Does solar generate more waste than nuclear? No.

There can be only one!

As I noted in an earlier post, one of the things you often come across in the energy blogging world is that supporters of one technology attack others.

This example takes the cake. It’s a somewhat old article (archived here) that claims to demonstrate that solar power generates more waste than nuclear.

Update: a reader pointed out a rather obvious error, and when I checked against my original calculations I found a whole section was missing. Both fixed.

The argument

The article starts off with this question:

But what of the volume of waste a solar powered world would create?

The article goes on to suggest that if the world were to switch to solar power, it would generate many, many times as much waste as if the world switched to nuclear power instead. It uses some basic math to support this claim.

But when you look at the numbers, even a tiny little bit, it all just falls apart. Let’s take this bit by bit…

Wow, that’s wrong

The very next paragraph in the article says:

Research on electricity generation and land use (paper, referenced here previously), estimates that solar power to produce 100% of world electricity generation would require 5,500,000,000,000 square meters (5.5 x 10^12 m^2).

The number in question can be found on page 96 of the linked paper, which states:

Some 5.5 million km2 of land area would be needed to supply 100 percent of current energy demand with solar photovoltaics.

Did you catch it?

Look carefully… notice how the first statement says you need 5.5 to “produce 100% of world electricity“, while the second one says “supply 100 percent of current energy“?

That’s right, the number they are quoting is not for electricity, but everything – transportation, cooking, heating, everything. What’s the difference? Well according to the EIA, global energy use in 2010 was 480 quadrillion BTU, or about 140 billion kWh, and electricity use was 17,388 billion kWh. So about 15% of all energy used is electricity.

So if you want the number the article claims to be presenting, you have to adjust: 17.4 / 140 x 5.5 million ~= 0.68 million km²

Now everyone makes mistakes, but we’re in the very first paragraph of the body, and we’re already off by an order of magnitude.

Numbers

Now let’s examine that area number and run it ourselves.

Using PVWatts on Scottsdale with a 20 degree tilt you get 1,979 kWh/kWp. If you think that’s unfair, Calgary is 1,650, and Toronto is 1,400, and if you think you get less sun than Toronto you should move here. Using a derate of 10%, we’ll call it 1800.

A typical modern panel is 270 to 280 watts and 1.6 square meters, so that’s (275 Wp / 1 kWp) / 1.6 m^2 * 1800 kWh = 310 kWh/m^2. So to produce 17,388,000,000 kWh with 310 kWh we have 17,388,000,000 / 310 = ~56,000,000,000 square meters, or 56,000 square kilometers.

They say 5.5 million, so now we’re off by two orders of magnitude.

By the way, I’ll take this opportunity to point to the wonderful Landart calculation from some time ago, which comes up with 496,805 square kilometers for all energy needed in 2030, which assumed a whole lot of new energy demand, on the order of 44% over 2008.

So if we take that number and assume 15% of that is electricity, that gives us ~75,000 square km. So we’re definitely both in the same ballpark in spite of using very different methods to calculate the area requirements. Both calculations suggest the original authors are wrong by about 100 times.

Have you actually seen a solar panel?

But what’s a couple of orders of magnitude between friends? An honest mistake, surely. So let’s forget that and move on. After a bit of handwaving we come to the meat of the argument:

Assuming that the land use is the approximate area of the solar panels, and that the panels are a minimum of 1 inch thick (2.54 cm), the volume of PV panels as waste would be 139,700,000,000 cubic meters.

A minimum of an inch thick? Has this person ever even seen a solar panel before?

A panel is about 1/8 of an inch thick. Most of them have a frame around the outside, but that isn’t solid. So it’s not 5.5 million times 2.56 cm, it’s 65,000,000,000 times about 0.3 cm. And that means it’s not 139,700,000,000 cubic meters, it’s 195,000,000 cubic meters.

We’re off by three orders of magnitude, and I’m only on the third paragraph!

Mistake? What mistake?

Ok, let’s look at the very next paragraph. They go to the same source I used for electrical use, the EIA, and then scale US nuclear electrical delivery to worldwide numbers:

Used nuclear fuel produced by all US reactors is 2,000 metric tons annually.  This is about 100 cubic meters per year.  Multiplying the amount of US nuclear electricity production to produce the same amount as worldwide electricity production (data from EIA), this becomes about 37,000 cubic meters worldwide per year.  After the same 30 years, a nuclear powered world would produce 2,220,000 cubic meters of used fuel.

Now you see the problem here, right? They’re comparing the volume of the fuel to the power plant itself. An apples-to-apples comparison would compare the fuel waste in both cases, in which case the solar waste is exactly zero. Or, more usefully, one can compare the total waste in both cases, the fuel and the power plant together. Let’s do that.

There are about 100 reactors running in the US right now. Most of these are in containment domes, but not all of them. For argument’s sake let’s pretend they all are in a dome and then apply a scaling factor (pick one, it makes no difference in the end, as you’ll see). On average, a typical containment dome is about 40 to 50 meters across, 60 to 70 high, and 3 to 4 meters thick. Taking the averages, that’s 45 x 65 x 3.5 ~= 10,500 cubic meters of concrete per dome.

Now that’s just one dome, providing some tiny amount of the world’s power. To be a bit more precise, 100 of those provides 20% of the US’s electricity and using the same numbers they do from the EIA, we’d need 370 times that to produce all the electricity in the world. So that’s 10,500 times 100 domes in the US times 370 = 388,500,000 cubic meters. And don’t forget the fuel waste on top of that, another 2,220,000, so we’re into the 400 million cubic meters range.

Well, there you have it. According to their own definition of “good”, solar is twice as good as nuclear, and I’m only on the 4th paragraph. Which brings us to their conclusion in paragraph number five:

This means that a solar powered world produces 63,000 times the waste of a nuclear powered world. 

Yeah, right.

Reduce, reuse, recycle!

Did I say half the waste? Let’s take a look at paragraph number six:

And used nuclear fuel actually still contains 95% of its energy- it is not just “waste”.  On top of this, radioactivity decreases with time, whereas solar panel components including cadmium and lead, do not go away.  Yes, there is the potential for both to be recycled.  There would be significant energy lost in recycling solar panels versus the energy produced from recycling nuclear fuel in fast reactors.

So according to them, recycling a solar panel is bad, but recycling a nuclear reactor is good. Oh really?

By weight, a solar panel is mostly glass – both the sheet of glass on the front and the solar cells themselves, which are basically processed glass. Next is aluminum, both on the backs of the cells as a conductor and wiring, and the frame around the panel’s perimeter. Then there’s a bit of silver on the front, those thin lines you see if you look close. On the back is a sheet of PET plastic and some glue. Finally, some copper wiring in a plastic sheath.

Now, look at that list… PET is the most recycled plastic in the world. Copper is very highly recycled. Glass and aluminum are also highly recycled, about half of it. If people stopped throwing their bottles and cans in the bush or garbage because they’re lazy, we’d have the same sorts of recycling rates as cars, 90% or more.

Recycling these materials costs 1/2 the energy of making it new. A solar panel produces about ten times as much energy as it takes to build. So if we build panels from recycled panels, they’ll produce 20 times as much energy as it used to build it.

So what we really need is some sort of recycling program that ensures these things actually do get recycled. You know, like this one. Or this one. Or practically any industrial electronics recycler, any of whom will come and pick them up for free. Why? Because it’s glass, aluminum, copper and silver. That old panel is a veritable goldmine.

What about nuclear? Not so much, First off, consider the fuel. As the quote notes, most of the fuel in nuclear fuel is still there when you pull it from the reactor. You can reprocess this once or twice to re-concentrate the good parts and put it back into the reactor, which reduces the waste by that amount. But this is actually very rare, reprocessing is expensive. Or you can, as the paragraph notes, use it in a fast reactor and breed more fuel. But that’s way too expensive, everyone gave up on that approach after Superphénix.

And then there are the reactors themselves. The problem here is that you can only recycle concrete a bit, currently about 28%. Crushing it costs energy, and it’s easy enough to scoop up new rocks from a quarry. Some concrete is used as aggregate for road construction and similar tasks, but that’s about it. In any event, you can’t turn old nuclear reactors into new ones.

But don’t look at the argument, look at the facts. People are recycling solar panels right now, today. Meanwhile, the cost of decommissioning nuclear plants is spiraling out of control, and we still have no idea what to do with the waste fuel.

Adding it all up

So let’s just put it all together. After 30 years (their number) when we decommission the power system that produced all the electricity in the world, we have something like:

• nuclear: 388 million cubic meters of concrete with a 30% recycling rate and 2 million of fuel we might recycle once
  • 115,000,000 cubic meters of waste, some of it radioactive.
• solar: 195 million cubic meters of panels with a 90% recyclable rate
  • 19,500,000 cubic meters of waste, all of it safe.

So solar is 6 times better than nuclear by their definition.

While all the fusin’ and a fuedin’?

Solar and nuclear are both going to play their part in the future energy mix. Neither can solve every problem, and there’s no reason anyone would even want that, they both have their strengths and weaknesses, and unlike most other sources (like wind), PV and nukes compliment each other. Nuclear supporters should be supporting solar as a peak play, and solar supporters should be supporting nuclear for night-time baseload.

So when I read stories like this one I’m left wondering at the “why” of it all. Why would someone who is clearly not familiar with solar (or more worryingly, is) make such a ridiculously silly argument? And when the arguments simply gloss over the real world problems of nuclear we should be discussing like adults instead of sweeping under the rug, how does that help anyone, least of all the nuclear industry?

Be afraid, be very afraid

Why?

Because it was written by an MIT nuclear engineer. One who’s thesis was on nuclear safety and now works for a company called UPower Technologies that wants to build tiny nuclear reactors.

Now that makes you stop and think.