Search sponsor
Search sponsor
The idea seems simple enough: by depositing tricky photovoltaic thin films on a small, wafer-like substrate, much better compositional control, film uniformity, and defectivity levels—and higher conversion efficiencies--can be achieved than on a large-area flat glass or flexible web surface. Throw in a reliable, low-cost processing method, with high-volume throughput capabilities proven in another industry that would be easily portable to PV applications, and the approach has merit as a way to quickly scale from lab to pilot to commercial production levels.
That’s the nutshell value proposition that copper-indium-gallium-(di)selenide development stager XsunX is working on with its equipment partner, Intevac, a well-established purveyor of hard-disk-drive and semiconductor manufacturing gear that has jumped into the PV tool sector. At least one other CIGS start-up, Applied Quantum Technology (AQT), is collaborating with the OEM on a small-area deposition process scheme similar to the one favored by XsunX, which announced last week that it had reached a critical milestone—the fabrication of a “fully functional” sample device.
Despite the intriguing possibilities of modifying existing HDD and IC systems and processes (and benefiting from the existing automation and other supply-chain infrastructure) for production of CIGS cells that resemble your average crystalline-silicon wafers, nagging doubts about one of those companies can’t be ignored.
But before addressing those concerns, let’s talk tech. I met with a C-level trio from XsunX and Intevac at the Solar Power International show a few months ago, where they briefed me on their progress and plans and offered details of the hybrid manufacturing technique in development.
XsunX CTO and veteran CIGSologist Robert Wendt ticked off some advantages that the HDD process technology brings to the PV party: similar deposition sources, well-developed sputtering techniques for back and front contacts, extremely uniform and thin films, really high throughputs, and very low costs.
He noted the difficulties of mastering evaporation-type CIGS deposition (XsunX’s method of choice, because of its world-class efficiency) on larger areas and how “the statistical distribution of defects on that smaller area is substantially less [than a larger area], primarily because the control of smaller sources is less of a challenge.”
Intevac boss Kevin Fairbairn pointed out that the HDD world has been producing very precise films (down to 8nm grain sizes with distribution of +/-1) and achieving low defect rates and excellent compositional control for years, and has shown the ability to produce a thousand discs per hour and hundreds of millions of units per year.
The capital tool firm examined the PV market, wondering if it could “provide equipment solutions that would be competitive with the best out there, looking at First Solar as a benchmark, and concluded that CIGS could be done on our machines, at a comparable cost, [which was] an important factor moving forward.”
The two companies started talking in November 2008 and began building their working relationship in February 2009, said XsunX CEO Tom Djokovich. A prototype equipment set is now largely operational at Intevac’s Fremont, CA, facility.
Fairbairn said the current XsunX pre-pilot set-up consists of two platforms: A refitted HDD sputtering tool puts down a back-contact bilayer of molybdenum and chromium films as well as the zinc-oxide and indium-tin-oxide top-contact TCO materials; and a reconfigured 200mm Lean Etch high-productivity platform from Intevac’s semiconductor business acts as the mainframe of the all-important CIGS absorber tool, where the 125mm (soon to be 156mm) “pseudo-square” stainless-steel disc gets flipped from a vertical to a horizontal position for processing, then flipped back to vertical for the ensuing steps.
Djokovich explained that “we won’t have a final toolset configuration yet until we’ve fully designed what a balanced system looks like. Once we complete all the processes and establish what all the baseline deposition rates are going to be and what kind of yield we’re gonna get from the stuff, then we can go ahead and design and engineer what a balanced system looks like. That’s probably something we’ll be determining by the March [2010] timeframe.”
(When asked in a recent follow-up email if that timeline had changed since October, XsunX would not provide an update or any other additional info.)
Wendt emphasized how the scheme results in a dramatically reduced development costs. “What this small-format, single-cell processing allows you to do is create a full-size module without having full-size equipment to do your piloting. You purchase pilot tools that are very similar to what Intevac created for the back and front contacts. It’s an adaptation of the Lean Etch tool for the evaporation, and with a reasonably small capital investment, you have the equipment to make a full-size module.”
“One of the problems and challenges that doing full panel processing has is the front-end capital equipment necessary to make that panel is substantial,” he continued. “So with this technique, you avoid substantial capital investment during the piloting phase. You’re able to actually fabricate what’s going to be fabricated in the production phase, on the exact same tools using the exact same processes.”
Fairbairn compared the approach with the roll-to-roll processing deployed by some CIGS companies, citing the favorable unit costs of about 20 cents during small-volume work that will drop to less than a nickel once production scale is reached. He also noted that “since we can run a lot more cells, the rate of yield and efficiency improvements goes up much, much faster.”
On the moduling side, Wendt characterized it as a “bifurcated pathway. Since the tabbing and stringing of CIGS is different than crystalline silicon--the cell cannot go to as high a solder temperature as crystalline—it’s not a direct drop-in process. So the process needs to be taken all the way to the end, so you can develop those processes, develop the reliability testing on a final product to make sure you have the UL, TUV, IEC capabilities.”
“Once the processes are developed, they may be portable over to existing back-end processes,” he continued. “We may be able to focus more on the front end and divert some the product down that bifurcated pathway.”
XsunX’s game plan calls for efficiencies of 16% for coupon-scale devices, 14% for cells, and 12% for modules, with the first 96-cell, 1 x 1.6 meter modules promised at 10% efficiencies for market entry, explained Wendt.
Line throughputs of at least 600 cells/hour are easily within reach, according to the partners. Although no nameplate has been attached to the production capacity of each panelmaking line, Wendt said modeling indicates that it will likely fall in the 10-14MW range.
While the company replied in a recent email that it is “building momentum and continuing to make excellent progress in the process development area,” Djokovich and his team will dish few additional details. When I interviewed them at the big show in Anaheim, the CEO said XsunX would start building out the pilot line in March and begin site selection for the first production facility between then and August/September of this year.
But no further details are forthcoming, which brings me back to those nagging concerns I mentioned at the top. XsunX has been in development stage for what seems like forever, with zero income generated on zero sales; in effect, its track record is that of an increasingly expensive science project of a company. Millions in investor dollars have been burned through, with no end to nonprofitability in sight.
For several years running (including for the fiscal just concluded), the over-the-counter corporation has had to file for an extension of its year-end SEC 10K reports. When the former CFO was let go in 2009, Djokovich donned the chief accountant hat as well. When I mentioned the company to a noted “been there, done that, seen it all” PV industry veteran at the Solar Power show, he scoffed and dismissed XsunX as “unimportant.”
The transition to a CIGS strategy marks at least the second time the company has changed direction. When Djokovich (pictured at left) joined the company in fall 2003, it was working on amorphous-silicon cell device and manufacturing technologies, and the potential licensing of said technologies, including a transparent thin-film IP picked up when it bought Xoptix.
Then in spring 2007, XsunX reevaluated its market positioning and decided to become a manufacturer and seller of a-Si TFPV modules. It leased a vacant facility outside Portland, OR, and said it would have its first 25MW line built by 2008 with as much as 100MW of production online by early 2010. A blog I wrote about the company during the 2008 Solar Power show went into detail (including some fancy-pants modeling data touting the superiority of a-Si) about the company’s then-amorphously inclined business plan.
Apparently, shortly after that enthusiastic interview, Djokovich was already starting to re-evaluate the gameplan, although none of the company’s final FY08 SEC documents issued in early 2009 offer any indication of a technological sea change. Once the decision was made, they pulled the plug on, and terminated the lease of, the fledgling Oregon production facility (and were none too happy with the broken promises of certain state politicians, according to the CEO) and cut expenses significantly.
He said that the move to CIGS “was actually welcomed by all the investor groups we spoke to”; they recognized the quaternary technology’s “ability to antiquate amorphous-silicon” via its greater headroom for conversion efficiency improvement. The presence of Wendt (pictured at right), a noted CIGS expert hired in early 2007, and the intriguing Intevac-involved model for rapid commercialization, he explained, were the other two parts of the trifecta that convinced the investment community.
Ultimately, XsunX doesn’t want to become a large-scale CIGS manufacturer on its own, but would rather have some production and then hook up with a bigger entity or entities, perhaps garnering licensing fees and other royalties along the way.
“Our real goal is to create joint-venture relationships with larger companies,” said Djokovich, noting the actions of Bosch over the past year or so in acquiring or investing in certain solar PV companies. “That allows us to continue to do what we do best, to take this technology to higher conversion efficiencies, and go to CIGS in 8-inch substrates.”
I can’t imagine XsunX would be able to survive as a going concern if it were to take yet-another 180 degree turn (or as Djokovich described the latest strategic move, “a 360 with a 180”). It's sink or swim time. The CIGS play will either be the one that pushes the company over the top and attracts a legit JV/equity partner/parent company, or it will be the final chapter in the drawn-out saga of yet-another failed solar PV start-up.
PHOTOS COURTESY OF XSUNX; CIGS SAMPLE PHOTO COURTESY OF NREL
