Math student builds fusion reactor at home with help from Claude AI and $2,000

[zohaibahd]

In a nutshell: An unlikely contestant has emerged in the race toward cracking the fusion code, which has seen the biggest labs and richest billionaires pouring billions into various projects: a university student who assembled a reactor right in his bedroom in just four weeks. He did this by spending around $2,000 on some off-the-shelf parts he ordered online.

Hudhayfa Nazoordeen, a math major at the University of Waterloo, created a mini fusor that resembles an actual tokamak and hooked it up to a humble 12kV neon sign transformer for power. The result was a setup that could successfully produce plasma, which is where the fusion magic happens.

With "zero hardware experience," he spent week one just sourcing and figuring out all the components needed from suppliers like McMaster-Carr. Week two involved the nuts and bolts of assembling the main chamber and rectifier circuit. By week three, he had the whole thing set up in his bedroom and began tinkering away to get that neon transformer integrated.

However, it was cracking the vacuum system in "week 3.5" that seemed to really put Nazoordeen's perseverance to the test.

"This was by far the most annoying part of this project," he admitted in a thread on X/Twitter describing the project. He had to hunt down and seal multiple tiny leaks to ultimately achieve a vacuum of 25 millionths of an atmosphere.

Achieving a vacuum environment is important because fusion requires ridiculously low pressure to let the nuclei get close enough to fuse. Nazoordeen opted for an MKS-901p transducer to monitor and control the vacuum continuously.

in a couple weeks, I built a nuclear fusor in my bedroom – with zero hardware experience

the secret? Claude sonnet 3.5 + projects

a glimpse into the process below pic.twitter.com/H4261f5bCy

– HudZah â (@hud_zah) August 23, 2024

Besides the help he received from other engineers on campus, Nazoordeen also gushed about Anthropic's Claude 3.5 AI chatbot, which also played a major role in the development of the reactor.

"I fed Claude all my datasheets, and it helped a ton with this," he said.

Of course, this homemade fusor didn't quite make it to inducing fusion this time, so it doesn't technically emit neutrons. Achieving that is extremely challenging in a small design like this. It would most likely require significant further engineering and potentially moving to more advanced reactor concepts.

However, Nazoordeen seems to have more planned for this little project and says he's waiting for funding for the "full fusor."

Nazoordeen's work builds on earlier exploits of Olivia Li, an engineer from the University of Toronto, who last year constructed her own fusion reactor in a New York City apartment using deuterium gas extracted from heavy water. Li praised Nazoordeen's achievements.

"A lot of people I've talked to have been excited about building a fusion reactor. Hudzah is the only person to have actually went on and executed!" she wrote on X. She also shared a link to a write-up that she says would help anybody else wanting to make their own fusor at home.

Image credit: Hudhayfa Nazoordeen

Permalink to story:

Math student builds fusion reactor at home with help from Claude AI and $2,000

 

[WhiteLeaff]

He looks like someone who never sleeps.

 

[Dr Roboto]

I want to give this guy a pat on the back, because it would be quite the feat given his lack of experience and accomplishing it in only 4 weeks. However, in this day of fake everything (including peer reviewed tech publications), I find it hard to believe this is true without an independent assessment.

Genius or internet hoax?

 

[wiyosaya]

Having gone through what it takes to "home build" a vacuum system many years ago, I empathize with his pain at hunting down leaks. If one is attempting to seal such leaks, you either need something like better gaskets, or if they are pinholes, you need an epoxy that does not "outgass" at such low pressures. I am not sure of the pressures I got to, but they were lower than 0.001-torr. I was able to "bury" my thermocouple gauge at the low end of the scale, but I also had a diffusion pump and a cold trap along with a mechanical pump similar to the pump he has. The gauge he has is more sophisticated than the one that I had at the time.

I cannot see his vacuum pump model, so its impossible to tell the limit pressure limit of a pump like that without knowing the model number. But 25-millionths of an atmosphere is approximately 0.019 Torr (which is a more common unit of vacuum pressure measurement for laboratory grade vacuum equipment) indicates to me that the pump is not all that great. The better pumps like this (direct-drive dual-stage pumps) can reach .1 millitorr or 0.0001 torr - so if he has one of those, his vacuum system likely still has leaks. I'd like to see a better picture of the pressure indicated by his sensor - which seems to be up to the task before making any more comments about his setup.

Producing plasma is one thing and is relatively easy - think neon signs and argon/krypton lasers. But creating a fusion is another task all together, and I'm skeptical of the ability of a setup like this to produce fusion. Regardless of what others might say about funding and grants, if creating fusion in such a small device were that easy, someone would have done it by now, IMO.

 

[mailpup]

That vacuum pump appears to be of the type that is used to pull a vacuum on A/C systems prior to the installation of the refrigerant.

It looks similar to a Kozyvacu brand vacuum pump but, of course, I can't be sure.

 

[wiyosaya]

That vacuum pump appears to be of the type that is used to pull a vacuum on A/C systems prior to the installation of the refrigerant.

It looks similar to a Kozyvacu brand vacuum pump but, of course, I can't be sure.

Pumps that are good enough for A/C systems typically are not capable of reaching the same pressures as pumps that look similar but are made for laboratories. Laboratory grade pumps (made by Edwards or Savant or a similar company) like this can typically reach a vacuum that is an order of magnitude (factor of 10) greater than a pump meant for evacuating A/C systems. They are also much more expensive and finding a used one on E-bay that is in good shape may have eaten his whole $2K budget.

 

[TomVA]

It looks like a simple hollow cathode discharge to me. Based on the color maybe argon or helium gas. With the pressure the cylindrical rods may as well be a solid cylinder. Anyway it is defined as a helical hollow cathode discharge in the second reference. I doubt that it would get much above a few hundred eV electron energies at that pressure.

High-Current Hollow Cathode Development | Electric Propulsion and Plasma Dynamics Laboratory

alfven.princeton.edu

Basic Mechanisms Contributing to the Hollow Cathode Effect

If a single plane cathode in a glow discharge is replaced by a cathode with some hollow structure such as a cylindrical or slit shaped hole, then, in a specific range of operating conditions the negative glow is found to be inside the hollow structure of the cathode....

link.springer.com

 

[Homer100]

This kid has figured out, and thwarted the sabotage that big oil, and big energy has done to the fusion effort. You see going big isn't the best way to do research. Get a small version working first. Iron out all the problems, then scale up. But not too big. If a small version reaches a good state, then double the size. Get that version working, then double the size again. Starting huge means your design had better be absolutely on point, or the whole thing fails. A smaller device is easier to change, or completely throw away for a totally new design. Building big commits you to failure. Building small allows for success, which can be built on.

 

[Win7user]

I'm just wondering what student has $2k lying around?

 

[erickmendes]

... and what could happen if it really achieve fusion on this home made fusion reactor and by some luck/unluck it output a LOT of energy? Am I wrong or could it blow his apartment?

 

[Dimitrios]

Anyone have Raja vibes?

 

[wiyosaya]

... and what could happen if it really achieve fusion on this home made fusion reactor and by some luck/unluck it output a LOT of energy? Am I wrong or could it blow his apartment?

Probably not enough mass to blow up his apartment or dorm room, but perhaps enough output to attract the attention of the DOE, in the US, and get the contraption confiscated. He's U-Waterloo which is outside of Toronto, so any "regulatory action" would be the responsibility of Canada.

Again, its too small and no where near producing a bonafide Fusion reaction.

 

[ScottSoapbox]

... and what could happen if it really achieve fusion on this home made fusion reactor and by some luck/unluck it output a LOT of energy? Am I wrong or could it blow his apartment?

We're talking about 48 watts of power according to one of his posts: "achieved plasma at 4kv and 12ma".

Even if he achieved actual fusion on this scale it wouldn't blow a fuse let alone blow up something.

You don't trip and accidentally create a reactor 1000X bigger than expected.

 

[ScottSoapbox]

This kid has figured out, and thwarted the sabotage that big oil, and big energy has done to the fusion effort. You see going big isn't the best way to do research. Get a small version working first. Iron out all the problems, then scale up. But not too big. If a small version reaches a good state, then double the size. Get that version working, then double the size again. Starting huge means your design had better be absolutely on point, or the whole thing fails. A smaller device is easier to change, or completely throw away for a totally new design. Building big commits you to failure. Building small allows for success, which can be built on.

First off, he isn't doing research. He created a small-scale model based on existing knowledge from previous discoveries from research. It is an impressively difficult model, but it did not create new knowledge in any way.

Second, researchers are familiar with the concept of smaller-scale prototyping. It's not a secret. It is definitely not being kept a secret by a powerful conspiracy.

 

[Eldrach]

I hope he keeps finetuning his device - and goes on to get some real funding - which hopefully leads to advancement. A real non-radioactive fusion reactor is what the world needs

 

[Endymio]

However, in this day of fake everything (including peer reviewed tech publications), I find it hard to believe this is true without an independent assessment.

Genius or internet hoax?

Neither. One of my college classmates built one of these as an undergraduate. It took him longer than four weeks, but unlike this one, his actually generated small amounts of fusion. They aren't terribly complex devices, but their "Q" factor is many, many orders of magnitude too low to be useful for power generation (modified versions of these have other uses).

Still, they're interesting to view in operation ... the fusion of a single pair of H2 atoms is energetic enough to view with the naked eye.

I hope he keeps finetuning his device - and goes on to get some real funding - which hopefully leads to advancement. A real non-radioactive fusion reactor is what the world needs

Fusion reactors still generate radioactivity, albeit at a significantly lower level than fission. Which, unfortunately, means that even when they become practical energy sources, environmentalists will still oppose them.

 

[Mark Fuller]

I just don't believe it. With such a magnificent problem presented as construction of a fusion reactor. How can any one make such a brag and not anticpate what the reaction would be. If it is true, there are many things that must have to be done such as produce a sustained net energy output for a long time. It is going be interesting indeed to see such a brag proved through scientific evidence. Its either sink or swim.

 

[B5S46M]

Looks like a simple glow discharge when trying to start an ion pump at too high of pressure (about 1E-7 torr). A fusion reactor it isn't. That rotary vane pump is a roughing pump. It's not going to get you much of anything in terms of vacuum, maybe 1 millitorr if you're having a good day. To get to real Tokamak vacuum levels you need to to pull vacuum with this roughing pump on a turbo pump, then start an ion pump, a titanium sublimator, a cryopump, and maybe also a non evaporable getter pump. Oh yeah, and need to bake it out for a week or more. Then you finally get UHV of about 1E-13 to 1E-14 torr, and can try running experiments.

 

[Endymio]

Looks like a simple glow discharge when trying to start an ion pump...

It's ordinary bremsstrahlung radiation.

.... at too high of pressure (about 1E-7 torr). A fusion reactor it isn't.

Depends on how you define it. If he purchased some heavy water and loaded it with H2, I'm sure he'd get a minimal neutron flux from it.

...To get to real Tokamak vacuum levels you need to to pull vacuum with this roughing pump on a turbo pump, then start an ion pump, a titanium sublimator, a cryopump, and maybe also a non evaporable getter pump. Oh yeah, and need to bake it out for a week or more.

None of that is necessary if you're simply trying to slam a few deuterons together to demonstrate fusion.

 

[TomVA]

It's ordinary bremsstrahlung radiation.


Depends on how you define it. If he purchased some heavy water and loaded it with H2, I'm sure he'd get a minimal neutron flux from it.


None of that is necessary if you're simply trying to slam a few deuterons together to demonstrate fusion.

The mean free path of an electron in helium at 0.02 Torr is about 0.3 cm. Plenty of interactions between electrons random atoms/molecules and cause the electrons not to reach energies above a few hundred eV. The light is from excited atomic states of the atoms relaxing and emitting light as well as recombination induced photo emission. If you can see it it is less than 10 eV which is well into the vacuum UV.

In a hollow cathode discharge the electrons bounce back and forth from one side of "hollow" part of the cathode to the other multiplying their interaction with the gas contained inside the coil. They also start to get some random metal atoms off of the surface involved. To the gas and electrons the coil, from an electromagnetic standpoint in the eyes of the gas and electron interactions looks more-or-less like a hollow tube. The plasma squirting out of the end is also a common feature of such a device.

 

[Endymio]

The mean free path of an electron in helium at 0.02 Torr is about 0.3 cm. Plenty of interactions between electrons random atoms/molecules and cause the electrons not to reach energies above a few hundred eV.

A fusor fuses deuterons -- ionized nuclei -- not electrons. And of hydrogen, not helium. Fusors can and do operate in the 0.02 Torr range. See link:

https://libjournals.unca.edu/ncur/wp-content/uploads/2021/07/2356-Garofalo-Aaron-FINAL.pdf#:~:text=The%20low-power%20fusor

The bremsstrahlung glow you see is primarily from the electrons as they collide with ionized nuclei and gaseous impurities in the chamber (less mass, same charge = greater acceleration) -- it is in the x-ray range (an ordinary x-ray tube isn't terribly different from a fusor chamber).

There's also some plasma glow which peaks in the UV. Both follow an ordinary Maxwellian distribution though, which means some visible spectrum light is emitted, and cross-scattering produces more.

 

[Bamda]

Having a personal power plant at my home would be a GOD send. Please make this happen.

 

[Hooda Thunkett]

This looks like a variation of a Farnsworth Fusor, which dates back to 1964. They're good for demonstrating fusion is possible and creating neutrons, but not so good at making energy. I'm much more impressed that he did this so quickly, and without a background in electronics or physics.

 

[mbk34]

The important thing about any project like this is not whether the fusion actually happens but how much publicity you get and how many key words you can utilise. "AI" and "Fusion" are likely to get him many interviews going forward though he really needs to include either Rust or Arduino in there to maximise interest.

 

[Nobbly]

Then he realised he'd just blue 2k on a lamp