Non-thermal ways nuclear power can generate electricity

[Cikomyr2]

I remember a few years ago I've watched a video about the history of nuclear power, and at one point the video asserted that "thermal plants are the simplest way you can generate power so most reactors use this method". And then the video went on to explain about heating power and rotating turbine.

I.e. once you have generated the heat, the rest of the power plant is exactly the same between a coal or a gas power plant.

I was curious of the other ways we know nuclear power can be converted into electricity, without going through boiling water?

[Maat Mons]

I would imagine photovoltaics, or something equivalent. Gamma radiation is part of the electromagnetic spectrum, same as sunlight.

Edit: Apparently, what I was imagining is called gammavoltaics. But there turn out to be some other options too. (https://en.wikipedia.org/wiki/Atomic_battery#Non-thermal_conversion)

[Rockphed]

a gas power plant.

As I understand it, you can use natural gas's exhaust as the working fluid in a turbine. You can also use the hot exhaust to boil water and then use the steam as the working fluid in a turbine. Based on five-minutes of research, it looks like natural gas extracts a little more energy from its fuel than coal and oil do and the ones without the boiler can be turned on and off very quickly (albeit they are less efficient) and so are nice for unexpected excess demand on the grid.

[Thomas Cardew]

In basically any nuclear reaction, the majority of the energy gets carried away as kinetic energy of the products. Since the energy of the reaction is much larger than the binding energy of the electrons, you tend to get a 1 or more charged particles. These particles dump their kinetic energy into the surrounding material by Coloumbic interactions. Direct energy conversion converts the particles kinetic energy into a voltage (electrical energy) directly instead.

There's been some research but it's not really done at scale anywhere. Basically, you use magnetic fields to direct and collect the charged particles. One design used thin fuel wires in parallel field. The fuel would fission, the fragments would separate and escape from the wire, drift to the end of the cylinder under the influence of the magnetic field and be collected. Other designs would generate a partially ionized plasma that would pass through a magnetic field and cause electrons to flow in electrodes. You'd pull part of the energy out directly in the electrode current, then past the rest of the heat off to gas or steam turbine to capture the rest.

There's also some photoelectron converters that basically involve the photoelectric effect, light striking a material releases an electron. By coupling a good absorber and some collection sheets you convert the photons into electric current. Basically a high energy solar panel.

[Fat Rooster]

It is specific to the nuclear reaction being used, as the conditions of each are generally already quite hard constraints. Nuclear batteries can use a thin film, so sometimes can use the charged particles directly. This does limit you to very low quantities though, which in turn limits the power. It can get quite high efficiencies though. Because of the low quantities you would want to use something with a pretty short half life, so this also limits duration. This can be good if you really don't need much power. They can be very effective in space, because they are basically foils separated by a vacuum. They can have near zero mass.

Nuclear lightbulbs is another possibility, again for nuclear batteries. If you embed your material in a matrix of graphite, and then reflect as much as possible of the infrared radiation back at it, you can get it to high temperatures, and just use photovoltaics. You can get higher temperatures than a turbine doing this, though it is still a thermal cycle and has other inefficiencies.

Fission is quite hard to get electricity from directly, because the products are varied, much of the energy is spat out in neutrons, and we need the reactor to be quite dense. A foil design might work in theory, particularly if you had it surrounding an existing neutron source (might be a viable first use of fusion). The charge, mass, and energy of the fission fragments can all vary, so it is far less efficient than more focused foil designs.

Fusion designs can be far more novel, because we are often working the the plasma phase already, and they have a much narrower range of products, energies, and charges. Getting it working at all is priority one, but arguably fusion needs direct to electricity to be viable. If we are stuck with thermal, then it has no major advantages over fission (Waste is not a problem for fission, that is why none of the megaproject 'final solutions' ever work out. It is a larger PR problem than technical one).

While technically true that there are other ways of getting energy out of nuclear power, calling thermal techniques 'simpler' understates the massive complexity gap involved, to the point that there are no other realistic solutions.

[Chronos]

Quoth Fat Rooster:

It is specific to the nuclear reaction being used, as the conditions of each are generally already quite hard constraints. Nuclear batteries can use a thin film, so sometimes can use the charged particles directly. This does limit you to very low quantities though, which in turn limits the power. It can get quite high efficiencies though. Because of the low quantities you would want to use something with a pretty short half life, so this also limits duration. This can be good if you really don't need much power. They can be very effective in space, because they are basically foils separated by a vacuum. They can have near zero mass.
And while that's a fairly narrow window of constraints, there are some things that fall into it. Artificial pacemakers, for instance, are sometimes powered by a betavoltaic battery based on tritium: They don't need very much power, and you want them to last decades without maintenance, but you don't need them to last beyond decades.

Another way you could directly harness energy from charged particles is to put your particle source in the center of a spherical capacitor. Hold the capacitor at a voltage just barely short of the stopping voltage of your particles. Whenever a particle is emitted, it'll give up its kinetic energy, and in the process slightly increase the charge on the capacitor. Bleed off the excess charge using conventional wires (that's your output), to keep it at your desired voltage.

[gomipile]

I realize that the actual question asked in the OP is about non-steam power sources. But, if I wanted to be a pedantic [redacted] about the thread title.....

How many of these methods don't count at all as a heat engine?

[Trafalgar]

An isotope that is a strong beta particle emitter could be used to produce small amounts of electricity. This is possible because beta particles are fast moving electrons. I believe the Soviet Union did this on some of their space probes but I could be wrong.

[Thomas Cardew]

If anyone is really interested in this, there's a professor at Trine University who literally wrote a book on direct energy conversion for his class. It includes a survey of energy conversion methods and theoretical ideas. It's targeted at junior level electrical engineers.

It's available free online here.

https://www.trine.edu/books/documents/de_text1.0.0.pdf (https://www.trine.edu/books/documents/de_text1.0.0.pdf)

[Chronos]

I mean, if you get sufficiently pedantic, everything is a heat engine. You can assign a temperature to any physical process or phenomenon, and then apply the laws of thermodynamics to those temperatures. Strictly speaking, when someone says that something "isn't a heat engine", what they mean is that it has an extreme ratio of effective temperatures and thus has an extremely high limit on the efficiency. But at some point, in doing so, you end up with concepts like "the temperature of thought", and the entropy increase needed to design the device, and so on.

[Rydiro]

I mean, if you get sufficiently pedantic, everything is a heat engine. You can assign a temperature to any physical process or phenomenon, and then apply the laws of thermodynamics to those temperatures.I've never heared of that. Can you explain further? And I also thought that temperature is an aggregate statistical way of looking at collections of matter. Can single interactions meaningfully descibed by temperature?

[gomipile]

I've never heared of that. Can you explain further? And I also thought that temperature is an aggregate statistical way of looking at collections of matter. Can single interactions meaningfully descibed by temperature?

It takes a lot of "single interactions" to generate a useful amount of power for most purposes.

[sihnfahl]

It takes a lot of "single interactions" to generate a useful amount of power for most purposes.
Seebeck effect, you thinking of?

[gomipile]

Seebeck effect, you thinking of?

No, I'm being far, far more general.

To get multiple watts of power from fission, there will be more than a few times 10 to the 10th fissions per second in the reactor.

That's far more than enough to generate meaningful thermodynamic statistics. It doesn't matter what method you use to extract energy from the gamma photons, it'll be so many more than a single interaction that thermal statistics will be a meaningful way to talk about them.

[Chronos]

The example that was used to bring it to my attention was in orbital mechanics. It's possible to travel from most places in the Solar System to most other places, while expending only a very small amount of energy, by applying that very small amount of energy as thrust in just the right direction at just the right time, and then waiting long enough. Add large tether structures and the like, and you can turn that "most places" into "all places". And it seems, at first glance, that you can get that "very small amount of energy" arbitrarily small, if you're precise enough and patient enough.

But the catch is that doing so would require a lot of computation, and that amount of computation itself generates entropy and requires energy input. So, no, you can't in fact get arbitrarily close to perfect efficiency. There's a limit on the efficiency, and that limit is a thermodynamic limit. You can get really, really high efficiency, but not perfect.

[Maat Mons]

I'm not entirely sure what the last few posts have been about, so what I'm about to say is probably, at best, a bizarre tangent, but...

Doesn't getting around the solar system "without using much energy" generally involve gravity slingshots? Those actually use the energy of the planet you're slingshotting around. It's just very difficult to measure the tiny changes in the planet's velocity due to the extreme difference in the masses involved. Thermodynamically speaking though, I don't think using energy from an external source counts as more efficient than using energy from an internal source.

[Chronos]

When you go to a higher orbit, you do so by stealing a little bit of orbital energy from some other object, but when you go lower, you dump energy in. You can make it all balance by just making sure you have equal-mass cargoes going in both directions.