III. On the state of fluids at their critical temperatures

Abstract

In carrying out the investigations which I commenced some years since upon the phenomena presented by the flow of different liquids through capillary tubes, the question as to what constitutes a liquid— that is in what way it differs from a gas, and how the great variance of the microrheometrical laws for the two fluids can be explained—again and again presented itself to me. Seeing that solids are soluble in gases as well as in liquids, one of the chief differences supposed to exist between the two states has disappeared ; and I have been compelled to adopt as the only definition of a liquid, that it is a fluid which has cohesion. Professor James Thomson, F. R. S., has suggested to me the use of the term contractility, instead of cohesion, and this term admirably defines the liquid state, but as it suggests (in a distant way perhaps) a voluntary power, and is used in connexion with organised structures, I shall retain the term cohesion at present. We have then the two states of fluids, first, the gaseous, in which the vis viva or heat energy of the molecules has entirely overcome cohesion, or their mutual attraction, and they are prevented from grouping ; and second, the liquid where the attractive power is greater than the vis viva, and the molecules are enabled to group themselves, but still are in sufficient motion to prevent the grouping from being permanent, hence we have cohesion, but no rigidity. We do not yet know that all solids are not also fluids, as many of them are known to flow, but this may be from other causes, but we know that the solid state is characterised by so much cohesion as to produce more or less rigidity. The most interesting point in the consideration of a liquid is that at which it approaches to the gaseous state, where its cohesion disappears, and we have what Dr. Andrews has termed the critical point, which is the termination of that property which distinguishes a liquid fluid from a gaseous fluid, or in other words the liquid becomes a gas. But a question arises. To observe this disappearance of the cohesion of a liquid, it is requisite that it should have a free surface, and this free surface has till now only been obtained by arranging the pressure that a portion of the fluid is in the gaseous state, and this only occurs at one pressure. Now, when the temperature of a liquid is raised while it is retained under very great pressure, so that it never has a free surface, but is always retained filling the vessel, does the liquid still lose its cohesion, and become a gas at the same temperature ; or, as the pressure is increased, does the temperature at which the cohesion of the liquid is overcome, also rise In the former case, the limit of the liquid state would be an isotherm, in the latter, a continuation of the boiling line. To determine which is the object of the work here described. With proper precautions, the loss of cohesion or capillarity can be noticed very accurately, and the level of the liquid in a fine capillary tube, seen to coincide with the plane surface of the liquid just before the final disappearance of the line of demarcation. One of the precautions to be taken is to obtain equable temperature, and while in my earlier experiments, X used a double air-bath, and considered this sufficient to obtain good results, I subsequently found that by the use of a triple bath of copper, every trace of irregularity of temperature disappeared, and I obtained results in which the line of division was admirably clear and sharp, and never became broad and hazy as in ordinary experiments. Another precaution to be taken is to have pure liquids, and this at first sight might appear to be an easy matter, but I find that in transferring a portion of a pure liquid to a tube, the momentary exposure to air, especially in the vicinity of the hands, hydrates the liquid sufficiently to render the line of demarcation rounded, and show a slightly greater refractive power in the lower part of the tube after the critical point has been passed. In the case of liquefied gases, such as carbon dioxide, ammonia, sulphur dioxide, and nitrous oxide, which are easily dried, the line is beautifully sharp, and the disappearing point easily noted. Alcohol cohobated over caustic lime for a week and transferred to a tube without contact with air, shows the disappearance of the line with great sharpness, and immediately after no difference in refractive power can be detected between the upper and lower portions. The least trace of moisture is sufficient to show such a difference. Whenever I notice any difference between the upper and lower portions after passing the critical point, I attribute it to moisture or other impurity, as careful treatment always removes the difference in density. In many organic liquids there is always a difference at the critical point, and sometimes before reaching this temperature, they form several layers, each having a different critical point as they seem to give rise on heating to new compounds, or form polymeric compounds having different critical points. Besides many organic compounds cannot be entirely freed from impurity they retain it even on repeated distillation. In the following experiments, therefore, such organic compounds were never used, and only perfectly anhydrous alcohol, or carbon disulphide,* or gases which can be obtained anhydrous, CO 2 , SO 2 , and NH 3 , being chosen. The apparatus used for obtaining pressure was that described in a former paper (“ Proc. Roy. Soc.,” No. 201, 1880, “ On the Solubility of Solids m Gases”). In order to determine, then, whether increased pressure applied above the critical point, would have the effect of reducing the gas to a liquid, as might easily be supposed, since the rates of expansion of gas and liquid become alike at the critical point, a new form of experiment was resorted to. It had been noticed that it was easy to determine whether the tube were filled with liquid or gas, by simply reducing the pressure somewhat quickly, when, if there were liquid present, it boiled, while if the contents were entirely gaseous, simple expansion was the result. The .boiling only takes place when the pressure is reduced so far as to be a little under the vapour pressure at that temperature, in other words, boiling cannot be observed, unless there exists a free surface, and this free surface cannot be obtained with the liquid alone above the “ critical pressure.” By the introduction of a quantity of hydrogen gas over the liquid, a free surface is obtained at any pressure, and the mixture of hydrogen and alcohol vapour being of so much less density than the alcohol, it remains divided from it by a line of demarcation for some time after the latter is undoubtedly gaseous. Now, let us see what takes place on lowering the pressure. When the temperature is even only 1°C. below the critical point, when the pressure is sufficiently reduced, the alcohol boils, showing that it still has cohesion, but if the temperature be 1° above the critical point, the fluid only expands, and no boiling is seen at any pressure, from 50 up to 200 atmospheres. Here the fluid above the critical point has just as free a surface as below it, and we see that the last trace of the liquid condition has disappeared. The line dividing the mixture of hydrogen and alcohol vapour from the pure alcohol is quite sharp, for a short time, and on altering the pressure, it moves up and down quite freely, and possesses exactly the same appearance and properties as hydrogen over carbon dioxide in a bell-jar.