An accurafu knowl edge of t he viscos it ies of liquids in absolute units is of fundamental importance in man y scient ifi c fi elds. Th e meas urement of t hese viscosities is almost universally based upon the absolute viscos ity of water at 20° C as a primary standard. During t he past 50 years t here has been an in creasing need for a more accurate d eterminat ion of t his s tandard. Consequen t ly, with t he coopera tion of t he Society of Rheology and so me financial assistance from t he Chemical Foundation , this proj ect was undert aken by t he National Bureau of Standards and has now been completed.Th e determination was made by t he method of capillary fiow. By means of a calibrated i nj ector, va riou s known constan t rates of flow were prod~c e d in capillaries of m eas ured dim ensions and obser vations were made of the co rrespondmg press ure drops through t he capillari es.' Th e eff ects of t he ends of t he capillaries were r endered negligible by t he s imu ltan eo us t reatment of da ta obt ained with pairs of capi llaries having essent ia lly t he same diam eters but differen t lengths.The value found for t he viscos ity of water is 0.010019 poise as co mpared wi th 0.01005 poise, which has generally been accepted for t he past 30 years. Th e estimated accuracy o f t he !lew determination is ± 0.000003 poise.As a r es ult of this work, begin nin g 0 11 JIIly 1, 1952, the Nat ional Bureau of Standards is plannin g to use t he value 0.01002 poise for t he absolu te v iscos ity of wa:ter at 20° C a s t he basis for the calibrat ion of viscometers and standard-sample otis. It] S r eco mm end ed t hat other laboratories adopt t his value as t he primary refere nce standard for co mparative m eas urem ents of viscosity.

No abstract

An apparatus designed for the precise measurement of absolute viscosity is described. The liquid being studied is caused to flow from one reservoir through a capillary tube into a second reservoir, the pressure difference between the two being measured with a differential manometer. Various rates of flow are caused by injection of mercury into the upstream reservoir from a cylinder by means of a uniform piston driven at constant speed by a synchronous motor through a gear train. A number of interchangeable capillaries of different bore diameters and lengths can be employed. The displacement apparatus and the viscometer are thermostated in baths designed for precise temperature control. A machine designed and built to lap the bores of capillaries into right circular cylinders is described.

In this paper equations are given for the direct treatment of experiments in which not only heat, but also masses, pass the boundary of the container of the system during the experiment. The theoretical development is correlated with the treatment of Gibbs, and certain difficulties, mentioned by others, in the physical interpretation of his equations, are incidentally removed. It was found possible, within a reasonable time, to cause a gas to expand slowly enough from a calorimeter to simulate a reversible expansion, and the special equation developed for the heat of expansion with the aid of the Beattie-Bridgeman equation of state was verified by the results for the slow expansion of carbon dioxide and ammonia. In the case of carbon dioxide the effect of the deviations from the ideal gas law was to make the heat effect in excess of that calculated for an ideal gas by a sufficient amount so that the excess itself could be calculated within about 7 percent. A series of expansions of carbon dioxide was carried out at varying rates of flow, some as fast as permissible. The results, correlated by means of an empirical relation, serve to show that the results of the slow expansions correspond practically to an infinitely slow expansion. They also indicate that the heat effect for an infinitely rapid expansion is not zero for a real gas, but possibly vanishes with the pressure. In the absence of a perfectly sound method of calculating the heat effect for an infinitely fast expansion, a method is suggested which has at least the merit of agreement with the present experiments. The bearing of the results on variable-pressure calorimetry, as practiced in experiments on the heat of adsorption, is briefly discussed.