As a licensed nuclear facility, ITER must limit the in-vessel tritium (T) retention to reduce the risks of potential release during accidents, the inventory limit being set at 1kg. Simulations and extrapolations from existing experiments indicate that T-retention in ITER will mainly be driven by co-deposition with beryllium (Be) eroded from the first wall, with co-deposits forming mainly in the divertor region but also possibly on the first wall itself. A pulsed Laser-Induced Desorption (LID) system, called Tritium Monitor, is being designed to locally measure the T-retention in co-deposits forming on the inner divertor baffle of ITER. Regarding tritium removal, the baseline strategy is to perform baking of the plasma-facing components, at 513K for the FW and 623K for the divertor. Both baking and laser desorption rely on the thermal desorption of tritium from the surface, the efficiency of which remains unclear for thick (and possibly impure) co-deposits. This contribution reports on the results of TMAP7 studies of this efficiency for ITER-relevant deposits