Building nukes from scratch

[DataPacRat]

I have a plotline I'm thinking of ending with a bang - more specifically, /lots/ of /really big/ bangs, in the form of a Project Orion launch. However, I also want to stay within real-world physics, and I'm looking for advice and suggestions.

Some decades after the Singularity ate all the cities and 99% of the population, Our Heroes, wandering around near Buffalo and Lake Erie, are likely going to get their hands on a 'factory seed', a small (non-superscience TL10ish) robofac that can Von Neumann itself into a larger one. Due to Plot, they are likely to decide that their best option is to get the bleep off of Earth; Ganymede is probably far enough away to avoid whatever gray goo gets released.

I've seen mentions a version of Project Orion which uses curium-245 as the nuclear explosive. By my back-of-the-envelope figuring, it would take roughly 15 tons of the stuff to get the necessary delta-v. Given that curium tends to be produced in microgram quantities these days, this seems problematic. Is there any theoretically feasible way to massively increase the production, or will I have to focus on the more usual radioisotopes?

Assuming that we have to work with either U-235 or Pu-239 made from U-238, and that any existing stockpiles were eaten along with the cities, the nearest significant source of uranium ore seems to be Elliot Lake in Ontario ( https://en.wikipedia.org/wiki/Uraniu...Canada#Ontario ). Any comments on what it would take to go from post-apocalyptic raw nature to several thousand nuclear pulse bombs ready to go?

[Anthony]

Originally Posted by DataPacRat

Any comments on what it would take to go from post-apocalyptic raw nature to several thousand nuclear pulse bombs ready to go?

Magic? Refining fuel for nuclear weapons is still basically a project of years for a nation.

[DataPacRat]

Originally Posted by Anthony

Magic? Refining fuel for nuclear weapons is still basically a project of years for a nation.

Taking years is fine, plot-wise. I'm hoping that the existence of the robofac can compensate for the lack of a globalized economic system.

[Anthony]

Originally Posted by DataPacRat

Taking years is fine, plot-wise. I'm hoping that the existence of the robofac can compensate for the lack of a globalized economic system.

Okay:
Step 1: build an economy on the scale of a nation state. It can be a modest size one, a GDP of $100B/year can probably manage an assembly line for a couple of nukes per year.
Step 2: Okay, you've got an organized government on the scale of a nation state. Why were you thinking of leaving in the first place?

[johndallman]

Originally Posted by DataPacRat

I've seen mentions a version of Project Orion which uses curium-245 as the nuclear explosive. By my back-of-the-envelope figuring, it would take roughly 15 tons of the stuff to get the necessary delta-v. Given that curium tends to be produced in microgram quantities these days, this seems problematic. Is there any theoretically feasible way to massively increase the production, or will I have to focus on the more usual radioisotopes?

It's made by neutron bombardment of plutonium. Making plutonium is just easier.

Assuming that we have to work with either U-235 or Pu-239 made from U-238, and that any existing stockpiles were eaten along with the cities, the nearest significant source of uranium ore seems to be Elliot Lake in Ontario ( https://en.wikipedia.org/wiki/Uraniu...Canada#Ontario ). Any comments on what it would take to go from post-apocalyptic raw nature to several thousand nuclear pulse bombs ready to go?

Either set up an economy and society of humans, as per Anthony's posts, or a smaller-scale robot society dedicated to this purpose. You need a fissile materials production infrastructure, which is a big job: you need to mine lots of uranium, and then either build a large-scale isotope separation plant, with associated power supplies, or a set of plutonium production reactors, either of which will require lots of mining and manufacturing infrastructure. Given that you don't care much about long-term pollution, the plutonium route is probably easiest, because the reactors can provide power, via an electricity grid that you need to build, for the rest of your infrastructure.

You may also need to set up a nuclear explosives R&D & Test infrastructure, because I suspect that compact nuclear bomb design is a field of engineering that might be omitted from the database of a factory seed, on safety grounds. Edit: It's an advanced field of engineering, and neither easy nor obvious.

Edit: Recommended reading: The Making of the Atomic Bomb, Richard Rhodes. Probably still the best account of what was required the first time round. Project Orion, George Dyson, gives you an idea of what's required for the spacecraft. Which includes quite a bit of R&D&T.

[malloyd]

Originally Posted by DataPacRat

Taking years is fine, plot-wise. I'm hoping that the existence of the robofac can compensate for the lack of a globalized economic system.

Well maybe. If you are actually building parts factories for each component you need, then you could probably build the industrial infrastructure of a small nation on a similar timescale. And probably need to since you are likely to need to establish mining colonies and plantations all over the place to acquire the necessary variety of raw materials. At which point you either have built one of the world's more powerful states (in which case why are you leaving again is a real question?), or one of them moves in and captures your valuable infrastructure well before you finish your spaceship.

If it's a slightly more superscience-y nanofactory that can make anything in a small package it may be more feasible - these sorts of things tend to have radical self restructuring abilities which mean you don't need nearly as many specialized factories, and a sort the atoms capability that can filter all the materials you need out of Lake Eire given enough time, which spares you from having to establish outposts all over the place. On the other hand, it does call into question why you want an *Orion* spaceship. Why not use it to make few copies of the seed and a launcher just big enough to get one to your destination and one to some suitable near Earth asteroid. You build the colony infrastructure at the destination before anybody arrives, turn the asteroid into the interplanetary transport, and move it to low Earth orbit. Now all you need to build on Earth is just enough lift to get your naked population to the transport alive, which isn't nearly as big a project.

[DataPacRat]

Originally Posted by johndallman

Either set up an economy and society of humans, as per Anthony's posts, or a smaller-scale robot society dedicated to this purpose.

Given that a similar industrialization process would be needed to create a self-sufficient colony off-Earth, getting some practice in the process while still on Earth seems as if it would be necessary anyway.

You need a fissile materials production infrastructure, which is a big job: you need to mine lots of uranium, and then either build a large-scale isotope separation plant, with associated power supplies, or a set of plutonium production reactors, either of which will require lots of mining and manufacturing infrastructure.

I'm thinking of something along the lines of NASA's paper on "Advanced Automation for Space Missions" ( https://en.wikisource.org/wiki/Advan...Space_Missions ).

You may also need to set up a nuclear explosives R&D & Test infrastructure, because I suspect that compact nuclear bomb design is a field of engineering that might be omitted from the database of a factory seed, on safety grounds. Edit: It's an advanced field of engineering, and neither easy nor obvious.

That seems a likely, and interesting hurdle. (There are a few possible solutions that come to mind...)

Edit: Recommended reading: The Making of the Atomic Bomb, Richard Rhodes. Probably still the best account of what was required the first time round. Project Orion, George Dyson, gives you an idea of what's required for the spacecraft. Which includes quite a bit of R&D&T.

I can get Rhodes' book quickly enough, but I can't seem to find Dyson's for less than $70, a bit beyond my budget for this project. (In fact, my early notes on a curium-based Orion come from someone else who read Dyson's book, which is why I already know the price...)

[DataPacRat]

Originally Posted by malloyd

At which point you either have built one of the world's more powerful states (in which case why are you leaving again is a real question?)

The setting includes entities that could be described as post-human super-intelligences. (It's complicated.) They occasionally release noticeable bits and bobs of advanced technology or custom biology or mysterious methods of influencing humanity and its offshoots. (The factory seed is one such bob - it's not quite molecular nanotechnology, but it's at least slightly better than could be built by pre-Singularity humanity in 2050.) There is a constant low-level looming threat that one of them will do something that wipes out the remains of humanity.

I could plausibly have one such entity's plans involve dropping some tech where Our Heroes could find it that would reduce the time required to build an Orion - say, a cache of computer chips that pass every anti-malware inspection check, or a database including relevant details that would be hard to find elsewhere. Nukes, however, simply require certain amounts of certain atoms that aren't so easy to get a hold of.

[malloyd]

Originally Posted by DataPacRat

Nukes, however, simply require certain amounts of certain atoms that aren't so easy to get a hold of.

The hyper-intelligences that rule the world are OK with the local insects building a nuclear arsenal? That seems like something that could potentially damage even a hyperintelligence, or at least something it cares about. "We *promise* they are for a spaceship, not offensive weapons" is probably not sufficiently reassuring.

Assuming obliterating a group trying this isn't the easy solution, they might want to hand out a space drive with less weapons potential. Though really *anybody* with a decent space drive in your own star system has to be considered some sort of threat, given what a even a moderate sized rock moved onto the wrong trajectory can do to a planet.

Maybe something they can't actually duplicate? "Here take these indestructible meshes of something not even made of atoms. They'll fuse hydrogen flowing through them from the orange to the purple side. Go build fusion torch ships and go away".

[DataPacRat]

Originally Posted by malloyd

The hyper-intelligences that rule the world are OK with the local insects building a nuclear arsenal? That seems like something that could potentially damage even a hyperintelligence, or at least something it cares about.

The local insects are arguably all the puppets of one or another of the hyper-intelligences, and their interactions form one of the HIs' battlegrounds, one that endures longer than the super-speed evolving-to-extinction risks many such HIs are focusing most of their attention on avoiding. The secret patron of the local insects in question is planning on sending a copy of itself to hitch a ride with them, in order to be the first one to make it through the grinder of the Kessler cascade; a feat that's required too much effort for any other HI to have successfully made the attempt to far.

(There's more to it than that; like I said, it's complicated.)