The Problem with Solar Energy

Over at The Island of Doubt, James Hrynyshyn has a post about solar skepticism on the part of some researchers, who think that claims of increased efficiency are often overhyped.

Of course, efficiency isn't the only issue. A couple of weks ago, we had a colloquium talk by Peter Persans of RPI, who is working on developing new types of solar cells using amorphous silicon and "quantum dots." He opened the talk with a fairly sobering description of the energy situation, though, which really puts the challenge of solar energy into perspective.

As best I can reconstruct it, the argument went like this: In order to meet the energy needs of the US entirely with solar power, we would need to cover 0.2% of the land area of the United States with photovoltaic cells, roughly equal to the area of paved roads in the US. And that's using solar cells with an efficiency of 50%, not too far below the theoretical maximum for a single-layer device.

But efficiency isn't the only problem: He pointed out that in order to build that sort of solar energy infrastructure, we would need to produce and install 2,000 square kilometers of solar cells a year for twenty years. To put that in perspective, we currently produce about 200 km2 of plastic film a year-- plastic wrap, garbage bags, etc.-- so we're talking about producing complicated solar cells at ten times the rate that we make plastic wrap. That's what they call a "significant technical challenge."

Of course, it wasn't entirely downbeat: he did note that great strides have been made in the production of solar cells, and described some intriguing possibilites using "quantum dots" embedded in silicon as a means of increasing the bandwidth of light that produces useful electricity. And there are other ways to use solar energy than direct photovoltaics.

Still, the plastic wrap number is really eye-opening. Research into better solar cells is absolutely critical (and that fact that funding was killed by Reagan in the early 80's is nearly criminal), but even if some brilliant scientist makes a dramatic breakthrough tomorrow, solar is not an overnight solution to our energy problem.

More like this

Yesterday's reason to love quantum was the CCD sensor, which relies on the photoelectric effect to take digital pictures. Sticking with the photoelectric theme, today's first quantum-enabled technology is the photovoltaic cell, the basis for solar panels. Photovoltaic cells convert light into…
A big step in improving the efficiency of photovoltaic cells in on the horizon. A paper published over the weekend in Nature Physics describes the ability of a substance called Graphene to convert a high percentage of the energy from sunlight into electricity. Graphene uses more of each photon's…
Perovskite (say it: pə-ˈräv-ˌskīt, -ˈräf-). It may never become a household word, and the chemical formulas (eg., CH3NH3PbI3−xClx ) are as long as the name. But if you find yourself, in the not-so-distant future, adding new-and-improved solar panels to your roof, they may well contain a…
Getting back to science, at least for the moment, I was puzzled by a press release from RPI, with the eye-catching headline Solar power game-changer: 'Near perfect' absorption of sunlight, from all angles. The article describes work published in Optics Letters (that I haven't been able to put my…

And, you have to find areas with sufficient sunlight to make installation worthwhile.

When subjects like this come up, I usually point out the tremendous damage caused by RR in this area, but you saved me the trouble. Imagine where we might be today if we had devoted enough time and effort into solving our energy supply problems for the last almost 30 years.

I'm not especially impressed by the plastic-wrap analogy. If plastic wrap could generate free energy, we could easily produce 10, 100, or 1,000 times more plastic wrap. We don't produce more plastic wrap because we don't need more plastic wrap, not because we don't have the industrial capacity.

"solar is not an overnight solution"

*ba-dum-ding!*

And, you have to find areas with sufficient sunlight to make installation worthwhile.

I forgot to note that the figure he was using supposedly took weather and general illumination issues into account. He didn't go into any great detail about how that number was arrive at, though.

When he mentioned that 0.2% is about the same as the area of paved roads, my first thought was "roofed highways!" That's probably not a sensible solution, but it's an amusing science-fictional sort of image.

I'm not especially impressed by the plastic-wrap analogy. If plastic wrap could generate free energy, we could easily produce 10, 100, or 1,000 times more plastic wrap.

Sure, but solar cells are at least 1,000 times more difficult to manufacture than plastic wrap. Plastic wrap is really easy to make, but solar cells are multilayer semiconductor materials, and you don't just rolls those out in sheets.

A better comparison might be to the total area of silicon chips produced in a year, but I don't know that figure. In any case, the point is that manufacturing is an often-overlooked bottleneck in this process.

At least in Desert areas, Solar thermal approaches may turn out easier, and they have more ability to load-follow and manage intermittancy (Not perfect, but a lot better than PV). Plus they should be easier to make; most of the area required consists of mirrors which are fairly easy to make.

Now, if they can produce PV systems on a roll via a continuous process.. and if this roll is robust enough to cover my south facing roof fairly easily, that's something useful..

By Andrew Dodds (not verified) on 18 Oct 2007 #permalink

It's been a long time, but I seem to recall the industrial processes required to make solar cells as being EXTREMELY nasty, and unless efficiencies have gone way, way up, you still don't reclaim the energy required to create the cell over the life of the cell.

Fuck solar. I want more nuke plants.

I think a good plan is to use modern-technology nuke plants to bide us time while we ramp up solar, wind, and other real renewables. But I still don't think the plastic wrap number should scare us. In 1908, someone was probably saying, "hey, we could make a car for every last person in the U.S." and the response was, "psch, we'd have to increase our steel production tenfold". The important number is cost. I've heard - and I don't know how accurate this is, that consistently the cost of solar drops about 20% for each double of solar capacity. If that's true, then solar that costs not all that much more than coal is in sight. And at that point, it's just a matter of putting the things on our roofs, which I'm sure cover .2% of the land area. Or just cover Nevada, it's a shithole anyways.

Regarding producing large areas of photovoltaic material, this is why I'm convinced that if there is a practical solution it lies in printing. Each sheet of paper in the New York Times is around 1 m^2, and there are about 50 sheets of paper in the daily times. Daily circulation is over 10^6, so we're talking 5x10^7 m^2 of printed material produced every day by just one newspaper. This fellow claimed we'd need 40000 km^2 (which is much higher than other numbers I've seen) of solar cells, or 4x10^10 m^2. We could do this in a little over 2 years if the material could be printed at the rate that the NY Times prints newspapers.

I was going to argue that Si isn't viable because it takes so much energy to purify it, but maybe I was wrong. Take Chad's number of 2,000 km^2 of cells a year, and just consider the silicon substrate. Say you manage to make very thin wafers, about 0.1mm. That's 2e5 m^3 of Si per year. Using the density of Si (2330 kg/m^3), and the fact that Czochralski growth of Si crystals takes about 35 kWh/kg, producing all that Si would take 16 GWh of power a year. Considering that the IAEA website says that the US used 4000 TWh in 2005, maybe that's not such a huge number.

I'd love to see organic photovoltaics take off, since there's great hope for some of the things mentioned in the above comments -- cheap roll to roll processing on plastic substrates, patterning devices by a printing process (people have done a lot with inkjet-type patterning, and I'd imagine a silkscreen-type process would be possible).

But, there are still a few huge hurdles for organic PVs that make them seem like a far-off solution. Efficiencies are still down in the 3-5% range, and the organic molecules being used are fairly reactive, and hence unstable in air. Even if good packaging serves to protect the device, I think all the UV hitting a solar cell is going to break bonds and kill the materials.

Chad,

If he did not ultimately talk about cents per watt (including both production and maintenance costs) then he did not take the analysis far enough to be of any practical use. As Dave notes, comparing the quantity required to another quantity of something completely different that we already produce is meaningless.

The serious report to read here (in everyone's copious free time-- I actually did skim the whole thing a few years ago) is here. Warning, it's a massive PDF file, from 2005:

http://www.sc.doe.gov/bes/reports/files/SEU_rpt.pdf

Note also that there will be no complete replacement of US energy needs with solar power until we figure out how to run cars and transport industries on solar power. (If you sufficiently redefine solar power to include bio fuels, then you can do that, but you're not talking about solar panels, you're talking about solar pond scum....)

However, if you can get 10% of the US, India and China on solar power, and another 10% on geothermal, and a third 10% on nuclear, which are all achievable targets over a 15-20 year timespan, that's a lot of stress relief on the oil industry.

By John Novak (not verified) on 18 Oct 2007 #permalink

Increasing the efficiency isn't the solution for most users. There are only few places where peak consumption coincides with peak sunshine. They tend to be where air conditioning is a major user.

Here at 60 degrees North we have abundant sunshine in the summer, but very little in the winter (that's why we have summer and winter...). Over here solar power will remain marginal as long as we have no way of time-shifting energy from summer to winter. There have been some small scale experiments in storing energy for months without significant losses, e.g. with Glauber salts, but they haven't been commercial successes.

So far the only energy storage that really works in big installations is damming a river - if you accept rains as an indirect form of solar power.

By Lassi Hippeläinen (not verified) on 18 Oct 2007 #permalink

Regarding Lassi's comment, one of the key points in any alternative energy use is finding the right application for any given energy source. For the far north or south, that might mean using passive solar for water heating in the summer, for example. It won't replace other energy sources, but it will reduce their consumption, and every little bit can help.

Well for one thing, paved roads are made out of roughly the same material as plastic wrap, and they tend to be a whole lot thicker. Without even stretching the comparison to asphalt, we use a lot of plastic in this country. And certainly there are innumerable wasted rooftops.

Second, for a single-layer solar cell, teh theoretial efficiency's considerably less than 50% (more like thirty and change), which can be beat by using multi-junction cells, or, at this point theoretically, by using quantum dots to generate multiple electron-hole pairs per photon. There's other stuff (wikipedia, sorry) that uses "quantum dots" which promises to be cheap anyway, and may well be. Certainly interesting.

Regarding sustainability, in addition to Novak's source (what a crowd, holy takeameback, Batman) this guy throws some interesting numbers together.

PV is limited but very useful for small-scale installations like houses or what not.

For commercial power generation, we will have to use solar hot water turbines. Those scale up just fine, and test plants are already working.

If we assume that manufacturing processes can be scaled in a big way without adversely affecting the price, then it is simply a matter of getting the price competitive. There are several Silicon Valley startups with that written into their business plans. If any of these are successful than rapid expansion of the industrial capacity is likeley. Most promising are BIPV Building Integrated PhotoVoltaics, because cost of mountint etc can exceed cost of cells, and concentrating photvoltaics. Some company names to checkout:
HelioVolt, SolFocus, and ASURA. None of these are using the traditional silicon crystal approach.

I'm not terribly worried about solar being a poor fit for high latitudes. The percentage of world population at such latitudes is pretty low. At such low population densities other approaches, such as hydro should aften be adaquate.

Does the 2,000 square kilometers of solar cells number take into account concentrated cells? A Spectrolab press release from 2001 (6 years ago!) sites 34% efficiency under a concentration of 400 suns. If the 2,000 km^2 refers to the amount of sunlight needed, then we can collect that much light and focus it down to 5 square kilometers of solar cells. Then, we just need to produce massive amounts of mirrors and lenses, and a lot of windex to clean them all.

http://www.spectrolab.com/com/news/news-detail.asp?id=137

If he did not ultimately talk about cents per watt (including both production and maintenance costs) then he did not take the analysis far enough to be of any practical use. As Dave notes, comparing the quantity required to another quantity of something completely different that we already produce is meaningless.

The focus of the actual talk was on the production of solar cells, and the analysis mentioned in this post was only included as background to the larger problem. He also mentioned that back when he used to work for a major energy company in the late 70's, they did a cost analysis and showed that even if the cost of the solar cells was completely negligible, they needed an efficiency of at least 15% in order to break even, after figuring in all the other infrastructure needed to use photovoltaics for energy generation. Obviously, that's almost thirty years out of date, which is why I didn't quote it, but it gives you another angle on the problem.

The plastic film comparison is useful, I think, not as an exact analogy, but because it gives you a sense of the scale. Plastic film is both fairly trivial to make and nearly ubiquitous (seriously, think about how many plastic bags and plastic-wrapped items you see in a day), and we're talking about increasing the production rate by an order of magnitude from that, for something vastly more complicated to make.

It's not impossible, but it's not trivial by any stretch.

Second, for a single-layer solar cell, teh theoretial efficiency's considerably less than 50% (more like thirty and change),

He wasn't saying that we had the ability to make those in any particular way, just that that's the efficiency you would need in order to get the necessary amount of electricity from tiling 0.2% of the land area with photovoltaics. If you want to use simpler panels, with lower efficiency, you need more of them.

He also talked about how an early estimate of the maximum possible efficiency was wrong, but I don't recall whether it was an estimate of about 30% that moved up to about 60%, or about 20% that moved up to about 40%. Anyway, there's some relatively recent calculation that shows a higher maximum than previously believed.

Anyway, I'm all in favor of putting solar panels on rooftops and that sort of thing, to reduce the need for electricity from fossil fuels. The point of his intro, and the point of this post, is that it's a mistake to think of solar power as a rapid cure for our energy woes. It'll help, but we're not going to be fully solar any time soon, and we really need to be thinking about a wide range of options.

I see. Your speaker Persans is probably citing the same numbers as two of the authors I cited. (A necessary 50% is what they estimated or stated as well.)

Also noting, teh longstanding 31% max efficiency is for single band-gap absorber drinking in the solar spectrum (it just so happens that silicon is peachy), but if you have more than one absorber in there, you can do better. That's how spectrolab has beaten it. (There's a lot of engineering involved, and I don't follow it very closely.)

Your speaker is probably using the theoretical numbers from quantum dot photovoltaics--quite likely if he's citing the same sources they are. They're claiming up to 60-something percent for that, I think, but not demonstrated so far.

A place I recently interviewed at told me they were manufacturing modules at something like $2/W, using fairly conventional technologies, but cleverly. The silicon supply is an issue.

This is the kind of thing I would like to see more information on. People in general seem to be excited about "renewable energy" as some kind of platonic ideal, but are much less interested in the somewhat difficult question of what happens when it makes contact with reality. People get excited when wind/solar installations go up, but they seem excited just on the basis of the fact that such installations are existing, they don't seem as interested in the question of "okay, but is it, or can it ever be, producing enough energy to make a dent in our carbon emissions?".

(I mean, it's not like renewable sources should be expected to solve the energy crisis right away-- r&d takes time, of course-- but it seems like if we don't at least ask the hard questions of whether alt-energy is putting us on the path to energy sustainability, then alt-energy will never get there!)

It's a bit relieving to see the kind of analysis Chad and James are reporting here being done-- not dismissing the new technology entirely out of hand, but still founded in reality rather than hope. I would be particularly curious to see how modern wind power tech stands up in this sort of "okay, but what if we want it to be MORE than a novelty?" analysis. Better, I take it?

Does anyone know, do there exist any blogs devoted to new developments in alternative energy technologies? That seems to be one of only a couple of science subjects that I've never been able to find a blog specifically covering, it seems like it would make sense for one to exist...

I'm not sure why "meet the energy needs of the US" is the metric here - we don't currently rely on a sole energy source, and I don't think anyone's proposing running the world on solar. (Because, as Novak says, cars.) Anyway, a solar cell is no more electrically complicated than an LED, just larger, and there's a great deal of materials research going to increasing the efficency with multi-layer devices and the like. I wouldn't put 50% efficiency at the hard and fast maximum.

(Oh, and you don't need to put quantum dots in quotes. It's not a dumbed-down description, it's what we actually call the things, just like quantum wells and quantum wires.)

Thanks for the post, Chad, and helpful follow-ups in the comments. I wanted to echo Craig's point (just above this one) that I don't believe the issue is that we should seek to replace our entire energy demands with solar. Most renewables folks will speak about how to promote more solar capacity and integrate that into a range of approaches.

I imagine the speaker you refer to noted that as well, so it isn't a new point here. My worry is about a slip from one argument to a second one and the style of thinking involved: that defining an approach to solving the technical problems of PV leads one to a conversation about whether or not PV can solve *all* of our energy needs. It then shifts easily into a new argument, which is to promote debate on just that, on whether or not it is the cure all. And since it is not a cure all (the second debate), then losing that corollary debate leads observers to think that the primary debate (about pursuing technical research and performance) is over too.

Where solar may make a real contribution is getting small shit like mobile phones and small scale consumer electronics *off* the mains grid, which would a/ remove the distribution load, b/ build the cost of the PVs into the manufacture costs of devices, c/ reduce the burden of producing chargers in the first place, d/ reduce the number of sockets required, etc etc. There are stacks of things that could be disconnected from batteries, mains, chargers, etc, with no loss of convenience and potentially significant energy savings in the round.

And you kids in the states need to learn to wean yourselves off air con! ;-)

By Amenhotep (not verified) on 19 Oct 2007 #permalink

Well, one of the things we should be considering is how not to use as much electricity as we do today. As one who is almost totally dependant on solar for all my electrical needs for a good chunk of the year, what is clear is that with a few solar panels you can supply all lighting you need and power to run things like computers and TVs if they are LCD not tube. With a few more panels you can also run a medium-seze fridge of freezer.
It gets harder if it comes to heat in things like room heaters and spin driers. As for the question about covering so much area in panels, try another calculation; suppose every roof we have today was covered in solar panels, how much electricity would we generate? I think you would be amazed.

I'm not sure why "meet the energy needs of the US" is the metric here - we don't currently rely on a sole energy source, and I don't think anyone's proposing running the world on solar.

I think it's a good metric because it provides a convenient upper bound which is easy to scale against. 100% of our energy needs is a pointless goal, but what if we wanted to see how feasible it is to, say, replace Coal with solar. Well, Coal provides for 23% of our energy consumption, so multiply all of Chad's numbers by .23 and you've got some idea.

Not that there's even any reason to demand that we be able to replace our coal consumption with solar! But surely there's some minimum contribution to expect out of solar if we're going to consider it important at all. I mean, if Solar [at least in some credibly feasible future version of the technology] can't provide enough power to cover, say, one percent of our energy needs, then I have trouble seeing what our electricity "approach" gains by having solar "integrated". Scaling Chad's numbers, just to get that 10% coverage apparently we'd have to take 20 years to do it and produce 2,000 square km of solar panels per year. Is that feasible, is it worth it? I have no idea, it sounds like a lot to me though. Following sailor's comment, one wonders-- how many square km of new roofing tile are installed in the U.S. each year?

Where solar may make a real contribution is getting small shit like mobile phones and small scale consumer electronics *off* the mains grid...

As one who is almost totally dependant on solar for all my electrical needs for a good chunk of the year, what is clear is that with a few solar panels you can supply all lighting you need and power to run things like computers and TVs if they are LCD not tube

This is something I am curious about. If we install solar panels on buildings which power the building on which they are installed, does the amount of energy produced by these panels get an effective boost because the transmission losses which would be present with any other energy source are eliminated? Does anyone know, how sizable is that boost?

Coin: People get excited when wind/solar installations go up, but they seem excited just on the basis of the fact that such installations are existing, they don't seem as interested in the question of "okay, but is it, or can it ever be, producing enough energy to make a dent in our carbon emissions?".

Yeah, that is definitely a problem. I've spoken on more than one occasion with very well-intentioned people, people I know from long acquaintance to be anywhere from not stupid to not fucking stupid and a lot of times it boils down to intentions over economics. That, and a blank, solid disbelief that current renewable technologies arent' enough, probably because most people either cannot or have not fully absorbed the sheer bloody magnitude of the North American energy infrastructure.

"Well, I just gotta believe that if we do this and this and this, ten times more than we are today, it's gotta make a difference!" And they're talking about 50 MW wind farms or 100 MW (physically ginormous) solar arrays or whatever. Um. Yeah. Fifteen of each of those combined is a midscale modern power generation system. One. Of which we already have hundreds, many of them two ro three times as big.

We use a staggering amount of power in North America and Europe.

My best hopes for the near term (which I think of as 10-25 years) is fission and biofuels of one stripe or another. I class most biofuels as solar, since we're really just using plants or microbes to process sunlight for us.

By John Novak (not verified) on 19 Oct 2007 #permalink

I class most biofuels as solar, since we're really just using plants or microbes to process sunlight for us.

There's an informative (though unfortunately somewhat one-track) blog called The Oil Drum, which I've seen occasionally express the opinion that there are only four sources of energy: Fusion; Solar; Fossil Solar; and Fossil Supernova.

They class "Solar" as including not just direct solar, but also biofuels, wind, and hydroelectric; they class "fossil solar" as comprising all fossil fuels (since all that stored energy eventually stemmed from photosynthesis anyway); and they class "fossil supernova" as comprising fission and geothermal (they also put tidal in this category, although really shouldn't that be a fifth category, "gravitational energy" or something?).

It's one way of looking at things, I suppose...

Several points:

1. solar cells typically achieve energy pay-back in about 18 months (out of a life-span of 8 + years). The idea that they use more energy than they produce is an urban myth that just won't die.

2. There's no reason to assume that we have to dedicate areas solely to solar panels - building-integrated panels reuqire no additional space. They also replace structural elements like shingles.

3. Current systems waste a great deal of power in transmission (up to 90%). By replacing power-stations with solar panels at the point of demand we avoid that waste - so we don't need to replace the entire current powerplant capacity even if we do want to go 100% solar.

4. In practice we are goign ot need a mix of all the available options - starting with energy efficiency and including solar, wind, biomass and nuclear.

5. speaking of nuclear, the figures on completely replacing the current fossil fuel powerplants with nuclear reactors in a 20 year period are every bit as daunting as for solar. After 50 years we've managed to build a total of around 450 nuke plants which provide around 6% of total world energy demand (and running cars on nuclear power is jsut as hard as doing so on solar. In fact its more difficult - we have prototype solar cars of a sort.) so in the next twenty years to go totally nuclear we'd need to build something like 8000 reactors - more than one a day. That's assuming no increase in demand over the 20 year period.

I'm also curious as to how people think nuclear plants are going to fare in countries like Congo, Afghanistan and Somalia (but I'm sure Iran and North Korea will back the idea enthusiastically.)

6. Scinetists tell us we need to reduce carbon dioxide emissiosn by around 70-80%. There are big differences in the efficiency of different fossil fuel plants. If we upgrade or close down the most inefficient power plants and replace them with current best commercial practice (not hypothetical clean coal technology), we can probably produce around 50% of world energy needs from coal indefinitely while still achieving the 70-80% redcution in emissions. That's before we consider sequestration.

7. If we set the price of emitting greenhouse gases at the correct level and get rid of market imperfections, the market will find the most efficient way to minimise emissions while maintaining living standards.

By Ian Gould (not verified) on 19 Oct 2007 #permalink

It´s a pitty most of the comments are US-centric.
Important advances in R&D&I are being carried out in Japan, Germany and even Spain, now a world class power in Eolics.
While you continue arguing "if" whatever, other important developed countries are going full ahead non-stop.
Europe and Japan for sure have taken the lead in renewables and it will take a major effort by the US to catch up.
Wish you do.

Ian Gould has summed up what I would have wanted to say.

So, some observations from my job. I work at a company that makes carbon bonded carbon fibre for high temperature furnace applications (Up to 3,000C in Argon). Our major and biggest growing market is the solar cell silicon market. Many of our customers are projecting solid growth for the next year to 18 months, and are adding crystal pullers as fast as they can. They think there is likely good growth for the next 3 or 4 years, but it is obviously risky and difficult to forecast that far ahead.
Also, one of our customers does not use the standard Czochralski process, rather they pull out sheets of silicon, in a very cunning way.

It might surprise Mike (#29) to know that the latest installed base figures for wind power ("eolics") in Spain and the United States was almost identical, at some 12.8 and 12.6 GW, respectively. Only Germany produces more. I don't think the US is exactly slouching, here, Ted Kennedy's efforts notwithstanding.

On the other hand, wind power has the same problem as many other renewable sources-- at a baseline of 10 MW/square kilometer for wind power, that's something like 2500 square kilometers of wind turbines for those two countries and almost as much in Germany. Now granted, you can do other things with the land, and the theoretical (even the practical theoretical) amount of extractable energy from wind power is very large, that's an awful lot of turbines on an awful lot of land.

By John Novak (not verified) on 20 Oct 2007 #permalink

"I don't think the US is exactly slouching, here, Ted Kennedy's efforts notwithstanding."

Ted Kennedy's efforts at slouching?

Chad:

For another perspective on the area of PV cells issue, I looked at the problem of PV power last January, and found (follow this link for calculations) that about 30 square meters of PV roofing of 25% efficiency for each person's share of roof would meet the current total power grid energy requirements of the USA. That is about 10,000 square kilometers, which is 0.1% of the US land area. Persan used about 50% efficiency to get 0.2% so our figures are at least in the same ball park.

The 1000lb gorilla that everybody is avoiding is the energy storage problem. A good storage system would allow households to sell power to the grid. A REALLY good storage system (e.g. a battery that stored around 10 MJ/Kg) might make roof PV power suitable for use in motor vehicles.

Ian, a few questions in response to your last post:

3. Current systems waste a great deal of power in transmission (up to 90%). By replacing power-stations with solar panels at the point of demand we avoid that waste - so we don't need to replace the entire current powerplant capacity even if we do want to go 100% solar.

Okay, that's quite large. So should we assume then, when looking at figures like Chad's at the top of this page, that the effective power output of solar is effectively doubled if we use the energy at the place where it is generated rather than placing big solar "power plants" off somewhere?

the figures on completely replacing the current fossil fuel powerplants with nuclear reactors in a 20 year period are every bit as daunting as for solar... After 50 years we've managed to build a total of around 450 nuke plants which provide around 6% of total world energy demand.

Hm. I think this is sort of a trick with numbers. The 6% you give us is for the world, and as I understand things that 6% "of total world" energy mostly comes from a small number of countries using a large amount of nuclear power and the rest using none. Of course this isn't going to average out to much.

Let's say we're only interested in, say, the United States (for continuity with the Solar numbers above), and we're not interested in displacing all energy production, just "some reasonable portion" of our energy production. Google seems to be telling me that 7% of American energy comes from nuclear, and 21% comes from coal. This seems to be telling me that if we set the more or less arbitrary goal of replacing all coal usage with nuclear, we don't have to do anything more than quadruple the amount of nuclear power we're producing. I'm sure that's not easy, but considering the U.S. hasn't built any new nuclear plants in what, thirty years... whatever the upper limit is on the speed with which we have the ability to grow U.S. nuclear capacity we're surely not exercising it.

I realize nuclear fission has nasty upper limits on what we can do with it-- it is, after all, based on a nonrenewable resource, uranium-- but in a straight comparison to most of the alternative energy technologies out here as a way of offsetting our carbon usage in a big way, it seems to me reasonable to class nuclear as being so much more feasible than the others as to not be even in the same class. When one option, if you want that option to make a major impact, involves scaling up what we're doing by a factor of four at a couple hundred sites with more or less established technology, while the other option involves scaling up what we're doing by a factor of a couple hundred with the "sites" being an entire nation's worth of roofing tiles and the technology being something that's not quite developed yet... I'm not trying to say pumping up our nuclear use would be easy, but "as daunting as solar" seems like hyperbole unless we set an unrealistically high bar for how much nuclear we want...

Anyway, my question: there clearly are some limits of the kind you are focusing on, on how much nuclear power we can produce and how quickly we can build up production capacity. As you see it, how do those limits change if we consider it a priority to reprocess nuclear fuel, or use "breeder" reactors or whatever? Do the limits become less daunting, or do you think the picture is just as bad?

6. Scinetists tell us we need to reduce carbon dioxide emissiosn by around 70-80%. There are big differences in the efficiency of different fossil fuel plants. If we upgrade or close down the most inefficient power plants and replace them with current best commercial practice (not hypothetical clean coal technology), we can probably produce around 50% of world energy needs from coal indefinitely while still achieving the 70-80% redcution in emissions. That's before we consider sequestration.

Something I keep hearing is that part of the problem here is that our air pollution laws are written to put very strict requirements on new power plants, while basically leaving old power plants "grandfathered" and exempt from anything. This, I'm told, creates a situation where coal plant owners basically just stopped building new, more efficient plants, and instead are just using the old plants-- going to great lengths to keep them in operation long past the date they might have lasted otherwise-- because the margins are better if you hang on to a grandfathered plant forever than if you have to start following air pollution laws.

How accurate is this?

Thanks!

"It might surprise Mike (#29) to know that the latest installed base figures for wind power ("eolics") in Spain and the United States was almost identical, at some 12.8 and 12.6 GW, respectively. Only Germany produces more. I don't think the US is exactly slouching, here, Ted Kennedy's efforts notwithstanding."

Spain's population is about 1/6th that of the US; Germany's is a bit more than 1/4 (IIRC), so in per capita, the US is indeed lagging.

When you take into account lower per capita energy consumption in Europe, the difference would be even greater: Germany and Spain get a much larger proportion of their energy from wind.

I agree about the problems with wind power. Current turbine technology is clearly reaching its limits. There are a couple of new technologies under development which may have a big impact in future.

One is the various designs to tap into the jet stream using kites or balloons. You reduce the footprint on the land and the wind at those higher altitudes is more reliable and has a lot more power.

The other is the small vertical-axis turbines for urban applications which can be cited on roadsides, besides railway tracks or on rooftops.

By Ian Gould (not verified) on 20 Oct 2007 #permalink

Hi Coin,

Yes, it's probably reasonable to assume that 1 unit of solar power at the point of use will displace, say, 2 units of power dispatched from a conventional powerplant. That's while it's working of course.

"Let's say we're only interested in, say, the United States (for continuity with the Solar numbers above), and we're not interested in displacing all energy production, just "some reasonable portion" of our energy production. Google seems to be telling me that 7% of American energy comes from nuclear, and 21% comes from coal. This seems to be telling me that if we set the more or less arbitrary goal of replacing all coal usage with nuclear, we don't have to do anything more than quadruple the amount of nuclear power we're producing. I'm sure that's not easy, but considering the U.S. hasn't built any new nuclear plants in what, thirty years... whatever the upper limit is on the speed with which we have the ability to grow U.S. nuclear capacity we're surely not exercising it."

A couple of points - most of the current US nuclear fleet is 30+ years old and will need to be replaced over that twenty year period as well.

The big bottleneck is likely to be getting adequate numbers of qualified nuclear engineers to design and build the new plants.

If the US is going to expand nuclear power significantly, you'll probably have to follow the French model - standardise on a couple of reactor designs and build really big multiple-reactor complexes. That minimises the design work and the supporting infrastructure needs.

"Anyway, my question: there clearly are some limits of the kind you are focusing on, on how much nuclear power we can produce and how quickly we can build up production capacity. As you see it, how do those limits change if we consider it a priority to reprocess nuclear fuel, or use "breeder" reactors or whatever? Do the limits become less daunting, or do you think the picture is just as bad?"

If anything breeders make things worse.

The first generation of commercial reactors had serious economic problems - they were expensive to build and had a lot of downtime.

The economics of nuclear power have improved dramatically since then for a couple of reasons:

1. many of the original operators went bankrupt and subsequent buyers bought the reactors below replacement cost;

2. as people became more experienced at running reactors, down-time declined dramatically.

Current breeder designs are much more complex than conventional reactors. Most designs use liquid sodium as the cooling liquid because water can't transport away the amount of heat generated. If the sodium is allowed to cool, it solidifies and then the plutonium overheats. The French Phenix prototype used so much energy keeping its sodium liquid while the core was offline that it was a net energy consumer.

You also need to transfer the heat energy from the liquid sodium to a fluid you can put through a turbine - typically water. Liquid Sodium explodes on contact with water so the heat exchanger has to be made to very high tolerances.

Shifting to breeders would probably put the nuclear industry back to where it was in the 70's with huge debt and balky reactors that work less than 50% of the time.

By Ian Gould (not verified) on 20 Oct 2007 #permalink

Coin:

"Something I keep hearing is that part of the problem here is that our air pollution laws are written to put very strict requirements on new power plants, while basically leaving old power plants "grandfathered" and exempt from anything. This, I'm told, creates a situation where coal plant owners basically just stopped building new, more efficient plants, and instead are just using the old plants-- going to great lengths to keep them in operation long past the date they might have lasted otherwise-- because the margins are better if you hang on to a grandfathered plant forever than if you have to start following air pollution laws.

How accurate is this?"

Reasonably accurate.

As a non-American I try to avoid party political comments about your domestic politics but I will note that the Bush administration has greatly expanded the extent to which old plants can be expanded without losing their grandfathered status.

If you adopt some form of carbon taxation or sale of carbon permits, a logical way to use any money raised would be to allow utilities to write-off the value of their older plants more quickly.

There are alos plenty of other reasons why there's been little new coal-powered capacity added in the US.

In the 80's and 90's it was simply cheaper to build gas-fired plants.

There's also the uncertainty over possible future regulation of carbon dioxide emissions - that's why even many utilities and big energy consumers now want the US to say how it will regulate CO2 emissions. They may not like the idea of regulation or the associated cost but they think its now probably inevitable and woudl like to know what's goign to happen.

By Ian Gould (not verified) on 20 Oct 2007 #permalink

Ian: Excellent comments in general, and you more than pre-empted most of what I would have said about transmission losses and such.

The problem I have regarding nuclear power is one that may be peculiar to the US -- given the known entrenchment of corruption and racketeering in our construction industry, I doubt our ability to build and run safe plants.

Note that I'm not saying that it's generally impossible to build safe nuclear plants -- but to do so, you really need to be able to trust that things will get built and maintained as designed. Then they need to be subjected to honest monitoring and inspections, backed by the power to shut down plants if needed.

Based on the scandals that have repeatedly popped up in our nuclear plants (substandard materials, paid-off or lazy inspections, etc), and which have been implicated in at least some "incidents" (IIRC, including the Three Mile Island meltdown), I don't think we have the political will to push through projects on that scale, without them getting parasitized by organized crime. And frankly, we got off lucky at Three Mile Island!

By David Harmon (not verified) on 20 Oct 2007 #permalink

Where is the 200 km^2/year number coming from? It seems very low: If it is true, that would mean that the per capita production is less than 1 m^2/year.

Valid point on the 100% solution solar can provide, but the economics of solar power still make sense for a lot of consumers (residential/commercial). We are definitely at the beginning of a big solar boom. Consolidation will surely follow suit. I wrote more about this in this article: "Solar Energy Consolidation Outlook".

I comment regularly on the business/investor side of alternative energy on Energy Spin: Alternative Energy Blog for Investors-Served Daily
Cheers,
Francesco DeParis

I wrote about your entry on my own blog and tried to send a trackback, but your server keeps returning me an error. Just thought you'd like to know that trackbacks don't seem to be working on your blog at the moment. (Specifically, the error is: 'HTTP error: 412 Precondition Failed')