Origin of the low-temperature saturation of the electron dephasing time

In quasi-$1\mathrm{D}$ disordered metal wires, electron-electron interaction gives $\tau_{\phi}\sim T^{-2/3}$ (Altshuler, Aronov and Khmelnitsky). Mohanty, Jariwala and Webb (1997) reported saturation of $\tau_{\phi}$ at low $T$ in Au wires and proposed an intrinsic mechanism; Pierre et al. (2003) found the interaction prediction obeyed in the purest Ag wires and traced saturation in other wires to magnetic impurities at the part-per-million level. Determine whether all reported saturation, including disordered alloy films, oxide films and $2\mathrm{D}$ semiconductor systems, is accounted for by magnetic impurities, radiation heating or other extrinsic sources, or whether a residual saturation remains once these are excluded.

$\tau_{\phi}$ measured well below $0.1\,\mathrm{K}$ in films and wires whose magnetic-impurity content is measured independently and whose electron temperature is fixed by noise thermometry, showing whether any saturation survives.