The State of Fusion Energy w/ Gerrit Bruhaug

然后那些中子可以去找锂，生成氚。

And then those neutrons can go find some lithium and make you tritium.

所以这样我们就得到了燃料。

So so that kinda gets us to the fuel.

我觉得在大众文化里，就像人们把氢想象成从水里来的燃料一样。

And I think in popular culture, in the same way that people sort of think about hydrogen as, like, it's fuel from water or something like that.

对吧？

Right?

我的意思是，氢燃烧确实生成水，但你知道，有很多简单的想法认为氢很容易替代甲烷，而实际上制造氢燃料虽然不是不可能，但肯定比通过电解或蒸汽甲烷重整来制造聚变燃料要容易得多。

I mean, it creates water as a product, obviously, but, you know, there's a lot of simplistic ideas about how it can substitute for methane really easily when making the hydrogen fuel is is, you know, not impossible, certainly more easy than than making, I think, fusion fuel through electrolysis or or steam methane reforming.

但跟我聊聊氘氚燃料吧，因为我觉得人们也会犯类似的分类错误，比如会说，嘿。

But talk to me about DT fuel because, again, people, I think, make similar sort of category error of, hey.

这不过是水的产物。

This is just kind of products of water.

它就是氢。

It's hydrogen.

直接把它们撞在一起就行了。

Just smash it together.

你刚才提到了制造氚的方法，具体是怎么做的？

How how do we you mentioned kind of this way of making tritium.

当然，我们也可以从加拿大的坎杜反应堆里获取一些。

Obviously, we can also give you a little bit from our can do reactors up here in Canada.

我们非常依赖坎杜反应堆。

We are very much relying on the candus.

整个聚变领域都指望坎杜反应堆能继续存在。

That is the whole fusion world is banking on candus existing.

但制造燃料似乎既棘手又昂贵。

But it's it seems that making fuel is is tricky and expensive.

比如，为坎杜反应堆制造重水，虽然不难，但成本很高。

Like, making heavy water for candus is, again, not not not hard, but it's expensive.

我最近了解到，为一个吉瓦级的坎杜电厂生产一批重水，大约需要消耗1.3太瓦时的电力，而且这可能是我们最可能采用的生产方式。

Like, I I learned recently that to make a a load of heavy water for a sort of gigawatt scale candu plant consumes about 1.3 terawatt hours of electricity and and probably the the method we're most likely process we're most likely to produce it.

所以这可不是免费的，而且氚看起来也很难搞。

So it ain't free, and and tritium seems hard as well.

所以，就像我们之前做过几期关于低浓缩铀燃料的节目，我们把它称为燃料中的瑞士手表，因为它的元素、锆合金包壳、几何结构等等。

So just maybe like, we we've done some episodes on just low enriched uranium type type fuel, and, you know, we've called it the Swiss watch of fuels in terms of the the elements, the zirconium cladding, the geometry, etcetera.

而且不像燃烧燃料那样，在发电厂里会被消耗掉并从根本上发生嬗变，燃料元件必须完好无损地离开，不能有任何瑕疵、泄漏或裂缝，否则裂变产物就会漏出来。

And unlike combustion fuels, which are destroyed and fundamentally sort of transmuted, when put through a power plant, like, the the fuel element has to kind of leave without any, you know, flaws or, you know, leaks or cracks where fission products should leak out.

那我们就稍微深入聊聊裂变燃料吧，而且惯性约束和磁约束的燃料显然会非常不同。

So let's just nerd out a tiny bit on fission fuel, and it's gonna be quite different obviously for inertial confinement versus magnetic confinement.

但让我们跳进那片浑水，试着搞清楚这个问题。

But let's let's jump into those stormy waters and try and figure this out.

为了呼应你之前关于铀燃料的几期节目，我快速提一下三种值得讨论的聚变燃料，因为肯定有人会在车里对着音响大喊，说我忽略了他最喜欢的古怪冷门东西。

To to kind of riff off some of your previous episodes with with uranium fuels as well, I'd real quick, let's touch on the three fusion fuels that are worth talking about because they do I'm sure there's somebody you can that's gonna yell into their, you know, car stereo that I'm ignoring their favorite weird esoteric thing.

但已知可行的反应是D-T。

But the reactions that we know work, d t.

我们知道D-T是可行的。

We know d t works.

这已经经过多次验证了。

Has been proven many times.

还有氘-氘反应，我们也知道它是可行的。

Deuterium deuterium, also, we know works.

这个其实就是重水里的成分聚到一起燃烧产生的效果。

That is just heavy you know, basically, the products of heavy water burning together.

至于氘-氦三反应，虽然还没人真正点燃过，但我们基本可以确定它是行得通的。

And then deuterium helium three, nobody has ever ignited that reaction, but we're pretty darn sure it works.

其实我们在一些设施里确实会经常进行这项实验。

There's a bunch of we do actually shoot it relatively regularly on some facilities.

比起氘氚反应，实现点火要难得多。

It's harder to get ignition than DT.

就像我之前提到的，氘氚是最容易的。

Like I mentioned, DT is the easiest.

但氘和氦三，我们确定只要达到相应条件，就能成功。

But deuterium helium three, we are positive, will work if you reach the conditions.

它会自热。

It will self heat.

它会产生净能量。

It will make net energy.

在这些里面，DT是最容易点燃的，而且，我觉得，它的燃料是最麻烦的，因为氘，就像你提到的，来自重水。

So of those, DT is the easiest to get to light, and has the well, I would argue the biggest pain in the ass fuel, because deuterium, as you mentioned, comes from heavy water.

制造重水有点贵。

It's kind of expensive to make heavy water.

不过我们不会烧那么多。

We're not gonna be burning that much of it, though.

氘的成本，我看过相关估算，其实还好。

The the cost of deuterium, when I've seen this specked out, it's like, it's whatever.

真的不会让你破产。

It's not really it's not gonna kill you.

相对而言，那东西挺便宜的，考虑到你能获得的能量。

That's pretty cheap stuff, comparatively, for, like, the energy you're going to get.

对吧？

Right?

所以这就是为什么，理想情况下，如果我们能忽略建造反应堆的所有复杂性，你只需要烧氘氘。

And so that's why, ideally, if we could ignore all the complexities of making the reactor, you would just burn deuterium deuterium.

那是你的低浓缩天然燃料，甚至可以说是聚变反应堆的入门选择。

That's your that's your low enricher natural that's even like your can do of fusion reactors.

如果你能烧氘氘，你实际上就是在烧水的产物。

If you can burn deuterium deuterium, you're literally burning the products of water.

处理起来很容易。

It's easy to deal with.

获取也很容易。

It's easy to get.

它产生的废气与氘氚略有不同，但差别不大。

It does make slightly different exhaust products than DT, but not enough to matter.

你会得到氦。

You make helium.

会产生质子。

You make protons.

还会产生中子。

You make neutrons.

直接排到大气里就行了。

You could just exhaust it out into the atmosphere.

没什么影响的。

Wouldn't matter.

如果是氘氚反应，产物是水，这个我觉得还能接受；但我们需要制造氚，因为它的半衰期只有十二年。

DT, we have our product of water that's, I'm gonna say, not that bad, and we have tritium that we have to make because it has a twelve year half life.

我们没法大量囤积它。

We can't store it up.

它主要是由宇宙射线产生的，反正自然来源就那么多。

It is made by cosmic rays, whatever.

靠从大气中提取是远远不够的。

You're not gonna pull enough out of the atmosphere for that.

我们还可以从加拿大那边弄到一些。

And we can get some from our friendly can dos.

不过上次我查了一下，全世界重水堆加起来，靠产氚提供的聚变功率大概也就20兆瓦左右，而且还得是持续不断的。

But last time I I looked into this, the entire world heavy water fleet could supply something like 20 megawatts of fusion power constantly with their tritium production.

所以这条路走不通。

So we're not we're not pulling that one off.

我们最多能做的，就是利用加拿大CANDU反应堆以及其他来源的氚，因为出于国家安全等目的，有些反应堆会直接增殖氚。

That's the the most we can do is we can use the the tritium from candus and other sources because there is tritium bred directly in reactors for, like, national security purposes sometimes.

我们必须用它来启动增殖循环，就像铀钚增殖循环那样。

We have to use that to kick start a breeder cycle, just like the uranium plutonium breeder cycles.

你知道的，先投入一点，然后理想情况下，它会不断倍增、再倍增。

You know, you put a little bit in, and then ideally, you're multiplying and multiplying and multiplying.

所以我们会进行前面提到的中子倍增，然后这些中子得去找锂，用锂来制造氚。

So we'll do that neutron multiplication we mentioned, and then those neutrons have to go find some lithium, and they can make tritium from the lithium.

现在，理想情况下，大家经常提到的是锂-6，它占所有锂元素的约8%。

Now, ideally, what you'll see a lot of people talk about is lithium six, which is about 8% of all lithium.

浓缩锂也不便宜。

It's expensive to enrich too.

这点我不骗你。

I'm not gonna not gonna BS you on that one.

部分原因是人们很少浓缩锂。

Part of that is people don't enrich very much lithium.

你可以说，如果有一个大型的锂浓缩产业，我们肯定能降低价格。

You could argue that we could definitely drop those prices if there was a large lithium enrichment industry.

现在，它只是用于调节压水堆pH值的小众产品，仅此而已。

Right now, it's bay it's a niche product for, like, balancing pH and PWRs, and that's about it.

但你知道，更好的选择是天然锂，有趣的是，这完全可行。

But, you know, even better would be natural lithium, and the fun part is, actually, that totally works.

直接从地里挖出来的天然锂也很便宜。

Natural lithium right out of the ground is also cheap.

你知道，人们会谈论锂作为关键矿产的冲突，但它和其他电池矿物完全不一样。

You know, people will talk about conflicts over lithium as a critical mineral, but it's nothing like some of those other battery minerals.

对吧？

Right?

到处都是。

It's all over the place.

而且，说到用量，你知道，聚变仍然是核反应。

And, when you're talking about the amount, you know, fusion is still nuclear.

说到聚变反应堆需要的量，我就不知道了，老兄。

When you're talking about the amounts you need for a fusion reactor, it's like, I don't know, man.

你可以从很差的来源获取，只要你的聚变循环能工作，那也没关系。

You could get it from a crap source, and it still wouldn't matter if your fusion cycle was working.

就像，我们假设我们已经有了一个能工作的反应堆。

Like, we're, you know, we're assuming we have our working reactor here.

但难点在于，你现在制造出了你的锂。

The hard part, though, is that you now you make your lithium.

你的锂是盐、固体或液态金属形式，不管你的反应堆概念有多疯狂，现在你有气态氚，或者它被化学键合了。

Your lithium's in a salt, or a solid, or a liquid metal form, whatever whatever crazy reactor concept you have, now you have gaseous tritium, or it's bonded chemically.

而且你得把氚提取出来，所以你得有个化学处理设施。

And you have to pull that tritium out, so you have to have a chemical processing facility.

这玩意儿肯定得是实时的在线操作，因为你不能囤积太多氚，还得把它重新注入系统里。

It is always going to be it's gonna have to be some sort of online thing because you're not trying to hold a lot of tritium, and then you have to reinject it back into your system.

这里头还有个讲究，就是看你手里要压多少氚，也就是所谓的氚库存量，跟你的加注速度怎么平衡。

And there's a whole game about how much tritium you have to hold versus how fast you put it into what's called the tritium inventory.

大家对此都特别紧张，因为氖这东西可不好弄。

And people are very, very worried about the tritium inventory because tritium is not easy to get.

那些能做到的确实能提供一些，但你们手里可没有几千公斤可以交出来。

The can dos will give us some, but you guys don't have, you know, thousands of kilos to hand off.

你们也不想手里握着几千公斤对吧。

You don't want to have thousands of yeah.

你们也不想囤积几千公斤的。

You don't want to have thousands of kilos.

它是气态的，一旦泄漏就会弥漫整个房间。

It's gaseous and likes to fill the room if it leaks.

而且，你知道，它真的只是因为它衰变而产生热量。

And, you know, it just literally makes heat because it's decaying.

它衰变得非常非常快。

It's decaying pretty pretty darn fast.

所以，在优化反应堆的整个过程中，另一个挑战就是如何降低你的氚库存。

And so that's another game in this whole trying to optimize a reactor is how can you lower your tritium inventory.

如何燃烧最多的氚，或者最有效地燃烧氚，以保持那个数字低？

How can you burn the most tritium or the tritium most efficiently to keep that number down?

但是，这就像，你知道，如果我们说燃烧氘氘是我们的“能做”型聚变反应堆，如果你能让它工作的话。

But this this is like you know, if you if we say the if we say the the burning deuterium deuterium is our can do type fusion reactor if you can make it work.

但与“能做”不同，它更难。

But unlike the can do, it's harder.

所以它就像是更好的燃料，但更难的反应堆。

So it's like the better fuel, but the harder reactor.

氘氚绝对是增殖循环。

DT is is definitely the breeder cycle.

它完全等同于增殖反应堆。

It's full on it's fully comparable to the breeder reactor.

它具备增殖反应堆的所有复杂性，但现在点火反应要容易得多。

It's all the complexities of the breeder reactor, but now the reaction's way easier to make to light off.

这更像是在处理高浓缩铀之类的东西，确实如此。

It's it's more like dealing with highly enriched uranium or something where it's like, yeah.

我不打算说‘连傻瓜都能做’，因为聚变其实很难，但实现DT反应比DD反应容易得多。

Kinda not I'm gonna say any idiot because fusion's really hard, but it's a lot easier to make a DT reaction work than a DT reaction.

然后快速提一下，因为肯定有人会问到，氘-氦3反应听起来像那种科幻里的胡扯，根本原因是目前根本没有氦3的供应。

And then just, like, real quick because it'll get brought up, d helium three is, like, weird sci fi nonsense because there is no supply of helium three.

那些说可以从月球上开采氦3的人，估计是脑子不清醒。

People who say you can mine it off the moon are smoking something good.

我其实写过一篇论文，专门分析为什么这种说法很荒谬，里面详细解释了从能量角度来讲这完全说不通。你要是感兴趣，我可以发给你看看。

I actually have a paper that I wrote about why that's stupid that I can send to you about why the energetics of that don't make any sense.

不过确实，氦3时不时就会在科幻作品里被拿出来炒一炒。

But, yeah, helium three is something that gets brought up for science fiction every now and then.

它既有处理氘氚反应的那些复杂麻烦，又得去深空里找燃料，乐趣十足。

It's it's all the complexity of dealing with DT, but all the fun of having to go find your fuel out in deep space.

好的。

Okay.

行。

Okay.

所以到目前为止，我们聊的应该是……我想想，相当于裂变的那部分，就是开采、选矿、浓缩之类的流程，呃，抱歉，是转化和浓缩，反正就是这些步骤。

So so so far, we've talked about, I guess, the equivalent to fission would be, you know, whatever mining, milling, and enriching and, you know, converting sorry, converting and enriching, whatever.

搞清楚燃料包壳里面到底是什么。

Making what's inside the the the fuel cladding.

不过咱们还是来聊聊实际的燃料元件长什么样吧。

But let's talk about what that actual fuel element looks like.

我知道惯性约束和磁约束的情况会大不相同。

I understand it's gonna be quite different for inertial confinement versus magnetic confinement.

那它具体是什么样子的呢？

So what does it look like?

具体是怎么制作的呢？

How do you make it?

那为什么需要把它冷冻起来呢？

And then why do you need to why do need to freeze it?

咱们先说个大概念，磁约束聚变虽然把很多复杂的细节留到后面再聊，但它的优势在于燃料处理起来会简单得多。

To to use real broad, terms here, magnetic confinement, they trade a lot of complexity that we can get into later by trying to make their but the the trade off is their fuel's gonna be a lot easier.

简单来说，你直接把气态燃料喷进去就行。

The the general gist is you could just spritz in gaseous fuel.

超级简单。

Super easy.

说白了就是个高级点的化油器。

It's, you know, a glorified carburetor.

他们确实会注入冷冻 pellet，出于各种有趣又复杂的等离子体物理控制需求，整个过程非常精妙。

They actually do inject frozen pellets for all sorts of fun, complicated, like, plasma physics control reasons, and it gets very, very neat.

他们用这种电磁炮往托卡马克装置里发射东西。

They have these rail guns that shoot into tokamaks.

不过总的来说，他们的燃料挺简单的。

But overall, their fuel is pretty simple.

他们用的都是那种特别、特别、特别简单的燃料类型。

They go for very, very, very simple types of fuel.

惯性约束聚变需要高品质的燃料，而且规模更大。

Inertial confinement needs high quality fuel, and it's bigger.

你想想，就像那种三结构同质颗粒（TRISO）大小的燃料球，但现在随着我们远离国家点火装置（NIF），尺寸还在不断增大。

You know, think like triso pellet sized, but now getting bigger as we're going away from the NIF.

这里有个挺奇怪的现象：激光越小，靶子反而越难做；或者说，压缩装置越小，难度也越大。因为Z机也属于惯性约束聚变，设备越小，容错空间就越窄，所以每一步都得做得更完美。

There's a whole there's a weird trade where the smaller the laser, actually, harder the target is to make, or the smaller the thing doing the compressing, because the the z machine also does inertial confinement, because you have to be more perfect with how you do everything because you have less margin.

不过咱们先聊聊激光约束聚变。

But so it's these little we'll talk laser confinement.

就是那种特别小的燃料丸。

There are these little tiny pellets.

燃料丸的外层材料会根据不同的设计方案而变化。

What the outside of the pellet is varies on your design.

可以是塑料的。

It can be plastic.

NIF喜欢用钻石。

The NIF likes to use diamond.

我说过它们必须完美。

I said they have to be perfect.

他们从德国订购钻石靶丸。

They get diamond pellets sent from Germany.

当然，肯定是从德国来的。

Of course, it's from Germany.

哦，那是唯一能制造它们的地方。

Oh, it's the only place that can make them.

对。

Yeah.

而且

And

而且这些，像是经过抛光的，表面公差能达到，嗯，到那个……

And and these are, like, polished, and the surface tolerances are, like, down to the

哦，非常非常好。

Oh, very, very nice.

对。

Yeah.

好。

Okay.

不过有意思的是，我查了一下。

Interestingly enough, was looking it up, though.

公差比三坐标测量仪要好，但也没有好得离谱。

Tolerances are better than trizo, but they're not like wildly better.

不是说，你知道，差四个数量级那种。

We're not talking, you know, four orders of magnitude.

也就是一到两个数量级左右。

We're talking one to two ish.

而且实际上，随着我们做得更大，我们会在更大的靶丸上达到类似三倍频级别的质量。

And actually, as we get bigger, we'll start getting to like triso levels of quality on a on a bigger pellet.

激光器变大，靶丸也得变大，一切都会变得更好一些。

As laser gets bigger, pellet gets to get bigger, and everything gets a little nicer.

那么，NIF的靶丸尺寸是多大呢？

So what what is the size of a pellet at the at the NIF?

是几毫米吗？

Is it a couple millimeters?

是

Is it

对。

Yeah.

大概就是毫米级别的燃料靶丸。

Like a millimeter scale pellet of fuel.

但是，你有这个金刚石外壳，然后我听说你需要钻孔、掏空，再充入气体。

But, you have this diamond casing, and then you'd I heard you need to drill it, hollow it out, and fill it up with gas.

是

Is

那个，实际上它出厂时就是空心的。

that So, it actually comes hollowed out.

别问我他们是怎么做到的。

Don't ask me how they do that.

我完全不知道。

I have no idea.

谢谢德国。

Thanks, Germany.

对。

Yep.

是的。

Yeah.

除了钻石还有其他材料可用，但钻石是最经典、最有趣的话题，而且在NIF上性能最高。

And there's other things other than diamond that are used, but diamond is the classic fun one to talk about, and it's got the highest performance on a NIF.

他们会掺杂一些东西，通常会喷上铍之类的，来获得合适的X射线吸收。

They dope it, so they'll often spray it with beryllium or something like that to get the right X-ray absorption.

然后，它上面已经有一个小孔，我们称之为充气管，我们就用这个来构造这个小靶。

And then, it's already got a little hole in it that we call the fill tube, and so we construct this little tiny target.

它要放进一个金制的小罐子里，因为NIF的工作方式要求这样。

It goes inside this little gold can because of how the NIF has to work.

它必须放在一个叫“黑腔”的金制小罐里。

It has to be in a little gold can called a hole realm.

我们有金。

We got gold.

我们有钻石。

We got diamonds.

对，我们都有。

We got yeah.

是的。

Yeah.

有趣的是，它就像一层纳米级的金，外面再包一层其他材料来固定。

What's funny, it's like a nanometer of gold with an outer casing of some other material holding it.

但，对，我们就是把所有贵的东西都堆在一起。

But, like, yeah, we just have to add every expensive thing together.

铍也不便宜。

Beryllium's not cheap either.

然后他们注入氢，也就是氘和氚，它以液态进入，然后被冻成一层非常漂亮的层，他们还会让它衰变一段时间，这样衰变产生的热量能让这层变得更平滑。

And then they inject in the hydrogen, the deuterium tritium, and it comes in as a liquid, and they freeze it into this really nice layer, and they actually let it decay for a while so that the the decay heat, like, makes the layer smoother.

如果你见过，比如，你是加拿大人，你肯定喜欢冰球，那种非常纯净、漂亮的冰，你得重新融化才能达到那种效果。

If you've ever seen think like, you're you're Canadian, you must love hockey, that really pure, beautiful ice that you kinda have to remelt to get it to do that.

他们做的就是这件事。

It's literally what they're doing.

你说冷冻它，我们不是把它降到零下四度或零下十摄氏度。

When you say freezing it, we're not we're not taking it down to minus four, minus 10 Celsius.

我们要去

We're going

哦，不是的。

to Oh, no.

是18开尔文。

It's 18 Kelvin.

哇。

Whoo.

好的。

Okay.

非常冷。

Very cold.

非常冷。

Very cold.

挺有意思的。

It's funny.

之前大家花了不少功夫，想研究怎么更快地做出那种完美的冰层。但因为NIF出于各种原因没法发射得那么快，所以干脆就让氚自己衰变产生的热量把冰层弄得更平整，这样反而更省事。

There was a bunch of stuff done to, like, figure out how to make that perfect ice layer faster, but because the for a the NIF can't shoot very fast for all sorts of reasons, it's actually just easier to let the tritium's own heat make the ice really nice.

所以这些靶丸，它们像酒一样陈化。

So the these pellets, they age like wine.

让我理清一下，因为我们之前讲过这个，如果大家觉得难以理解，应该回去听那一集。

Just to get my head around it, because I didn't have we we have covered this before, and people should go back to that episode if they're feeling overwhelmed.

但我认为让大家感到有点复杂是很重要的，因为如果你只了解表面，你会错过森林里的很多树木。

But I think it is really important to kind of overwhelm people because if you only get this at a surface level, I mean, you're missing you're missing a lot of trees inside the forest.

对。

Yeah.

聚变的细节决定成败。

The devil's in the details on fusion.

完全正确。

Totally right.

所以你有一个两毫米的胶囊，里面含有，我想想，数不清的单个氘和氚原子。

So you got this, like, two millimeter capsule with some, like I think it's quintillions of individual deuterium and tritium atoms.

然后激光加热它的中心，开始融合其中一些。

And then the laser is kind of heating the center of it and starting to fuse a few of those.

我不确定，好像是喹啉之类的。

I'm not sure, like, quinoline or something.

不是。

No.

它没有开始点燃。

It's not starting to ignite.

激光没有加热任何东西。

The the laser is heating nothing.

这是有趣的部分，因为我知道你喜欢引擎。

This is the fun part because I know I know you love engines.

哦，X射线，宝贝。

Oh, x rays, baby.

对。

Yeah.

对。

Yeah.

所以激光进来，打到金罐上，产生X射线，整个装置，它的德语词是烤箱。

So the laser comes in, hits the gold can, and it makes x rays, and the whole realm is it's German word for oven.

这是一个X射线烤箱。

It's an x-ray oven.

然后，关于如何制造X射线驱动，还有一段内容，X射线实际上在烧蚀外部。

And, there's a whole bit about how you make that x-ray drive, and the x rays are actually ablating the outside.

它们在烧蚀掉那层钻石，这就像火箭一样。

They're ablating that diamond off, and it's a rocket.

这简直就是一个360度的火箭，从四面八方压缩这个，然后发生柴油反应。

It's literally a 360 degree rocket compressing this in, and it's a diesel reaction.

我们实际上就是在做，你知道的

We're literally just doing, you know

压缩这个。

Compression this in.

压力压缩做功，产生热量，把密度推到极限，中心开始达到劳森判据。

Pressure compression work that heats it up, that drives your density through the roof, and the very center starts to get to that Lawson criterion.

它开始发光，然后燃烧波向外扩散，就像柴油发动机一样。

It starts to light, and that burn wave goes out, and it's just like in a diesel engine.

你做的就是聚变柴油，对吧。

Is it is fusion diesel is what you're doing.

所以，就像等压的，我不想太深入

So, like an isobaric I don't wanna get too much

这个。

into this.

你们是在追求最完美的压缩。

Are you are trying to do the most perfect compression possible.

我们花了大量的金钱、时间和精力来思考如何尽可能完美地压缩这个东西，因为这关系到能量输出。

We spend an absurd amount of money and time and effort thinking about how to compress this thing as perfectly as possible because that drives your energy.

压缩得越完美，需要投入的能量就越少，反应效果也越好。

The more perfectly you can compress, the less energy you have to put in, and the better the reaction is.

好的。

Okay.

那么说到点火，就像，我是说，想想看，比如火，比如篝火。

And then so in terms of, like, when we're talking ignition, like, I mean, just thinking about, like, I don't know, like a fire, like a campfire.

这是大家最熟悉的能量形式。

Most familiar with in terms of of energy.

你知道，我是个困在城市里的牛仔。

You know, I'm a cowboy trapped in a city.

不管怎样，说到NIF的成就，我觉得人们的印象可能是，它产生了能量，然后持续了……你知道，你听说托卡马克创纪录，持续了几百或几千秒，我相信。

In any case, in terms of what was achieved at NIF, like, I think people's perception might be that this, you know, produced energy and it went on for you know, you hear about records being set at Tokamak, and it's going for hundreds or thousands of seconds, I believe.

二十二分钟，我想，是法国的世界纪录。

Twenty two minutes, I think, is the world record in France.

但是，我们通过X射线，开始融合一些氘氚原子，然后它就像滋滋作响，蔓延到其余那些千万亿个氘氚原子或核……我不知道。

But so we were we were we through these X rays, starting to fuse some of these DT atoms, and then it was sort of sizzling and spreading through the rest of those quintillions of d t atoms or nuke I don't know.

核原子。

Nuclear atoms.

我有点搞混了。

I'm getting confused here.

然后就结束了。

And then it's over.

那么，这个点火持续了多久？是纳秒还是皮秒级别的？

So so in terms of, like, how long that ignition lasted for, are we talking, like, nanoseconds, picoseconds?

激光开启，粗略来说，大约十到十纳秒。

The laser turns on, round numbers here, ten ten nanoseconds.

那是驱动圆柱体的能量。

That's the drive for the cylinder.

然后反应，所有能量在大约五十皮秒内释放出来。

And then the reaction, all of that energy is released in about fifty picoseconds.

哇。

Wow.

就像爆炸一样。

Just boom.

这确实是一场爆炸。

It is an it is an explosion.

那东西的峰值功率简直离谱。

The peak power of that thing is insane.

它爆炸时的峰值功率高达好多好多拍瓦。

It's like many, many petawatts of peak power as it goes off.

所以，基本上，我们就是在引爆一堆小氢弹。

And so, again, basically, we're setting off a bunch of little hydrogen bombs.

差不多吧。

Pretty much.

对。

Yeah.

如果我们想持续几秒或几分钟，每秒或每分钟需要有多少个这样的惯性约束燃料胶囊被激光击中？

How many do we need to settle if we were gonna try and sustain this for seconds or minutes, how many of these little inertial confinement fuel capsules do we need to be hitting with the laser per second, minute?

我不知道。

I don't know.

所以我们不会像NIF那样做。

So, we wouldn't do it like the NIF.

就像我说的，NIF只是勉强够用。

Like I said, the NIF is marginal.

我们会建更多的公司，或者尝试建一个更大的设施，更大的设施会获得更多的增益，所以投入同样的激光能量，你会得到多得多的聚变。

We would build as many companies, or trying to do, a bigger facility, and a bigger facility would get more gain, so you would get a lot more fusion for the same amount of laser you put in.

那可不是1980年代那种低效的激光。

That's or a They're a not 1980s efficient laser.

不仅仅是那样。

It's not only that.

这是一个非线性过程。

It's that this is a nonlinear process.

所以当你稍微加大压缩，我们把激光做得更大一点，把靶丸做得更大一点，我们就能得到多很多倍的能量输出，这正是NIF让人头疼的地方之一，NIF就像是在悬崖边上。

So as you compress it a little bit harder, we made our laser a little bit bigger, we made our capsule a little bit bigger, we get many, many times more energy out, and that's that's one of the pain in the asses of the NIF is the NIF is, like, right on this ragged edge.

你只是让引擎咳嗽着启动，而不是像一台平稳运转的大型柴油机。

You're just you're getting this coughing start on the engine rather than, like, a big, smooth diesel running.

所以你会把它做得更大。

So you'd make it bigger.

说实话，你得把它做得大得吓人。

You'd make it frighteningly bigger, to be honest.

要建发电厂的话，每次点火就相当于引爆几百公斤TNT的爆炸。

To make a power plant, oh, you're going we would be setting off, like, the equivalent of hundreds of kilos of TNT worth of explosion every time this thing fires.

每次发射。

Every shot.

每次发射。

Every shot.

所以到那时候，你面对的就是非常大的爆炸了。

So you're you're talking about very big explosions at this point.

那现在的燃料胶囊有多大？如果是一毫米的话……

So how how big are are fuel capsules now if they're a millimeter

现在我们要说的不是一毫米，而是几毫米的直径。

So now we're now we're talking, say, rather than one millimeter, it's a couple millimeters across.

好的。

Okay.

那没问题。

That's Okay.

不需要大太多，但这时候你开始谈论的就不是八兆焦耳了，而是吉焦耳级别的能量。

Doesn't need to be that much bigger, but now you're starting to talk, you know, like, gigajoules of rather than eight megajoules, gigajoules.

给你个概念，四吉焦耳就相当于一吨TNT的爆炸能量。

So for some context, four gigajoules is a ton of TNT.

而且，想想也挺疯狂的，一个普通的核电站每秒钟释放的能量也差不多相当于一吨TNT。

And also, that's kind of a wild thing to think about, that a a normal nuclear power plant is also releasing a ton of TNT about every second.

所以规模其实是一样的。

So it's the scale is still is the same.

只不过你现在得用一台差不多就是改装过的柴油发动机来实现这个。

It's just you have to now do this with a a glorified diesel engine, effectively.

所以你得每秒或每两秒来这么一次。

So you're doing this once every second to couple of seconds.

关于如何设计这个系统，人们正在研究很多不同的方案，看看哪种最合理。

There's a big, like, trade space that people are looking at as to what what makes the most sense on how you design the system.

我是说，我们从来没做过这种事。

I mean, we've never done this.

对吧？

Right?

我们有爆炸驱动的发动机，但规模完全没这么大。

We've we have we have explosion powered engines, but nothing quite of this scale.

最接近的可能是大型游轮的柴油机。

The closest you could look at is, like, big cruise ship diesels.

船用柴油机。

Marine diesels.

对。

Yeah.

对。

Yeah.

船用柴油机。

Marine diesels.

但即便是那些，我们还得在能量释放上再高出一个数量级。

But even those are still we still gotta go like an order of magnitude above those in energy release.

再说，你的能量释放大约有20%是阿尔法粒子，它维持着点火，维持着聚变反应，但最终所有能量都会释放出来。

And again, your energy release is about 20% alpha, which sustains the ignition, sustains the fusion reactions, but you still all comes out.

单个燃料靶丸持续大约一秒钟，然后你就需要下一次激光脉冲。

The of the individual of the individual fuel pellet that lasts for a second, and then you need that next laser pulse.

所以总是有，好吧。

So there's always okay.

就像，激光传递能量转化为X射线，触发每个新的靶丸点火。

Like, there's energy being delivered by the laser transformed into X rays, which triggers each new capsule to go.

它们正在点火。

They're igniting.

它们产生大量阿尔法粒子，你知道，帮助其余的燃料靶丸滋滋作响并聚变。

They're producing lots of alpha, which is, you know, helping the rest of the the fuel capsule sizzle and and fuse.

相比之下，磁约束的情况是喷入氘氚气体，你仍然有很大的寄生负载，大量电力用于冷却磁体、控制系统、运行你的增殖层等等。

Compared to a magnetic confinement situation where you're spraying in deuterium, tritium gas, you still have a big parasitic load, a lot of electricity going into cooling magnets, control systems, running your breeding, whatever

也把等离子体加热了。

heat the plasma too.

就像，磁系统，嗯，我们回头再谈。

Like, the magnetic so the magnetic systems, yeah, we'll circle to those.

但一旦启动了，你就不需要持续输入新的激光脉冲了。

But once it's started, you're not needing to continuously input new kind of laser blasts.

点火似乎能靠自身维持更久，还是不是这样？

The the ignition seems like it lasts longer under its own power, or no?

这就是很多磁约束系统的秘密。

So that's that's like the dirty little secret in a lot of these magnetic systems.

理论上，你可以达到所谓的无限Q值，就是你的容器完美，等离子体加热后，自己维持燃烧，就像一个完美的小型聚变炉。

In theory, you can get what's called q infinity, where you could get you could have your your bottle be perfect, and you get your plasma hot, and your plasma just self eats and runs, and you have your perfect little, balanced fusion furnace.

没人知道怎么做，甚至不知道是否可能，因为等离子体自身会引发很多不稳定性，聚变也会在等离子体中引发不稳定性。

Nobody has any idea how to do that, if that's even possible, because there's so many instabilities that the plasma drives, that the fusion will drive in the plasma.

目前的做法是注入燃料，喷入气体，然后进行一系列加热步骤，用微波、粒子束、大电流等等。

The current way that they're going for it is you inject in your fuel, you spray your gas, you do a bunch of different steps of heating, you're using microwaves, and particle beams, and big of current, and all this other stuff.

哦，对。

Oh, yeah.

对。

Yeah.

对。

Yeah.

这挺疯狂的。

It's pretty crazy.

你把它加热，然后保持住，不管多少秒。

You get it hot, and you hold it for however many seconds.

说实话，当你进入自加热阶段，也就是它要输出能量的时候，脉冲会变得更短。

And to be honest, as you get to the self heating portion where it's gonna give you energy out, it's going to be shorter pulses.

你看到的那个二十二分钟，比如托卡马克的记录，或者其他磁约束聚变的记录，通常可以这么理解：其他条件都很糟糕。

That twenty two minutes that whenever you see, like, a Tokamak record, or any kind of MCF record, it's usually a good rule of thumb that the rest of the conditions are pretty crap.

可能温度很低，密度也很低，离真正的聚变还差得远。

It's probably pretty cold, pretty low density, very far from doing fusion.

所以当你要达到这种自热状态时，他们会启动它，把它提升到那种炉膛条件，然后保持住，希望他们不需要加热太多。

So when you're gonna get to this self heating, they're gonna turn it on, they're gonna get it up to that furnace condition, they're gonna hold it, and they hopefully don't have to heat that much.

他们可能还在加热。

They probably are still heating.

某种加热还在进行，但等离子体会自热，如果设置得当，它产生的能量会超过你输入的能量。

Some aspect of heating is still going on, but the plasma will self heat, and it will make more energy than you're putting in if you set it up right.

而且随着人们对这些设计越来越现实，我们到了这样一个地步：我们知道的足够多，以至于发表的设计、正在建造的机器都是可信的，那么他们就会得到他们认为会得到的数字。对于惯性约束聚变，任何惯性约束都能获得非常高的增益，是输入能量的数百倍，这是你需要的，因为你的驱动器、激光或脉冲功率可能效率不高。

And the the kind of general rule of thumb, as people are getting more realistic with these designs, like, we're we're getting to the point where we know enough where the designs that are being published, machines being built are, you know, believable, that then they're gonna get the numbers they think they're gonna get, ICF, any kind of inertial confinement gets very high gains, hundreds of times the energy you put in, which you need because your drive your laser or your pulse power is probably more on the inefficient side.

磁约束聚变，增益相当低。

MCF, magnetic confinement, gets pretty low gains.

比如麻省理工学院的SPARC，我想他们目标是5倍，也许10倍于输入，而对于惯性约束聚变，我们说的是200、300倍的增益。

Spark, the guys out of MIT, I think they're going for five, maybe 10 x over that, versus, you know, for ICF, we're talking like 200, 300 sort of gains.

而这一切都回到了真正的关键问题上。

And, this all circles back to the real killer.

我认为聚变能源的两个关键问题，能告诉你是否能取得进展的，是氚库存和增益。

I think the two killers on fusion energy that are gonna tell you if you can get anywhere are tritium inventory, and gain.

你实际的工程增益是多少。

Your actual engineering gain.

你提到的那个寄生负载是什么样的？

What is that parasitic load like you brought up?

因为电厂可能会有寄生负载。

Because a power plant can have a parasitic load.

我去查了一下。

I went and looked this up.

AP1000大约有8%的功率用于寄生负载。

An AP1000 puts about 8% of its power in a parasitic load.

所以，是运行反应堆冷却剂泵吗？

So, running reactor coolant pumps?

对。

Yeah.

是的。

Yep.

好的。

Okay.

而且，你可能还能再加上，大概可以凑个整，再说个2%给设施的其他部分。

And, you could probably put another You could probably round up and say another 2% into the rest of the facility.

所以，如果你在大约10%的水平，你就能撑过去。

So, if you're at about 10%, you can survive it.

但这会给系统的其他部分带来一些严峻的数字。

But that's that puts some harsh numbers on the rest of your system.

对吧？

Right?

这意味着你的效率必须很高，或者你的增益必须很高。

Means your your efficiency has to be high and or your gain has to be high.

这两者的乘积必须相当高，因为你仍然是以热功率的形式输出这些。

The multiplication of those two things has to be pretty darn high because you you're still getting this out as thermal power.

对吧？

Right?

说到底，你就是在加热冷却剂。

In the end, you're heating coolant.

你得经过涡轮机，所以你可以先扣掉

You have to go through a turbine, so you can already knock

好的。

Okay.

对，差不多。

About Yeah.

是的。

Yeah.

你可以扣掉大约三分之二的能量，那部分就是没法利用的热量。

You can you can knock, like, two thirds of the energy off is just heat that you don't get to use.

靠。

Shit.

所以你看，要是真打算这么干，经济上是很残酷的。

So that you can see why, like, the the the economics of this are brutal if you're gonna do it.

你必须尽可能提高效率增益。

You you have to drive that gain in efficiency as high as possible.

好的。

Okay.

而且，我的意思是，再次谈到你的能量转换，那就是中子被某种东西吸收，无论是陶瓷还是熔盐。

And, I mean, in terms of, again, your power conversion, that's neutrons being absorbed by something, whether it's just like a ceramic, a molten salt.

对。

Yeah.

因为我们必须阻止中子来制造氚。

Because we have to stop our neutrons to make tritium.

我们不想让任何快中子泄漏。

We don't wanna let any Very neutrons fast.

对。

Yeah.

对。

Yeah.

我们得在增殖层里把它们拦住。

We gotta stop them in our breeding blanket.

有些中子会加热其他东西，但理想情况下，它们应该在增殖层里停下来。

Some of them will heat other things, but ideally, they're stopping in the breeding blanket.

剩下的能量，你提到的α粒子，占能量的20%，还有你输入的能量。

The rest of the energy, you talk about the alphas, the 20% of the energy that's alphas, and whatever you put in.

当然，你输入的能量最终都会变成热量。

Of course, whatever you put in does come back as heat.

对。

Right.

是的。

Yep.

是的。

Yep.

这些基本上都会变成X射线，所以必须被壁面或者面向等离子体的表面吸收。

That basically all comes back as X rays, so that has to be absorbed into the wall, or some sort of surface that's gonna be facing the plasma.

他们称之为等离子体面对的表面，或者是一个大的研究领域。

They call it plasma facing surfaces, or like a big area of research.

所有这些都转化为热量。

All of that turns into heat.

每一部分都是热量，你之后必须进行正常的热循环处理。

Every bit of it is heat that you then have to do a normal thermal cycle on.

排除其中一些热量难吗？

Is it hard to reject some of that heat?

听起来中子可能比阿尔法粒子更容易处理。

Like, it sounds like the neutrons might be easier versus the alphas.

难以置信地难。

Unbelievably hard.

是的。

Yes.

好的。

So Okay.

不是

It's not

衰变热。

a decay heat.

当你关掉它的时候，它不会持续下去，也不会成为管理效率低下的噩梦。

When you shut it off, like, it's not gonna last and be the nightmare that management is inefficient.

还有衰变热，我们可以讨论一下。

There is also decay heat, which we can get into.

但如果我们说的是初始加热，初始热负荷，那么如果我们用ICF，我说过它的峰值功率非常高。

But if we're talking about the initial heating, the initial heating load, so if we go ICF, I said it's a really high peak power.

对吧？

Right?

所以你必须非常小心墙壁的距离，以免墙壁被烧蚀回系统。

So you have to be very careful on how close your walls are, so you don't ablate the walls back into the system.

ICF可以有一个非常简单的腔室。

ICF gets to have a very simple chamber.

我们可以就用一个大的钢制容器。

We can just have a big steel vessel.

我们不需要做什么特别复杂的事，因为我们不用维持磁场。

We don't have to do anything crazy because we're not holding magnets.

激光从孔里穿过来，诸如此类。

The lasers come through holes, that sort of thing.

对吧？

Right?

激光是安全的。

The lasers are safe.

它们离得很远。

They're far away.

但是腔体要承受巨大的爆炸，所以你得考虑怎么保护它免受那种可怕的X射线冲击。

But the chamber is withstanding a big explosion, so you have to think about how you're going to protect it from that horrible X-ray blast.

中子也是个问题，但X

The neutrons are also a problem, but the X

这不是压力波爆炸。

It's not a pressure wave explosion.

是X射线冲击。

It's X-ray blast.

是X射线和中子冲击，而且很多设计都会在腔室内使用流动冷却剂，这样可以把X射线冲击转化为压力冲击，因为压力冲击比X射线冲击更容易应对。

It's an X-ray and neutron blast, and then you might have a lot of these designs involve flowing coolant in the chamber, in which case you can turn the X-ray blast into a pressure blast, because you would rather have a pressure blast than an X-ray blast.

X射线冲击，如果只是承受X射线冲击，比如在NIF上，我们知道NIF本来就不是为了永久运行而设计的，也不是为了作为发电厂。

An X-ray blast, if you're just taking an X-ray blast, like on the NIF, right, we know the NIF was never meant to run forever, and it's not meant to be a power plant.

但我们知道，随着我们从中获得越来越多的聚变能量，它自身也会受到损伤。

But we know as we get we get more and more fusion out of it, it damages itself.

而且它有一个使用寿命，可以设定它能承受多少次射击，直到它的壁被烧穿。

And there is a lifetime that you could put on it as to how many shots before it will have just, like, burnt through its walls.

所以你不会想那样做。

So you you don't wanna do that.

磁约束还有一个额外的问题，就是等离子体废气必须排出，也就是那个氦灰。

Magnetic confinement has an extra problem of the plasma exhaust has to come out, that helium ash.

所以在惯性约束聚变中，就是全部向外喷发。

So in ICF, it's just all blowing out.

你其实不太在意。

You don't really care.

它均匀地向各个方向扩散，和X射线相比不算什么大问题。

Goes evenly in all directions, and it's not a big deal compared to the X rays.

磁约束聚变有中子，这些中子本身就是个噩梦，因为它们的磁铁离得很近。

MCF has the neutrons, which are, their own nightmare because their magnets are close.

你的磁铁现在不得不应对中子。

Your magnets are now having to deal with the neutrons.

X射线会击中各种东西，四面八方都有。

X rays hitting stuff, which you get in every direction.

虽然不是爆炸，但量很大。

And it's not a blast, but it is a lot.

对吧？

Right?

比如说，一吉瓦就是一吉瓦，不管你怎么做都是这样。

Like, a gigawatt's still a gigawatt, no matter how you're gonna do this.

然后等离子体废气本身，必须排出来才能让反应持续运行几秒钟，你得把那些氦气排到某个地方，而且它是以一股可怕的射流形式喷出来的。

And then the plasma exhaust itself, which has to come out to keep let the reaction keep running for seconds, you have to throw that helium somewhere, and it's it's coming as a horrible jet.

总是有这些闭合的磁力线，等离子体沿着它们走。

There's always these, like, closed field lines where the plasma goes.

在偏滤器那里，那是他们给等离子体射流撞击的部件起的名字，比如在托卡马克或仿星器上。

And at the diverter, that's the the component they call where that jet of that's what they call where that jet of plasma goes on, like, a tokamak or a stellarator.

那里的热负荷相当于甚至超过航天器再入大气层时的水平，但那是等离子体射流在撞击它，而且还有辐射。

The heat loads are equivalent or worse than reentry from space, but it's a jet of plasma hitting it, and there's radiation.

所以这确实是一个很有挑战性的工程问题，这么说都算是轻描淡写了。

So it's it is a challenging engineering problem, to put it lightly.

这是个有趣的工程问题，但也确实有挑战性，而且正是这类事情影响了我——你知道，我之所以研究我搞的那种聚变，是有原因的。

Is a fun engineering problem, but it is a challenging engineering problem, and it's things like that have colored my you know, there's a reason I work on the kind of fusion I work on.

你已经做出了选择。

You've made your choice.

对。

Yeah.

对。

Yeah.

对。

Yeah.

我看着同事们讨论这些，觉得挺有意思的，我也支持他们，但天哪，有些问题真的很难。

I I am very entertained watching my colleagues talk about this stuff, and I am rooting for them, but boy, some of those problems are hard.

没错。

Right.

好。

Okay.

我觉得我们已经聊了很多，虽然这个话题可以无限深入，但我们已经深入讨论了聚变是什么、燃料的细节、装置本身等等。

So I think we've I mean, there's there's like an infinity of depth we could go into, and I think we've we've gone into a good amount of depth in terms of, you know, what what fusion is, what and and the devils of the details of of the fuel and and the the devices themselves, etcetera.

我想稍微换个话题，聊聊发展方面的事情。

I wanna switch switch a little bit into talking about, I guess, like, development.

比如，我们之前谈到的那些初步的台架实验，如果我们可以这么称呼它们的话。

Like, we talked about the inertial sorry, the initial sort of bench experiments, if we wanna call them that.

托卡马克装置已经存在很长时间了。

Like, we've had tokamaks for a long time now.

对吧？

Right?

苏联人在五六十年代就有了托卡马克装置？

Like, the Soviets had a tokamak going fifties, sixties?

我相信是五十年代末发明的。

Late fifties, I believe, when it was invented.

那真是不可思议。

Like, that's wild.

它们已经存在很久了。

They've been around for a long time.

好的。

Okay.

那惯性约束这方面，我们是从什么时候开始有的？

And in terms of inertial confinement stuff, we've had that since?

这得看你想从什么时候算起，不过，是的，差不多同一时期。

Depends when you wanna put the start date, but, yeah, similar.

我是说，那篇著名的约翰·纳克尔斯论文，就是讲惯性约束的那篇，也是那个年代的。

I mean, the the the in the famous John Knuckles paper, which, like, talked about friendly inertial confinement, is also that that same time frame.

说实话，激光直到六十年代才发明，而第一次激光聚变——这对你来说是个有趣的历史故事。

To be fair, the laser wasn't invented till the sixties, and the first the first laser fusion this is like a fun historic thing for you.

第一次激光聚变是由一家私人公司完成的，也是第一家聚变初创公司。

First laser fusion ever was done by a private company and the first ever fusion startup.

KMS Fusion在威斯康星州做到了这一点，因为当那些最初的激光聚变论文发表时，所有人都觉得——那是在我们了解到这有多难之前。

KMS Fusion did it out in Wisconsin, because when those initial laser fusion papers came out, everyone looked at this is before we learned the realities of how hard this is.

他们看到后，心想：天哪。

They looked at it, they went, oh my god.

我们要搞激光聚变了。

We're gonna make laser fusion.

我们大概十年内就能搞定。

We're gonna do it in, like, a decade.

这玩意儿会特别容易成功。

This is gonna work so easy.

这会让所有磁约束的东西看起来像个笑话。

This is gonna make all this magnetic stuff look like a joke.

结果到了2026年，这事儿没成，但当时确实激励了一些人押注聚变能源，还成立了一家初创公司，后来跟原子能委员会闹出了麻烦。

And here we are in 2026, and that didn't quite work out, but it inspired some people to actually take a bet on fusion energy back then and make a startup that got in trouble with the AEC.

那段历史特别疯狂，CEO在国会作证时去世了。

There's a whole, like, there's a whole huge crazy history there, and the CEO died testifying to Congress.

我是说，这太离谱了，但那就是……

I mean, it's it's nuts, but that was the Died

在国会作证的时候去世了？

during test testimony to Congress?

对，就是在作证的时候去世的。

Died during his testimony.

他是不是被中情局用那种带毒镖的，你知道，那种果冻状的毒药暗杀的？

Did he have, like, a CIA dart with the, you know, the jelly poison ergo?

不是。

No.

他是个大块头，喜欢吃汉堡，最后身体吃不消了。

He was a he was a bigger man who liked his hamburgers, and it eventually caught up to him.

哦，哇。

Oh, wow.

对，确实。

Yeah, it

太离奇了。

says wild.

这故事太离奇了。

It's a wild story.

我强烈建议你去查查。

I would highly recommend looking it up.

亚历克斯·韦勒斯坦在他的某本书里提到过，就是那个KMS Fusion。

Alex Wellerstein has it in his in one of his books, the KMS Fusion.

所以，我们在这个项目上已经投入了几十年的努力。所有这些项目都很早就开始了，而最初的情况是，我们都以为……不，你和我那时候还没出生呢。

So, we have decades decades of work on this, this initial All of these projects started a long time ago, and what it really was is initially, we all thought it was no, we, you and I weren't alive.

大家都以为这会很容易。

Everyone thought it was gonna be easy.

他们当时因为裂变和原子弹的成功而信心满满，甚至包括早期的氢弹工作。

They were riding so high off fission and the bomb, and even, like, the initial hydrogen bomb work.

他们觉得，哥们儿，这我们搞定了。

They were like, dude, we got this.

这根本不算什么。

This is gonna be nothing.

我们最迟到八十年代就能搞出聚变反应堆。

We're gonna have fusion reactors by, like, the eighties at the latest.

还有谢伍德计划，那是美国关于这个的机密项目，因为我们当时觉得，聚变很快就能实现。

And there's project Sherwood, was which the classified US project on it because we went, oh, we're gonna get fusion so fast.

我们最好把它保密，这样才能领先于那些该死的共产主义者。

We we better keep it classified so we can keep ahead of those damn dirty commies.

然后我们发现等离子体有很多挑战，最终，你知道，大家不得不开始越来越多地分享、交流和合作，就在冷战结束时，这些积累成了里根和戈尔巴乔夫之间的协议，后来变成了ITER，因为，你知道，随着资金越来越紧张，我们可以讨论的聚变反应堆有无数种。

And we learned that there is a lot of challenges in plasmas, and eventually, you know, everyone had to start sharing and talking and collaborating more and more, and right at the end of the Cold War, that accumulated in the agreements that later became ETR between Reagan and Gorbachev to make this because, you know, as as money got tighter, there's there's, like, infinite versions of fusion reactors we can talk about.

我们可以讨论，但他们开始逐渐缩小范围。

We could talk But for they started to kinda narrow down.

托卡马克开始达到所需的性能。

Tokamaks were starting to get the performance that was needed.

激光聚变看起来仍然很有希望。

Laser fusion still looked promising.

它与武器计划有联系，所以这两个是将会不断进步的。

It had its connections to the weapons program, so it was going those two were the ones that were just gonna keep eking up.

对。

Right.

是的，他们达成了这些重大协议。

And, yeah, they made these these big agreements.

KMS Fusion 很遗憾地倒闭了，他们在激光聚变方面也没走多远，但后来，你知道，我们进入了2000年代，那时候第一批聚变初创公司才真正出现，你开始看到更多在 Eater 之外的工作，更多人参与进来。我觉得人们对 Eater 和 NIF 有点失望，它们好像吞噬了所有聚变的兴趣，我认为这很可能推动了很多这些初创公司的成立。

KMS Fusion sadly went under, and they didn't get that far on laser fusion, but and then they you know, we we started getting into the aughts, and that's when the first Fusion startups actually came out as and you started to see more work outside of Eater, more people getting I think people were getting frustrated with Eater and Nift kind of, like, eating all of Fusion interest, And I I think that's probably what drove a lot of these initial startups.

有加拿大的 General Fusion，还有加州的 TAE Technologies。

There's the one in general general fusion up in Canada, TAE Technologies down in California.

那些都是2000年代早期的。

Those were like early aughts.

那些科学家喜欢一些旧的想法，认为我们应该重新审视它们，并且能够说服投资者去承担一个，说实话，疯狂的风险，尤其是在那个时候。

Those were scientists who liked some of these old ideas, thought we should revisit them, and push that, you know, were able to convince investors to to take a, honestly, crazy risk, especially at that point.

这些年来一直在积累，非常有趣，就像在这个领域成长起来，看到资金在积累，兴趣在增长。

And it's just been building over the years, and it's been very interesting, like, being like, kinda coming up in this field and seeing as the as the money's building, as the interest is building.

积累了多少资金？

How much money is building?

有几百亿吗？

Like, it's it's tens of billions?

让我想想。

Let's see.

那是多少？

What is it?

我昨天看到的数据是174亿美元，自2010年以来，私人领域对聚变能的投资已经达到了这个数。

17,400,000,000 is what I saw yesterday since 2010 has been invested into Fusion on the private side.

那这算多还是少？

Like, is that a lot or a little?

因为你听说的，比如AP1000的开发成本。

Because you hear, like, you know, the development cost of AP 1,000.

我的意思是，他们开始建Vogtle的时候设计还没完成，但成本已经是几十亿美元了。

Mean, I the design wasn't done when they started Vogtle, but was in the billions.

这个问题问得好。

That is the correct question.

既多又少，多的是居然有这么多私人资金投入进来。

It's both a lot and a little, and that it's crazy that there's that much private money.

而且很酷的是，它推动了很多之前被搁置的项目。

And it's cool that it's pushing forward so many things that got dropped.

有很多这样的想法，你也可以说是出于政治原因被搁置的。

There's a lot of these ideas that you could argue got dropped for politics as well.

我可以告诉你，大科学项目充满了大政治。

There's big science I can can tell you, big science is full of big politics.

所以从这个角度看，看到这些进展确实令人兴奋。

And so it's exciting on that hand to see this.

我还觉得，基本上所有的钱都集中在美国和中国。

I think also because basically all the money is in The US and China.

美国大概投入了100亿美元，中国大约50亿，当然还有我们的政府项目。

There's like about 10,000,000,000 has been invested in The US, 5 ish billion in China, and then we have our government programs, of course.

这些项目并没有完全消失。

Those haven't exactly gone away.

我觉得这也有点像是一种声望的较量，而且在美国和中国，私营部门和公共部门之间有一种有趣的互动，有时候很难分清是谁在推动谁前进。

I think part of it is a prestige play too, and there's an interesting interplay between the private sector and the public sector, both in The US and in China, where it's a little hard to tell, like, who's pushing who forward.

不过，是的，我们确实继续看到大量的资金投入。

But, yeah, we're we're continuing to see lots of billions.

具体来说，我想特别提一下，我相当确定CFS，就是麻省理工那个托卡马克团队，是资金最充足的私营核能公司，也就是私营核能初创公司。

And, specifically, I wanna call out, I'm pretty sure CFS, which is that tokamak group out of MIT, is the most well capitalized private nuclear company, like, private nuclear startup This

是联邦聚变系统公司。

is Commonwealth Fusion Systems.

联邦聚变系统公司。

Commonwealth Fusion Systems.

对。

Yeah.

他们刚刚又筹集了十亿美元。

They just raised yet another billion.

他们这些年总共筹集了大约560亿美元。

They're at it's like $56,000,000,000 that they've raised over the years.

用于磁约束。

For magnetic confinement.

对。

Yeah.