Anything you put in the loop will attack the metals. The issue is some of them will attack the metals very slowly indeed. Distilled is fractionally better than deionised, but they'll both pick up co2 during bleeding anyway and end up much of a likeness.
silver/nickel/copper/brass are generally considered fine together as the electrochemical potential between them is relatively small. Certainly you're very unlikely to see corrosion issues when using these, most barbs are nickel coated brass and most radiators are brass with copper fins. PT nuke is a marginally better idea than silver as it's copper sulphate, which is pretty unlikely to attack copper. Kill coil is more convenient though.
Mixing metals is an issue if you have aluminium and copper in physical contact, i.e. Swiftecs famous **** up. Jokester maintains that corrosion can't occur if the pieces are electrically isolated from each other, say connected by plastic tubes rather than copper piping. Again this will be an approximation but a reasonable one, corrosion will just go a lot slower if the metals are separated and insulated.
The conclusion most people reach is to not have aluminium anywhere in the loop. This is the approach I take but I'm hardly experienced. I believe a lot of people use antifreeze to stop algae and this is considered to reduce corrosion, but I've not seen a source for this.
