INTRINSIC CRITICAL VELOCITIES IN SUPERFLUID HE-4 FLOW-THROUGH 12-MU-MDIAMETER ORIFICES NEAR T-LAMBDA - EXPERIMENTS ON THE EFFECT OF GEOMETRY

Citation
Gm. Shifflett et Gb. Hess, INTRINSIC CRITICAL VELOCITIES IN SUPERFLUID HE-4 FLOW-THROUGH 12-MU-MDIAMETER ORIFICES NEAR T-LAMBDA - EXPERIMENTS ON THE EFFECT OF GEOMETRY, Journal of low temperature physics, 98(5-6), 1995, pp. 591-629
Citations number
82
Categorie Soggetti
Physics, Applied
ISSN journal
00222291
Volume
98
Issue
5-6
Year of publication
1995
Pages
591 - 629
Database
ISI
SICI code
0022-2291(1995)98:5-6<591:ICVISH>2.0.ZU;2-8
Abstract
We report superfluid He-4, flow measurements at temperatures from 1.2 K up to T lambda - 3 mK in three orifices of different mesoscopic geom etry. Under conditions of out experiments, the flow usually reaches a temperature-dependent intrinsic critical velocity, where dissipation i s believed to occur by thermal (or quantum) nucleation of individual q uantized vortex rings or loops. The nucleation rate should be sensitiv e to the wall geometry of the flow channel and to any local velocity e nhancement at the most favorable nucleation site. According to the Ior danskii-Langer-Fisher (ILF) theory, the radius of the ''critical'' vor tex ring, the threshold size which can grow freely by extracting energ y from the flow, increases inversely as the superfluid density on appr oach to the superfluid onset temperature, T lambda. Thus sufficiently near T lambda the critical ring should be large enough that the geomet ry relevant to the nucleation process and local velocity enhancement c an be studied by scanning electron microscope (SEM). We examined our t hree orifices by SEM. One, a standard optical pinhole, has a relativel y smooth taper on one side and a sharp lip on the other. The second is similar, but contains a 1-lambda m flake perpendicular to the flow, w hich should provide additional velocity enhancement at its edge. In th e third, the sharp lip is beveled to reduce the velocity enhancement a t that site. Contrary, to expectation, the intrinsic critical velociti es are the same, within a relative calibration error of 10%, in all th ree cases.