Ceramics can be really hard, but still very brittle, wouldn’t work.
Composites not very likely, carbon based composites are conductive, so scratch that; fiberglass isn’t, but then again it is much weaker than carbon composites.
Almost certainly some way could be found of isolating train wheels from each other. But in engineering, the question is never “can it be done”, but rather “is it the most practical way to do it”. And adding a non-load-bearing conductor is a more practical solution to the problem.
Industry standard has been, with exceptions, 25kV AC since the 1950s. Off the top of my head, I don’t remember how far that can go. Here in Sydney, we are still using a 1920s legacy system of 1500V DC, and that requires the placement of substations feeding the network approximately every 8 to 10 kilometres along the line.
1500V DC is actually fine for commuter trains, but is stretching the friendship a little for freight. Our network used to use powerful electric locomotives to haul heavy freight trains (ironically perhaps, in terms of this thread, these were replaced by diesel units twenty years ago on economic grounds), and if you wanted to run two of these heavy trains one after the other, there was a mandatory waiting period of, IIRC, twenty minutes, to allow the first train to clear any given electrical section, otherwise there was a real risk of tripping out the substation.
Here, they often load locomotives onto flatbed trucks and send them by road (very slowly) because the track access charges of ‘Notwork Rail’ make it too expensive to use the railway.
Must be very large trucks.
The electric motors - are they single-phase or 3-phase?
Generally true, but I’m sure if this was a problem we absolutely needed to solve, we’d find or create a suitable material or mechanism.