointed to current examination of one cell at 17.5 efficiency. For historical reference: 1366 achieved 14 percent cell efficiencies in August 2010, 15 percent in July 2011, and 16 percent in March of last year. Having in-house solar-cell-making capabilities, including characterization, is basically a quality-control effort, but it also "has allowed us to progress much faster," he said.

Scaling Up

Right now 1366 is putting "finishing touches" on what it calls its "Generation-1" equipment, improving upon its previous work on early "Generation-0" tools in which many processes are in sequence — for example, a wafer is cooled in the same physical space in which it is made, creating a bottleneck for the next wafer, Sachs explained. Gen-1 improvements will focus on automating the process, with processes done in parallel; as in the previous example, a wafer is created and then moves to a different cooling system. Separating the steps also will let the company better tweak the process control for each step, he added. The company is already putting together design elements for the next Generation 2 equipment (target date: up and running in 18 months) that will have the same production capacity as Gen-1 but with streamlined design elements and fully automated production that is six times faster, Sachs said.

1366 has accumulated roughly $47 million in equity backing, with VC partners including North Bridge, Polaris, and Ventizz, plus an investment from Korea's Hanwha Chemical. It also has had a $150 million DOE loan guarantee in its back pocket since Sept. 2011, but is reserving that money for when it is ready to build its second facility with 1 GW capacity. Among the key criteria for that DOE-backed loan is that the technology "works flawlessly," it has "firm customer commitments at prices which are profitable," and that it gets private investors to match the loan dollar-for-dollar, explained van Mierlo. All those should come together by the first half of next year, he said.

The real yardstick of success will be how 1366's technology translates to a production-environment scale. The new 25-MW facility is basically a "proving ground" — the real proof will be how it scales up in the next 1-GW factory, assuming all criteria are met to get there. van Mierlo lays out 1366's true argument here: Today a fully loaded cost of legacy wafers is $0.29/W, vs. a sales price of only about $0.20/W, i.e. it's a money-losing proposition. Once 1366 ramps in its big 1-GW factory, it will deliver a wafer cost of $0.10/W, a third of today's fully loaded cost, he said.

In his own opening comments at the open house, van Mierlo said the opening of this new facility puts the company on a "three-year runway" to prove its technology and business. In a follow-up interview he declined to offer specifics about the company's actual revenues now, but he did note the company has been cash-flow-positive from operations for the past three years. "We have real cash commitments," he said, which in fact paid for this new $6 million facility. The eventual 1-GW factory would be a profitable operation from the start, he added. "Financially, we're in really good shape."

Lead image: 1366 Technologies wafer after metallization, in the company's new facility in Bedford, Mass.