[en] In this paper, the results of postmortem analysis of mirror samples exposed for the one-year campaign inside the Tore Supra (TS) tokamak and for shorter time inside the TCV tokamak are presented, together with results of laboratory mirror experiments. In TS, although all mirrors were eroded, the observed erosion rates do not correspond with the published sputtering yields. In TCV the thickness of the carbon-based film deposited on the mirror samples was found to depend on the mirror material, being several times higher on Si than on W and Mo substrates. An attempt is made to attract for explanation of results the difference in stickiness of carbon to SS and to Cu surfaces. The laboratory experiments qualitatively support such explanation