[neilw20]

Often, something softer can grip better better as it conforms/deforms to the surface.
I had a hand made mild steel taper in an old machine and it gripped exceptionally well, but I had to careful not to damage it.
Another method to rough things up microscopically is to get a piece of carbide, a 12v power pack with a capacitor across the output and spark many times on the hardened surface. Stronger sparks? Larger value capacitor. Put a 20 or 30W bulb in series with supply so as not to damage with a full short. Put the capacitor after the bulb.
This also works very well on lathe chuck jaws when needing to hold hard material, that would normally slip with heavy cutting.

[mike sr1]

Often, something softer can grip better better as it conforms/deforms to the surface.
I had a hand made mild steel taper in an old machine and it gripped exceptionally well, but I had to careful not to damage it.
Another method to rough things up microscopically is to get a piece of carbide, a 12v power pack with a capacitor across the output and spark many times on the hardened surface. Stronger sparks? Larger value capacitor. Put a 20 or 30W bulb in series with supply so as not to damage with a full short. Put the capacitor after the bulb.
This also works very well on lathe chuck jaws when needing to hold hard material, that would normally slip with heavy cutting.

You think a lot like I do Neil, that should work well.
I am thinking about a rig we used to electro etch stainless lines with in the dairy, I would think anything that would take the high polish finish down a bit would really help.
I also thing getting the collet to seat over the entire surface would help as well, I havent checked mine but having it pinch at the bottom 1/8" inch would put all the clamping pressure on the first 1/8" leaving the rest of the surface not clamped tightly and then it could wobble slightly, and if anything moves it will come loose eventually, an old machinist told me that years ago.

BTU pretty much covered this in his post and in much better detail.

It would be interesting to see if dykem on the clamped shank surface would show an uneven pattern as well.