Moiré materials and the Mott transition

There are back-to-back papers in Nature this week, one out of Columbia and one out of Cornell, using bilayers of transition metal dichalcogenides to examine the Mott transition.  (Sorry for the brevity - I'm pressed for time right now, but I wanted to write something....)

As I described ages ago in here, imagine a lattice of sites, each containing one electron.  While quantum statistics would allow each site to be doubly occupied (thanks to spin), if the on-site electron-electron repulsion \(U\) is sufficiently strong (large compared to the kinetic energy scale \(t\) associated with hopping between neighboring sites), then the interacting system will be an insulator even though the non-interacting version would be a metal.  Moving away from this half-filling condition, you can get conduction, just as having an empty site allows those sliding tile puzzles to work.

As discussed here, in bilayers of 2D materials can lead to the formation of a moiré lattice, where the interlayer interactions result in an effective periodic array of potential wells.  The Columbia folks got a moiré pattern by using a 4-5 degree twisted bilayer of WSe2, while the Cornell folks instead used an aligned bilayer of MoTe2 and WSe2 (where the moiré comes from the differing lattice constants).  In both cases, you end up with a triangular moiré lattice (encapsulated in hBN to provide a clean charge environment and protection from the air).  

The investigators are able to tune the systems in multiple ways.  With overall gate voltage, they can capacitively tune the "filling", the ratio of number of "free" charges to number of moiré lattice sites.  By adjusting top gate vs. back gate, they can tune the vertical electric field across the bilayer, and that is a way of tuning interactions by pushing around localized wavefunctions for the lattice sites.  

Both groups find that they can tune in/out of a Mott insulating phase when they're at one carrier per moiré lattice site.  Interestingly, both groups see that the Mott transition is continuous (second-order) - there is no sudden onset of insulating response as a function of tuning either knob.  Instead, there appears to be quantum critical scaling, and regions of linear-in-\(T\) temperature dependence of the resistivity (a possible indicator of a strange metal) on either side of the insulating region.  The Cornell folks are able to do magnetic circular dichroism measurements to confirm that the transition does not involve obvious magnetic ordering. 

This is very pretty work, and it shows the promise of the moiré lattice approach for studying fundamental issues (like whether or not the Mott transition in a triangular lattice is continuous).  I'm sure that there will be much more to come in these and related systems.