The US uses over 25 percent of the world's energy. Most of it comes from fossil fuels, with about 8 percent from Nuclear, and 7 percent from renewables (mostly hydroelectric). Fossil fuels are obviously finite. Production of oil in the US peaked in the early 70s and has declined ever since. This trend has been or will be seen in every oil producing area of the world. The story will be similar for natural gas. Already, new wells in North America produce less than old wells did, and they decline faster. The trend will be followed by coal as well.
Prior to the use of fossil fuels, humans depended on solar energy, mainly in the form of wood. Does the end of fossil fuel mean the end of industrial society? Or, could it be powered by renewable sources?
There is considerable discussion these days of a "Hydrogen Economy". The vision is to use Hydrogen as a clean fuel: when it burns (or is used in a fuel cell) it produces only water. Actually, a clear distinction needs to be made. Because Hydrogen must be extracted from other things, and the extraction process requires energy, Hydrogen is really an "energy carrier" (of course, fossil fuel is really just an energy carrier, a transformation of sunlight from millions of years ago, but we only have to dig it up and use it, which requires much less energy than was required to create it).
Supposing we wanted to use Hydrogen as a "universal" energy carrier, where would the energy to create it come from? We would hope from renewable energy sources. One way of producing Hydrogen is to use an electric current to split water into Hydrogen and Oxygen. One possible source of the electricity to do this is from Photovoltaic (PV) panels.
So, we start by asking the question, "what would it take to replace all of the fossil fuel currently burned in the US with solar-generated Hydrogen?".
The US currently consumes 96.6 Quads (Quadrillion BTU) per year. This is a power measurement that can be converted to watts:
P = 96.6E15 BTU/Year * 0.293Watt/BTU/Hr * 1 Day/24 Hr * 1 Year/365 Days = 3.23E12 Watts. (3,230,000,000,000)
This is our average rate of consumption of all fuels. What would it take to generate this much Hydrogen from electrolysis, powered by photovoltaic panels? Assume that the sun shines on average 5 hours a day, that a panel can produce 100 watts per square meter in full sunlight, and that it can be converted to H2 using electrolysis with 80 percent efficiency (a number I have heard tossed out by H2 folks)?
How large an area of PV panels would be required?
Area = 3.23E12 Watts * 1m**2/100 Watts * 24hr/5hr = 1.55E11 square meters
This is about 155,000 square kilometers. If put all in one place, it would be a square almost 400 kilometers on a side. This is also about 60,000 square miles.
At today's prices, the panels would cost some $1.5E13 ($78,000,000,000,000). About 78 trillion dollars. As daunting as the expense is the fact that this amount of PV is hundreds of thousands times the Current annual output of the PV industry. *(note that we are talking not just about personal energy use, but about all industrial and commercial energy use in the US).
Obviously a huge area, but tiny compared to the area of the US. Let's think of some other ways to put the problem into perspective.
The first is that sunlight arrives in diffuse form. People also happen to exist in diffuse form across the land. What sense does it make to put all of the collection in one place?
What does the problem reduce to on a human scale? Dividing the area of panels by the 270,000,000 people in the US, we come up with 574 square meters per person, an area approximately 24 meters on a side. At today's prices, the cost would be about $1.3 million per person. Spread over 20 years, the cost would be $65,000 per year (ignoring the massive problem of producing the stuff).
While this is certainly high, it is not so high that the idea should be dismissed entirely. Remember that in 1970, the cost of PV was $100 per watt (in 1970 dollars), compared to less than $5 today. In inflation adjusted dollars, a decline of a factor of over 100. Cost of the technology continues to decline.
It is quite reasonable to expect that, especially if PV enters extremely large-scale production, cost will drop dramatically.
What else could be done? The most obvious answer is conservation. Per capita, the US uses twice the energy of Western Europe and Japan. Should we reduce our array to 287 square meters? We will assume that all parts of American life are equally energy inefficient, and that halving our consumption will be painless (do Americans wear energy waste like some sort of badge of honor?).
26 percent of our energy use is for transportation. Of our 287 square meters, this amounts to 74 square meters. Today, this liquid fuel is burned in internal combustion engines at about 15 percent efficiency or less. With conversion to fuel cell drive, at 40 percent efficiency, the same motive effect is obtained using only 10 percent of our energy. Our per capita PV area to generate the H2 for this goes down from 74 to 28 square meters, a reduction of 46. Our 287 square meters are now 241.
Now consider electric power generation. In the US, this uses 39 percent of our primary fuel. Do you know that 2/3 of this is sent into the environment as waste heat? This is 26 percent of the energy production of the US. Let us consider an alternate scenario: given a developing infrastructure for distributed energy processing, the electric power generation will be very distributed as well. Much of the electric power (perhaps 1/3) will come directly from the PV arrays, converted at about 90 percent efficiency (of course, the arrays are only 10-15 percent efficient at converting sunlight). The remaining 2/3 will be generated in fuel cells from H2, at 50 percent efficiency. Overall, about 58 percent of the energy from the PV is converted, instead of the 33 percent with today's technology, we achieve an effective conversion efficiency of 58 percent, though being able to use some of the energy directly instead of converting to H2.
The result is that our PV array can be further reduced. Whereas it originally took 38 percent of 287 square meters, (109 square meters) with improved conversion efficiency and much used directly, we can do our job with only 60 square meters. This takes our overall array down to 192 square meters per capita.
Now, there is another beneficial side effect of doing distributed electricity generation: the waste heat is available to be used instead of being thrown away. Today, a large amount of fossil fuel burning is simply to produce low-grade heat for process or simply for space heating. With distributed electric power generation, the waste heat can often be put to good use (residential and commercial space heating and water heating are good examples).
The availability of local waste heat, combined with good passive solar building design and inexpensive solar water heating should all but eliminate the need to size these tasks into the PV/H2 fuel system. How much should the PV array per capita be reduced to accommodate this fact? I can only make a guess, and take a conservative 25 percent off of it. I think it will actually be more.
This takes us down to 144 square meters per capita. Assuming the technology becomes half as expensive, instead of spending $1.3 million per capita, we might be talking about spending $160,000 per capita, and this over a 30 year conversion period.
With numbers like this, there is reason to believe that it would be possible to run our economy with H2 as the main energy storage/transport mechanism, and with the energy coming from photovoltaics. It certainly will not be easy, but neither does it appear impossible.
This discussion has investigated the use of solar energy to meet all needs. In reality, other renewable sources will be used for large portions of the capacity. Large wind turbines are currently competitive with fossil fuels, far cheaper than solar technology. In the future, they will be the most appropriate source for a huge portion of the energy supply.