Doping a Spin Liquid

One of the dreams of the great P. W. Anderson was to use a spin liquid as a substrate for superconductivity. Starting with a highly entangled quantum state (In his original proposal, the resonating valence bond state) and adding charge carriers would give you a superconductor or at least a highly-correlated metallic state. Unfortunately, usually some form of localisation wins and the dreams are shattered.

There is a recent exception – a hyperkagome iridate Na4Ir3O8 is an insulating spin liquid, but one can also create a Na3Ir3O8 compound, which is a hole-doped system and it conducts! The big mystery is what happens to the original spin-liquid magnetism?


We used NMR to access the intrinsic magnetic properties of this doped spin liquid and found that the susceptibility is governed by the semimetal electronic structure. The dynamical properties were studied by measuring the relaxation of the nuclear spins. InĀ  Na3Ir3O8, there are two sites for the 23Na nucleus – One has two Ir-triangles connected to it, whereas the other one has only one full triangle and the remaining Ir atoms come from three different triangles. This local structure is crucial – it allows us to contrast the dynamical fluctuations in two different environments.

By comparing the relaxation of the two nuclei, we found that the antiferromagnetic fluctuations persist even in the semimetal compound.

To find out more, see our preprint: