Not enough attention is being paid to Japan's contrarian "red hydrogen" strategy.
This entails using a novel helium cooled fission reactor to generate very hot (950C, 1750F) process heat that is then fed into an Sulphur-Iodine cycle hydrogen plant to create very cheap hydrogen without any feedstock but air and water.
Beyond fuel, hydrogen can be used to replace coal in iron smelting, Haber-Bosch fertilizer, and other chemical processes that require hydrogen made today via fossil fuels.
They already have a 30MW pilot reactor in operation, and are just about to turn on the S-I hydrogen plant. Could be a very interesting addition to global energy mix.
Yeah, a chemical cycle involving vaporizing and decomposing sulfuric acid at 850 C is totally credible. Yes, let's avoid turbines and instead make our equipment out of ... tantalum? /s
The maximum thermodynamic efficiency of this cycle is 50% (and likely will be lower). This is not much better than making electricity with the reactor and driving electrolysers. And when the levelized cost of energy from renewables is very low (especially for the surplus energy that could be fed into electrolysers as needed) it's hard to see how this scheme competes. Yes, it doesn't have turbines, but it does have all sorts of high temperature chemical reactors that must survive corrosive conditions.
>The maximum thermodynamic efficiency of this cycle is 50% (and likely will be lower). This is not much better than making electricity with the reactor and driving electrolysers.
The numbers I've found elsewhere dispute this. The World Nuclear Association puts the thermal -> electric conversion efficiency of nuclear plants at around 35% (current) to 40% (best) [1]. A similar but far less corrosive copper-chlorine process is being developed in Canada with preliminary analyses claiming as much as double the efficiency of electrolysis [2]. The maximum thermodynamic efficiency may also be exceeded if waste heat can be converted to electricity simultaneously.
With that said, tests of the sulfur-iodine process continue to find that corrosion and damage to the (chemical) reactor components remains limiting, with 20% efficiency lost in H2SO4 splitting over just four days [3]. H2SO4 at 900 C is much more corrosive than HCl at 500 C.
Being twice the efficiency of electrolysis with nuclear is not enough, because the levelized cost of electricity ratio between nuclear and renewables is more than 2 (and even more so the ratio between nuclear and surplus renewable output.)
Looking at the solar generation LCoE alone, you're correct. The numbers I found for battery LCoE are all over the place, which makes it hard to judge. Assuming $100/kWh and a cycle life of 1000 (close to existing batteries), battery LCoE is high enough to justify nuclear-to-hydrogen. With longer cycle life, which has gone up to 10,000 cycles at lab-scale, that number comes down substantially. Assuming technology improvements, of course, would also have to be applied to nuclear.
I was originally interested in thermochemical hydrogen for a solar concentrator source, but a lot of current research seems to focus on nuclear because the thermal energy input is more stable, with an outlook towards solar as the technology is worked out.
A life cycle of 1000 is very short, and I don't think it's reasonable to expect that for currently installed batteries. I understand current utility scale batteries are shooting for an eight year lifespan (3000 cycles) with the hope of getting to 10-15 years (but data is lacking to justify that.) Longer cycle life should be proved by the time any nuclear thermochemical hydrogen system could be up and running.
I probably shouldn't have mentioned batteries at all considering that we're comparing hydrogen generation technologies. But the review I found shows cycle life is variable in existing commercial batteries, and generally not much higher than 2000:
The interest in hydrogen is for thermal processes (cement, alumina) and weight-sensitive applications as far as I can see. For electricity storage the fuel cell efficiency is prohibitive.
Assuming a daily charge/discharge a 1000 cycle battery needs to be replaced every 3 years. That would suggest that the majority of installers that guarantee the batteries for 10-15 years are going to lose their shirt on warranty replacements.
Batteries tend to longer if you don't completely drain them in each cycle. I strongly doubt the typical driver covers 200 miles in a day — otherwise cars would only last three years. IIRC you can extend the useful life of a typical Li-ion battery significantly by only charging it to 80% most of the time.
I think the point is that we should try a lot of different approaches, as there may be good fits that aren't possible to do via electricity - ie is electro smelting iron > using hydrogen? Unsure we really know what that looks like at scale yet.
Either way there's an unproven process step with lots of efficiency questions for plants at scale, and I'm impressed to see Japan seriously trying something totally different.
The high cost of Gen 3 nuclear power is largely due to the very large steam turbine and heat exchangers that are required if you are using water as a working fluid, coolant, and moderator.
From the 1950s to the 1990s the interest in fast reactors has been in the 60x better fuel economy and reduced waste problem. (e.g. less radioactive than the uranium ore in 1000 years) It was believed back then that a fast reactor coupled to a steam turbine would have a higher capital cost than an LWR. There also was a lot of concern that it takes a lot of uranium or plutonium to form a critical mass and that would be an expense.
Recently fast reactors and other high temperature reactor types are of interest because getting rid of the water could allow miniaturizing the whole system and get the cost competitive with natural gas not to mention solar and wind on the days that solar and wind feel like supplying power. The stockpile of plutonium in spent fuel is getting bigger and bigger every day and the public seems entirely uninterested in throwing away 98% of the energy content of the spent fuel away in a place like Yucca Mountain. Thus high-quality fissile material seems a lot less scarce than it did in the EBR I - Superphenix era.
> There is a reason we run nuclear reactors at 400C rather than 1000C.
There are several reactors design that work at high temperature, HTGR are not new.
> A novel nuclear reactor which operates just below the temperature where our most exotic alloys mechanically fail
We have plenty of materials that can withstand higher temperatures than that, both refractory alloys and ceramics. Do you know the temperature in an aircraft engine? If weight is not a factor, plain old tungsten melts at 3400°C.
https://atomicinsights.com/chinas-high-temperature-reactor-p... has a summary of the past experimental reactors (skip half way down to the heading “Brief high temperature reactor history”), and explains some of the reasons past reactors have failed to be commercialised.
It also explains one central problem with using gas: “Even with higher temperatures and higher efficiency, each core can produce 1/10th of the electricity of light water reactors, but [China’s] HTR pressure vessel is described as ‘the world’s largest and heaviest pressure vessel.’ Pressurized gas has a far lower capacity to move heat than pressurized water.”.
Well yeah, they did not displace PWRs (the heat capacity issue is one of the reasons why molten salt and sodium fast reactors were investigated). But the problem is not the lack of high temperature materials (which is false).
I had never before considered, but how do they melt/cast tungsten? How do you get temperatures that hot in a controlled way, when basically nothing is solid anymore.
Melting tungsten in atmospheric conditions is not really useful. You start with tungsten oxide as a gas and it reacts with hydrogen and tungsten powder precipitates out. The powder is then heated and compressed to form a polycrystalline billet (sintering).
Technically speaking, the first step in tig welding aluminum is to melt the tungsten and it forms a little ball. It is not hard to reach the required temperature. You can just put a lightbulb filament in 220 socket and it will melt the tungsten in a split second. That being said, a crucible of molten tungsten is not a thing.
It's that the pressure of water either in liquid or supercritical form gets too high.
Material issues for fast and/or high temperature reactors are not trivial but don't look insurmountable. 1980s literature seemed to think fast reactors could be longer-lived than LWRs, LWRs look longer lived today than they did back then.
I imagine a reactor operating at 1000C is built almost entirely on graphite and ceramics... But I have no idea how they plan to use it with water, as graphite doesn't behave well on those temperatures with water around. (Maybe there is a heat exchanger at some point that makes failures on the wet side safe.)
It certainly looks possible. You will probably end up needing some amount of metals on the hot portion, but they can be minimized to a very large extent. Cheap, on the other hand, is not a property I would guess from that description.
This entails using a novel helium cooled fission reactor to generate very hot (950C, 1750F) process heat that is then fed into an Sulphur-Iodine cycle hydrogen plant to create very cheap hydrogen without any feedstock but air and water.
Beyond fuel, hydrogen can be used to replace coal in iron smelting, Haber-Bosch fertilizer, and other chemical processes that require hydrogen made today via fossil fuels.
They already have a 30MW pilot reactor in operation, and are just about to turn on the S-I hydrogen plant. Could be a very interesting addition to global energy mix.
Strategy: https://www.csis.org/analysis/japans-hydrogen-industrial-str...
Reactor: https://www.world-nuclear-news.org/Articles/Japanese-gas-coo...
S-I Process: https://en.wikipedia.org/wiki/Sulfur%E2%80%93iodine_cycle
Video (hypey): https://www.youtube.com/watch?v=_uTZWaJU6ho