25 thoughts on “Fusion Power”

  1. They use deuterium and tritium. The cross-section of the D-D reaction is “theoretically feasible” for power production, but D-T is a great deal easier.

  2. One problem is that you need those neutrons from the reaction to breed more tritium – you can’t lose them in a metal liner unless that metal liner is a Li6 blanket.

    Also – if the plasma is in contact enough with the metal to gain energy from some sort of compressive shock, it’s going to be in contact enough with the metal to lose thermal energy via electron transport extremely fast. Possibly faster than you could get from the fusion reaction.

    The magnetic field in a tokomak is only partially there to keep the vessel-wall from being damaged. (Nevermind your divertor, what about my poor plasma?) It’s to insulate the plasma so it loses particles slowly enough that they have a decent chance of fusioning before running into a wall. Electrons generally go where the ions go due to electrostatic charge neutrality, but they can run back and forth along fieldlines 60 times faster carrying away energy.

    1. Just read one of their presentations. The wall is a lithium blanket. (I think a breeding blanket will have to be pretty thick though.)

      I’d have to think about the magnetic diffusion rate of the confining fields into a hot metal though (it’s not a superconductor): Why the plasma wouldn’t just run into the walls.

    2. Another semi-frustrating thing about fusion experiments: If they aren’t monster stadium sized constructions, you rapidly run out of *room* on the chamber to put all the things you need to diagnose, poke, prod, heat, and manipulate the plasma.

  3. As to space propulsion:

    All fusion reactors to date are extremely heavy things. There is a certain minimum viable size for a fusion reactor governed by the Lawson criterion:

    You invest energy into a bit of plasma to get it to fusion temperatures. You must hold onto the plasma long enough for a proportion of it to fusion to pay back the energy invested. (x 1/(heating_efficiency * conversion_efficiency)). This is known as a “confinement time”. In addition, you’ll be losing energy from Bremstrahlung from the volume of the plasma.

    All methods of plasma confinement we’ve devised to date are “leaky” in some way – usually the losses scale with something like the area of the plasma while the power gained (and Bremstrahlung) scale with the volume. These set a minimum viable size for a fusion reactor – so far that size seems to be very large.

    Spacecraft propulsion needs to be lightweight. (I suppose it can get as big as you want if you’ve got in-space construction ability, but it can’t be heavy).

    Fusion reactors (in the conventional Tokomak or stellerator versions) are unfortunately very heavy, due to complexity, due to needing the breeding liner, due to needing the confinement magnets, the gas plants to cool the confinement magnets.

    1. If your objective is to get to Alpha Centauri within a reasonable time, pulsed nuclear propulsion (akin to Project Orion) would get you 10 million seconds Isp at a huge thrust-to-weight ratio, using fusion bombs currently in the arsenal. No need for a steady-state, controlled fusion reactor to ever be developed.

      1. …pulsed nuclear propulsion (akin to Project Orion) would get you 10 million seconds Isp at a huge thrust-to-weight ratio, using fusion bombs currently in the arsenal.

        Nonsense. An Isp of ten million seconds would require the conversion of about 5.5% of the reaction mass to energy. The reactions in existing weapons are less than a tenth of that and that is not including the much greater mass of the non-reacting material.

        1. Yup, made a rookie mistake. It’s more like 350,000 seconds.
          There goes my trip to Alpha Centauri…

          1. I’ve heard of one interesting fission fragment scheme.

            There are always beamriders and laser-lightsail-craft. Robert Forward discussed how you can decelerate lightsail-craft with a beam from the sending star.

            There will be ways, but we’ll have enough problems holding civilization together on this planet long enough to settle the Solar System.

    2. Liquid metal held in place via rotation. I wonder how that is induced. Sounds like yet another source of instability to me. Gee, the liquid jacket didn’t quite behave as we expected in the under unity demonstration unit. Subtle vortices form during attempts at plasma compression, that were not apparent during validation of the individual components. Oops. Oh well under our ‘accelerated’ development model, a small additional $4 billion investment should speed development along for another 3 years. As we can see the $/length-of-time is rapidly accelerating with an expanding numerator and shrinking denominator. Driving investment capital well over-unity. “Often it’s not the rocket science that gets you. It’s the nuclear science.” “Still the flurry of private sector activity is exciting as it is sure to put to fire tons of paper investment capital”

      1. Given that the liquid metal blanket is many, many orders of magnitude more dense than the plasma that it contains, I suspect that the spinning blanket won’t be much harder to control than your standard washing machine spin cycle. But YMMV.

        1. Speaking of washing machines, I made the mistake of buying one of those high efficiency units that you can’t control the water level of unless you place a brick in with the load of laundry. Thanks again to Federal Regulation. I’ve had to do some research to find models that actually perform work-arounds. I won’t make that mistake again.

  5. Rand,

    Can’t remember the name of the Lady Professor at UC Northridge from whom we all took her “Update on Space” seminar in the early eighties. I *do* remember Robert Bussard telling us his plan for INESCO was first to develop a deuterium-tritium reactor for Earth-based power generation, *then* with the profits generated thereby, a deuterium-deuterium reactor for space propulsion. He never got there, of course, but…just sayin’! 😉

  6. Another well-funded start-up enjoying some success is TAE Technologies. I believe they have raised $880 million in total. Most of these companies are hoping to build a subscale demonstrator within 5-10 years. More and more, it appears that the physics is less of an obstacle than the engineering. That said, a lot of development remains to be done. It’s great to have so many paths being pursued today, with many attracting substantial investment. Unfortunately, the government continues to put almost all its money into a single, giant tokamak demonstrator, ITER.

    1. ITER unfortunately suffers from the same problems as SLS. SLS doesn’t disprove that rockets are a viable way to get into space any more than ITER proves tokomaks are unworkable.

      Commonwealth Fusion Systems at MIT is trying to build their own burning-plasma-demonstrator tokomak. They’re using superconductors for their toroidal field coils – the field being higher means the plasma pressure can be higher, which means their reactor scales can be smaller than ITER. Apparently they’re absorbing a good fraction of the YBCO production capacity of the world right now spinning their magnets.

      Tri-alpha energy’s name suggests to me that they were interested at one point in time on the p-B11 reaction. It’s an interesting reaction, aneutronic – all products are charged meaning you can do things you can’t in a purely thermal reactor. There is one major problem that needs to be overcome though: At any temperature, bremstrahlung from a fully ionized B11 plasma is greater than the fusion power rate – without some way of preventing hard x-rays from leaving your plasma carrying away energy, you can’t get the reaction to work for a powerplant.

      Last I heard, they’re back to looking at D-T.

      1. You’re talking about Dr. Ritter’s PhD thesis. There ought to be special recognition for those who disprove the field of study they wanted to succeed. The academic equivalent to a Darwin Award.

        Such a pity. Charged reaction products cycling at 60Hz gives you direct electric conversion. No messy thermal transfer mechanism needed.

    2. Certain aspects of the physics (the reactions, their cross sections, radiation rates) are well understood.

      Certain physics problems (details of plasma stability, all the strange things that can happen in plasma) are still open. We didn’t expect H-mode confinement until it happened in a test device.

      Physics says we’re good: Fusion reactions giving net energy are totally possible.

      But a lot of it is engineering, and the engineering problems are severe. Atomic sputtering of the divertor poisoning the plasma, and turning tungsten into metal fuzz. Any impurities in the plasma at all causing extremely fast radiative cooling fo the plasma: That leading to requirements for ridiculously clean high-vacuum in what we hope to some day be an industrial device.

      This is one of the reasons why I’m a “fission or shut-up” guy wrt the Greens. Don’t wait for fusion. Fusion isn’t “X years away”, it’s “X engineering breakthroughs” away. Very important to research, not to be used as an excuse to avoid developing practical power sources *right now*.

  7. For Earth based power, my personal opinion is that we should be building fission plants as fast as we can. A fission powered civilization won’t last forever*, but it will easily last quite long enough for us to work out fusion, or any other arrangement.

    *even that might not be so: I’ve heard there is more movement on getting uranium from sea-water sustainably for net positive energy-returned-over-energy-invested. If so, fission can support industrial civilization indefinitely.

    1. We’ve got more Th-232 than we know what to do with. We don’t have to rely solely on U-235 and Pu-239 when we can breed U-233. Sure let’s have fusion, but fission will be cheap for a long long time. My guess is that a fission/fusion hybrid will be attractive for space propulsion before pure fusion drives, but hey, whatever works.

  8. I’m happy that the tokamak monopoly seems to have been broken. GF’s approach is novel to say the least, using mass-inertial does solve a lot of problems but I’m skeptical that they can make the timing work and handle the nonlinearizing impact of shock waves.

    In general I think inertial confinement of some sort is the right way to go. The statistics of thermal plasmas makes tokamak style confinement extremely difficult. In order to get enough really hot nuclei bumping into to each other for net energy you need to keep a long tail of really really really hot nuclei from hitting the walls. My money is on non-steady state concepts like GF, z-pinch, and deep plasma focus, but I could see electrostatic confinement being feasible eventually too just due to how strong the EM force is (that is essentially what GF is using when it’s all said and done).

Comments are closed.