We don’t burn solar modules to make electricity. If we could recycle 100% of them forever, we would produce an infinite amount of electricity per gram of material. Even in practical use, we use less PV material per kWh than uranium per kWh when we make PV electricity in comparison to nuclear electricity. The amounts of active materials used in PV are tiny.
Just how tiny? For an up and coming thin film, CdTe, we use about 12 gm of CdTe to make a square meter module. In a year in an average US location, we harvest about 11% x 1750 kWh/m2-yr, or 154 kWh/yr (after accounting for another 20% in losses, but not for an additional, but small annual loss). Thus in one year, we need 0.08 g/kWh. But wait! We don’t burn PV modules, and they don’t die after one year – warranties are about 30 years, so this is really one thirtieth of that, or 2.6 milligrams per kWh. Let’s make a table:
Amounts of CdTe Used with Different Recycling and Lifetime Assumptions
|Assumptions about PV
|30 years operating life||2.6|
|90 years and 90% recycling (after 180 years)||0.5|
In comparison, we burn:
|Uranium||24 (from http://www.stormsmith.nl/report20071013/partB.pdf )|
So the ratio of the use of CdTe to these fuels is as follows:
|Assumptions about PV
||CdTe/Coal Use per kWh||CdTe/Uranium Use per kWh|
|30 years||Five millionths||A tenth|
|60 years||Two and a half millionths||One twentieth|
|90 years and 90% recycling (after 180 years)||A millionth||One fiftieth|
So even without fancy assumptions about lifetime and recycling, today’s PV systems will use CdTe more conservatively than nuclear will use uranium by a factor of 10. But with reasonable recycling assumptions, made more realistic when one understands that CdTe manufacturers already recycle their modules, CdTe will use 20 and 50 times less material than nuclear per kWh of output. Compared to coal, of course, the numbers are out of this world. These differences in resource needs bear on the ultimate sustainability of the PV in comparison to other more resource-intense energy technologies.
Paula Mints is a respected member of the elite of the PV community. Her article in Solar Industry (November 2010) is not wrong, it just is not deep enough. Yes, to a great degree, thin films failed to revolutionize PV. But there are too many interesting things to discern from her data to stop there.
Perhaps Mints’ most convincing tabular evidence of the failure of thin films to revolutionize PV is her Figure 1, thin film share of total shipments. She shows that thin films’ share peaked in 1988 at 32% and plunged from there to 5% in 2004. Then it only picked up slightly to 17% by 2009. She didn’t have numbers from 2010, but given the growth of Chinese silicon, thin film share may even fall in 2010. Is this the characteristic of a revolution?
Having been part of the development of thin films, I’ll give Paula her due and say we once thought it would be a revolution. But the surprise wasn’t just about thin films – it was how good crystalline silicon could be. That’s a message I would agree with. This has to be said – crystalline silicon was better than we (the thin film community) realized; and it still seems to have plenty of potential for continued progress. There hasn’t been and there will not be any overall revolution of PV by thin films.
However, that is not “ ’nough said.” That falls far short of ‘nough said. Because Paula’s own data shows one thing in favor of thin films and in favor of CdTe specifically – it has grown explosively, an d even more rapidly than any other technology, including crystalline silicon. And it has made a difference, along with Chinese silicon, in adding zest to PV competition, with the very desirable effect of reducing typical prices and keeping the pressure on for continuing to do so. This is shown in Mints’ Figure 2, adapted here to show comparative CAGRs graphically:
The lesson is thin films are not monolithic. Pick out the growth rate of amorphous and thin film silicon (purple Xs). Despite the opening made by expensive silicon, thin film silicon grew only marginally since 2004 – 38%. By comparison, CdTe (blue *) grew at 181% since 2004, three and a half times higher than the growth rate of all module shipments. This is what success looks like. To ignore the success of CdTe is to fall short in terms of understanding a crucial factor in the future of PV and to diminish unnecessarily the success of thin films to-date.
Thin films are not created equal. A bit simplistically, it’s like the old story of Goldilocks and the Three Bears. For amorphous and thin film silicon, the porridge is not hot enough. The efficiency is too low. For CIS alloys, the porridge is too hot – although CIS cells can be very efficient (20% cells!), CIS is hard to manufacture. For CdTe, the porridge is just right – efficient enough and easy to manufacture. There may be a dozen documented ways to make 10% CdTe cells.
The concept of thin films makes sense – less material, less handling through large area substrates – but it has to be executed. So far, CdTe is the only technology that has executed. But when the concept of thin films is executed, it can stand up to any competition in PV, including Chinese silicon. Can anything else in PV say the same?
Beyond the status quo, only CIS looks like it might, someday, be a serious competitor of CdTe. And this would be good, because we need more competitors to keep the progress in PV robust and the competition lively.
There is something else to be learned about the failed thin film revolution, and this is very important, especially here in DC. A good new idea may take 30 years to pan out, and by the time it does, the status quo will have changed radically. There will very likely be no revolution, and there may not even be progress.
The technologies now in place have tremendous up-side. They will get much better and deserve the Federal support needed to keep them moving (and keep the US leading technologically). Tens of new PV ideas have been examined and dropped in comparison to them, but newcomers to PV (and I include almost everyone who comes to DC with every new Administration) make the mistake that they (1) can have a revolution in PV and (2) that we need one. We can’t, and we don’t.
For usually, when newcomers have new ideas, they are old ones – old and discarded ones that are very boring to the PV community. Plastic solar cells? Dye cells? Cells made on flexible foils? Give me a break. They exist and can be barely sold. These technologies are not terawatt-worthy. Yet often, it is just these kind of cell technologies that people new to PV fund and research when they talk of the next revolution. Someday we will have a mature PV program, when we have the patience to learn about the one we have.
So, Paula, pardon me for using your excellent story as a springboard. Your data tells a deep and important story, even beyond the necessary recognition that thin films did not revolutionize PV.
I get sick and tired of constantly reading off-the-charts high prices for solar PV quoted as the only price. You see this in articles from all sorts of media, and it is swallowed hook line and sinker by everyone who reads them. So what is the real price of PV?
The answer is “prices.” There is no real price, because there are many prices. Price varies by the amount of local sunlight and by the system size and type.
And there are two kinds of prices – dollars per watt, which is price per instantaneous power. And cents per kWh, which is price per unit of delivered energy. Dollars per watt is hard; cents per kWh is harder.
So with this in mind, let’s do some prices!
Big systems are cheaper than small systems; tiny systems, like those on your house, are more expensive again. If we assume the biggest systems can be installed (sans delays and all sorts of undefinable costs) at $3/W; then large rooftop systems on WalMart might go in for $4/W; and residential systems for $5/W. These would be “perfect” systems, with no delays and other setbacks. For more typical ones, you can add a dollar or even $2/W. These are all fixed arrays; if you want tracking, add another 50 ¢/W to a dollar to the large system price (but you get 25% more output).
Then there’s location. In comparison to the desert southwest US, putting up systems in average US locations are about 20% more expensive because you get 25% fewer photons; the East Coast is about 35% worse. (These are slightly corrected for reduced temperature losses versus the Southwest.) So if a large fixed array is three units of cost, what are the others?
Table 1. Relative Costs for Different Sizes and Locations (ratios to $3/W in Southwest)
|Largest Systems||Large Rooftops||Residential|
Let’s put this into something we can understand, cents per kWh (¢/kWh). There is no easy way to do this, but all we are looking for is a sense of these prices. Even these prices are estimates, and real prices vary all over the place. So let’s assume $3/W is 16 ¢/kWh, and take ratios to get the others.
Approximate Prices by Location and Size in ¢/kWh (no incentives)
|Largest Systems||Large Rooftops||Residential|
|Great Sunlight (US Southwest)||16 ¢/kWh||21 ¢/kWh||27 ¢/kWh|
|Average (Kansas City)||20 ¢/kWh||27 ¢/kWh||33 ¢/kWh|
|Northeast (NYC, DC)||22 ¢/kWh||29 ¢/kWh||36 ¢/kWh|
Thus you can see where someone can correctly quote nearly 40 ¢/kWh about PV costs, for example in a less sunny place like Germany; and I can quote 16 ¢/kWh in the US Southwest for large systems. We have not yet come to the point where different locations, applications, and sizes are distinguished in PV.
With 30% investment tax credit, these would calculate about 30% cheaper, and many have made the argument that this is only compensation for solar’s environmental, economic, and security values, which are otherwise unmonetized. Just to validate this somewhat, Jens Meyerhoff, head of First Solar’s utility systems business, stated in sworn testimony before Congress September 23: “First Solar is capable of providing solar electricity at a cost between $0.12 and $0.16 per kilowatt-hour” (presumably after application of the investment tax credit). This is consistent with the numbers in the above tables.
Is there a more lovable energy source than solar power?
Its fuel is light from the sun, which is free and available almost anywhere. It’s entirely clean to run. It’s easy to install and maintain. The big utilities can use it to generate green grid power, but you can also put it on your own roof to become your own utility or to go off-grid.
And did I mention that it’s powered by THE SUN?
(Full disclosure: I do some work for a solar power company in Virginia called Secure Futures).
It’s no wonder that public support for solar power is high, with 92% of Americans in a 2009 poll saying that it’s important to develop solar energy resources. In particular, environmentalists love solar power best of all energy sources. It’s as clean as wind power, but solar is much less controversial.
Wind turbines make noise and spoil hillside “viewsheds” that bother the neighbors and the turbines’ huge spinning turbine blades can kill birds and bats. But aside from a need for land and some use of toxics in manufacturing, solar has few environmental impacts. And so far, there has been very little NIMBY opposition to solar installations.
Itsy bitsy teeny weeny
But many experts in energy beg to demur. They admit that yes, solar power sure is cool. But they say that solar is not a practical source of electricity today. And they predict that it will probably never become practical in the future.
Skeptics have three main problems with solar. First, because it’s intermittent (the sun doesn’t always shine), solar power can’t provide the always-on power that we’re used to. So, every time you put up a solar power installation you also need to build or pull in a dirty fossil fuel or nuclear plant to back it up. That’s not so clean, is it? And it’s wasteful too, since you basically have to keep those backup plants running on standby 24/7.
Second, solar power is also expensive, not only because you need all those other plants just sitting around as back up, but also because making solar panels requires fancy-dancy materials like rare-earth minerals and costs money for many other reasons. Even with government subsidies, solar power still can’t compete today with coal or nuclear power rates.
Finally, even if you could store solar power at night or ship it over from sunny areas like Arizona to places that need the juice like New York City, the battery and grid technology is so far in the future that solar power won’t be able to scale up in any meaningful time frame to replace coal or nukes.
So, critics say, no matter how neat solar panels and reflector mirrors look gleaming in the noonday sun, solar power always seems to be the energy source of tomorrow. Put together, after decades of development photovoltaics and solar thermal power still can’t produce even one percent of America’s juice. Doesn’t that prove that solar will always be rinky-dink?
In support of this view, Tad Padzek of the University of Texas at Austin told the ASPO-USA conference in October that if you measure all electricity sources by the number of days worth of usage per year that each provides, solar is microscopic. If coal covers 176 days, nuclear power covers 72 days and wind power covers 5 days, solar power would account for only one puny hour of America’s electricity usage.
Another skeptic, Robert Hirsch, who also spoke at the ASPO event, referred to solar power in his book The Impending World Energy Mess as the “emperor’s underwear,” an energy source that is not a total fraud and does have some value, but whose power comes only at a very high price.
Growing, but without much love
“You have to start somewhere,” says Ken Zweibel, director of the GW Solar Institute at George Washington University.
Zweibel told the ASPO-USA conference that, although the US doesn’t have much more solar power today than we did ten years ago, we have yet to see a nationwide emergency program to ramp it up. Quite the opposite, in fact. Most of solar’s growth has taken place in a start-again-stop-again policy atmosphere where incentives were intermittent and investors had difficulty planning the true costs of a project. “We’ve seen a 3000-fold increase in solar capacity without really trying.”
“Recently, the numbers have started to grow, doubling over the year before,” Zweibel told me. “It doesn’t take many doublings for things to get pretty astounding.” Read more…
I have never actually seen the inside of a calculation of how much it costs to avoid CO2 using PV. So I thought I’d do it myself and see if it’s as horrible as some people seem to imply.
Now a watt of PV installed for a year can produce from about 1-2 kWh/yr. Maybe that watt cost $3 to install. One kWh produced normally in the US, by the EPA estimate, is about 0.7 kg CO2 / kWh. So using an average value of 1.5 kWh/W-yr, we could assume about 1 kg/W-yr.
But this forgets two pretty significant things – those kWh are valuable; and there is a loan that adds cost to the PV. So for round numbers, let’s assume the loan doubles the amount of money spent on the PV; but the value of the electricity offsets the original amount. Voila, we are back where we started – $3/W of added cost. This is obviously very rough! But it is assumptions like this that make all these calculations ‘rough’.
So if you calculated the cost per MT of CO2 on the basis of one year PV output, it would be $3/1 x 10-4 MT = $3,000/MT – not a pretty picture.
But if you assumed that PV would last 30 years, then PV would cost $3000/30 = $100/MT. This seems to be the number you hear the most often.
But at a hundred years, this might settle down to about $30/MT plus some degradation and for O&M. So here’s my first ansatz at calculating the $/MT CO2 of PV:
Cost per MT of Avoided CO2 from PV (First Approximation, $/MT CO2 avoided)
|Assumes 1.5 kWh/W-yr Sunlight||$3/W||$1.5/W|
If the PV were cost effective at $2/W, say, then the $3/W number would be higher; but the $1.5/W number might be almost zero (and the the details of the loan, the degradation, extra cost for variability, and O&M would need to be included).
Here’s another way to look at this, based on levelized electricity cost. We have to make some broad assumptions, so let’s assume that solar electricity is worth 10 ¢/kWh. And let’s assume that $3/W system has a levelized cost of electricity of 16 ¢/kWh with a 30-year loan (this is about right for a sunny location). That means it needs a 6 ¢/kWh for 30 years subsidy to be installed. Each kWh avoids 0.7 kg, so this is 9 ¢/kg, or 90 $/MT (rounding up). Each year, society would have to pay an additional $90, but would also get another MT of saved CO2.
This doesn’t count the fact that after 30 years, society would actually start saving money (the PV system is running for 1 ¢/kWh, so we would be getting 9 ¢/kWh savings). So that means in another 20 years, or 50 years total, our CO2 would have been avoided for free. (This connects up with the previous analysis if the savings after the loan were included.)
But no one ever counts the post-30 years part, usually saying it is “discounted” away. So much for the Hoover Dam – it’s just a phantom of our collective imagination. (Actually, the savings are real; it’s only the economists who think business-defined discounting is “real,” when in fact it is only used for business to maximize short-term profit. It is not societally defined the same way.)
A few years ago, I jumped on compressed air energy storage (CAES) as a means of shifting solar electricity to nighttime.
More recently, I figured out that as long as wind was less expensive than PV, wind would be stored first in CAES.
Under the circumstances that wind is cheaper than PV and wind blows more at night, we won’t be seeing much PV stored for shifting to nighttime.
It’s going to be hard for PV to get cheaper than on-shore wind. Wind today seems to have about 50% more output per installed watt than PV does. So that means PV must be 2/3 of the cost of wind per watt to be equivalent to it. If on-shore wind is about $2/W, then PV would have to reach $1.33/W – maybe $1.5/W because wind has higher O&M costs. This is tough but eventually likely about 2020. Even when it is achieved, however, it only means PV becomes about the same price as on-shore wind is now. (Of course, solar is produced during the day, when it is more valuable than wind. And PV is already about the same price as off-shore wind.)
According to recent DOE studies, up to deployments of about 20% PV or wind, we can handle the variability. But above something like 20% electricity each, there will start to be too much PV electricity during the Spring and Fall days, and too much wind on many nights.
So here we have it – a world where we push the limits of the grid with 20% PV and 20% wind, at something up to 40% of our electricity. And sometimes, on the days or nights of the lowest demand or the highest wind or PV, we have too much wind at night or too much PV during the day. Does this limit us? Is 20% each the end?
That’s the point where people start talking about storage. But storage about doubles the cost for the electricity that is stored (roughly!), and it isn’t even all that proven at scale.
What about moving electricity as an alternative to storage? Move PV electricity from midday to evening by sending it east from the west coast. Or move it from daytime to nighttime by sending it from the Sahara to New York City under the Atlantic with high voltage DC. Or move wind from nighttime to daytime. You can build more and more PV and wind as long as you can send it further and further away where conditions are different.
Surprisingly, the economics of shifting electricity are about the same as storage. 10% loss for three thousand miles is like batteries; 20% for 6000 miles is like pumped hydro; 40% loss is about 12,000 miles, half the earth away – and is like compressed air energy storage (CAES). Then you have to compare the capital costs, and interestingly, they tend to favor transmission, except for the very longest distances (and depend sensitively on whether you can use the transmission both ways, i.e., fully use their capacity to offset their cost).
It comes down to caverns and transmission on land for CAES; or transmission on land and underwater for shifting wind and sun to where it’s needed. It’s transmission either way, but shifting is more proven at cost and scale than storage. Yet it is rarely compared to storage.
Transmission isn’t as much a technical problem (it is already way further along than storage) as a societal one (i.e., NIMBY); and political, since it connects different parts of the Earth. So it doesn’t support the idea of energy independence; and it is vulnerable to culture clashes and isolated revolutionaries. But perhaps proper redundancy and keeping the fossil fuel power plants ready and storing some fossil fuels needed to overcome an emergency might be enough (I want to acknowledge that I heard something like this idea for backup fossil fuels first from Arnold Goldman; and the idea of international transmission goes back at least to Buckminster Fuller). After all, as everyone knows, it’s a lot easier to store fuel than store electricity.
It isn’t time to get too squeamish about solar and wind beyond 40% of our electricity. As we know from other blogs, this is enough to eliminate all our imported oil, if we had electric transport. So it’s quite a good amount. But it is time to think a bit beyond the box about alternatives.
This must be the byword of photovoltaics. Could anything be more out of tune with its age?
Invest big, now; make money slowly, later.
Invest big, now; get nearly free electricity, later.
Invest now, reduce CO2 later.
Work now, sit on your laurels, later.
Even worse: subsidize now, reduce costs for later.
So many things about PV have the form: pay a large amount now (invested money, energy, CO2), get paid back slowly but in a big way, later:
Reward(t) = Slow payback(t) – One Big, Up-front Cost
Just as a reminder of some other things like this, let’s recall the beauty of:
- Bridges and tunnels
- Water supplies from distant sources
- Cars and trucks
- Manufacturing equipment
- Houses and buildings
- Flood control and dams
- Laws and regulations
- Standards and codes
- The Constitution
- Public health policy
Are you vague about standards, codes, laws, regulations, the Constitution being like this? Think of Haiti – think of the costs of building codes in San Francisco, and compare that to a lack of codes in Haiti. Up front cost. Think of signing and living by a contract instead of stealing.
These are the bases of civilization.
If PV is out of tune with our times, we are out of tune with civilization.
We can do better.