Oscillations In Ionized Gases

VOL. 14, 1928 oPHYSICS: I. LANGMUIR 627 cisely the plane grating formula. It will be noted that, with the exception of two, all of the points fall to the right of the regions of "anomalous dispersion," and that none of them falls in this region. It is due to this circumstance presumably that the displacements of the electron diffraction beams from their x-ray analogues display no marked abnormalities. It will be noted also that although the values of ,u calculated from the diffraction beams are rather scattered they are not inconsistent with the dispersion curve constructed from the more precise data of the reflection beams. I Davisson and Germer, Proc. Nat. Acad. Sci., 14, 317 (1928). 2 Davisson and Germer, Nature, 119, 558 (1927); Phys. Rev., 30, 705 (1927). 3Eckart, Proc. Nat. Acad. Sci., 13, 460 (1927). 4Bethe, Naturwiss., 15, 787 (1927). c Bethe, Ibid., 16, 333 (1928). 0 Andrewes, Davies and Horton, Proc. Roy. Soc., 117, 660 (1928). OSCILLA TI&NS IN IONIZED GASES By IRVING LANGMUIR RESEARCHLABORATORY, GsERALu EIXCTRIc Co., SCH NsCTADY, N. Y. Communicated June 21, 1928 In strongly ionized gases at low pressures, for example in the mercury arc, the free electrons have a Maxwellian velocity distribution corresponding to temperatures that may range from 5000° to 60,0000, although the mean free path of the electrons may be so great that ordinary collisions cannot bring about such a velocity distribution. Electrons accelerated from a hot cathode (primary electrons), which originally form a beam of cathode rays with uniform translational motion, rapidly acquire a random or temperature motion which must result from impulses delivered to the electrons in random directions. In this laboratory we have been studying these phenomena' in detail during the last 4-5 years, but the mechanism underlying the Maxwellian distribution and its extremely short time of relaxation have not been understood. At an early date it occurred to me that electric oscillations of very high frequency and of short wave-length in the space within the tube might produce a scattering of the kind observed, but calculation showed that average field strengths of several hundred volts per centimeter would be necessary and this seemed an unreasonable assumption. Experiments capable of detecting oscillations of the electrodes with amplitudes greater than 0.2 volt failed to show such oscillations. Ditmer2 although unable to detect oscillations, concluded that oscilla- 628 PHYSICS: I. LANGMUIR PROC. N. A. S. tions of frequencies higher than 108 probably caused the scattering but was not able to suggest why there should be such oscillations. Penning3 detected oscillations of frequencies from 3 X 108 to 6 X 108 per second and found that the electron scattering and the oscillations nearly always occurred together. No cause was assigned for the oscillations. Dr. Tonks and I have repeated and have confirmed Penning's observations. The amplitude of the oscillations is small, less than 0.2 volt, and frequencies up to 1.2 X 109 have been observed. These waves, which are of well-defined frequency, appear to depend on the presence of the ultimate electrons (low velocity electrons) and can be observed in any part of the bulb. Other oscillations of much lower frequency (2 X 107 to 20 X 107 per second) can be observed in regions transversed by the primary electrons, which are injected into the tube with velocities corresponding to 25 to 70 volts. These oscillations and the scattering of the primary electrons are observable only when the current of primaries is raised to 10 milliamps. or more. It seemed that these oscillations must be regarded as compressional electric waves somewhat analogous to sound waves. Except near the electrodes, where there are sheaths containing very few electrons, the ionized gas contains ions and electro