The required proliferation of solar energy provision that I spoke of on Monday, "A Slim Chance for Solar Energy", begs many questions about the resources actually in the earth and the necessary scale of their extraction from it as raw materials to provide "renewable energy". Most "solar cells" are based on silicon wafers with a thickness range of 180 - 350 microns (one micron is one thousandth of a millimeter, and for comparison the thickness of a human hair is about 70 microns), which are connected into solar panels. Though the technology is as yet unproven on the grand scale, research is ongoing into "thin-film" cells, which are anywhere down to one hundredth of the thickness of the current technology, and which in principle would reduce the quantity of pure "solar grade" silicon required to meet global electricity demand which I estimated at 52 million tonnes. Given that the current production of "polysilicon", currently in short supply, is around 30,000 tonnes, a mere doubling of that production capacity would be required per annum for a 20 year programme of solar electricity supplementation, rather than the hundred-fold increase that would be necessary to produce the world's 14.85 trillion kWh of electricity using current commercial silicon based solar technology. "Thin film" devices on a widespread scale are it must be said, a long way off, as far as the eye can judge.
It is not only silicon that is in the minds of researchers into photovoltaics. Silicon at least as a raw material should be found in sufficient quantities for whatever uses humankind might envisage. Around 25% of the earth's crust consists of silicon, albeit mostly in the form of "clay" and other aluminosilicate materials. For silicon production, it is silicon dioxide (SiO2) that is needed, mainly as mined in the form of sand. This is "silica sand", as formed from the weathering of silicate rocks, as part of the natural CO2 cycle, where silicate in contact with CO2 and water is eroded into silica, and calcium carbonate, an inorganic and geologically sequestered form of CO2. Desert "sand" is often misunderstood to be the kind of (silica) sand that one finds on beaches, whereas it is in fact dried earth (clay). A simple way to tell the two kinds apart is to spit into the palm of your hand, add a small quantity of the "sand" and then rub it: if it is clay, it will turn into brown mud, whereas if it is silica, not much will happen beyond the material becoming damp and clumping together. Thus, as the Sahara Desert advances because of lack of rainfall, more sand appears, but really it is just the earth turning into dust.
It seems likely that there is sufficient silica in total to extract from it all the silicon we are likely to need in any foreseeable period. Other imponderable events are more likely to hit us before we run out of sand. But if it is not just silicon devices that are being considered, what other materials are, and how much of each is there thought to be as a world resource? I have looked into the following elements (metals and semimetals) that might find a place in future solar technology (photovoltaics, rather than roof-based water heating systems).
Cadmium. There are 6 million tonnes of cadmium, of which world production amounts to 16,000 tonnes.
Selenium. Reckoned at 80,000 tonnes worldwide, and currently extracted at a rate of 200 - 300 tonnes per year.
Arsenic. There is somewhere over 1 million tonnes of arsenic wordwide, of which 54,000 tonnes is produced annually. This is predominantly in the form of arsenic sulphur compounds, although another 11 million tonnes is thought to occur along with copper and gold ores.
Gallium. Also reckoned at a global resource of 1 million tonnes, and extracted at a rate of only 61 tonnes per year.
Indium. The reserve base (that's everything - economically minable ore and that which is uneconomic according to the present stock-markets) is around only 6,000 tonnes, of which 350 tonnes is extracted annually.
Nickel. Plenty of nickel, at 130 million tonnes worldwide.
Zinc. 1.8 billion tonnes overall - loads!
Copper. 1.6 billion tonnes as a world resource - loads too!
Interestingly, in the production of silicon by the reduction of SiO2 using carbon (high-grade charcoal), it is thought that around 1.5 tonnes of CO2 is produced per tonne of silicon that is manufactured. However, since generating 2,283 kWh of electricity produces 1 tonne of CO2, and 13 MWh (13,000 kWh) of electricity is required to maintain the temperature of 1,700 degrees C, in order to make the reaction (SiO2 + 2C --> Si + 2CO) go, that means another 13,000/2,283 = 5.7 tonnes of CO2 is released just to provide the electrical power, and hence 1.5 + 5.7 = 7.2 tonnes of CO2 are released overall to make each tonne of "solar silicon". However, even the 374 million tonnes of CO2 that would be produced if we could manufacture the 52 million tonnes of silicon, over a 20 year period, fades into the background of 24 billion tonnes of CO2 released in 2004, from all our fossil fuel combustions, being less than 0.1% of that per year.
Thin film devices that use gallium arsenide, and composites of indium, copper, selenium (and other chalcogenides) would undoubtedly reduce any potential strain on the extraction and ultimate availability of some of these materials (particularly indium and selenium). There is the further and as yet unsolved problem of energy storage. Storing hydrogen and transmitting it enmeshes a technology that is fraught with problems: low energy to weight ratio; embrittlement of metallic containment vessels and pipes, with leakages of hydrogen gas and ultimate wholesale fractures; energy losses in compression or liquefaction of hydrogen etc. etc. Therefore, since renewable electricity, e.g. from solar, can't be produced at a constant rate (the sun doesn't always shine - not at night! - the wind doesn't always blow, or not at a constant speed), there is the need to store it, and if we can't perform this task using hydrogen, we need some other means. The energy that we want supplied in the form of electricity is best stored in the form of electrons, and so some type of battery technology comes to mind. I note there is plenty of nickel and zinc and so zinc-nickel batteries might supply part of the solution, whereas nickel-cadmium (only 6 million tonnes of cadmium worldwide) batteries are of limited use, considering the stupendous number of them that we would need.
There are other ways of storing "energy", e.g. flywheels, pumping water "up" into reservoirs, as low-tech devices, and high-tech means such as confining electrons in superconducting storage rings, which would probably consume more electricity to run than it could hold: liquid helium to cool the superconducting magnets to provide the magnetic fields, power supplies etc. Indeed, all such energy storage systems, including hydrogen, incur large energy losses, both intrinsically and when that residual energy is used to generate electricity once again.
In principle solar might work very well on a local level. For example, an average "terrace" of 8 houses in this neighbourhood has a total roof area of around 800 m2. Taking the "British" sunshine level of 150 W/m2, and using solar panels at 10% efficiency, this could produce 800 m2 x 150 W/m2 x 0.1 = 12,000 W. i.e a generating capacity of 12 kW, and hence 12 kW x 8760 (hours/year) = 105,120 kWh. Dividing by the 8 households, that gives a little over 13,000 kWh, which is about half the total energy reckoned to be used by the average U.K. household for everything: electricity (3,500 kWh) plus space heating (20,000 - 25,000 kWh). Energy efficiency could cut this by about 50%, according to the Oxford Environmental Change Institute, and so we would be about spot-on, just from solar. I accept that there are issues of energy storage which we must meet if we are to use solar as our exclusive energy source. However, even if it could power our "terraced" dwellings during the day only, in combination with an efficient CHP (Combined Heating and Power) system for the terrace, we might be close to energy self-sufficiency, without needing a national grid system (which is also energy wasteful). Much of industrial power provision might be provided by some similar "localised" means, once again leaving-over the problem of profligate transportation use. However, using less cars, more sustainable living, with activities and food production being carried out locally, and using electrified tram-style systems on the scale of a town, say the size of Reading with its population of 150,000, to link together local "pods" as necessary, could work. Probably a "county-sized" grid system could be used to supply such "meso" infrastructure, drawing in electricity from off-shore wind farms and sea-power (wave and tidal stream) installations, and even for (ideally throium fuelled) nuclear power.
Cutting energy demand through such efficiency-driven strategies must be our first move in the future energy game.