But the dream of capturing the sun's energy directly has been elusive. The problem is that energy needs are unevenly distributed and usually peak at night. Fuels and reservoirs provide inexpensive storage of the sun's energy making it available when and where we need it.

Though the solar industry is making rapid progress, it is still by far the most expensive form of alternative energy. A recent NYU study found the following actual 2005 costs in cents/kWh:

Geothermal                         3.1  –  4.3  

Concentrating Solar Power    11   –  15   

Photovoltaic                         18.8 – 31  

Wind                                    4.3 –  5.5  

Coal                                     1.2  

Natural Gas                          1.35  

Of course these costs will come down some day but for now solar is basically a subsidized research project.  The new CSP plants with heat storage can keep the power flowing when clouds pass over and in the evening but that doesn’t help costs. The problem is that the sun only shines part of the time. Capacity factor even in the California desert is still only 25%, which means that a 4 MW solar plant only delivers an annual average of 1 MW.  Unfortunately the custom of rating solar plants based on their peak output on a clear summer day at noon leads to some dangerous misconceptions.  Cost per Watt, for example, understates the actual cost by a factor of 4 even in the desert. Growth figures and land use in acres/MW are similarly grossly misstated.

If we look at land use of some real projects now on the drawing boards we find that the latest photovoltaic, parabolic and tower projects all use about 5-6 acres per peak MW. The Saguaro 1 MW parabolic trough plant near Phoenix for example, generates 2000 MWh of electricity annually, using 15.8 acres.

It’s interesting to compare this sun-capturing performance to a field of biomass. Miscanthus is perennial grass that yields 15-20 tons/acre on marginal land.  That’s about 250 million Btu/acre which is 73 MWh/acre. If you use a 85% efficient combined heat and power (CHP) plant to convert the biomass to power, it would take only 2000/(73X.85) = 32 acres to grow the same amount of power. I’d rather mow and haul 32 acres of grass over the year than keep all those shiny troughs clean and working. And the one-time grass planting is a lot cheaper! 

So the race is on and only time will tell whether nature’s storage of the sun’s energy in plants can keep up with man’s best mechanical efforts. The nice thing about the biomass is that you can keep it around until you need it. The hot-oil thermal storage at Saguaro is only good for 6 hours.

The specifications for the Saguaro solar plant illustrate another messy thing about the specifications on solar power. The spec shows a capacity factor of 23% now, but with the 6-hour storage added the capacity factor jumps to 40%. This seems to be common practice. When storage is added the capacity factor spec goes up apparently to indicate the % of time that power is available. Power is sold by the kilowatt-hour, so perhaps it would be better if we stopped talking about Watts and used GWh/yr instead. 

Our comparison to biomass was a little unfair because we used an 85% efficient CHP plant for the biomass and Saguaro throws their waste heat away using an evaporation pond.  By locating solar thermal plants in places where heat is needed, they can be efficient too. The waste heat is simply sold or put to use near the plant running a cold storage warehouse, a kiln, etc. Hotels, industrial parks and apartments should have their own solar thermal CHP plants for hot water, air conditioning and pool heating. We have to break the “giant power plant” habit.

Solar thermal is often supplemented by natural gas at night. The boiler is simply kept going as needed with gas. Since heat loads are often variable, CHP plants lose efficiency if the waste heat must be disposed of. A good approach is to size the solar collectors for minimum heat needs so that efficiency is always high, and then use natural gas to make up the difference. This minimizes the investment and maximizes efficiency.  In fact, just one collector to preheat boiler water can cut gas consumption and CO2 emissions significantly.

A mass-produced solar thermal-CHP system sized for large homes or apartments could be much more cost effective than the typical overpriced home photovoltaic installation we often see. Most homes and buildings use more energy for hot water, heating and cooling than for electricity. Instead of electric air conditioning, waste heat can power the heating and cooling. With decentralized small CHP plants scattered all over the map, power transmission losses are almost eliminated and we don’t need to spend billions adding transmission corridors.

In places that often have cloud cover, thin film photovoltaic power works much better than polycrystalline because the clouds scatter the light in all directions more than actually blocking it. Thin film panels usually have 3 layers to cover a wider spectrum of light. Evacuated tube solar collectors also work well with clouds.

Concentrating PV needs a sharp sun image to be efficient. It is best done in deserts where there are no clouds or haze. Concentrating PV lenses and mirrors work by focusing an image of the sun’s disk onto the solar cell. When haze scatters that image, efficiency plummets. Deserts in the US Southwest, Mexico and in North Africa have the potential to supply less sunny northern areas if we make the investment in massive HVDC transmission lines. Since solar panels produce DC naturally there may be a savings in electronics. Also the significant amount of waste heat can be used for energy intensive industries and to make fresh water from the ocean.

But don’t bet against solar as a long-term winner. Silicon development for computer technology fooled everyone with their “Moore’s law” doubling capability every couple of years.  This has gone on for decades and will probably continue. For example, proton plasma beams can now cut wafers as thin as 20 micrometers thick. This cuts material costs to 1/3 and the wafer flexes like thin sheet metal.  Another group at University of Delaware just announced a cell/concentrator combo with 42.8% efficiency. Clearly exciting developments will make this a fascinating race with many winners. We must pursue all ideas and let the winners be chosen by the marketplace.