SummaryMeasurements of the drift velocity, the ratio of diffusion coefficient to mobility, and the "magnetic drift velocity" for electrons in helium have been made at 293°K in the range 1· 8 X 10-19 < E/N < 3 X 10-17 V cm 2 • From an analysis of the drift velocity data, an energy-dependent momentum transfer cross section has been derived for which an error of less than ± 2 % is claimed over the central portion of the energy range. The cross section agrees with the theoretical cross section of La Bahn and Callaway to within 2% over the whole energy range. The agreement with the cross section derived by a number of procedures from the total elastic scattering cross section measured by Golden and Bandel is less satisfactory. The drift data are sufficiently accurate to enable a search to be made for the effects of fine structure in the cross section at low energy. The results do not support the existence of such structure.

The momentum transfer cross section for electron-argon collisions in the range 0-4 eV has bel(n derived from an analysis of recent measurements of DTIIl as a function of EIN at 294 K (Milloy and Crompton 1977a) and Was a function of EIN at 90 and 293 K (Robertson 1977). Modified effective range theory was used in the fitting procedure at low energies. An investigation of the range of validity of this theory indicated that the scattering length and effective range were uniquely determined ,and hence the cross section could be accurately extrapolated to zero energy.It is concluded that for 8 ,;;; O· 1 e V the error in !he cross section is less than ± 6 % and in the range 0·4 ';;;8 (eV) ,;;; 4-0 the error is less than ± 8 %. In the range 0·1 < 8 (eV) < 0·4 the presence of the minimum makes it difficult to determine the errors in the cross section but it is estimated that they are less than -20 %, + 12 %. It is demonstrated that no other reported cross sections are compatible with the experimental results used in the present derivation.

Negative differential conductivity in gases has been studied using simple models of elastic and inelastic collision cross sections for electron scattering. The use of such models has demonstrated features of the cross sections that lead to the phenomenon, and shown that it can occur without a Ramsauer-Townsend minimum (and even without a sharply rising momentum-transfer cross section) or a special combination of inelastic cross sections.

SummaryThe use of data from swarm experiments for electron energies approaching those corresponding to thermal equilibrium demands results of greater precision than has hitherto been available. In order to examine the possibility of producing such data, the swarm method for detennining WID has been extensively examined over a range of values of the parameter Elp where the agreement between the results of recent investigations is not good. A number of factors influencing the accuracy of measurements of this type are discussed. The results for hydrogen which are presented are considered to be subject to an error of less than 1 %. I. INTRODUCTIONThe results of swarm methods for determining the ratio WID of electron drift velocity to diffusion coefficient have found application in a number of recent papers in which collision phenomena between low energy electrons and gas molecules have been discussed (Gerjouy and Stein 1955;Huxley 1956Huxley , 1959Shkarofsky, Bachynski, and Johnston 1961;Frost and Phelps 1961). For some of these applications, more especially those dealing with collision phenomena for electrons with mean energy of several electron-volts, the accuracy of existing data is sufficiently good. Not only are the results of a number of investigations substantially in agreement for this energy range but the degree of accuracy to which they have been obtained is adequate for most purposes. On the other hand, for those applications where the difference in energy between the electrons and gas molecules is important, small errors in the determination of WID become significant as this difference approaches zero. Unfortunately it is in this energy range that the measurements become more difficult and the agreement between the results from various laboratories is not good.In this paper an account is given of a systematic investigation of the swarln method for determining WID using an apparatus which enables the dimensions of the diffusion chamber to be quickly and simply varied. A number of possible sources of error has been investigated both theoretically and experimentally, as a result of which it has been possible to determine the experimental procedures which lead to results of maximum accuracy. The results which have been obtained where these procedures were followed show good agreement over a wide range of values of the experimental parameters and it is considered that the values of WID in hydrogen from O'1

Ab8tractDrift velocities of electrons in helium at 76· 8°K have been measured for 8 X lO-20 <: EjN <: 2 X lO-17 V cm2• From these data, and the earlier measurements of Crompton, Elford, and Jory made at 293°K, the energy-dependent momentum transfer cross section has been determined for electrons with energies between 0-008 and 6 eV. The present cross section agrees with that of Crompton, Elford, and J ory to within 1 %. The extension of the energy range to 8 me V permits a direct determination of the scattering length, for which a value of 1·19ao is obtained.