But there's more than that... you'd also need a body of large thermal mass opposite the halogen bulb to absorb the heat, and then re radiate it, so it could bounce off the CO2.
I was thinking of a 12" glass sphere with an 11.5" black sphere inside of it. The black sphere would be the thermal mass (maybe full of water) and it would have temperature sensors all around. Then in the space between, some air. Get a baseline temperature reading. Then pump in a tiny amount of CO2. Check temperature delta.
The amount of change you would see from that is going to be really way too small to get an accurate reading on. You'd also need to run a control experiment with heating the ball with just plain air (rather than air + extra CO2), because the glass is also going to act like a greenhouse trap anyway.
I wonder if something like that has been done.