Silhouettes of the bomb: what can we learn from the shapes of atomic weapons? | Hacker News Reader

Silhouettes of the bomb: what can we learn from the shapes of atomic weapons?

blog.nuclearsecrecy.com

86 pointsrdl34 comments

Relatedly, I find it deeply troubling at some level to know that the exact shape of the Earth (specifically, mass and magnetic distribution) is highly classified, since it's important to ICBM guidance.

Doesn't ITRF2008 and (to a lesser extent) WGS-84 give you what you want?

Source? I'd like to read about this.

Could be something to do with this:

http://nsarchive.gwu.edu/NSAEBB/NSAEBB509/

"NASA's Applications Satellites and National Security Requirements:

NASA planned on launching the Geodynamics Experimental Ocean Satellite-3 (GEOS-3) with a radar altimeter whose data would generate more accurate gravity models than any measuring devices previously flown. The data would be used to meet civilian requirements and by the U.S. Navy to improve the accuracy of its submarine-launched ballistic missiles (SLBMs). Because it also had the potential to be used by the Soviets for the same purpose, the DoD requested NASA to encrypt the data, limit the acquisition of data over the ocean areas where Soviet ballistic missile submarines operated, or delay the planned April 1975 launch until the issue could be resolved. It also requested encryption of the data from the even more capable radar altimeter planned for SEASAT-A. NASA rejected these requests but quickly agreed to restrict the dissemination of the GEOS-3 data collected over the regions where Soviet ballistic missile submarines operated (Document 38)."

However is it still the case that NASA continue to withhold this data, or is it data collected by the NSA/NRO/et al's satellites that's considered classified?

they should be keeping that secret; good gravity models are a major part of the difference in accuracy of American missiles vs Russian ones.

Yes because the last thing you want is for the "soviets" to be able to hit their target. Surely them hitting instead a city near by or just using bigger nukes to compensate for lesser accuracy is a much better approach.

No, the more inaccurate their targeting, the harder it was for them pull off a counter-force first strike. No need to lead them into temptation....

Once they're launched, you've already lost catastrophicly, no matter exactly where they come down.

And why the scare quotes for our adversaries who build 40,000 nuclear warheads, most of them aimed directly at us?

For important things like ICBMs, you don't want to trust easily spoofable navigation like GPS. Instead, you double integrate over the acceleration data obtained from very accurate accelerometers and gyros.

since these sensors measure acceleration, and gravity is acceleration, any change in gravity over the missile's path will cause it to think it had moved differently than it did. Knowing the precise earth gravity then you can account for these changes.

Changes in the magnetic field will affect compass readings, which help you decide what direction the missile is facing. Knowing the (fixed over time) change will help account for measurement error in the compass. I'm not sure how much that helps though, since the compass' earth field is constantly changing anyways due to the sun and magnetisation of the missile itself.

The actual surface shape has an effect on the gravity experienced above the ground, but is also important to make sure you know what height you're aiming for, and that there's nothing in the way.

Magnetic info is sensitive more for detecting hidden equipment, like submarines.

Cool :)

But also see Rapatronic photographs, at <10ns after detonation: http://www.freerepublic.com/focus/f-chat/3104424/posts

I still don't understand how this stuff is so hard. The US pulled it off in a few years, in the 40s. Total cost of a few billion. The biggest issue is enrichment, but again, they did it within that time and budget. Several other countries/physicists were on to the same idea.

Even a simple gun type with a 20kton yield is enough to severely harm a city, right? So even without the complexities of implosion or thermonuclear, it seems any dedicated actor, even a private one, should be able to pull it off. Delivery systems aside; I'd guess testing ICBMs would be difficult enough by itself.

What am I missing?

Making a gun-type bomb from enriched uranium isn't that hard. Any shop capable of building an auto engine could do it. Even machining uranium isn't that hard; there's a Union Carbide tech note on that. It requires a respirator and protective gear, but not remote controlled machine tools. That's why anti-proliferation people are so concerned about enriched uranium.

But enriched uranium is hard to make. The old gaseous diffusion process was incredibly inefficient - plants over a mile long to make tens of kilograms. Centrifuges have changed that. Centrifuge plants aren't all that big, about Wal-Mart sized. They're still expensive. Highly enriched uranium is only made for bomb purposes; power reactors don't need it (although some submarine reactors have used it), so it's not made without good reason.

There's some concern about laser enrichment, which apparently works quite well and is seldom discussed. The US has at least one laser enrichment plant.

Plutonium bombs are hard to get right. A gun bomb with plutonium will pre-detonate. It takes an implosion. Making a perfectly symmetrical implosion is hard. The explosive lenses have to be very uniform in density and accurate in dimension, while being made out of a soft plastic explosive. There are tricks to the joints where the lenses meet, some of which are still classified. Then there's the whole detonator thing, although getting the big pulse required is easier than it was 50 years ago.

Plutonium is a by-product of reactor operation. Thus, there's more of it around than is really needed. The US and the former USSR made way too much (tons). But working with plutonium is a huge headache. The dust is radioactive and poisonous. The PUREX chemical process to refine it from reactor fuel rods is difficult; places which did it tend to be toxic waste sites now. Its machining properties are strange; it expands when heated but then doesn't contract fully. Plutonium bombs usually have a neutron generator to get things started, and surrounding shells to perform neutron reflection and tamping. While the general principles are well known, it takes a lot of engineering R&D to get it working right, including many non-nuclear test explosions.

That's why we don't see terrorist groups making bombs from spent fuel rods.

It's even harder than that. Unless you remove the fuel rods quickly, in 1-2 months, you get too much thermally hot plutonium-238, and too much plutonium-240, which is the isotope that prevented using reactor bred plutonium in a gun assembly bomb and required developing the implosion design. Another isotope quickly decays into a fierce gamma ray emitter.

At best, from what I've read, you'd have a bomb that prior to detonation would dissipate 100 kW, requiring serious refrigeration and therefore size, and you'd be lucky to get a yield greater than 1 kT. That's of course quite capable of ruining your whole day, but it's perhaps more accurately viewed as a super dirty bomb, that's likely how you could achieve the most damage with it. Not a city killer, but a contaminator.

> Plutonium is a by-product of reactor operation. Thus, there's more of it around than is really needed. The US and the former USSR made way too much (tons).

I was under the impression that there's not that much plutonium kicking about, unless it's the Russians and former USSR states that have the lion's share:

http://archive.is/UleJ1#selection-2455.625-2467.68

(src: http://www.wired.com/2013/09/plutonium-238-problem/all/)

"So, we depend on plutonium-238, a fuel largely acquired as by-product of making nuclear weapons.

But there’s a problem: We’ve almost run out.

“We’ve got enough to last to the end of this decade. That’s it,” said Steve Johnson, a nuclear chemist at Idaho National Laboratory. And it’s not just the U.S. reserves that are in jeopardy. The entire planet’s stores are nearly depleted. The country’s scientific stockpile has dwindled to around 36 pounds."

Previous HN discussion: https://news.ycombinator.com/item?id=6414717

There are many isotopes of Plutonium. Pu-238 has a half-life of 88 years and is used in Radioisotope thermoelectric generators. That's the isotope that's running out.

A lot of the other isotopes will be around for many many years. We have more of those than we need. https://en.wikipedia.org/wiki/Isotopes_of_plutonium

So, of course, we come up with schemes to get rid of them, like blending some in with other nuclear fuel. Such as what we did in Unit 3 at Fukushima Daiichi. And then hydrogen gas caused massive explosions, releasing radioactive material. Oops.

Importantly, Pu-238 isn't fissile (it's not the isotope you need for a weapon).

Are there any naval reactors that don't use HEU?

Super carriers have a much bigger space budget than submarines, so I wouldn't be surprised if they used less enriched uranium. On the other hand, refueling is a total pain, so that might put a premium on the total U-235 in the fuel rods (granted, some of the U-238 will breed into plutonium and some of that will burn up, but...).

The US and UK use 90% enriched uranium in naval reactors. China, India, France, and Russia use lower levels of enrichment. The US really ought to have converted over by now. There are people lobbying for this for future nuclear vessels.

The only genuinely hard part is the fissile material. It would be trivial for a country like Japan to do this in days/weeks (going from materials they have to weapons); a country like, say, Sweden with less of a nuclear industry could also become a nuclear state fairly quickly. Plenty of companies or individuals could do it.

The difference is -- if the US/Russia/etc. detect that someone is trying to make a nuclear weapon, all means up to an including a preemptive strike (nuclear or conventional) is on the table. Diplomatic pressure is the first step, but unless you're in a position to essentially blackmail/stalemate everyone, it would not run to completion. If, say, Nigeria were taken over by a crazed regime which wanted to make a nuclear weapon, and it didn't have either the US or Russia protecting it (or possibly China), all national means would be expended to prevent this from happening.

(Stuxnet was a horrible Pandora's box opening, but IMO the US/Israelis behind it probably saved the world from a nuclear conflict, which is win/win for basically everyone. If it were a private individual or group, it would be even easier to go after them.)

Theoretically, it may be possible to do "fission free fusion"; if someone can develop a reasonable, cheap, undetectable procedure for that, we are utterly fucked.

The other thing which terrifies me would be laser separation/enrichment. Basically anything which could be done surreptitiously or in a widely dispersed way, vs. at a small number of detectable, centralized plants.

(My money is on bioweapons destroying civilization, if not humanity, within 50 years, though, before nuclear weapons have a chance. Bioweapons are far easier to develop surreptitiously, and with very small scale initial capital requirements.)

> all means up to an including a preemptive strike (nuclear or conventional) is on the table

>all national means would be expended to prevent this from happening

What you're saying is overly dramatic to say the least.

If any of that were actually true, the US would have nuked or invaded North Korea in the mid 1990s. It was extremely well understood they were developing a nuclear weapon. We knew it for two decades. The same is true for Pakistan, we knew they had a program, why didn't the US invade / nuke / attack / subdue them in response (even afterward)? The US hardly did anything of consequence to Pakistan in response, they got a slap on the wrist and the runway to develop a hundred nuclear warheads.

NK has leverage a random group wouldn't (alliance with China, and more importantly, a huge, dug-in conventional force which can obliterate Seoul in hours no matter what the US does, even a pre-emptive nuclear strike).

Iran does, too.

A hypothetical rando country probably doesn't. A company or terrorist group unallied with a major state certainly doesn't.

> My money is on bioweapons destroying civilization, if not humanity, within 50 years

Oh, humanity — i.e. the Platonic quality of being humane — will survive, but mankind may not.

The cost ($2bn in 1945 dollars) ends up being over $30bn in present-day terms, which isn't really in the reach of many private entities.

Close to 1% of the total US labor force ended up working for the Manhattan Project in the period up to August '45. It was pretty big.

The US government also wasn't exactly paying market rates for what they bought during the war, and the Manhattan Project also had the highest priority call on materials, after the super-Mulberry we were building for the invasion of the Kanto plain in Honshu, the main island island in Japan and where Tokyo is located (look up Operation's Downfall (the general one) and Coronet for more details).

You can't disregard delivery systems, that's a major piece of the picture. Nuclear bombs are most effective when airburst. If you are going to deliver by ICBM or with a military aircraft then you need an implosion design with plutonium. You might be able to sneak it in with a civilian aircraft, or via a ship or something (obviously then you can't airburst).

If you want to fully open up your delivery options, you want an implosion design with plutonium. This involves running a reactor for a couple years and reprocessing the waste. And you'll have to avoid regulatory scrutiny from the IAEA to do so (which is also true of a uranium bomb).

At this point the biggest problem with building a nuke of either type is getting the material. As an experiment, some college students were asked to design a nuclear weapon back in the 70s, and they apparently got it close enough that the FBI freaked out and tried to suppress their work. It would be even easier nowadays, but refining requires a lot of energy and specialized materials (eg maraging steel) that are watched really closely, and so are the materials that would be used as feedstock.

> some college students were asked to design a nuclear weapon

Also known as the Nth Country Experiment. See https://en.wikipedia.org/wiki/Nth_Country_Experiment for details, also the final report [1] is interesting.

[1] http://nsarchive.gwu.edu/news/20030701/nth-country.pdf

Creating U-235 takes a lot of work by almost all chemical/physical methods.

https://www.osti.gov/opennet/manhattan-project-history/Event...

Of course, once you can bootstrap a nuclear reactor, things get a lot easier because now you have Plutonium. Of course, Plutonium is rather toxic and working it metallurgically is problematic.

Edit: Plutonium also seems to require an implosion device. Implosion engineering is quite hard.

This brings to mind the story of John Coster-Mullen - the first person to build an accurate historical replica of Little Boy. I'd recommend reading the New Yorker article on it. It's a pretty fascinating read.

http://www.newyorker.com/magazine/2008/12/15/atomic-john

"You have the long-barrels that look like gun-type designs."

No. That setup was abandoned very early on, as in even before the end of WWII. The OP should have spent more time reading this https://en.wikipedia.org/wiki/Thermonuclear_weapon and less time speculating.

Not in artillery. It was a narrower diameter than either implosion or Teller-Ulam until probably the 70s or 80s, at which point we'd given up on nuclear artillery.

(I am confident the author of the blog is knowledgeable about the US arsenal.)

Fascinating; per Wikipedia, e.g. https://en.wikipedia.org/wiki/W79 and https://en.wikipedia.org/wiki/W48 they used a "linear implosion" technique, but it sounds like it had about the same inefficiency as the Little Boy gun assembly design. Which for safety reasons alone are iffy to field, and even more I'd imagine in a quickly accelerated artillery shell....

From reading Command and Control, a lot of the US arsenal was manifestly unsafe for a long time. Based on the number of accidents, and lack of criticality, I suspect there was some super secret design choice or design flaw which meant many classes of weapon would never go critical at all. Someone probably made the decision that only a deterrent was important, so a bunch of unsafe weapons should instead be rendered largely inert while making other gestures of readiness. I don't have any data to support that other than just the lack of unintentional nuclear detonations.

http://www.amazon.com/gp/product/B00C5R7F8G/ref=dp-kindle-re...

Please consider reading about the author: http://blog.nuclearsecrecy.com/about-the-blog/

Reply on news.ycombinator.com