Xenon is great for low to medium energy density discharges as I described above, but when you drive the flash lamp harder and harder the discharge volume becomes more and more like a black body. The type of material becomes irrelevant as a radiator, but its quantity, pressure and electronegativity remain significant.
Your CO2 system is basically the same. The tremendous energy density of ablating wall lamps (100-150J/cm3) sputters even tungsten and erodes the walls (very, very slowly). The flowing gas cleans the tube and so each discharge cleans the debris from the previous one and the envelope remains pristine.
For initial testing you might use simple, but very inexpensive (and temporary) dual reservoirs with siphon tubes. Place several L of dye mix in a polyethylene container and allow it to gravity flow through the laser to its twin sitting at a lower level on the floor. As the dye empties from the upper container reverse the positions of the containers. This seems crude, but it has lots of advantages. Aside from the low cost for a system that may have other serious problems and not be worth it to you to pursue, this method avoids contamination in pumps and radiators, generates minimal scattering bubbles, operates at low pressures that will not spray across the room if your system leaks and allows you to drain, fill and vary the dye concentrations as you work to optimize the system.