Comments: The New California Solar Plants, In Context

My first response would be you should do two calculations, one based on "need" and one based on current "demand".
Subtract the difference and you'll have a nice tidy figure to work with when it's time to change everything about how we live.
Rather than configure our hardware to meet the ungoverned volume of current usage, strip usage to its sustainable minimum, and go from there.

Posted by Roy Belmont at August 18, 2008 02:38 PM

Bingo. I agree with Roy Belmont; need calculations should be predicated on future needs not current waste.

McMansions anyone? Suburbs? Abandoned strip malls? Focus on economic growth that benefits a few at hugely inefficient models. Profits privatized, risks socialized?

Posted by Labiche at August 18, 2008 03:38 PM

You might also look at a calculation of area used. These solar plants generate only 1/30,000th of the current world energy needs, but they sit on 12.5 square miles, or 1/4.6e6-th the world's landmass, or 1/280,000th the area of the Sahara Desert. In other words, we could generate energy to the entire world with only about 1/10th the Sahara, a region that nobody is using anyway.

The big problems for our energy supply are distribution and storage. Generation of energy is a solvable problem, and these solar plants prove it.

Posted by Remus Shepherd at August 18, 2008 04:47 PM

hmmm... its interesting to think that with less than 1/10th of the area of the sahara devoted to solar, plus already necessary changes in the lifestyles of those in the rich countries (and of the rich in the "developing" ones), enough energy could be generated sustainably for an economically far more just world. of course, this is only an illustration: other forms of generating energy sustainably in different parts of the world would be far more efficient. but still, a provocative illustration, nonetheless.

the big problem, of course, would not be of landmass usage; it would be of money. the cost of 30,000 solar plants of this type would be of course extraordinary, or even of, let's say, 10,000 of them (which would in no way whatsoever be a conservative estimate of energy need in the case of reduced worldwide consumption).
then again, if one totalled the cost in todays dollars of building every oil drilling rig, pipeline, refinery plant, etc., I'm sure a much huger number would be generated.

if we've gotten to here, we can get to there. the difficult part is that we have such little time now in which to do it...

Posted by matthew c at August 18, 2008 05:34 PM

No, that sounds about right. But, as Remus says above, doing the area calculations gives you a better idea of what's involved since solar panel arrays require lots of space.

A big hurdle to rolling out solar PV on this scale is the materials used--lots of silicon and glass. Also, something to look out for in the future is increased efficiency of solar PV. Current efficiency values of what's on the market today are around 15%. Experimental panels can get over 30% and there's been one or two with 40-42% efficiency. Don't know if those are commercially realizable, but doubling the efficiency should roughly cut the area of needed panels in half.

Something that I don't see brought up often (maybe its not important) is the question of what sort of environmental effects enormous solar PV installations would have. I know the Sahara is sparsely inhabited and barely has any life to begin with, but covering 1/10th of it (or 1/10th of any place) with huge quantities of heat-absorbing metal would probably do some drastic things to the temperature there, which could in turn effect precipitation patterns all around it. I live in Phoenix right now and the urban heat island effect from all our concrete has had a huge effect over the last few decades.

Posted by Bolo at August 18, 2008 05:41 PM

You're about right, but so what?

Assuming that the power has to come from somewhere (it does) and assuming that the cost of fossil fuels is going to continue to rise over the next century (it will), renewables are going to become more economically viable, which means people are going to build them because they can sell the power. Therefore, if we need 30,000 solar plants to generate the world's energy, we'll end up with 30,000 solar plants. The world currently obviously needs 30,000 nuclear, gas and coal (etc etc) plants and they are all going to need replacing sooner or later. I know it's a big scary number but, well, duh.

Realistically, however, that's a silly way of looking at it. The biggest advantage of PV is going to come in microgeneration. Technological advances being what they are, we're less than 25 years away from a new-build house being able to generate all its power from PVs in the roof, the windows, scattered on the lawn, what-have-you, at a cost that's comparable to hooking up wires to a power wholesaler and paying by the KW/Hr. Not all houses will do this in 25 years, but some will, and this microgeneration will take down the demand on national and international grids. Just as many currently "developing" countries have bypassed the 19th-century copper-wire analogue phone system and jumped straight to a more efficient cellular system, we're more likely to see other countries forgoing upgrading power grids in favour of just building houses with PV, boreholes and batteries. We can already produce PV in some quantities at $1/Watt, which is the "Holy Grail" of price-competitiveness with fossil. The competition is fierce and we're likely to see that drop still further as the economies of scale kick in. At $.20/Watt (my guess - ten years) people are going to be covering their roofs with this stuff in Africa and India.

More to the point, we're not going to use all solar for grid generation (which we will still need, and in significant quantities, no matter how much microgeneration we do). As the fossil cost rises, we're more likely to switch to a mix of nuclear, geothermal, tidal/gravitational generation because they're all baseload technologies that don't rely on intermittent energy sources. Saharan Africa will almost certainly end up dotted with shiny silver mirror arrays pumping daytime electricity to, at the very least, Southern Europe, as will Arizona and New Mexico for the Western USA, but people have started to notice that the Earth's core doesn't turn off at night and, hey, maybe we can take some of that heat and use it to drive turbines. My guess for HDR Geothermal to hit or beat economic parity with fossil is two decades at the outside.

Basically, your point is that we're going to need a lot of this stuff. But since we're already building it and technology rarely gets more expensive over time, it's a very good guess that we'll get the quantity that we can pay for, when we need it. It won't, of course, be in time to stop the cumulative effects of a century and a half of global fossil emissions on the world's climate, but then we already knew that. We'd have had to start in, say, 1980 in order to ameliorate the damage, but there were apparently more important things to do back then. Oh well, we live and learn, don't we?

Posted by McDuff at August 18, 2008 06:26 PM

Solar thermal has the bonus of relying on it's own cheap energy system storage to power sterling engines - the stuff you boil with said sun: water, salt, whatever - so it can also serve as a non-intermittent source.

Posted by buermann at August 18, 2008 07:45 PM

First of all, you say California is building these two solar plants. So, why do you then leap from California power plants to energy consumption by "mankind" across the entire earth?? Since when have California power plants ever provided energy for the entire earth?? Your calculations would have been more honest and helpful if you stuck with California and assessed what percentage of the state's energy needs will be provided for, and how much smaller our carbon footprint will be.

Posted by Andy at August 18, 2008 08:16 PM

why do you then leap from California power plants to energy consumption by "mankind" across the entire earth?

Because these are the biggest solar plants ever constructed -- more than ten times larger than the runners up. And the stories about them emphasize that and don't provide any context for how much more of them we need.

I certainly am not down on these solar plants -- just the opposite. I would happily support any plan to have the US (or world) generate all its power from solar, which is clearly feasible.* But it's important to understand the scale of the challenge.

* Actually, that's not quite true. I'd rather have a big chunk of our power coming from biochar, which generates energy in a way that actually takes CO2 out of the atmosphere (as opposed to solar, which is merely carbon neutral).

Posted by Jonathan Schwarz at August 18, 2008 08:23 PM

You may have missed this article from MIT:

In a revolutionary leap that could transform solar power from a marginal, boutique alternative into a mainstream energy source, MIT researchers have overcome a major barrier to large-scale solar power: storing energy for use when the sun doesn't shine.

Until now, solar power has been a daytime-only energy source, because storing extra solar energy for later use is prohibitively expensive and grossly inefficient. With today's announcement, MIT researchers have hit upon a simple, inexpensive, highly efficient process for storing solar energy.

Requiring nothing but abundant, non-toxic natural materials, this discovery could unlock the most potent, carbon-free energy source of all: the sun. "This is the nirvana of what we've been talking about for years," said MIT's Daniel Nocera, the Henry Dreyfus Professor of Energy at MIT and senior author of a paper describing the work in the July 31 issue of Science. "Solar power has always been a limited, far-off solution. Now we can seriously think about solar power as unlimited and soon."

I hope this is more real than the cold fusion, or was that fission, in a glass of water thing.

Posted by Rob Payne at August 18, 2008 10:05 PM

Because these are the biggest solar plants ever constructed -- more than ten times larger than the runners up.

I know there is a 5,000 Mw construction project in the works in India when it comes online it will dwarf this.

Last I heard Algeria was looking at doing exactly with the Sahara what was suggested.

Posted by Ed Marshall at August 18, 2008 10:24 PM

I apologize for the former physics teacher who is about to appear.

Rob Payne - The MIT thing is real. There's a lot of research in that field, it's quite mainstream. I'm also a big fan of another recent MIT result - a solar collector design utilizing an ordinary window with a waveguide coating and solar cells embedded in the frame. With this design, it's possible to utilize a collection area of ~ 1 sq meter or so with a solar cell area of ~ 10 sq cm or so. A great idea.

McDuff - kW/hr makes no sense, don't use it. It's kW-hr. This seems like nit-picking, but it isn't. These numbers are not complicated, and it would be good if non-technical people took the opportunity to play with them more often. But that's impossible if there's confusion about what power is (energy divided by time) and what energy is.

Bolo - you're correct about the efficiency numbers but you should dig a bit deeper there. The maximum theoretical efficiency for a single-junction cell of any type is about 30% - there is a little controversy about this, but it's largely settled. The only type of conventional cell anyone will ever put on their roof will be a single junction cell. I think it is unlikely that efficiencies exceeding 25% will ever be available commercially with this technology.

There are quite conventional solar cell designs with 45% or better efficiency. They use exotic materials (InGaAs, Ge, InP, and others) and multiple junctions, and are fabricated very slowly by molecular beam epitaxy. If you have them on your home, you live on a space station and don't care what they cost.

Jon - don't say "average of 16 TW". It's not terribly incorrect, but the statement makes terawatts sound like energy (which is TW-hrs) and it crosses units. (That is, you average a basket of TW-hr apples and arrive at TW oranges on the other side of the equation. That's not an average.) It would be proper in my opinion to say that "Human energy consumption is 140,000 TW-hrs, which corresponds to power consumption of 16 TW." It makes a great deal of sense to average numbers which convey energy usage.

It's a bit trickier for power. Although there is certainly such a thing as average power it's kind of a different concept.

Posted by Aaron Datesman at August 18, 2008 10:43 PM

I Agree with McDufff, solar panel roofing or even modular solar panel roofing shingles. Two jobs in one. Just plug 'em together.

Posted by Mike Meyer at August 19, 2008 12:43 AM

yeah, it's a big operation, this industrial world. we build and recycle a fleet of almost a billion cars every 15 years or so, too, roughly. the number of miles of pipe we use for water, fuel, etc is pretty crazy. and how much we add to all the trash piles and so on. those ~1200 terawatt-hours we blow on incandescent lighting. the 20,000+ cargo ships at sea today.

la, la la, la la, la la. here we go, here we go.

Posted by hapa at August 19, 2008 01:03 AM

I'm astonished physics hasn't updated its energy units since the mid-80s. No wonder we have an energy crisis, it's like we're stuck in the dark ages of the 19th century!

"The MIT thing is real"

I don't understand this MIT thing at all. They're doing electrolysis with a cobalt and phosphate catalyst at the anode, and the usual platinum catalyst at the cathode, so they're apparently cutting the material cost by nearly half, but does it make the reaction more efficient? I was under the impression you'd lose at least 30% of the electrical power input to heat, and some similar amount in the fuel cells, in which case you're better off with existing batteries, or, supposing we have two such promising lines of future research for electricity storage, ultra-capacitors.

I don't really understand such direct comparisons to photosynthesis, either. It's not storing solar energy in hydrocarbons like plants or algae, it's just generating H2 with electrolysis. Crude oil has a stronger claim to being "inspired by the photosynthesis performed by plants" than that.

Posted by buermann at August 19, 2008 01:22 AM

So, what I want to know, and what I haven't been able to find easily anywhere, is just what the net energy yield of these things is over their lifetime. I.e., how much energy does it take to produce, say, a square metre of photovoltaic cell, complete with wiring, mounting, off-line energy storage, etc., etc., versus how much energy does the thing produce during its useful lifespan?

The (obvious) point being, if it takes more to produce the things than we get out of them, then it's a downhill slide into oblivion.

Oil works because the energy to produce it was applied over millions of years, and it costs less to extract it than it yields in energy. Ditto for uranium.

Hydrogen, on the other hand, is useless as a fuel source (unless you can fuse it), because it requires more energy to crack it out of water than you get back when you burn it. It's only useful as an energy transfer mechanism.

I seem to remember reading somewhere that wind power fails this test, because the resources required to build and maintain the wind turbines and towers exceed the energy extracted from them over their useful lifespan (certainly in wet, salty environments such as Scotland — YMMV in a dry, windy area).

Anyway, the answer to the first question would be appreciated, if anyone knows it. It also seems somewhat essential if one is going to invest any significant resources in deploying the things.

Posted by Mike at August 19, 2008 03:21 AM

@mike: nobody would be buying this equipment for real in the present if it didn't provide an energy gain.

It also seems somewhat essential if one is going to invest any significant resources in deploying the things.

exactly, and since people are investing heavily, that should have been a red flag in your head that your information was wrong or out of date.

current home and industrial PV run a range from 3:1 to 20:1, basically, with something like 8:1 being the midpoint, and very fast improvement.

current industrial wind operates at about 20:1 or better. hence all the new wind farms.

pardon the harsh tone but you're saying -- in the face of one of the world's greatest ecological crises -- that environmentalists and clean tech investors don't know how to count -- when the evidence points toward the polluters as the math-challenged.

Posted by hapa at August 19, 2008 04:08 AM

First off, let's get one thing straight: I'm all for anything that generates enough power to keep civilisation going without destroying the damned planet in the process. If that means paving Saudi Arabia and the entire southwestern US with photovoltaics, I say, let's do it.

nobody would be buying this equipment for real in the present if it didn't provide an energy gain

Bullshit. People invest in lots of things that are long-term losers, if they see a short-term gain. Look at automobile manufacturers that still make SUVs. And the last time I checked, people were still buiding petrol stations. The fact that people are investing in something does not, a priori, mean that the something in question makes any sense.

current home and industrial PV run a range from 3:1 to 20:1, basically, with something like 8:1 being the midpoint,
current industrial wind operates at about 20:1 or better

Really? Up to 20 times the energy out of them that it took to mine the materials, refine and process them, manufacture the components, assemble them, distribute them, install them, maintain them, and then dispose of them when they're done? And that includes moving all the workers around, feeding them, etc. I'm not questioning your numbers, per se, I just want to make sure you understand my question, and those numbers look pretty good.

in the face of one of the world's greatest ecological crises -- that environmentalists and clean tech investors don't know how to count

There are a lot of environmentalists and tech investors who don't know the first thing about thermodynamics or nuclear physics, and I don't necessarily trust them when it comes to analysing power production. There's a lot of politics involved in this stuff, and that means many of these people are not behaving rationally. Certainly that's the case with nuclear power and waste disposal, and I see no reason to assume it's not the case with other technologies as well.

Posted by Mike at August 19, 2008 04:33 AM

Oh, here we go:

A number of detailed studies on the energy requirements on the three types of photovoltaic (PV) materials, which make up the majority of the active solar market: single crystal, polycrystalline, and amorphous silicon were reviewed. It was found that modern PV cells based on these silicon technologies pay for themselves in terms of energy in a few years (1-5 years). They thus generate enough energy over their lifetimes to reproduce themselves many times (6-31 reproductions) depending on what type of material, balance of system, and the geographic location of the system. It was found that regardless of material, built-in PV systems are a superior ecological choice to centralized PV plants. Finally, the results indicate that efficiency plays a secondary role to embodied energy in the overall net energy production of modern solar cells.

Posted by buermann at August 19, 2008 05:11 AM

Great! Let's go!

Posted by Mike at August 19, 2008 06:12 AM


The fact that people are investing in something does not, a priori, mean that the something in question makes any sense.

mmm. and by this, if people are not investing in something, that doesn't mean that the something in question doesn't make any sense.

and from that, one gets a renaissance, without any real new productivity?

Posted by hapa at August 19, 2008 05:49 PM

There's no reason to restrict considerations to PV solar. The heat engine solar is more efficient - quite a bit more efficient - in its use of land area, with less environmental downside in its manufacture. And it scales up well.

Posted by hercomed at August 20, 2008 06:24 PM

Just looked it up: there are more than 50,000 coal fired electrical power plants alone, currently in operation.

In case anyone was frightened off by the 30,000 figure.

And the California location is not ideal - smaller plants would do as well on better sites. As has been noted above, it's storage and distribution that bottleneck, not production.

Posted by hercomed at August 20, 2008 06:44 PM

@ buermann -

A couple things to note about the MIT discovery.

1) Photosynthesis is so ubiquitous that it's taken scientists many years to bother to examine it closely. (It's also a matter of the tools to examine it only recently becoming available.) The discoveries which have been made are truly amazing. For instance: photosynthesis is a process which consists of 3000 (!) individual steps. And it also seems that the efficiency of the process (which is amazing) depends completely upon quantum mechanics. It can't be understood according the laws of physics as they were construed just 100 years ago - when everybody who was anybody thought they were totally complete.

2) At its basis, photosynthesis describes a very complicated and involved problem of photon-assisted charge transfer. You can view this mechanistically to some extent - the electron hooks on here, this flexes, the electron jumps, it's sitting there, it pops over on to this other molecule, etc. It's obvious that this includes all of the elements we want for PV systems (sunlight and charge transfer). But the basis is chemistry, not semiconductor physics.

3) It's understanding the charge transfer mechanics which is interesting. Advances in this field could push forward development, for instance, of plastic solar cells. Imagine if you could roll solar cells out on your roof like a carpet! I've seen recent NREL results for plastic cells in the 3-5% range, not bad for a field that didn't exist ten years ago.

4) The Nocera invention described in the MIT press office blurb also involves charge transfer dynamics. This is a more interesting description of the research motivation here:

So in my opinion the comparison to photosynthesis is perfectly appropriate.

5) Your remark about electrolysis is not totally correct. Electrolysis involves ionic fluids, a list of which does not include (pure) water. When I used to demonstrate electrolysis for students, for instance, I poured salt into the water. In this case, the ions assist the charge transfer process. The Nocera discovery is qualitatively different.

If you consider how corrosive salt is, you begin to see why this discovery is important. If you believe (as many do, including me) that distributed systems are the way to go, then a storage mechanism is very important. This is a big step.

6) You might be surprised to learn that one location where much innovative solar energy research is being conducted is Sweden. But they are interested only in solar-to-hydrogen conversion. Because they have a great solar resource for 3-4 months out of the year, and none at all for about the same length of time, they focus all of their efforts in this direction.

Posted by Aaron Datesman at August 20, 2008 08:54 PM