Frictional motion of normal-fluid component of superfluidH3ein aerogel

Abstract

The superfluidity of liquid 3 He in a high-porosity aerogel has been studied using a fourth-sound resonance technique. This technique has two significant advantages: it can directly determine the superfluid density and it can derive the transport properties of the viscous normal-fluid component. The temperature dependence of the resonance frequency revealed suppression of superfluidity and that a finite normal-fluid fraction exists even at T = 0. The motion of the normal-fluid component has also been investigated. As T → 0, the energy loss becomes very small, despite a finite amount of the normal-fluid component remaining. This implies that the normal-fluid component is highly constrained by the aerogel, and hence the dissipation mechanism cannot be described in terms of the conventional hydrodynamic model. We have succeeded to explain these results by introducing a frictional relaxation model to describe our observations, and found that the flow field changes from being parabolic ͑Hagen-Poiseuille viscous flow͒ to flat ͑Drude frictional flow͒ on introducing an aerogel. Numerical calculation of the relaxation time using the quasiclassical Green's-function method reproduces experimental results.

Dissipation signatures of the normal and superfluid phases in torsion pendulum experiments withHe3in aerogel

“…5 shows this for the 29.1 bar data along with the data from Ref. 38. We find a good agreement between the two experiments, implying that the observed dissipation in the 50 kHz sound attenuation experiment and the torsion pendulum Q −1 in the superfluid B phase probably have a similar origin.…”

“…We expect the fluid to be in a Drude flow regime, 10,37 where velocity of the fluid with respect to the aerogel is constant across the channel, with the exception of a small region of size δ d = (2η/ρτ f ) away from the edges. 37,38 The frictional relaxation time τ f is related to the friction force coupling the helium with the aerogel matrix: [39][40][41]…”

“…The data plotted is for 29.1 bar. We also show the data from Ref 38,. which is deduced in a similar way.…”