scholarly journals The effect of temperature on the viscosity of air

The Kinetic Theory of Gases leads to a number of relations between the diffusion, conductivity and viscosity coefficients of gases, and the large measure of confirmation of these has been the greatest triumph of that theory. Most of these relations have been shown by S. Chapman and Enskog to be independent of any particular model of the molecule. In the case of the dependence of viscosity upon temperature, however, the theory gives different results for different molecular models, and the determination of the temperature coefficient of viscosity can therefore be of service in the elucidation of molecular forces.

The experimental method employed was that described in earlier papers. A slider having a spherical face is made to slide over a plate in an atmosphere of rigorously clean and dry air. The friction measured is static friction and the object of the experiments the determination of the effect of temperature. This has now been studied over a range of 15°C. to 110°C., and it may be said at once that the relations discovered are of a totally unexpected character. More than one attempt to study the effect of temperature was defeated by the fact that lubricating vapours were given off from the walls of the chamber in which the plate and slider were enclosed. This difficulty was completely removed by using a chamber with double walls, the inner wall being a continuous sheet of nickel. Between the walls were placed the electric grids for heating the chamber. The stream of dry air with which the chamber was flooded was also heated by being passed through a tube of silica, which was maintained at the required temperature by a coil of wire through which a current was passing. The temperature of the stream of air and the temperature of the chamber were recorded electrically.


1890 ◽  
Vol 16 ◽  
pp. 65-72 ◽  
Author(s):  
Tait

(Abstract.)In the preceding part of this paper I considered the consequences of a special assumption as to the nature of the molecular force between two particles, the particles themselves being still treated as hard spheres. My object was to obtain, by means of rigorous calculation, in as simple a form as possible, a general notion of the effects due to the molecular forces. My present object is to apply this general notion to the formation and interpretation of the Virial Equation.


1944 ◽  
Vol 17 (2) ◽  
pp. 412-420
Author(s):  
La Verne E. Cheyney ◽  
Robert W. Duncan

Abstract Temperature coefficient of vulcanization may be defined as the increase in time of vulcanization necessary to produce a given property in the vulcanizate per unit range of temperature decrease, the latter being taken usually as either 10° F or 10° C. Coefficients of vulcanization for natural rubber stocks have been determined by several investigators. Although the data vary somewhat with the worker and with the stock investigated, the general range of temperature coefficients is close to 2.0 per 10° C. This is regarded as evidence of the chemical nature of the vulcanization reaction. The values obtained from physical test data do not always agree with those from combined sulfur analyses. This has been interpreted as an indication that the chemical reaction between the rubber and sulfur is not a simple bimolecular one, and that the rate of change of physical properties is not directly related to the rate of chemical combination of rubber and sulfur. A number of studies were published recently on the effect of variables on the vulcanization of Buna-S (now called GR-S in the United States) and on the properties of the resulting vulcanizates. In addition, compounding reports have been issued by manufacturers of rubber chemicals, as well as confidential reports submitted to the Rubber Director's office by rubber manufacturers. None of the published investigations, however, have been concerned with the determination of numerical relations among the properties of vulcanizates obtained at various temperatures. The properties of Buna-S vulcanizates differ markedly from those of rubber in certain characteristics, while possessing certain similarities in others. The only published mention of the effect of temperature on Buna-S stocks was in a release from the office of the Rubber Director, giving tables for conversion of cure to a standard temperature. These tables are based on a temperature coefficient of 1.43 per 10° F. The source of this information is not available, however.


Until our knowledge of the disposition and motion of the electrons in atoms and molecules is more complete, we cannot hope to make a direct calculation of the nature of the forces called into play during an encounter between molecules in a gas. It is true that a step in this direction has recently been made by Debye, who has investigated the nature of the field in the neighbourood of a hydrogen atom, assumed to consist of a negative charge in motion in circular orbit about a positive nucleus, and has shown how the pulsating eld gives rise on the whole to a force of repulsion, as well as one of attraction n a unit negative charge. But it is difficult to see how this work can be extended to more complex systems. At present we can only hope to derive information by more indirect methods. One such method is to assume a definite law of force, and then by the methods of the kinetic theory to deduce the appropriate law of dependence of the viscosity of a gas on temperature. Comparison with the actual law, as observed experimentally, serves to support or discredit the assumed law of molecular interaction. Unfortunately, the calculations involved in the application of be kinetic theory are so complicated that progress has been made only in certain simple cases. Thus, the original investigation by Maxwell applied only to molecules repelling as the inverse fifth power law. His work has since be generalised by Chapman and Enskog and formulæ have been obtained: the coefficient of viscosity in the case of (i) molecules, which repel according an inverse n th power law, (ii) molecules which behave on collision like rig elastic spheres and (iii) molecules which behave as rigid elastic spheres with weak attractive field of force surrounding them. Of these models the la generally referred to as Sutherland’s model, is found to give the best agreement between theory and experiment. But the agreement is by no means perfe As Schmidt, Bestelmeyer, Vogel, and others have pointed out, there considerable divergence from observed values a t low temperatures.


Although in the two preceding papers it was not possible, either from viscosity measurements or from isotherm measurements, to arrive at conclusive information about molecular fields, it is shown in this paper how some finality can be reached by using as well X-ray measurements of crystals. All that was possible in the other investigations (at any rate in the case of Argon) was the determination of a series of molecular models, each of which satisfactorily explained the experimental facts. With this information it is now possible to make theoretical calculations of the interatomic distances in crystalline Argon, and as crystalline Argon has recently been obtained and its structure measured and examined, a comparison of theoretical and observed values serves to fix the molecular field of Argon. Thus here, for the first time, information derived by the methods of the kinetic theory is applied to explain interatomic distances in crystals. Attempts have already been made to determine atomic and ionic fields from crystal measurements alone, as they have from viscosity and compressibility results alone, but the results have not been wholly satisfactory. By regarding atoms as rigid spheres, Bragg and Wasastjerna have obtained an atomic “diameter” characteristic of each element. But the values so obtained are in conflict with those obtained by other methods on similar assumptions. As Rankine has observed, they are considerably less than those derived from viscosity measurements. Such representations of atoms are recognised as only rough and approximate and are used in lieu of something better.


1969 ◽  
Vol 47 (12) ◽  
pp. 2197-2203 ◽  
Author(s):  
H. B. Spencer ◽  
J. M. Toguri ◽  
J. A. Kurtis

An experimental method for the determination of the diffusivities of metal vapor – gas mixtures has been examined. An investigation of the mercury vapor – argon system in the temperature range 180–340 °C indicated that the experimental diffusivities agreed well with the diffusivities calculated from the kinetic theory of gases. The temperature dependence of the diffusivity was found to vary with the absolute temperature to the 1.7 power.


The theory of liquids formulated in part I and applied to the equilibrium state in part II is here extended to liquids in motion. The connexion between the macroscopic and microscopic properties is revealed by the derivation of a set of generalized hydrodynamical equations, of which the fundamental equations of hydrodynamics are a special case; the more general equations describe the mean motion of clusters of molecules in the fluid. It is shown that the pressure tensor and energy-flux vector in a fluid consist of two parts, due to the thermal motion of the molecules and the intermolecular forces respectively, of which only the first is found in the kinetic theory of gases, but of which the second is dominant for the liquid state. A method is evolved for the study of those ‘normal' non-uniform states which relate to actual monatomic fluids in motion. It becomes apparent, as in the case of equilibrium, that there is a region of temperature and density where analytical singularities arise, closely associated with the process of condensation. Rigorous expressions for the coefficients of viscosity and thermal conduction are then derived which apply equally to the liquid and the gas. They consist of two parts due to the thermal motion and molecular forces respectively, of which the first is dominant for the gas, and the second for the liquid. By approximating to the rigorous formula, an expression for the viscosity of liquids is obtained, comparable with certain other formulae, previously proposed on quasi-empirical grounds, and giving good agreement with experiment. An integro-differential equation is derived for the determination of the distribution functions relating to the non-uniform state. A full discussion is given of the simplest case, and the velocity distribution in, non-uniform liquids and gases examined.


In a recent paper Prof. A. 0. Rankine has put forward a number of criticisms of the results obtained from, and the experimental method employed in, the determination of the temperature coefficient of viscosity of air by the present author. In the first place, a comparison is drawn between the author’s results and those of other observers in the lower part of the temperature range, and the conclusion is drawn therefrom that there is a possibility of an error of 3 percent, in the author’s measurements throughout the whole range of temperature used. This inference is reached from the figures quoted in Table II of Rankine’s paper, in which the temperature range from 15° to 183° C. is considered. That some difference exists between the author’s results and those of other observers in the lower part of the temperature range is clear, but it must again be emphasised that the values given for low temperatures are not experimental values, but were obtained by an extension of the graph (fig. 2) for higher temperature measurements to the value of the viscosity as given by Millikanj for room temperatures. A slight alteration of the curvature of this extension would make an appreciable change in the ratios η100/η15 and η183/η15 , but this would not be sufficient to account for the curvature at B in fig. 3 of the original paper. If the values of T ⅜ /η for Breitenbach’s results at 182° C. and 302° C. are plotted on this curve, they lie above the present results and on a curve which would intersect AB at about 600° C. That part of the difference is due to this cause seems to be indicated by the fact that the difference diminishes as the temperature rises. Thus at 300° C. the following values of η300/η15 are obtained by Breitenbach, the only other worker at this temperature, and the author. The figures used are those given by Rankine.


Sign in / Sign up

Export Citation Format

Share Document