Which of these solutions is more space efficient?
If you assume 5.0 watts (peak) per square foot for thin film photovoltaic, you end up requiring 4.6 acres per peak megawatt-hour (MWp), about the same as Nevada Solar One’s solar thermal farm (4.7 acres per MWp). Thin film PV panels now have a factory cost of about $1.00 per watt, which is quite cost competitive with solar thermal. Remember there isn’t nearly as much balance of plant with photovoltaic. With solar thermal, the solar field is just the beginning – you need the heat concentrators, the boiler, the turbine, and the condensing loop.
While Ausra and others have designs somewhat more space efficient than Nevada Solar One (Ausra’s proposed Carrizo plant will only require about 3.6 acres per MWp), more efficient PVs require far less space. A high-end PV will generate 20 watts per square foot (peak), which equates to 1.15 acres per MWp. Needless to say the costs per watt for these higher efficiency PV panels are 2-3x the costs for the thin film panels – at least at present, polysilicon may decline in price, and who can say how the electrical input required to bake monocrystaline PV is costed? Also, we don’t have good history yet on the longevity of thin film PV.
Which of these systems delivers lower lifetime costs per kWh?
The costs for photovoltaic solutions are better documented, although utility scale photovoltaic installations are just beginning to be implemented. The most significant variable isn’t peak output, which is somewhat easier to calculate, but kilowatt-hour output per year. Variables affecting this yield, of course, are the latitude and the weather conditions at each site. Notwithstanding the fact that some latitudes are cloudier than others, in the higher latitudes, say beyond 50 degrees north, you can multiply your peak hourly output by about 1,200 to get your yearly output. If you have an array that delivers 1.0 kilowatts in full sun, then that array will deliver about 1,200 kWh per year. In the temperate latitudes, around 35 degrees north, that multiple increases to around 1,600, and in the tropics, it can get over 2,000x.
Another way to stretch the output of photovoltaic systems is to point them at the sun, the same way solar thermal mirrors track the sun. This can increase yields per collector area by 50-100%, more if optical concentrators are used, but this also increases costs.
Calculating costs per kilowatt-hour are also impacted by how long the system will last, and how much output will degrade each year. Solar thermal systems are more expensive to maintain, but can last 50 years or more. Monocrystaline photovoltaic systems only degrade about 0.5% per year, which means they still produce at around 80% of their original efficiency at age 25. But by age 50 their output will be significantly degraded. Nobody knows yet how long thin film arrays will last.
At this point it isn’t clear which systems are cheaper. Taking the total installation price divided by peak megawatt output to get a installation cost per 1.0 megawatt output (peak), is a cost that is easier to objectively calculate, To cite a recently commissioned solar thermal plant, Acciona’s Nevada Solar One plant cost a reported $260 million and will output 64 megawatts at peak, or about $4.1 million per megawatt. For an example of a utility scale thin film plant, Optisolar’s 40 megawatt Sarnia facility is reported to cost $300 million, or $7.5M per megawatt.
Getting to cost per kWh from peak output is not easy. Annual kWh yield is influenced by latitude, weather, and whether or not the array is tracking. Lifetime maintenance costs are a factor, and the life of the system and whether or not there is annual degradation are factors. At the end of the day, the solar thermal industry is claiming they can immediately deliver power for about $.12 per kWh, and they think they will eventually get that cost down to around $.07 per kWh. The thin film folks haven’t disclosed their costs – but with the incentives of feed-in tariffs this technology will continue to be developed.