Under the suggestion of M. Was, instead of comparing the temperature of the point absorbers against the background of the same image, I calculated the difference between the thermocamera image with the ITF locked in CARM_NULL and the image with the ITF unlocked (background reference). This analysis in principle is more accurate w.r.t the previous one as the background pattern is completely removed. First, both images were calibrated in temperature using the colorbar of the thermocamera app (to keep in mind that there is an uncertainty in this calibration since it is taking a 2D image for that purpose) and then the relative temperature of a given area will be the variation ΔT = T_pa(ITF_locked) - T_pa(ITF unlocked), where T_pa is the temperature estimation on the point absorber pixels coordinates. The complete analysis can be found here: VIR-0667A-25.

The results of this analysis is the following:

• Old WE: ΔT = 0.75 °C (Fig. 1);
• New WE: ΔT = 0.61 °C (Fig. 2);
• NE: ΔT = 0.47 °C (Fig. 3);
• WI: no clear point absorber was identified but a wider hotter area w.r.t the surface of the mirror, as shown in Fig. 4.