It could be that you are inputting the entire qub file of 10 bands.  The thermal atmospheric correction tool in ENVI only works for IR band ranges between 8 and 14 µm and some of the bands of themis data are not in this range.  Looking at the header file of a qub file I have, it also looks like the radiance units are W/cm^2 while the thermal atmospheric correction tool expects data to be in W/m^2, so you likely need to scale your data.

Normal 0 false false false MicrosoftInternetExplorer4 I've limited the data to bands 4-9 (8.56-12.57) and I've set the data scale factor to 10000 to account for the w/cm^2 to w/m^2 unit conversion. The Themis radiometrically corrected RDR file exhibits negative values for radiance, that seems quite odd. The darker (cooler) regions tend to exhibit the wildest deviations from a 0-1 range, while lighter (warmer) regions tend to have what appear to be valid emissivities. i.e. absorption features of a depth of 0.1—0.5 rather than absorption features >=1 in darker regions. This is consistent over several scenes at different acquisition times. I hope the issue is not over at ASU's end. Any other ideas? Thanks,   ~Matt