In 1961, a year after my birth, the world’s first nuclear powered automatic weather station was installed on Axel Heiberg Island in the Canadian high arctic. Unlike nuclear fission, which is right now providing the electrical output shown in Tables 1 and 2 in the left hand sidebar, the nuclear power for the weather station came from the heat generated by disintegrating radioactive material. That material was strontium-90, a fission product which is made in nuclear reactors.

The location of the world’s first radioisotope-powered weather station. A return flight for one passenger to Axel Heiberg from Resolute (on the southern part of Cornwallis Island) costs roughly $2,100. That is why the station was powered with radioisotope decay heat: the station had to run with 100 percent reliability. Decay is an inherent physical property of radioisotopes: it happens no matter what.
Sr-90 is an unstable, or radioactive, isotope of the element strontium. “Unstable” or “radioactive” means it has a finite existence. Sooner or later every atom of any amount of any radioactive isotope will transform into an isotope of another element. This transformation happens in its nucleus. It so happens that Sr-90 will transform into an unstable isotope of the element yttrium (yttrium-90). Every atom of Yt-90 will itself eventually transform into a stable isotope of zirconium.
Sr-90 has a half life of roughly 29 years. It decays by ejecting a Beta-minus particle, essentially an electron that is produced when a neutron morphs into a proton. That Beta leaves the nucleus at a relatively high energy: 200,000 electron volts. That is why a mass of Sr-90 will generate heat: when those Beta particles slam into the first material they encounter, the friction of that collision creates heat. It is exactly the same principle in the case of electrical resistance heating. Your cookstove coil turns red-hot because of the innumerable electron collisions that occur in the coil material, which resists, rather than conducts, electric current. (The analogy ends there, as your stove uses the electromagnetic force to generate heat, while a radioisotope device loaded with Sr-90 uses the weak nuclear force.)
The heat from disintegrating Sr-90 is harnessed using a radioisotope thermoelectric generator (RTG). An RTG incorporates a series of thermocouples, which use heat differences to generate an electric current.
Thermocouple-based RTGs have no moving parts and are therefore extremely reliable. Essentially you design them based on the electric current required by your application, taking account of the length of time the application will require that current. Then you load them with enough fresh radioactive material to deliver the thermal power required to produce the electric current for the duration of time needed. The decay heat will generate current at that level until the material has decayed to the point where there is not enough heat to maintain that current.
The RTG that powered the Axel Heiberg weather station was loaded with less than half a kilogram of Sr-90 in the form of strontium titanate (SrTiO3), a non-soluble compound. It generated 5 watts at 4 volts—1.25 amps—without fail, for two years.
As I said in “Inuvik running out of gas,” there are numerous applications in which Canada could make use of the Sr-90 that it produces in nuclear power and research reactors.
So why haven’t we Canadians used this material? My guess is, because it is nuclear material, and we have been taught to fear it. Somehow, we have been bamboozled into believing that stuff like strontium-90 is irredeemably dangerous.
It is not. Strontium-90’s decay product, ytttrium-90, is used to fight various types of deadly cancers. As you can see in the fact sheet on strontium published by the Argonne National Laboratory, the biggest health threat that Sr-90 poses is ingestion: it must be taken into the body before it can do harm. The Sr-90 in the Axel Heiberg RTG was, as mentioned, in the form of strontium titanate, a solid, non-soluble compound. That means for it to enter your body, you would have to chip a piece of it off, and swallow it.
There are a lot of other materials that can also do grievous harm if you swallow them. Fiberglass insulation, for example. It would be fairly easy to eat: just go to a construction site, find a bale of rock-wool insulation, pull out a tuft, and eat it. It will rip up your insides, and you could die a horrible death.
I personally cannot think of a reason why anybody, other than Frank Rizzo, would do such a thing.
But I also cannot think of why anybody would go through the trouble to dismantle a heavy RTG and chip off a piece of strontium titanate and eat it. Especially when it would be much easier to eat some rock wool insulation, or drink some chlorine bleach.
Somehow the Sr-90 in the RTG has been portrayed as far more dangerous.
I think it is time we stopped indulging our silly paranoid fantasies. We should use the useful, carbon-free heat that Sr-90 safely provides.
Steve a question – I’m just trying to get a rough-cut grasp (i.e. order of magnitude) of the economics here. How much thermal power does 1 kg of Sr-90 generate? It would be useful to relate output from 1kg of Sr-90 to the cost of delivering the same amount of heat via oil or natural gas over a given period. I’m assuming the thermal power is considerably higher than the electrical power output referenced in your article as an RTG probably has a fairly low conversion efficiency…
Steve, thanks, good question. Dave Weber of TalkAtomic.com estimates that a gram of pure Sr-90 emits just over half a watt of thermal heat per hour (see http://www.talkatomic.com/idesign/default.html). Factor in the properties of SrTiO3, and you’ll get something less than that. Bottom line is that thermocouple-based RTGs have low efficiency in converting heat to electric power — something like 6 percent. But the tradeoff is reliability: they have no moving parts. When you send an RTG-powered probe up to Mars and the RTG fails, it is quite expensive and inconvenient to go up and fix it! (I know Curiosity’s RTG is fueled with plutonium-238, but the principle is the same.)
There is a lot of R&D today into Stirling engine based RTGs, but they’ll have to be double proven before anybody sends them up into space.
Steve: I thought I’d try and figure it out and I get a number that is twice as much!?
In the top link on Sr-90 (ANL publication, 2001) it lists activity of SR-90 at 140 Ci/g.
This is a 2-stage decay: Sr-90 -> Y-90 -> Zr-90. The first stage is a beta decay releasing .20 MeV and the second stage is a more energetic beta decay releasing 0.94 MeV. Since the Yttrium-90 activity is 550,000 Ci/g, this will soon reach equilibrium, i.e. the rate of Sr-90 decay into Y-90 will equal the rate at which Y-90 decays to Zr-90. So, for all practical purposes we can consider each decay of Sr-90 to become Zr-90 with an energy release of the combined 1.14 MeV.
1 Ci = 3.7E10 decay events per second, so the activity of 140 Ci/g translates into 5.18E15 decay events per second per kg of pure material. Each decay results in 1.14 MeV that ends up as heat, which is equivalent to 1.826E-13 Joules. Multiplying this out gives 5.18E15 decay/s/kg X 1.826E-13 J/decay = 946 J/s/kg!
So, my math shows roughly 1kW per kg… not sure where assumptions or math part company! Anyway, onward:
The atomic weight of SrTiO3 is 186, so 1kg of this material would equal almost 1/2 kg of pure Sr-90.
Lets look at the heat output of 100kg of SrTiO3 for the sake of argument. This would contain roughly 50kg of Sr-90 and produce about 50kW of continuous heat or 1200 kWh of heat per day. One barrel of oil contains about 5.8 million BTU of heat potential, which is equivalent to 1700kWh. So our hypothetical lump of radioisotope material would generate the heat equivalent of about 0.7 barrel of oil per day, worth about $70 (probably a lot more if on as-delivered cost basis to a remote northern community). The 5-year present-value of that heat source would be ball-park $100,000 depending on assumptions.
That rough order-of-magnitude estimate tells me that this could be an economically very useful resource, when 100kg of something can net a 5-year economic benefit of roughly $100K under the right circumstances.
Although lossy when converting to electricity, there are many applications that could benefit from cheap, 24×7 low-grade heat supply that can run for year after year without any maintenance – district heating for remote communities, keeping infrastructure defrosted in harsh climates, etc.
Very interesting!
Steve, I don’t see any problem with your calculation of the specific power of Sr-90 — I get the exact same number (0.946 W per gram), and essentially the same number (0.938 W/g) from an ORNL publication which gives Sr-90’s specific power as 0.0067 watts per curie.
Your 100 kg of SrTiO3 might put out 17 percent less power at the end of year 5 than at the beginning of year 1, thereby making it worth 17 percent less than the original amount of petroleum, but that’s a quibble. The bottom line as you put it is that there are savings — not just in the cost of the petroleum but also delivery, maintenance, etc., leaving aside the GHGs. It depends on the cost of making the SrTi03 (i.e. the cost of separating Sr-90 from used fuel/targets, plus chemical synthesis, etc.).
Yes, very interesting. One application could be using Sr-90 heat to accelerate anaerobic digestion of municipal waste: digestion is most efficient at around 50 °C, a difficult temperature to maintain in polar climates.
The anaerobic digestion can be promoted by burning some of the evolved methane, but that’s wasting a resource that could be used elsewhere. If I was designing a dumb-as-a-rock system for you, I’d have a 200 kg SrTi03 source producing about 100 kW(th), coupled to an Organic Rankine Cycle (ORC) boiler producing vapor to drive a turbine at perhaps 15% thermal efficiency (15 kW) with the waste heat going to keep the heavily-insulated waste digester tanks warm and the town’s water supply and sewage drainage lines from freezing. Some of that 15 kW of power would run compressors to filter the methane via pressure-swing absorption and compress it into tanks for use as vehicle fuel. Why should you have to import gasoline to run snowmobiles, when you have locally-sourced methane?
You may say I’m a dreamer, but I’m not the only one… I’m just one of the elite few who are also hard-core geeks.
The sky is the limit for this kind of application. Imagine if we were not all so propagandized into thinking radioisotopes are the most dangerous things in the cosmos. Another sewage application involving fission products, possibly well suited for coastal communities that still dump raw waste into the ocean, is one where Sr-90 heats the stuff up and Cs-137 sterilizes it prior to disposal.
Just as long as you don’t have people clueless enough to cut open an old one and play with the pretty glowing stuff inside….
In a sewage application, I just don’t see that happening. I recently replaced the cottage sewage pump and I can say from first hand experience that while it was nothing if not pure fun, I hope I never have to do it again.
The USAF recently replaced the Burnt Mountain RTGs (Sr-powered) with a propane-powered system. Is there any guarantee that some dumb sap won’t break into that while puffing on a lit cigarette? I don’t see how being blown to bits or at least severely burned in a propane accident is preferable to what happens if you swallow a piece of SrTiO3 or put some CsCl in your pocket. Either way the outcome is the same: you are hors de combat, probably permanently.
The difference is, people use propane and recognize its tanks and hazards. People can be quite clueless about Cs-137.
Not that I don’t think it’s a great idea, but there would need to be a lot of public education (including in K-12) to make sure that people understood what was being used, what it did and why it is dangerous if misused or tampered with… and suitably armored and alarmed to make ignorant or malicious mischief hard to get away with.
I would look at optimization in a different light. What is the likelihood of failure, what is the cost of backup power, what is the cost of lost mission duty cycle, what is the ability to procure repair parts, what is the cost to maintain indigenous repair capabilities, what is the cost of building a repair infrastructure, etc.
From my experience dumb systems that just work and are inefficient and expensive are preferred over highly complex and evolved systems. The complexity requires much more support. In remote applications like you cited, the dumber the better.
I operated submarines easily 3,000 nm from any repair station with it taking several weeks to transit to a facility that can conduct the repairs, and that was even a remote island on the opposite side of the world from our home port. I was responsible for the maintenance and logistics of executing those repairs. even routine things like painting were a challenge. I hated the wiz-bang systems defense industry shoved down our throats.
Keep it simple and keep it dumb, because the resources to correct failure are much more than you may initially consider. The value of the design is in how little infrastructure is required to operate and maintain the component.
Perhaps it’s time to resurrect the Slowpoke-3. It faced a regulatory hurdle in that it was an unmanned, remotely-controlled device. A small crew of operators who are rotated in and out of northern communities would address that and possibly make it a viable economic competitor to fossil fuels. It would certainly be a better choice on an environmental front.
In my day job, I have been trying to help some people who are expending vast quantities of brain cycles trying to provide reliable, long lasting back-up power that a Slowpoke-3 would fit the bill and allow elimination of a whole lot of lead acid storage batteries and multiple back-up diesel generators with large fuel supply systems.
So far, no traction, but I will continue to press the issue.
In nuclear reprocessing and waste managment plants we use steam for evoparating radioactive process fluids using high pressure steam either through jacket or coils. These evoparators needs maintenance. As 1Kg of Strontium-90 provides around 1000KCal of heat i.e evoporation of 2 Lit water per hour. Although evoparation rate is very slow but it will be a totally maintenance free system. Can it be used for such application.
I’d say that a major reason Canada doesn’t use strontium heat sources is that 90-Sr is typically recovered from reprocessing of reactor fuel, the cost of which is normally defrayed principally from the Pu & 235-U contained in the discharged fuel. CANDU spent fuel doesn’t contain enough either to make reprocessing economically attractive, at least with PUREX-type techniques (although the Indians do it on a small scale, to provide Pu for their fast reactor program), so there are no by-product isotopes to be had.
The USA recovered large quantities of radiostrontium & radiocaesium from military wastes, with the declared intention of finding commercial uses, but there was never enough government support for that part of the program, so the capsules sit in warehouses in Washington State or Idaho or someplace.