Flow and torsional oscillator measurements on liquid helium in restricted geometries under pressure

When a solid surface is in contact with liquid helium, the strong van der Waals forces between helium and the solid substrate result in a layer of high-density helium localized on the substrate. It was expected that, upon raising the pressure, further solid layers would grow out from the wall. Many experiments have now shown that this usually does not occur, and even at solidification pressures ${}_{}{}^4{}_{}{}^{}\mathrm{He}$ liquid, rather than solid, wets a disordered surface. This implies that, in the neighborhood of some solid surfaces, liquid helium can exist at pressures above bulk solidification pressure.

We have devised two experiments, using dc superflow and torsional oscillator methods, to investigate this effect in the confined geometries of Vycor and fine powders. In the dc flow method, which relies on the detailed form of the He I–He II–solid-phase diagram, the outer ends of a rod of Vycor or other material are left in the superfluid region while the center is in the bulk solid region. We confirm that superflow takes place through this system, and observe new phase diagrams showing where superflow can occur in several materials of different pore size. The torsional oscillator consists of a Vycor cylinder encapsulated in a beryllium copper container with a Be-Cu torsion tube through which helium can be admitted. In the bulk fluid region we confirm that the superfluid transition in Vycor occurs at temperatures $T_0$ which are parallel to, but lower than, the bulk λ line.

At higher pressures, in the bulk solid region, $T_0$ is found to propagate up into the bulk solid region as high as 25 bars above the bulk solidification pressure. The superfluid mass in Vycor cannot be calculated unambiguously, but it can be inferred that at absolute zero it decreases slowly between zero pressure and 30 bars, and thereafter more rapidly, becoming zero at about 50 bars. We suggest that the failure of solid to nucleate at the surface is due to its amorphous nature which discourages the formation of crystallites, and that a high-density disordered phase nucleates within the fluid in the pores according to a well-known nucleation theory which gives the correct order of magnitude for the overpressure required. This implies the existence of a high-density localized state of helium which is disordered even at 0 K. An observation which we have not as yet been able to explain is a falloff in period of the torsional oscillator at the solidification curve as the temperature is reduced at constant pressure, starting in the normal fluid phase. This phenomenon is observed with ${}_{}{}^3{}_{}{}^{}\mathrm{He}$ as well as with ${}_{}{}^4{}_{}{}^{}\mathrm{He}$.