The extent to which various liquid drops may be supercooled has been observed by examining the appearance of clouds of such drops in a beam of light a t different temperatures. For most of the liquids studied the great majority of the drops did not freeze until a comparatively narrow range of temperature had been reached far below the melting point. In this range where the drops freeze spontaneous homogeneous nucleation is assumed to occur. This is discussed in the light of current theories of nucleation phenomena in which a key role is played by the interfacial tension between solid and liquid. An attempt is made to determine values of this quantity, and correlate them with the corresponding heats of fusion.IN recent years measurements have been made, particularly by Turnbull and his colleagues, on the extent to which liquid metals will supercool in the absence of catalysts (Turnbull and Cech, J . AppZ. Physics, 1950, 21, 804). An essential feature of these experiments was that observations were made on numerous, isolated drops, in the expectation that while some of these might have occluded impurities the majority should not. Experiments on the same lines with molecular liquids with the single exception of water have not been described. In this paper we present some observations on the supercooling of some common organic liquids and of some liquefied gases. An attempt will be made to interpret these on current theories of homogeneous nucleation. EXPERIMENTALThe problem was to find the temperature a t which a large number of isolated drops of liquid froze. One may in principle either examine a cloud of drops simultaneously or else a large number of individual drops. An unsuccessful attempt was made to follow the latter course, by adapting Rumpf and Seigl's method (2. Physik, 1938, I l l , 301), in which individual drops were held stationary and observed in a cold stream of air flowing upwards. It was found that it was not easy to hold a drop stationary, that the drops steadily evaporated, and that freezing could not readily be detected.The method had been used for water, though few experimental details were given (Schaefer, Bull. Auner. Mat. SOC., 1948, 29, 175), and it was reported that the water clouds froze completely over the range -39°f0-10.Consequently, a cloud method was adopted.

No abstract

No abstract

No abstract

A differential method is described for measuring the second virial coefficient of a gas from approximately the normal boiling-point of the substance to room temperature, given the value of the virial coefficient at one temperature. The imperfection of the gas is compared directly with that of a reference gas (helium) for which the second virial coefficient is known over the whole temperature range. Measurements have been made on argon, krypton, and methane, and on a sample, approximately equimolar, of each of the three binary mixtures which can be formed from these gases. The results for the pure gases have been used to demonstrate the superiority of the 18-6 form of the Lennard-Jones potential over the 12-6 form in accounting for low-temperature virial coefficients, while the figures obtained for the mixtures have been applied to examine the validity of three combining rules for estimating the intermolecular energy parameters for a pair of unlike molecules from the corresponding parameters for the two pairs of identical molecules.