Central claims: Langmuir-Irving (Tonks and Langmuir, "Oscillations in Ionized Gases", Phys. Rev. 33, 1929)
Source gap note: there is one document, and its text layer is 100% OCR (17 pages), which garbles every equation. The theory part (PDF pp.1-9) was re-read from page images and the equations transcribed [read: page image]. Of the experimental part, pp.5-6 and Figs. 8-10 (pp.13-15) have since had deep reads from page images, and the Step 2 reviewers read pp.11-17 from page images; the QE agent checked the OCR text of pp.8-17. Still not read from page images by any dedicated deep read: pp.9-10 (apparatus, detector calibration), the 100-300 MHz section (pp.10-12) and Figs. 2-7; these rest on the digest and OCR text (QE deep reads DR3 and DR4 not run). Digest errors to avoid: no epsilon in the formulas, the dispersion term is c^2/lambda^2, and the sound speed is (kT_e/m_p)^(1/2).
- An ionized gas has a natural electron oscillation ("plasma-electron oscillation") with frequency nu = (n e^2/(pi m))^(1/2) = 8980 n^(1/2) cycles/s (Gaussian units), which is about 9e8 Hz at n = 1e10 cm^-3. [read]
- In the model used, these oscillations have zero group velocity, and electromagnetic waves in the plasma obey nu^2 = n e^2/(pi m) + c^2/lambda^2. [read]
- With Boltzmann-distributed electrons, ions support "plasma-ion" oscillations. These lie at about 1.5 MHz and below for mercury and pass to sound-like waves with v = (kT_e/m_p)^(1/2) (about 6.5e4 cm/s) at long wavelengths, with a Debye-length-scale changeover. [read]
- Mercury arcs emit radio-frequency oscillations from about 1e7 to 1e9 Hz as noise. Wavelengths in the 300+ MHz range scale as relative ionization to the power -1/2, as the plasma-electron formula predicts (the authors: 13 of 15 points within error; digitised Fig. 10 gives a log-log slope of -0.60 +/- 0.16 over a factor 3.4 in ionization, consistent with -1/2 and not with 0; the ionization values are relative only and the authors say not to rely on them too much), and the arc shows two stable scattering states with frequencies of 1080 MHz (low scattering, 26 V anode) and 660 MHz (high scattering, 28 V anode), a ratio of 1.64 that the authors call "almost twice". The 300+ MHz frequency is independent of collector potential from -80 V (to cathode) to 5 V above anode potential. [read: pp.13-15, page images]
- The authors suspect that these oscillations explain the anomalous electron scattering and energy gain in low-pressure Hg arcs, but they call their correlation runs "inadequate", say the oscillation voltage amplitudes "appear to be very small" (crystal readings of about 9 to 90 mV), and report that the beam-oscillation identification is problematic: the body (p.16) says the 55 MHz line cannot be identified with the computed 18 MHz beam oscillation, although the abstract still offers "50 to 60 megacycles may correspond to oscillations of the beam electrons". The 1.38 MHz ion-range resonance was not confirmed with a tuned amplifier. They also give a qualitative thermal-electron argument (fast electrons drop out of the effective density, so oscillations below a "minimum size" may be impossible, with frequency falling as wavelength falls, the opposite sign to later Bohm-Gross dispersion) and a beam-to-plasma-electron resonance idea, without calculations. [read: pp.1-2, 5-6, 12-16]