I was interested in your post "What is the best way to sort melee diamonds of all colors, clarity and sizes?"
About this problem, I accidentally found out on the Internet in March of this year, when I was developing design of the Device for freezing CZ's without harmful hydrofluoric acid (see my eBay http://www.ebay.com/itm/142387010973). According to your post, you sort the gems obtained as a result of dissolution of jewelry in Aqua Regia with subsequent treatment of gems with hydrofluoric acid. I am also familiar with this process very well (see my eBay http://www.ebay.com/itm/131919322542).
Conceptually, I have already decided the problem of sorting CZ's from diamond melee. This is a simple desktop Device, where the water separates stones. It works and sorts diamonds from CZ's by 100%. According to your message, you want to separate 2500 carats at a time. To do this, you need to design and test the Device with batch of 500 g of stones. If diamond bulk density is approximately 2 g /cc, the loading should be approx. 1 cup. Compare, now I'm experimenting with one teaspoon. In addition, about a problem with Zircon - ZrSiO4 (d = 4.73 g / cc) and Moissanite - (d = 3.21-3.22 g / cc), I learned only from the text of your post.
Thus, for scaling of the process, it is are necessary to make experiments. However, I do not have and will not have enough material without your participation in the project.
About me: I am a researcher and I have been engaged in design of chemical and hydromechanical processes for many years. I'm PhD of Chemical Engineering. If you wish, I show you my Diplomas, Scientific Articles, and Patents. For the last 7 years, I have been developing and selling desktop devices for extracting precious metals by non-professionals in rooms that do not have fume hoods (see my eBay).