Hi Zog,
It's been a while. I expect Silex will continue to take off, as the signs look promising for the future.
I wanted to respond the the MIT thesis that you linked to in your earlier post. There's really not anything new there, and it doesn't discuss the physics of the process in the same way that Snyder's 2016 paper does.
First, it's important to remember that neither Eerkens or Kemp (advisors on this thesis) are pure physicists. Snyder is and that matters. This doesn't mean that engineering and experimental details are unimportant. They can sometimes be the foundation upon which an analysis depends, but the physics behind SILEX is not well documented anywhere, and Snyder attempted to deal with it and show where Eekerns comes up short. This is in his technical supplement available online:
https://www.tandfonline.com/doi/suppl/10.1080/08929882.2016.1184528/suppl_file/gsgs_a_1184528_sm4908.pdfMany analyses of enrichment factors of Silex are dependent on a lot of equipment parameters, but they are likely higher than the factor of 2 that Snyder uses in his analysis once the parameters are overcome. This is one advantage that Snyder has as a physicist compared to an engineer. He wasn't tied down to engineering details. He wanted to know how nuclear weapon could be constructed once another group just decided to adjust the engineering design to take advantage of the physics behind the condensation repression process.
He discusses this in his supplement linked above. Silex Systems seems to confirm this by saying the enrichment factor is between 2 and 20. He also discusses the laser challenges with the CO2 laser but there is enough evidence on how to overcome them, as Snyder discusses. This is almost certainly the laser than Silex began with, but they have likely moved up to something else, probably a solid state device. If they have not yet, they likely eventually will, or someone else will. The evidence for Silex moving on is that they themselves have said that the laser is new technology.
I'd also raise some objections with your statements about the higher efficiency of Silex at low 235-UF6 assess and less efficient at higher assays.
The idea at low assays is that fewer uranium-235 atoms exist so that less laser light will be absorbed as a laser propagates through a Silex separation unit. This is true, but it's not much of an efficiency advantage. It's there, but it's hardly anything. The pressures in the vessel are so low that not many photons will be absorbed compared to initial power needed (due to spatial needs about where an unexcited uranium-235 atom could be located). Snyder discusses this somewhere but he may not have accurately said that the advantage is so small that it's negligible.
On the question of high assays, the MIT thesis cites the Lyman paper from Livermore (
https://sgp.fas.org/othergov/doe/lanl/docs4/silex.pdf) incorrectly as stating an upper limit of 50% enrichment, where Lyman only suggests an upper limit of the enrichment level due to an engineering arrangement in the plant. Snyder cuts though this as a physicist and shows that none of this matters. In fact, because there will be less uranium-238 to collide with uranium-235 at higher enrichment levels (and also less condensation pressure) enrichment could become easier as the enrichment of the feed material goes up. It's true that slightly more energy would be needed to excite all of the uranium-235, but this is negligible. It's in the n_235*z term in equation 7 of page 8 on Snyder's main paper (not the supplement). No matter what n_235 is, that terms will always be small compared to the required laser power, so it just doesn't matter much. The bottom line is that enrichment to 90% likely becomes easier as you move up in enrichment levels, unless there's some factor that I am missing. The pressure will stay roughly the same, and there will be less uranium-238 to impede the movement of uranium-235 to the outer rim of the jet, and the uranium-238 will always be left behind in the center. This looks like why this is a serious proliferation problem. If you've come across an argument that suggests something different, I'd be curious to hear it.
The problem with the MIT thesis (other than not attempting to discuss the physics is sufficient detail) is that it still falls back on international agreement and technical monitoring to manage the proliferation problem. But no one knows how to detect a facility using this technology, and solid state lasers are not export controlled, and they almost certainly never will be. Far too many people, including Michael Goldsworthy (but that's not a surprise given his financial interest), simply suggest that Silex is too complicated to use to produce HEU. People once said the same thing with centrifuges.
But my question is why should you consider it too complicated? I don't think anyone has provided a satisfactory answer.
If you have any insight into this, I'd appreciate your argument. I just don't see how anyone's argument can rest on Silex's complexity. Any secret will be revealed to any group who is patient enough in asking the right questions. And it looks like there are no secrets left.
(20min delay)