I did thermal simulation with different laser power for two materials to obtain thermal induced birefringence, for a 11 cm plate. The beam size was 5mm (radius) and the plate was heated for 2hrs, to reach thermal equiilibrium. You might see that the birefringence of entire plate increases with higher laser power, because the temperature rises consequently. I attach fig 1 and fig 2 corresponding to two different materials. 

The plot has two data, one is birefringence at the center (where the laser is incident), second is at the edge of the plate. You can see that if we use too much CO2 power, we will be heating the plate so much that even if there is no laser, we will have a large retardation. 

For now, I don't include the stress induced birefringence. The contribution is only from thermo-optic and expansion.