I am not an electrostatics (or an electronic) engineer. I am only a hobbyist, albeit one who has dabbled in electronics from before transistors became available. However, I am an avid denizen of independent local used book stores and one of my treasured finds is, "Electrostatics and its Applications," edited by A.. D. Moore, (John Wiley & Sons, 1973, ISBN 0-471-61450-5). I urge you to find a copy - preferably through your local independent bookstore (they need our support) but alternatively through a search service like Abe Books. I have used the information in the book to do home-brew electrostatic painting of steam radiators. It was a hair-raising experience but it did get the paint into the intricate inner recesses of the radiators and I lived to write this.
It has a chapter on Electrostatic Coating which includes a two page section 11.4.1, "Electrostatic Formation of Sandpaper." Two pages is too much to attempt to reproduce here, but I'll mention a couple of its main points.
VOLTAGE: It mentions about 100 KV applied between two metal plates separated by about 4 inches.
MATERIAL: "A continuous belt of semiconducting material passes over a group of rolls and is in contact with the upper surface of the lower electrode plate." Both electrodes are "...usually ... metal sheets encapsulated in a semiconducting material." The lower belt is the carrier for the abrasive grains to be transferred to the glue-coated upper belt.
I interpret the function of the semiconducting material to be to have enough conductivity to allow enough current to flow to charge the abrasive particles on top of the semiconducting belt without allowing enough current to flow laterally away from the charging area to either short the charging electrode to ground or leak the charge off the abrasive grains before they can be transferred to the glue-coated belt above. The text says nothing about the nature of the semiconductive materials. You might experiment with spraying your lower belt with CRT screen "anti-static spray" to provide a little conductivity.
The accompanying figure (in the book) shows the lower electrode charged to high voltage (either positive or negative) and the upper plate, above the glue-coated belt, grounded.
The text mentions that the primary advantage of the electrostatic method over simple mechanical spreading of grains is that the electrostatic transfer leaves the particles embedded endwise in the glue with sharp ends protruding most effectively.
I don't think the type of metal used in the plates has any influence beyond its compatibility with the semiconductive encapsulant of the plates.
I'm sure there are other books available on the topic. I presume that you have done a Google search. Another thought would be to contact the abrasive product manufacturers like Norton. Sometimes they have informative booklets bragging about their technologies.
awright