In March 2007 I wrote an article for this blog called “Techno Optimism“. It was about global warming and technological solutions and it started as follows:-
The prospect of global warming induced by carbon emissions scares a lot of people. There are two reasons why I’m not scared. One is that I think the more alarmist predictions are overblown. The other reason is that I am very much a technology optimist. I think that technological advancement has an enormous capacity to solve most of the problems predicted by the likes of the IPCC.
Before trying to pick some winners I offered two caviets:-
- “Firstly my track record at picking winning technologies is not great.”
- “Secondly very few attempts at futurology prove successful (thats why venture capitalists have broad portfolios).”
In that article I looked at two interesting technologies. One was the thermal Solar Tower technology being promoted by Enviromission and the other was a battery by a company called Eestor. Four years on I’m less optimistic about these technologies succeeding in the market place and even if they do I’m quite sceptical about them offering any form of meaningful solution toward the global warming issue. However always the optimist I have some new tips to offer.
However first I should outline briefly why I continue to discount the hot favourites of wind power and solar power. A lot of subsidies have been poured into wind and solar on the assumption that with scale the cost of production would fall and these technologies would get cheaper. These technologies have in fact got cheaper so the proponents of subsidies can make a respectable case that they have been right. However the problem with this argument is that even as the efficiency of these technologies has improved and the costs of these technologies has reduced there are still scale issues that are in my view insurmountable. Cheaper and more efficient due to the scale of production is arguably a case of global economies of scale. Scale delivers what is known as a learning effect not to mention other cost reduction benefits. However from the vantage point of a single grid there are still massive dis-economies of scale. Converting our grid to 90% wind power will cost a lot more than nine times converting 10% of our grid to wind power. In short dis-economies of scale in grid operation will trump any economies of scale in the production of these machines.
In the last few years I have become quite interested and impressed with nuclear technology. It is a solid proven base load power source that can scale. France get’s over 80% of it’s electricity from nuclear power plants. Despite the bad publicity it has an incredibly good safety record as I discussed in my article here called “Safety and Electricity Production“. The recent events in Japan have done little to dent that view. However in spite of being safe and viable nuclear power contributes less than 15% of the worlds electricity. The issue is cost. Nuclear power remains expensive for a host of reasons.
Part of what makes nuclear power expensive is sovereign risk and excessive regulation. That is the risk that the government will under pressure change the rules half way through the construction of a nuclear power plant or during it’s operating life or that regulators will make it hard to achieve construction deadlines or to operate profitably. The argument is that governments driven by sections of the community with a hyper unrealistic view of safety drive regulators to extreme levels of interference and the costs mount up until nuclear plants are not cost effective. Those that argue for this explanation point to the fact that nuclear construction in the US halted after the melt down at Three Mile Island in 1979. However whilst I think there is some truth in this argument the construction of nuclear plants was halting before the Three Mile Island accident because of higher than expected construction costs.
For nuclear power to be a solution to the global warming issue it really needs to remain safe but it needs to be cheaper than coal. Given that a lot of the costs are tied up with safety systems these two goals seem to be in some contradiction. What is required is a radical rethink of the design of nuclear power plants. Which is where we get to my latest attempt at picking a winner.
The Liquid Fluoride Thorium Reactor (LFTR) has never operated commercially but it has an impressive set of design characteristics that suggest that it should. However to understand them we need to understand some of the features of conventional reactors. The list of advantages is long so I’ll only touch on some of the major ones.
Conventional reactors heat water to high temperatures (hundreds of degrees). The hot water is used in a heat exchange to drive a turbine system. The problem with this approach becomes apparent when you think about how it is that water can be heated above 100 degrees Celsius and remain in liquid form. The answer to that puzzle is pressure. The water in a typical reactor is compressed to around 100 times atmospheric pressure. That means very think reactor walls. It also means if there is a leak the water will flash to steam and rapidly exit the reactor. In doing so it will carry out radioactive elements. These will be captured by a containment building but that building needs to be very large because when high pressure water converts to steam the volume increases massively.
What if the water in such a reactor could be handled at normal atmospheric pressure and still remain liquid? And what if when there was a leak the water oozed out slowly and then froze solid? And what if we could still have it at very high temperatures in the reactor? This highlights the first big advantage of the LFTR. Of course water does not have these properties and the coolant used is not actually water. It is molten salt. The reactor runs at atmospheric pressure which means the construction does not need to handle high pressure, the molten salt is a very good thermal conductor so it works well as a coolant, and if the salt leaks out it freezes to a solid and tends to block the leak.
The second big advantage with the LFTR is that the nuclear fuel is not solid like in a normal reactor. The nuclear fuel is actually dissolved in the coolant. There is a clever way to add fuel at any time without shutting down the reactor. So rather than stop the reactor and add 18 months worth of fuel in one top up, the reactor can be fueled continuously as required. The stock of fuel in the reactor is in general vastly smaller than the stock of fuel in a normal reactor. Likewise the reactor does not need to be stopped to extract the waste products. This approach also removes nearly all the costly fuel fabrication processes that current reactors require. In fact the business model today for nuclear companies is to basically give the reactor away at cost and make all the profit on fuel fabrication.
Even just in terms of the raw conversion ratio between fuel and electricity the LFTR requires about 200 to 300 times less fuel than a conventional reactor. And it can be fueled using existing stockpiles of nuclear waste. At the end of the process the longest lived waste products are incredibly tiny in volume and are radioactive for only 2 to 3 hundred years. Many substances such as glass are chemically stable for much longer than 300 years and chemical encasement becomes extremely viable.
The raw fuel for these reactors is Thorium. Thorium is much more abundant than uranium.
There is a lot more to these reactors than I have outlined here. I have not gone into the construction and operating costs but they are anticipated to be very favourable. I’ll leave you with this TED talk by Kirk Sorensen.
For an encouraging although very poorly informed debate on the LFTR in the British House of Lords see the following:-
I’m a fan of LFTRs but if anyone has read this post and wonders why we aren’t building them right now, well the historical answer is that they can’t be used as breeders for weapons development, but the present day answer is that there are a bunch of technical issues:
– Problems with extraction of proactinium (a waste product) from the fuel.
– Formation of radioactive tritium which then diffuses through the metal piping.
– Corrosion of the piping carrying the molten salt.
– Damage to the graphite moderator over the course of reactor operation.
None of these are necessarily insurmountable problems, but they all need to be fixed before a full-scale LFTR gets built, or the builders risk losing a ton of money. So it’s a risky investment. The good news is that if the problems do get fixed, the reactors are a candidate for mass production. Because they operate at atmospheric pressure, you don’t need to manufacture huge, over-engineered pressure vessels, which is where a lot of the cost for conventional reactors comes from.
China has recently announced a research project aimed at eventually building commercial LFTRs. If I were running Australia I would be offering assistance in return for the sharing of technology.
I thought the proactinium issue was solved within the two fluid system and was only a problem in single fluid system.
Don’t know. Maybe. I thought single-fluid was still the dominant design.
It seems to be convered under point one of “Technological disadvantages” in the relevant Wikipedia article and doesn’t seem to be a show stopper.
In short Protractinium will act as a neutron poison but not a serious one. Removing it will create some long lived nuclear waste (231Pa) so better to leave it in the mix and let it suck up a few neutrons and then burn it up.
As I understand it the most serious neutron poison# in a nuclear reactor is xenon-135. However this is where the LFTR has a big advantage. Because the fuel is in liquid form the Xenon will bubble off and can be readily extracted from the mix. It will then decay after about nine hours.
# A neutron poison is a substance that slows the nuclear reaction by taking up neutrons.
The Chinese seem intent on building a Liquid Fluoride Thorium Reactor;