Oh my god! I don't understand. Couldn't you just have used a piece of steel? This sounds like something out of Frankenstein's lab. I bet it was a nice shade of orange though.
My thinking re. beam quality is that as the power goes up (never too much) the diode solution always suffers from beam degradation. The beam either enlarges in the near or far field and co-alignment can never be perfect.
You probably never saw the blending of the colors due to the high quality of the beams; their diameter was substantially smaller than their proximity
With regard to 637 diodes some Russian researchers a few years ago pushed these down to 610nm at near LN2 temps. and then with pressure tuning (I think something like 200,000 PSI) sent them down into the green. Along with the beam quality issue you have identified the other problem which is cost. To reach temps. below -30C I think will involve significantly more expensive staged chillers and TEC or maybe even cryogens. I think a dye solution will easily give substantial power from 580-610nm (your choice).
Yes neoprene does flow. I don't understand the mechanical properties well enough to explain this,but after tightening some screws against neoprene sealing washers and returning several days later the screws were much looser. This may not continue indefinitely, but probably will cause a slight focus as well as pointing shift.