I dealt a bit unconventionally with solar energy in my last post.  I artificially finagled a way to express the energy density of a photovoltaic panel in Joules per kilogram simply so I could include it in my table and compare its energy content to gasoline.  I’d like to delve a little deeper here.

Based on the mean distance between the earth and sun and the average solar output, the earth receives about of 1361 Watts per square meter at the top of our atmosphere.  If there were no additional attenuating factors, that is the maximum energy you could harvest.

But there are many other factors to consider.  Even with a clear, cloudless sky, our atmosphere absorbs about 25% of the incident radiation, leaving about 1000 W/m2 maximum at the earth’s surface.

Next come azimuth and elevation.  The sun is strongest when directly overhead, falling off sinusoidally toward dusk and dawn.  This pattern is further modulated by an additional sinusoid as we move seasonally between solstices.  And, of course, we get no direct energy at all during the night.  The result is that the average spot on earth receives only 250 W/m2 before cloud cover is taken into consideration.  Average cloud cover cuts that down to 168 W/m2.

Now that we know what we have to work with, how do we actually capture this energy?

There are 3 basic approaches to harvesting solar power.  The first was pioneered by Mother Nature and consists of letting sunlight drive a chemical reaction like photosynthesis.  Viewed this way, wood, coal, and gasoline are all forms of solar energy!  Other modern incarnations include biofuels like corn ethanol and switch grass.  The point here is that the relatively weak supply of solar power gets integrated over time to produce the high energy content in the final product.

Photovoltaic cells directly convert light into electricity.  Incoming photons produce electron-hole pairs near the surface of a semiconductor material.  The electrons are then routed through an external circuit before being allowed to recombine.  The process is clean and dry and has no moving parts.  Unfortunately, because of its quantum nature, there are another series of losses resulting from spectral  and quantum inefficiencies, as well as  purity and impedance issues.  Practical, single-crystal photo cells have efficiencies of about 25%.  This brings us down to about 42 W/m2.

In an effort to circumvent these inefficiencies, the third approach is to convert sunlight to heat and then harvest the heat using traditional boiler technologies.   The new Ivanpah generating plant in California works this way.  It is a joint venture between Google and NRG and is supposed to produce 392 MW, though, so far, it has only produced a quarter of that amount.  If it does reach its potential, this 3500 acre project will be producing 27.7 W/m2, or about 2/3 the energy of the equivalent photovoltaic array.

None of this should be taken as disparaging solar energy.  Our relatively weak solar supply simply is what it is.  And it is terrible for some applications and great for others.  It excels in low to moderate power applications, especially if they are hard to connect to the grid (LED road signs, remote sensors, spacecraft, etc.).  It is terrible for large industrial or automotive uses where it simply cannot provide the real-time power (smelting aluminum, running presses, powering cars).

This is not the way it is portrayed in the popular press, of course, so let’s examine a single example: automobiles.

A car with 3 square meters of usable surface area could typically produce 126 Watts if it were covered in photocells.  That equals about 0.169 horsepower!  (At least a real horse could produce 1 horsepower!)

Could we not compensate for this dearth by running our car off batteries whose energy is ultimately supplied by a land-based solar array?  After all, we can make the array as big as we want.

According to the LA Times, there are 253 million cars and trucks on the road.  For purposes of our exercise, let’s assume they are all 150 horsepower Toyotas.  That means we need to supply 3.8 x 1010 horsepower or 28 million megawatts to run our vehicles.  Using the Ivanpah approach, this will take 253 million acres of generating stations.

My home state of Indiana is only 22.9 million acres.  (The nation of Egypt is 247 million acres.  Perhaps they will volunteer their space.)

To return to my starting point, I’m not sure where solar power should actually appear on my list, but I think I was probably generous to place it as high as I did.

AuthorMalcolm McClure