* The output is greater than the energy *in the lasers*, but the lasers deliver 1% of the energy required to power them. Need 100x improvement to break even.
* Converting the generated energy into electricity would cut the output in half. We need a further 2x improvement here, so it's ~200x to break even end-to-end.
* The scientific equipment requires immense & expensive maintenance.
* Plus the $3B facility around the equipment, that theoretically could deliver just 2.5 MW.
So we might be as close as 10-20 years away, as always!
No, not as always. The laser confinement mechanism works, it has been shown, lasers that are more than 20 times as efficient as these NID lasers are now available, so the improvement needed to scale and "commercialize it," whatever that really means looks more like 10x than 200x. In the world of fusion, that counts as really good progress. For one thing, perhaps a lot of the research money will move to lasers now.
I mean, you nail it on the head. It's not "congrats on limitless free energy" but more "looks we might still get value in the future if we keep pouring money into this." Positive indicators at milestones are good. Onward.
> so the improvement needed to scale and "commercialize it," whatever that really means looks more like 10x than 200x. In the world of fusion, that counts as really good progress.
Yes it's good progress, but an order of magnitude is not nothing. Squeezing another order of magnitude efficiency out of the lasers will be very difficult. It took 30 years or so to go from 1% efficiency to 20%, and law of diminishing returns applies.
That is trivially true at the extremes of energy input. If you input an infinite amount of energy you will not get an infinite amount out.
But that’s not what we’re talking about. This is a physical process which is known to be exothermic for the energy ranges we care about.
As another example, raising the temperature of a flammable material 1 degree from room temperature will probably not light. Ditto with 2 degrees. But eventually, if you raise the temperature high enough, you’ll get more energy out than you put in. That’s the type of process we’re talking about now.
Diminishing returns usually applies if you assume there are no major breakthroughs. Can we assume that there won't be any major breakthroughs in this field?
Diminishing returns describes a trend. A breakthrough describes a single data point that bucks the trend. I'm not sure these are mutually exclusive, as after any breakthrough the diminishing returns trend is reestablished.
I wouldn't bet on no breakthroughs happening in laser efficiency, but more efficient lasers doesn't look like it will be enough to get to net energy given other inefficiencies in the system.
"Fast ignition and similar approaches changed the situation. In this approach gains of 100 are predicted in the first experimental device, HiPER. Given a gain of about 100 and a laser efficiency of about 1%, HiPER produces about the same amount of fusion energy as electrical energy was needed to create it (and thus will require more gain to produce electricity after considering losses). It also appears that an order of magnitude improvement in laser efficiency may be possible through the use of newer designs that replace flash lamps with laser diodes that are tuned to produce most of their energy in a frequency range that is strongly absorbed. Initial experimental devices offer efficiencies of about 10%, and it is suggested that 20% is possible."
This is irrelevant - each shot also requires a highly precision engineered piece of metal called a hohlraum to be destroyed.
With current technology, running an ICF plant would cost literally hundreds of millions of dollars per hour in hohlraums, since a single one costs millions, and you need to shoot several times per minute to produce energy.
That's why ICF is not even close to being a plausible electricity generation technology, so it is only being researched by nuclear weapons research labs like LLNL.
hohlraums are not expensive because of base materials, but because today we generally produce them as one offs and the process is incredibly man hour intensive. The DOE "roundtable" on the announcement today addressed this.
They are one-offs, but even if they were to be mass-produced, they require extraordinary precision. I very much doubt claims that one can be built in the range of a few dollars that each shot is worth in terms of generated electricity.
The reports you quote actually mention the target costs very clearly. The IAEA one talks about needing 500,000 targets per day, and sets a target of 0.30$ per target. At the time it was written, it says that a target costs 1000$, which is probably before NIF found put just how much more stringent the requirements for the shape of the target were (since the numbers I saw last time NIF achieved ignition were closer to hundreds of thousands of dollars per target, though maybe I am misrembering).
It's also worth noting that that report was expecting NIF to achieve the current milestone within 3-6 years, and it actually took 13. So I feel their numbers can well be considered optimistic.
The NIF is using old laser technology. Current tech can get above 20% efficiency. Sure, that still means more improvement is needed, but 200x is probably an overstatement by an order of magnitude.
> So we might be as close as 10-20 years away, as always!
I don't really get the cynicism here. This is a huge milestone that's been passed. Maybe with this, we actually will be 10-20 years away. Or maybe it's more like 30-40, who knows. But this experiment shows that net-positive energy is actually possible to do with our current understanding and technology; before this, I believe much of the skepticism was based on a belief that it may not actually be possible to get more energy out than put in, at least not without technology that's significantly out of reach.
Anyone have insight into how this new development differs from this article from back in 2014 about the NIF, entitled: "Fusion Leaps Forward: Surpasses Major Break-Even Goal"
Back then they were comparing to the energy actually absorbed by the fusion fuel. This is indirect drive, the laser hits a metal container first and only some of the energy gets to the fuel pellet.
This time, they're comparing to the total energy in the laser beams.
They're ignoring the inefficiency of the laser devices, but that kinda makes sense because they're using really old, inefficient lasers and much better ones are available now.
> This time, they're comparing to the total energy in the laser beams.
How do you know? Nothing has been published yet; it’s science through press release. In the past, published papers from NIF have been a real wake-up call after absorbing the misleading hype (the papers are most honest than the folks taking to the reporters).
> The fusion reaction at the US government facility produced about 2.5 megajoules of energy, which was about 120 per cent of the 2.1 megajoules of energy in the lasers
I guess we'll see how things develop. But from a quick google, 2.1 megajoules is about what the lasers deliver, unless they've significantly increased their power recently.
Right. Livermore has been working on this since the 1970s, with increasingly powerful lasers. Now, they claim "theoretical breakeven" - slightly more energy came out of the reaction than went into the reaction. But 100x less than went into the lasers, let alone the whole facility. Nor is energy being recovered.
This was never expected to be a power plant technology. It's a research tool, for studying fusion.
"Technical breakeven" is when the plant generates enough energy to run itself. This is at least 100x below that.
"Commercial breakeven" is when it makes money.
How's that Lockheed-Martin fusion thing coming along?[1]
I wasn't diminishing the achievement but clarifying its place in the context of how far we are from commercialization. The director of LLNL who announced the breakthrough discovery said she expects we are 3-4 decades away from commercializing it.
Well, they can operate on just DD, so they can start from no 3He and make 3He.
This video of a presentation by Helion's Kirtley at Princeton has a slide where the reactivity vs. energy loss is shown for a DD system at beta=1. That system will make 3He directly, and also make tritium by two modes (directly from DD, and by capture of neutrons on 6Li in a blanket.) The net result would be production of 1.5 3He nuclei per DD fusion, on average. It takes a while for some of those 3He to be produced though, as the tritium has to decay (halflife of 12 years.)
Even if the lasers were currently 100% efficient, the Q still needs to be increased by 2 to 3 orders of magnitude. That's because they're making less than a penny's worth of energy here, and the system cannot be economically feasible with that little energy per expendable target.
Probably 5-10 years if this turns out to be the key unlocking it. If it is, the floodgates will open for funding, public and private, and we'll see a race to build the first reactor. Similar to how the first COVID vaccine was predicted to take 2-3 years and it took 8 months instead because it was a priority.
It took only 8 months because covid has been in existence for decades. Covid 19 strain was new and the vaccines had to be adjusted to new strains not created from ground up
Not so; the mRNA technology used to develop and deliver the vaccine has been in progress for decades. The hardest parts were done before SARS-CoV-2 ever existed, but it's wrong to claim that "the vaccines" needed to be tweaked - they never existed.
For people confused about this, there were prior commercial attempts at coronavirus vaccines, with mixed success. They were not RNA vaccines. The COVID-19 vaccines built on that research (regarding what proteins to target, in particular), but the COVID vaccines that were rolled out were completely novel technology.
To be honest, looking at those numbers, that doesn't look 10-20 years away. We'd need Moore's law style improvement in efficiency and to productionize it. So we're really saying 20 years at best for the technology, and then let's look at quickly we can build Nuclear power plants today... uh oh. In the UK for example it has taken 12 years to even agree to build a new Nuclear plant on a site that already has Nuclear plants!.
My TLDR (from a layman):
So we might be as close as 10-20 years away, as always!