Definitely fit superheaters, and make them the radiant type. The theory and practical results are proven, see http://modeleng.org/articles/hall01.pdf, and an article by a chap called Wallbank in ME a couple of years ago.

As regards how big, I think the critical factor is the area exposed to the radiant heat of the fire, so as big as you can get in (probably 5/16" in your case). Put them side by side, not on top of each other, otherwise the top run is shielded from the fire by the bottom one.These will be well big enough to carry the steam.

It was qualitative, no numbers as I recall. Howvere, if you do the heavy lift of going through the index to find the issue number I will have it in the attic and can copy it for you. Bill Hall did the numbers, and his program 'perform' might help, but I don't think he ever wrote it up so mere mortals could understand it. I might have a copy of the code if you can read C. As I recall the bit inside the flue didn't do a lot, most of the heat transfer took place in the bit exposed to the fire. This is actually easier to work out, you need to look up 'view factors' to find how much of the radiant heat liberated by the fire finishes up in the superheater, but the first problem is working out the fire temperature. Bill's program overestimaes it I think because he doesn't allow for combustion being split between the firebed and gas phase above. I'm getting well out of my depth now! Now that infra red pyrometers are reasonably available it would be a nice experiment to measure the relationship between grate load and fire temperature, but I'm not volunteering, too many other irons in the fire (sorry!)

When I get time I'll go through all the stuff Bill left and see what there is, but don't hold your breath, my recods system is even ess organside than my workshop.