Hacker News Reader

Comment by spawarotti | Hacker News Reader

How long will the fuel last? Sabine Hossenfelder references studies in which we have enough Uranium for 50 years max.

Uranium and thorium fuel on earth can make 100% of today's primary energy for about 4 billion years, using breeder reactors, which were proven in 1952 near Arco Idaho at the Experimental Breeder Reactor 1.

Writeup explaining this with lots of actual scientific references at the bottom here:

https://whatisnuclear.com/blog/2020-10-28-nuclear-energy-is-...

People who say uranium will last 50 years either aren't aware of breeder reactors, which have been the long-term plan for nuclear fission since the 1940s, or they're misleading you. We found a lot more uranium than expected in the interim so they have been put off for a while. But we know they work and have built many.

This is so depressing. It's nearly impossible to form a confident opinion these days without a ton of effort. I heard Sabine, I said to myself...makes sense...and it looks like she's put a lot of work into understanding all of this, and she is a professional. So that was it.

Now you're saying she's wrong. Very wrong. And you're a nuclear reactor physicist (thanks for comment!) Am I going to go read all of the relevant references for myself and study the state of the art well enough to understand it all? No, I'm not because I'm not a policy maker or advisor.

I feel like there is a legitimate problem with science communication, especially where it can influence government policy.

It's worth mentioning that being a theoretical physicist working in a niche subfield doesn't necessarily qualify you to talk broadly about other fields and assess their societal impacts. I think these types of videos work best when it's clearly just a smart person sharing what they've learned to engage your interest rather than as a substitute for a lecture by an expert.

It’s always been this way, and it will always be this way.

Don’t get discouraged. This is all part of the process of finding and disseminating knowledge.

FWIW, as someone with a degree in physics, I don't know anyone with a similar background that watches her content. It's not that she is frequently wrong, it is that the opinions are pretty biased and cherry picking. She seems to be more focused on content creation than the actual physics. Honestly I think PBS does a much better job, and importantly stresses that their presentation is overly simplified. A lot of science is extremely nuanced and a first order approximation can lead you in the complete opposite direction, so it usually is a good indicator at who to trust. Are they telling you the way it is or are they attempting to convey a complex topic as simply and accurately as possible? The difference is often subtle.

Definitely. Her content gets posted here a lot but the quality of her videos is just not very good and she’s not great at explaining things either.

Sabine is very opinionated, which is fine for a theorist, but I do find that she sometimes slightly misrepresents the views she disagrees with in her science communication.

Sabine I've found to be spreading nothing but FUD personally. I've only watched a few of her videos but she seems to enjoy inflating minor issues into major ones and then proclaiming everything can't work.

She address fast-breeders and thorium in the same video: https://youtu.be/0kahih8RT1k?t=734 (concluding that they are too complex and expensive)

> (concluding that they are too complex and expensive)

She states "To make a long story short, they didn't catch on, and I don't think they ever will."

She states her opinion, and there's nothing wrong with that, but (a) others may have a different opinion, and (b) it may be possible to make them more practical if more effort is put into them that has been in the past.

Everything I know about breeder reactors comes from the Radioactive Boyscout, which says they sounds great but are insanely unstable and not yet workable.

Is that true? What’s the state of breeder reactors today?

Russia has been operating big ones for decades just fine. The US had a few excellent and beautiful prototype ones (EBR-2 and FFTF), but Clinton shut them down. France built 2 big ones, the first (Phenix) was great and the 2nd (4x larger, SuperPhenix) had various non-nuclear problems and shut down with a poor overall history. China and India both have small ones running fine and big ones under construction.

Many Breeder type reactors also exhibit natural safety characteristics, where they can both shut down and remove afterglow heat with no external power or user intervention at all. This is because of low-pressure coolants like liquid metal or molten salt. They can handle loss of heat sink, loss of flow, and tranisent overpower (e.g. rod widthdrawal) without the control rods going in. Normal water cooled reactors could not survive such events without melting.

So there's a strong argument to be made that while regular reactors are very safe, breeder reactors can be even safer.

https://www.sciencedirect.com/science/article/abs/pii/002954...

France has abandoned their fast breeder effort. As in, put the research on the shelf and stopped funding. That should tell you what they think its prospects are.

Can you name a single industrial satisfyingly-working breeder reactor? AFAIK, and after ~70 years of research (juge investments in many nations) there is none.

This is making a lot of assumptions about what uranium is economically viable to extract. Specifically, this link appears to be assuming that it is possible to filter the entire oceans and Earth's entire crust for all the uranium and thorium they store. Those are both obviously unreasonable assumptions.

For seawater extraction, you just put enough uranium capture fibers in a few places and the uranium is delivered to you slowly over billions of years via ocean currents. This is well supported by the various articles and entire scientific issue featured in the See Also section.

But if you don't buy seawater extraction, check out the Weinberg 1959 reference (https://doi.org/10.1063/1.3060564), which contains a calculation for how much earth would need to be moved to power the entire world on granite. They calculate that we'd need granite mining from the crust about the same order of magnitude of the fossil fuel mining operations at that time. Of course, mining granite is far less destructive than mining fossil fuel, so it's totally acceptable.

Recall that there is 20x more nuclear energy in average crustal rock than there is chemical energy in coal, per kg. So to a breeder reactor, it's literally as if the entire earth's crust is made of pure coal, 20x over.

Will that last long enough for ya? :)

And with that kind of energy density, it's all economical to extract.

this has some serious problems in the analysis. First of all, about half of the earth's crust is under an ocean. Second, of the remaining half, it is on average about 10 miles deep. There is no way that digging up 10 miles of rock to get to some scattered uranium atoms is net positive from an energy perspective. The deepest mines in the world are 2.5 miles under ground, and is in a location with a very high concentration of gold. Most of the uranium in the earth's crust is at concentrations of less than 1 part in 1 million, and one ton of uranium can only (being maximally optimistic) lift 1 million tons of rock by about 1 mile, so any uranium lower than that (not in a major vein) will produce negative energy to mine. Also, just because it's technically net energy positive, doesn't mean it's efficient at all. If we want mining to be at least somewhat efficient, we will only be getting roughly 1/3rd of that (since by the time you are lower, you will be losing too much energy to be cost competitive).

This very basic analysis suggests that your link is off by at least a factor of 100, which doesn't inspire much confidence in their results.

Ocean floor bedrock is basalt, which has much lower U/Th content than granite. Granites are continental rocks.

Geologically, U and Th have been concentrated over the billions of years by about a factor of 1000 in the minerals that have accumulated in continents. Were this not the case, fission power would be completely impractical.

What are the disadvantages of thorium reactors, making them so rare in practice? Cost?

There's no particular upside that at the moment would justify commercializing the technology.

Except for breeder reactors are even more uneconomical, have even higher proliferation risk. So yes they are a solution if you ignore all the other issues with them.

If you fully fission the U + Th in an average chunk of crustal rock (using breeding so you can use the 238U and 232Th) then said rock releases fission energy equal to the combustion energy of 20x its mass in coal. So with breeders we can in effect treat the entire Earth's crust as something an order of magnitude more energy rich than coal.

(Whether breeders are practical or competitive is another matter, but then fusion looks pretty challenged in that respect also.)

This shocking factlet is repeated, verified, and credited to you in the writeup in my sibling comment.

So you did, thanks! BTW, the Earth becomes uninhabitable in about 1 billion years, not 4.

We could get there even quicker with a little effort.

Yea the efforts of people like you pushing anti-nuclear views. Can't wait for the greenhouse gases to make warming even worse while we use NAT gas and coal to smooth over wind and solar.

Reply on news.ycombinator.com