Grab your coat and get your hat
Leave your worries on the doorstep
Life can be so sweet
On the sunny side of the street
     —Jimmy McHugh and Dorothy Fields (1930)

You could not step twice into the same river; for other waters are ever flowing on to you&emdash;Heraclitus

In the era of global warming, peak oil, and the perilous petroleum dependency of the OECD countries, solar energy is now the hot ticket. Venture-funded companies compete for investment and market share. The scientific community is doing its part, proposing grand visionary plans for meeting most of humankind's energy needs by mid-century by tapping the original source of our fossil fuels—the Sun.

The late Nobel laureate Richard Smalley spent his last years in a quest for terawatts. A posthumous summary of his work published in the Houston Chronicle, Imagine a world that's energy rich, states the central themes of the new solar movement.

I have been on a personal journey for the past year and half in a search to find some happy answer to the energy problem. I believe the problem is, simply stated, that we have to find a new oil. Oil was, unquestionably, the basis for prosperity for this country and the planet in the last century — particularly the last half of the century.

But it is very clear to many of us, including leading scientists and policymakers, that if oil remains the basis for prosperity for the world throughout this century, it cannot be a very prosperous or happy century.

And so, if you have one word in this scenario to describe this new oil, it would not be ''oil," it would be ''electricity." That is the key conceptual insight that makes things work.

In A Solar Grand Plan, Ken Zweibel, James Mason and Vasilis Fthenakis flesh out a scheme by which "solar power plants could supply 69 percent of the U.S.'s electricity and 35 percent of its total energy by 2050" (Scientific American, January, 2008). In Powering Civilization to 2050, The Oil Drum's Stuart Staniford describes a scenario—not a forecast—which depends on a huge, anomalous rise in marketed primary energy production from solar photovoltaics and other renewables after 2020 (graph left). How plausible are these kind of stories?

Such sweeping visions of our solar future beg the central question of humankind's destiny on this planet—

Do we live in a world of ever flowing abundance, or do we live in a world of limits to growth?

If your answer is "abundance", your approach to the future requires a shift in direction in a context of business as usual. If your answer is "limits", your approach requires a shift in behavior in a context of living within your means. What follows examines possible constraints on the expansion of solar energy in the 21st century.

A Long, Long Way to Go

All visionary schemes start with the really large numbers¹ measuring the available solar energy. Using a calculation based on the solar constant, which ≅ 1366 watts per square meter, the available energy is 3844 zettajoules per year. Another source calculates that the Earth receives 89,000 terawatts (TW) of solar power. All of humankind's power consumption amounted to only 15 terawatts in 2004, so converting less than 0.02% of the Sun's power delivered to the Earth would have replaced all of it (graph below). Any way you slice it, that's a lot of energy.

According to the IEA's 2007 fact sheet Renewables in Global Energy Supply, solar provided only 0.039% of the world's primary energy in 2004, and their latest data for electricity consumption in the OECD indicates that only 2% came from geothermal, solar, wind and other sources combined. Year on year growth in January-November 2007 compared to 2006 was an impressive 20.2% for renewables, amounting to 166 terawatt hours altogether. However, 4.1% growth in electricity from combustibles (fossil fuels) amounted to 245 terawatt hours during the same period, which equals 147% of all electricity from renewables.

If our electric future is solar, there is a long, long way to go.

Potential Limits to Solar Cell Manufacture

Solar power is all the rage² in the energy markets because of the thin film revolution, which promises low-cost mass production of photovoltaic (PV) cells (Technology Review, July 27, 2007). Leading the way is First Solar, which uses cadmium-telluride thin-film PVs in its solar panels. Scientific American's solar "grand plan" cites cadmium telluride as the cheapest option today.

To provide electricity at six cents per kWh by 2020, cadmium telluride modules would have to convert electricity with 14 percent efficiency, and systems would have to be installed at $1.20 per watt of capacity. Current modules have 10 percent efficiency and an installed system cost of about $4 per watt. Progress is clearly needed, but the technology is advancing quickly; commercial efficiencies have risen from 9 to 10 percent in the past 12 months...

First Solar has its detractors, who "assert that the company could be hurt by limited supplies of raw materials in the future and increased competition." It turns out that tellurium is one of the nine rarest elements on Earth. Here's the smoking gun from altenergystocks.com—

In 2006, First Solar's 60 megawatts of production consumed 4% of the world's annual supply of [tellurium]. In 2008, analysts expect revenues of approximately 4x the 2006 number, meaning they will need approximately 16% of new annual Tellurium supplies. 

60 megawatts is nothing, a drop in the bucket. So much for cadmium-telluride thin film. But what of the many other alternatives?

Companies using silicon-based photovoltaics still have the largest market share by far. Silica, or silicon dioxide, comes from sand or quartz. The largest component of glass, silica is not rare at all, but appearances can be deceiving. The semi-conductor and solar industries, which are now in direct competition for processed materials, require highly refined polycrystalline silicon (polysilcon) or monocrystalline silicon for making chips or solar cells. Almost all this refined silicon is still made using the complex, expensive and energy-intensive Siemen's Process (1st graph left, with a 2nd simplified version from the Wall Street Journal).
For four years now, purified silicon has been in short supply. See the Wall Street Journal's The Silicon Shake-Up (subscriber only, September 21, 2007). Polysilicon prices soared from $25/kilogram in 2004 to around $200/kilogram in 2006, and PV production growth has come to a virtual halt. Sharp, the largest solar cell manufacturer in the world, "produced panels that could generate 434 megawatts of electricity [in 2006], or the equivalent of a single gas-fired power plant, about the same amount it made in 2005." The shortage has spurred the thin film craze because investors are leery of the polysilicon supply & demand imbalance, even though thin film PVs are only 7-10% efficient in converting sunlight, as opposed to about 15% for commercial silicon PVs.

How long will the purified silicon shortage last? An optimist will say at this point that the supply crunch is only temporary—high prices will stimulate supply and the market will come into balance again. However, in any steep exponential growth situation like those required by Staniford or the Scientific American authors, manufacturing capacity, even where potential breakthrough technologies exist (for pure polysilicons) will periodically lag behind demand, inhibiting steady growth over and over again as the market overheats and cools down. Large price volatility is destructive to markets. And then there is the question of scale. Should using purified silicon remain the method of choice, just how much built-out capacity would be necessary to create enough solar cells to meet 5% of our electricity needs? 10%? 20%? There is no free lunch.

In the case of depletion, as with cadmium telluride, the optimist might also respond that there are other alternatives, such as the copper-indium-(gallium)-diselenide (CIS or CIGS) thin film technology being pursued by Siemens, Nanosolar or Global Solar Energy. The risk is that the current bullish commodities market may become a permanent feature of the economic landscape. Take the indium used in thin films. Resource Investor reports that—

The Earth is estimated to contain about 0.1 ppm [parts per million] of indium which means it is about as abundant as silver. However, bullish supply-demand fundamentals have propelled the price from US$70/kg in 2001 to over US$1,000/kg today.

Indium is produced mainly from residues generated during zinc ore processing but is also found in iron, lead, and copper ores. In recent years, supply has decreased after a number of Chinese mining concerns stopped extracting indium from their zinc tailings.

At present, about 75% of the indium in the world is used in the manufacture of Liquid Crystal Displays (LCD’s), in computer screens and the new generation of flat screen TVs. The LCD industry is expected to achieve growth rates exceeding 30% over the next three years.

The flat screens will only be for the super rich some day because unlimited growth always entails scarcity eventually. Substitutes are not always readily available for elements on the periodic chart. It will always be something. There is no way out of this maze.

Bigger is Not Better, or Even Believable

Richard Smalley saw energy storage as the biggest challenge for solar (and wind) because each day the Sun shines on only half of the rotating Earth and thus it does not shine all day anywhere. Clouds also impede insolation , further dimming our ability to collect solar energy. One solution, discussed by Staniford, involves building out a slightly modified global power grid as proposed in Project Genesis (Global Energy Network Equipped With Solar Cells and International Superconductor Grids). Staniford, who does not want to depend on future breakthrough technologies, would replace the superconducting cables with high voltage direct current (DC) transmission lines, which are efficient over long distances. Such a grid gets around the storage problem by shifting the Earth's global electricity generation focal point to follow the Sun.

It is hard to directly criticize something like Project Genesis, for it seems so outlandish in its gigantism, requiring such a huge strain on Earth's limited resources and such an unprecedented level of co-operation amongst disparate peoples, that it is hard to imagine how it could ever get off the ground. A global electricity grid based on an enormous network of solar farms located in the Earth's deserts seems akin to science fiction fantasies like terraforming Mars or mining Saturn's moon Titan for its hydrocarbons. It's even hard to imagine how a "proof-of-concept" implementation would work. The burden of proof lies with those advocating such schemes. "Make it so," says Star Trek's Jean-Luc Picard.

Driven by their hopes and fears, people often can not distinguish between what they can imagine doing and what they can, or will, actually do. Why do engineered solutions need to be implemented on a planetary scale? The more grandiose the solution, the more unlikely it will ever be implemented. Bigger is not always better.

A somewhat more down to Earth proposal put forward by the "Grand Plan" authors envisions vast solar farms in America's southwest where daytime insolation (kilowatt hours/meter²/day) is high. If you view the Genesis project as a network graph, the solar farms in Arizona would represent one node of that graph which is not connected to vast solar farms elsewhere on Earth. This standalone electricity generating site thus requires large amounts of energy storage to supply reliable current when it is dark or cloudy.

The Scientific American authors envision that compressed air in underground caverns would provide that storage. "Electricity from photovoltaic plants compresses air and pumps it into vacant underground caverns, abandoned mines, aquifers and depleted natural gas wells. The pressurized air is released on demand to turn a turbine that generates electricity, aided by burning small amounts of natural gas."

Again, the problem is scalability. Small compressed-air storage has been used successfully in Germany and Alabama, but Zweibel et.al. are talking about supplying 69% of America's electricity consumption by 2050. This is another Big Engineering Project that is unlikely to ever be implemented. Richard Smalley preferred a "small is beautiful" approach that makes more sense.

The biggest single problem of electricity is storing it. When we are trying to find a way to store electrical energy on a vast scale, as we generally need energy in gigawatt power plants, there are very few options that one can imagine on that large scale for energy storage.

But if you imagine attacking the energy storage problem locally, at the scale of a house or a small business, the problem becomes vastly more solvable because there must be many more technologies that are accessible at the smaller scale.

If 100 million local sites each had their own local storage (based on improved batteries, hydrogen conversion systems and fly wheels), and they locally decided for their own particular sociological, economical reasons to use a particular technology to give themselves an hour of buffer or a day of buffer, five days of buffer, however long they have decided to do it — then the electrical grid could afford to be fairly erratic.

Only this kind of smaller-scale, distributed solution to generating the electricity we need in the future has any chance of being implemented.

The first obstacle we must overcome in solving the power problem of the 21st century is acknowledging that there are inherent limits to growth and thus what people can or can not do. A lesson in humility is required here. There are limits at the small-scale, illustrated by the problems with solar cell manufacture discussed above, and limits at the large-scale, where planetary engineering is unlikely to ever lead to real solutions for real people.

Residential solar electricity and heating is good. Vast solar farms in the world's deserts are unworkable and inherently risky. Imagine New England connected to the solar grid by only a few very large DC transmission lines emanating from Nevada. Tomorrow's power grid will not look like today's. Everybody is jumping on the Solar Bandwagon, but solar energy growth fantasies that allow us to power our plug-ins and keep the lights on without disruption are not helpful in making the stepwise changes we will need to get us to 2050 all in one piece.

Contact the author at [the original site].

Notes

1. Think of Energy (Work) = Power x Time. Or equivalently, Power = Work / Time. Power is measured in watts, energy in joules.

The standard metric unit of power is the watt. As is implied by the equation for power, a unit of power is equivalent to a unit of work divided by a unit of time. Thus, a Watt is equivalent to a Joule/second. For historical reasons, the horsepower is occasionally used to describe the power delivered by a machine. One horsepower is equivalent to approximately 750 Watts.

A terawatt (trillion watts) = 3.6e+15 joules/hour. 1 zettajoule = 10²¹ joules. Electricity is standardly measured in kilowatt hours, where 1 watt hour is equivalent to 3,600 joules. A kilowatt is thus the amount of energy "produced, transmitted, distributed, or consumed in a hour", which = 3,600,000 joules.

2. See Silicon Insider: Solar Companies Glow Despite Economic Slump for an example of bubbly hype in the venture capital-funded solar power industry.