Comments On Supersolidity

Abstract

Assuming that the well-confirmed non-classical rotational inertia (NCRI) effect in solid $^4$He, suggested by Leggett, indicates supersolid behavior, we make a number of remarks about both theory and experiment. (1) The long-wavelength, low-frequency ("hydrodynamic") part of the theory of Andreev and Lifshitz has nine variables, and thus must have nine modes. We find a new mode associated with lattice point diffusion (and thus vacancy diffusion); it may explain the absence of supersolid behavior in low-frequency pressure-driven flow. (2) The observed upper limit for the NCRI fraction (NCRIf) of about 20%, in disordered samples, is more-or-less the same as the already predicted upper limit for the superfluid fraction of a well-ordered crystal; we argue that this may not be a coincidence. (3) The negative experimental evidence for a second propagating hydrodynamic mode (expected to be fourth sound-like) may be due to the long relaxation times $τ$ at low temperature $T$; only for frequencies satisfying $ωτ\ll1$ does the hydrodynamic theory apply. (4) The fundamental principles of quantum mechanics imply that Bose-Einstein condensation is not necessary to define a quantum-mechanical phase; therefore the absence of a finite condensate fraction $f_{0}$ does not necessarily imply the absence of superfluidity. (5) Just as vortices should avoid occupied lattice sites to provide a vortex-lattice interaction, the lattice should interact with the vortices to provide a lattice-vortex interaction; thus dislocations should interact with vortices, whose motion is affected by rotation. We discuss some experimental implications for the vortex liquid model, shear response, hysteresis, and relaxation.