Here are a couple of interesting papers I've seen in the last week or so:
Doh et al., Tunable supercurrent through semiconductor nanowires, Science 309, 272 (2005).
Very pretty use of InAs semiconducting nanowires grown by the now-usual VLS approach, and contacted by aluminum pads. When the Al goes superconducting, carriers in the InAs (controlled via field effect with a gate) become superconducting, too, via the proximity effect. Basically this results in a tunable Josephson junction with the InAs nanowire as the controllable weak link. Gorgeous, as is most of the stuff that comes out of Delft.
Ghosh et al., Zero-bias anomaly and Kondo-assisted quasi-ballistic 2d transport.
A preprint out of the Cambridge folks that looks at very clean mesoscale 2d GaAs/AlGaAs systems and argues that there is evidence (a temperature-dependent zero-bias peak in the differential conductance) that points to Kondo-assisted transport in these systems. The remarkable thing is that the orthodox Kondo effect relies on localized spin degrees of freedom, and there shouldn't be any in these materials. The authors suggest a more exotic (2-channel!) Kondo effect involving localized two-level defects. Very intriguing data, though the interpretation is likely to be controversial, if past 2-channel Kondo reports are any indication.