I think there is such a thing as UHMW (ultra high molecular weight) High Density PE, which is "both". At least that's the catalog listing...what do I know? That's about what I know.
At any rate, according to Doctor C. Willis, it's the total density of both hydrogen and carbon, which does some of the moderation too. And this stuff is heavy stuff (goes to see if it even floats),
and it does, barely, a 5" long 4" diameter piece in a bucket of ice cold water sticks about 1/4" out of the water, so your density numbers appear to check. Interesting that it was stable long way down in the water, which was quite cold (the bucket was brought indoors to thaw out from solid ice).
The point being that the smaller it is, the shorter the "moderation length", the less scattering
out of it there will be as well -- important here since we only have neutrons coming from one direction. At any rate, the ~1.5" thickness was calculated by Carl to be about the best you could do to put the most neutrons into the silver/indium resonances as possible at a few ev, and I check that with actual measurements here (it's not critical at all, though). There is so much plastic in that paper design that I'd guess most of the neutrons are down to thermal, well below the max sensitivity point for either silver or indium for most of it -- the Ne110 is also a moderator plastic.
I doubt there's a serious difference between them other than the value of the scraps around the shop. The UHMW stuff is very like super strong teflon and has a lot of uses for things that slide, or need HV insulation, as well. I can tell you it is a son of a gun to work with in sizes over 4" diameter. An extremely sharp Japanese wood saw simply skips over this stuff, doesn't even leave a mark. My metal bandsaw cuts it, kind of, but tends to melt it and get stuck in the cut, which is bad for both. I finally cut a piece off that long rod I got (for a B10 tube) with a chainsaw, and nearly ruined
that when the molten plastic jammed up in the clutch housing and re-hardened. And basically, if you can't melt it, you can't cut it. Drilling that same piece ruined my arm running the drill press table up and down to clear the molten chips before they seized the drill and to put on ever longer drill extensions -- took about 4 hours solid hard work to make an 18" long hole through it -- about half an inch at a time.
I have some offhand near-evidence of how good this is as a moderator compared to wax. I have my 22" long 3He tube in wax, 6" diameter, and almost touching the tank where the action is.
On the floor, about 4 feet from the action on the opposite side is a 16" B10 tube in this HDPE stuff, also 6" OD. It counts 1/2 or 1/3 as fast during a run. We know the B10 tube isn't supposedly as good as the 3He....I have them on stereo channels in a little amp with speakers I built into my equipment rack, which is how I got that -- pure "feels about like 1/2 or 1/3", no real numbers yet. (In fact, I've been building tube to ttl preamps for both this weekend so I can put them into my data logger stuff for the next run -- as they are, they won't work a logic level counter and I've just been using them on audio for realtime feedback as I adjust things). Very interesting the things one's ears can tell you about time correlations in bursting.
I was scratching my head about their claimed 50cps background in that article until I remembered they had set threshold above the positron annihilation energy. Else I was wondering how they got it so low - here a tube with a 4" sq scint set to "normal energies" counts thousands/second sometimes, and about 100 always -- with a lot less square area involved. They are incorrect about being able to count 10 meg cps reliably with that scintillator, it's not fast enough, the tube's not fast enough and so on for there to not be very significant pile ups at that rate. Merely having tubes and logic able to go 10x the count rate isn't good enough by far; you're already missing some counts due to pile ups. The linearity curve starts to look like log(x) instead of x.
Doesn't matter to us. We can always run shorter times, wait longer to start counting (and get the more reliable longer half life stuff alone), move it farther away....
I've been running 5 minutes mostly as a convenient runtime, and taking about 20 sec to get to start of count (which time is logged so I can know it exactly). I should be able to put up some plots soon, but I'm going to finish the neutron tube counter electronics before I run again. When the tech guys were here, I had to do a short run due to equipment issues, but got to the 700 cpm region with the system as is in about 4 minutes of running. No fancy counting for 10 minutes to curve fit a scatter plot needed at all for that, it was way above the background.
Note
data at this link, which shows what happens when an H atom captures a neutron (it happens enough to require enriched U to make a reactor with regular water). That's one hot gamma, and might itself be used as a neutron detector just fine, indirectly! I'll soon be reporting on what I see with one of my NaI heads during a run -- I'll set it the same way so the 511 kev gammas from Cs137 don't fire it, and see what we see. I couldn't tell much last run (the tech guys were here) but it looked fairly promising.
Pretty much everything that captures neutrons does that, so borated wax and a high threshold gamma counter might be ideal -- if the fusor itself didn't make so many, and enough EMI to make using phototubes troublesome.