Online information source for semiconductor professionals

Although I haven't finished with my post-Intersolar North America/Semicon West-inspired blogging yet, I'll be on the road again the beginning of this week, visiting a mostly solar PV cast of companies in the anachronistically named Silicon Valley. I also have a backlog of interviews and coverage to plow through that I collected before the recent double-banger event, including interviews with execs from Suniva and Skypoint Solar, as well as a report about my recent visit to Cymer's extreme-UV lithography source project.

In the meantime, here's a tasty leftover from the big show to start off the week.

The Prometheus Institute's loss is certainly SunPower's gain, if the calibre of the rest of Elizabeth "Libby" Wayman's work is of the same high quality as her talk comparing the different types of concentrating solar technologies--thermal and PV--presented during Monday's CSP conference (cosponsored by the institute and Greentech Media). I admit to being photovoltaic-centric for the most part, so in addition to a well-rounded overview of the different CPV options, her succinct, clear explanations of the parabolic trough, linear Fresnel reflectors (LFRs), tower, and dish-engine thermal side of concentrating solar rounded out a solid, informative primer of the irradiance-loving sector.

One chart, which compared land use for the eight major kinds of CST and CPV technologies, piqued my topographic interest. It showed average land area measured in acres needed per 1 MW of installed capacity. As one might expect, both thermal and PV, depending on the flavor, can gobble up some pretty large chunks of real estate--or not.

At the large end of the spectrum are tower-style thermal concentrators, which need between 13 and 14 acres per megawatt. The second biggest land user is tracking PV at around 12 acres, with low-concentration PV just a tad less voracious, needing about 11 acres.

The least land-intensive CSP system is also one of the least mature--LFRs--which can generate a megawatt on less than 4 acres. Nontracking PV is the next least land hungry, requiring about 5 acres per MW. In the middle of the pack, ranging between 6 and 9 acres are parabolic trough, dish or lens CPV, and dish-engine.

At the end of her presentation, Wayman summarized with a complete system comparison by technologies, thermal and PV. Trough and LCPV are the most commercially established in their respective categories, while tower thermal and dish and lens CPV are in the commercial or prototype/commercial stages. LFR remains in the prototype phase, while nontracking CPV has started into early manufacturing.

In terms of markets served (or potentially served), the thermal varieties score highest in the utility category, although trough, LFR, and dish-engine have limited commercial-sector application potential as well. On the CPV side, all score well in both the commercial and utility markets, and lens CPV and especially nontracking CPV have residential market upside as well. In fact, only nontracking CPV shows strong potential across all three market segments.

But the residential sector doesn't generate alot of interest among concentrating solar companies, which seemed to be much more interested in the larger-scale installations: Data shown in a subsequent presentation counted 46 CSP systems development companies targeting the utility sector, 38 looking hard at the commercial space, and only 10 firms expressing an interest in the residential market.

Her table made clear a striking difference between the thermal and PV schools of concentrating solar: Dispatchability may be a strong suit for the trough, LFR, and tower solutions, but in terms of modularity, they are no match for every type of CPV. Only the still-experimental dish-engine technologies offer modularization.

CSP systems' levelized cost of energy is all over the place, with trough and LPCV showing the lowest LCOE, and dish-engine, dish CPV, and nontracking CPV having the highest. Tower and lens CPV end up in the middle of the cost pack. National Renewable Energy Lab data presented later in the conference showed LCOEs in 2020 projected to reach 6 cents per kilowatt-hour for power tower and trough concentrating thermal solar, 12 cents per kilowatt-hour for concentrating PV, and 10 cents per kilowatt-hour for plain-wrap, nonconcentrating PV.

Cost reduction potential looks a bit better on the PV side, but not by much. Nontracking CPV and tower thermal have the highest upsides in possibly bringing costs down, with LFR, dish and lens CPV also having decent shots at pushing the dollars per watt lower. Trough, dish-engine, and LPCV seem to have the least cost reduction potential in the CSP arena.

In some ways, CSP is a microcosm of the larger solar industry, with a range of technologies that each has potential niche and broad-swath applications to go after, but still needing some economies of scale to get the price points to more attractive (AKA grid-parity and below) levels.