History and prospects of research on regelation

Abstract

The paper reviews experimental and theoretical studies on regelation. Cohesion of contacting two pieces of melting ice was found in 1850 by Faraday. Although its mechanism remained undefined, a new experiment using a weighted wire through ice was conducted (Bottomley, 1872). This penetration phenomenon was explained by pressure melting and refreezing. Experiments using ice slabs were then conducted over years, until a new method using a semi-cylinder of ice was introduced in 1973 (Drake and Shreve). Quantitative analysis has advanced the understanding of the relationship between the wire speed and the driving stress, the wire diameter, and the thermal conductivity of the wire. The speed was in proportion to pressure at pressures greater than 0.13 MPa. However, analysis revealed that the speed was extremely low in a region under pressure of 0.13 MPa. On the other hand, the speed was derived theoretically under the assumption of uniform water thickness around the wire, and pressure melting and refreezing (Ornstein, 1903; Nye, 1967). (i) For nylon wires, heat conducts mainly from ice, and the estimated speed was twice the observed speed. (ii) For low thermal conductive wires such as chromel wires, the estimated and observed results were in good agreement. For both (i) and (ii), the speeds were roughly independent of the thickness of the water layer. (iii) For high conductive wires such as copper wires, the speed was remarkably higher than the observed speed. The temperature balance was not achieved because the water layer was too thin, and the viscosity η was too small. If the value η was 125 times greater than that of the bulk water, the layer would become five times thicker, and the calculated speed would agree with the observed speed. Water vapor was often observed on the upper side of the wire, indicating that atmospheric pressure never affects the wire. The load is supported only on the underside, and then pressure melting will occur from a tripple point, 0.01 °C．A transition region for speed can be interpreted by no or partial melting. Future tasks include the measurements of the distribution temperature and pressure around wire and the viscosity of the thin water film.