Energy loss of relativistic electrons and its fluctuation in gas proportional counters

Absolute measurements of the energy loss and its fluctuation of relativistic electrons in this gas layer over the momentum region of 50-800 MeV/c have been extensively achieved with high accuracy using a 100 cm proportional counter, operated with a conventional mode or a special mode (the central wire is divided by a small glass bead into two sensitive lengths of 10 cm and 90 cm). The counter is filled with Ar + 10% CH₄ mixture at various pressures between 0.1 and 2.0 atm. The experimental distributions of the energy losses are compared with those predicted by the Landau theory, the Blunck-Leisegang theory, the Chechin-Ermilova theory and the Monte Carlo method (Ispiryan et al., Cobb et al. and Ermilova et al.). The observed fractional widths agree relatively well with those of the Monte Carlo calculations. The observed values of the most probable energy losses for small thickness (≤3 × 10^-2 g/cm²) of gas layer are 10-25% smaller than those predicted by the Landau-Sternheimer theory and they approach to the theoretical ones with increase of the thickness. A part of such a discrepancy between the theory and the experiment may be explained by the fact that the inner shell electrons of argon atoms in the absorber material do not contribute to the most probable energy loss, when the energy loss is not larger than the binding energy of K-shell electrons. The explanation, however, does not give an understanding for the fact that even when the thickness of gas layer is equal for the short and long counters, the difference in the most probable value between both counters is larger than that predicted by the Landau-Sternheimer theory. This suggests that the density effect correction given by Sternheimer may be insufficient to the thin gas layer. The observed most probable values are also compared with those predicted by the Chechin-Ermilova theory and the Monte Carlo method (Cobb et al. and Ermilova et al.). These predictions do not give a satisfactory explanation for the decrease of the observed values. However, the Monte Carlo results calculated by Ermilova et al. show the tendency close to the observed data.