They key to efficient heat transfer using a granular dipping media is very fine media size and a high thermal coefficient. Copper powder comes to mind. Aluminum or brass powder might also work well, though oxidation will definitely be a factor. Perhaps a reference of the thermal coefficient of the oxides would be in order.
I didn''t get into the science of it as far as you did Ian, but I did think that inserting into a solid would be handicapped to some extent from the start because it doesn't have the same uniform contact as a liquid. But I did think that you could optimize your medium by picking something like... well, your copper pots and pans. There is this (and finer) available:
HOWEVER, I actually started looking into this and it appears there may be some issues. Regarding the molten salt bath,
https://www.ampannealing.com/articles/52/salt-bath-annealing--does-it-work-/ .
Yes, I saw that quite a while ago when I was wondering what an AMP machine cost - and saw the $1500 pricetag or whatever it is before you start buying the extras and add ons. Not to attempt to throw shade on their technological approach, which is laudable, considering how much you pay for their machine, but there were a few things about that which made me curious.
First, they claim the guy (who's a member of several cast bullet forums as well as others) who makes and sells those kits says salt bath annealing is the best and easiest way to anneal. But he doesn't say that, so I wonder why they infer that he does? And their competition from the flame annealing variants vastly outstrips this one guy's kits made in his spare time. He's a P.Eng in the metals end of that profession; those kits for him are like others making PLC/PIDs - he does them in his spare time.
So they analyze the hell out of just this very, very small competitor - but unless they've done it recently, they've never put up an analysis comparing flame analysis to their induction annealing. Now why is that - wouldn't you want to show people just how superior your induction annealing is to flame annealing... especially when you can buy (or make) something like an Annaleaze for less than 1/5th the cost of their machine?
Second, I'd assume without them even doing any comparative analyse that a computerized induction annealer is almost certainly going to give better and more repeatable results than salt bath annealing (and probably flame annealing). The question for me is whether my rifles and my diminishing skills will see any difference on paper. I'm not a world class F-Class competitor with a $4,000 rifle with a $4,000 scope on top of it.
So, back again to Analyzing Made Perfect, in my retirement biz, we do a lot of statistical analysis on data done by testing samples i.e. randomly taken samples of earth sent to labs for analysis on the presence of heavy metals from industrial smokestacks. After we do that, before our professional stamp goes on the work and it goes to who contracted with us, we head back out and ground truth our analysis by collecting more samples for analysis at new spots. If those results don't fit into what our predictive analysis says they should be as far as those heavy metals... then we have a problem.
So I would think AMP would have simply started with a quick and easy comparison e.g. ten cases firing ten shot groups ten times at 1000 yards while being salt bath annealed, versus another ten being annealed with an AMP. Same rifle, same load, same bullet - what an easy and graphic way to show the world just how poor the results are out of salt bath annealing! They've done similar long string testing of their annealing, but not here.
Now why is that? I don't know, but if it really mattered to me, I'd ask them both questions: why they don't do comparative testing to flame annealing, and why they avoided doing real world testing of salt annealing simply by comparative shooting. And despite that, if I was in the high dollar shooting sports, I'd probably already own one of AMP's machines.
Anyways, there's enough stuff to bat around on that subject already, and for what I'm doing, so far fingers 'n alcohol lamp flame has been doing a pretty good job. Immersion annealing just looks like it might take up very little extra space if done in the right way, and possibly I might see an improvement on paper. Using another medium than a salt bath falls into that same analysis.
Any granular media will have a recovery time for the area just used, so there is no guarantee of consistent temperature even if a pattern of dipping is followed in an effort to keep the temperature of the media equalized.
Probably true... but perhaps not. A solid media like sand, the copper above, tiny steel beads, etc should allow a case pushed into it case mouth first to stand on it's own without a frame to support it. That would mean placement in the granular media would be randomized - chances of picking the same spot twice in a row is pretty unlikely, especially if you're mindful of that.
What I wonder is how long it would take fine copper to be covered in oxides, and how much that would diminish it's ability to transfer heat.
I'm still musing about giving this a try using a pot within my casting pot. I'm trying to diminish the reloading associated stuff I have in and around my bench - not increase it.
Thanks for the response and the rational as well.