Index of content:
Volume 35, Issue 10, 01 October 1964
35(1964); http://dx.doi.org/10.1063/1.1713106View Description Hide Description
Velocity modulation of a one‐dimensional electron beam is treated relativistically, starting from the collisionless Boltzmann equation and using first‐order perturbation theory. A general expression for the perturbation current densityj 1 is applied first to a mono‐energetic unperturbed beam. Compared to the non‐relativistic case, the magnitude of j 1 is reduced by the factor (1−u 2/c 2)3/4, where u is the unperturbed beam velocity and c the velocity of light. The distance between the maxima of j 1 is increased by the inverse of the same factor. Relativistic effects decrease the efficiency of modulation η, which is defined as the energy carried by the electric field generated by the perturbations, divided by the energy flux in the unperturbed beam. For an unperturbed beam with a rectangular velocity distribution of narrow width w and a fixed mean velocity um , the results are independent of w to first order. Expressions for j 1 and η are also found for an unperturbed beam with a rectangular momentum distribution of arbitrary width. When this width is small and increasing as um is fixed, j 1 increases in second order if the beam is highly relativistic or if its plasma frequency is larger than the frequency of modulation, while it decreases otherwise. Under either condition η decreases in second order.
35(1964); http://dx.doi.org/10.1063/1.1713107View Description Hide Description
The predictions of the Archard diffusion model for electron trajectory within a specimen are compared with experimental data. The predicted distributions in depth of primary x‐ray emissions are calculated. The applicability of the model to the absorption corrections in electron microanalysis is discussed in light of these calculations.
35(1964); http://dx.doi.org/10.1063/1.1713108View Description Hide Description
Only a few experimental techniques permit the determination of inhomogenities in the electron density of a plasma, and even these lose their usefulness near the walls of the plasma container. A microwavesurface wave technique using the container boundary as a propagating surface is described. Under the assumption that the plasma is azimuthally symmetric and axially uniform, perturbation relations for the microwave field permit an integral transform to convert phase shifts into radial electron density. This technique applied to a 2‐Torr neon rf‐generated plasma in coaxial geometry indicates an electron density near the inner wall as having a square‐law distance dependence. This is not in accord with the well‐known theoretical models of diffusion‐limited plasmas, but can be explained qualitatively.
35(1964); http://dx.doi.org/10.1063/1.1713109View Description Hide Description
The electron ionization coefficient α/p 0 (cm×Torr)−1 has been measured over an E/p 0 (volts/cm×Torr) range from 20 to 100 in deuterium subjected to various purification procedures. The results obtained lie slightly below those previously reported by Rose; however, the agreement is within the combined experimental error. Comparison is also made with the measurements of the ionization frequency ν i/p (sec×Torr)−1 of Cottingham and Buchsbaum. The present results, at a given E/p 0 value, lie above their data. The discrepancy between the ac and dc results is discussed.
35(1964); http://dx.doi.org/10.1063/1.1713110View Description Hide Description
35(1964); http://dx.doi.org/10.1063/1.1713111View Description Hide Description
Some plasmas contain condensed species as liquid or solid particles which may be hot enough to emit electrons thermionically into the plasma. The charge and voltage distributions in the vicinity of the particle are determined from the Poisson—Boltzmann equation and from a plasma sheath theory that is modified to account for thermionic emission. There is a critical temperatureT 0 at which the sheath changes from an ion sheath to an electronsheath. At temperatures greater than T 0 an electron cloud is injected into the plasma, but it is confined near the particle surface within a distance less than the Debye shielding distance. Some typical voltage and charge density profiles are given for a rocket exhaust plasma.
35(1964); http://dx.doi.org/10.1063/1.1713112View Description Hide Description
The image quality obtained with neon as an imaging gas was studied. Neon can give stable images of non‐refractory metals because its ionization field is lower than the evaporation field of these metals. Field evaporation occurs in helium and in neon at slightly lower fields than in vacuum when the incoming nonionized gas molecules which have both thermal and dipole attraction energy collide with the surface atoms. Localized field evaporation leads to dynamic end forms of the emitter tip which are characteristic for vacuum, neon, and helium. The ion image obtained with helium containing a few percent neon has a reduced contrast and shows more details around the low‐field areas. Field evaporation in this mixed gas produces surfacevacancies inside closed net planes. These effects indicate that adsorbed neon atoms act as intermediate collision partners to allow a more efficient transfer of the energy from the light helium atom to the heavy metal atom.
35(1964); http://dx.doi.org/10.1063/1.1713113View Description Hide Description
A dc probe was placed behind a small slit in the collector electrode in each of several cesium‐vapor thermionic diodes. These systems were used to study the velocity distribution of the small fraction of the emitted electrons which passed through the collector slit. This technique overcomes several of the inherent disadvantages in the use of the dc Langmuir probe in cesium vapor. The error due to momentum defocusing in the collector slit is calculated with the presence of a plasma in the emitter—collector interspace. The probe log‐current vs voltage curves could be accurately fitted to a Maxwell—Boltzmann velocity distribution over several cycles in log‐current. In this manner, effective temperatures were defined as a measure of the randomness of the velocity component normal to the collector surface. No fully randomized, high‐temperature electron group was detected in the unignited mode of these diodes. The effective electron temperature in the unignited mode became increasingly larger than the emitter temperature as the latter temperature was increased, but the departure was always less than 25%. The effective electron temperature increased rapidly with diode current as the ignited mode was entered and rose to be 470% higher than the emitter temperature.
35(1964); http://dx.doi.org/10.1063/1.1713114View Description Hide Description
A set of general equations describing the time‐dependent behavior of a laser is derived. The derivation is based on a density matrix description of a homogeneously broadened laser transition and a semiclassical treatment of the radiation field, consisting of an arbitrary number of modes. The resulting equations resemble ordinary rate and energy conservation equations but contain additional interference terms which, in general, couple all of the modes. The strength of this coupling depends on the inhomogeneity of the population inversion as well as on the mode frequencies. The energy conservation equations imply that absorption and stimulated emission in such a system can apply, strictly speaking, only collectively to the system of modes as a whole. In addition, the threshold inversion is affected by mode coupling. Stability analysis carried out in the two‐mode case indicates that a spatial inhomogeneity in the pumping rate is required for undamped oscillations. It is also concluded that coupling of off‐axis modes is more likely to lead to undamped spiking than coupling of axial modes.
35(1964); http://dx.doi.org/10.1063/1.1713115View Description Hide Description
The thermionic work function of aluminumelectrodes with anodized aluminum oxide coatings of 30 and 2000 Å was determined for each electrode from the contact potential difference of each electrode with a well outgassed and aged polycrystallinetungsten wire emitter. In vacuum, the thermionic work function of the 30‐Å coated electrode was 4.3±0.1 eV, relative to a polycrystallinetungstenwork function of 4.6 eV and a tungsten emission constant of 120 A/cm2· °K2. In cesium vapor at reservoir temperatures up to 70°C, the 30‐Å coated electrode had a minimum observed work function of 1.4±0.1 eV relative to the same tungsten reference electrode. The resistance of the 2000‐Å film coated collector electrode was too large to allow accurate determination of the work function of this electrode. In both vacuum and cesium vapor, however, there was an indication that the work function of the 2000‐Å coated electrode was a few tenths of an electron volt lower than that for the 30‐Å coated electrode. It is suggested that similar but perhaps less stable oxide‐coated electrodes could occur naturally on the collector of a thermionic energy converter due to oxidation of the collector surface or buildup on the collector surface of outgassing products from the emitter.
Auxillary experiments were performed on the secondary‐electron emission from these oxide‐coated electrodes. No lingering secondary electron emission current corresponding to a Malter effect was observed. The true secondary yield from the 2000‐Å coated electrode was 3 in vacuum and 7 in cesium vapor for a primary electron energy of 150 eV. The secondary yield for the 30‐Å coated electrode was similar in cesium vapor but was not studied in vacuum.
35(1964); http://dx.doi.org/10.1063/1.1713117View Description Hide Description
High‐frequency electromagnetic waves guided by an air‐bounded, axially magnetized, plasma column of radius a are studied through the technique of parameter space mapping. The parameter space is 3‐dimensional with coordinates ω p /ω, ω b /ω, and c/ωa. In this 3‐space, cutoff and resonance loci form surfaces which enclose approximate domains of existence for various classes of waves.
Three wave classes, characterized by distinctive cutoff and resonance surfaces, are identified for a magnetized plasma column, and their relation to waves in a boundless plasma, and on a column of small radius (quasistatic waves) is investigated. Also, evidence is found for the continued existence of the surface wave class into the regime ω b >ω p .
Parameter space analysis leads to a determination of salient dispersion characteristics for all wave classes and all order of azimuthal variation by a method contributing to physical insight but which avoids exact but laborious numerical computation. Waves are found which change from forward to backward or vice versa with increasing wave number.
35(1964); http://dx.doi.org/10.1063/1.1713118View Description Hide Description
Breakdown by thermal instability of a field emitter is analyzed, taking into account the temperature dependence of field emission and of resistivity. Beyond a certain temperature the emission increases while the necessary field drops. It is shown that for a whisker‐like emitter this instability occurs when the emitting tip is only several hundred degrees centigrade hotter than the bulk of the cathode.
35(1964); http://dx.doi.org/10.1063/1.1713119View Description Hide Description
H2 + ions produced by a Philips ionization gauge (PIG) type discharge in hydrogen were injected into a cylindrical electric field which was varied in magnitude to give a maximum strength ranging from 65 to 450 kV/cm. H2 + ions in highly excited vibrational states were dissociated by the electric field. The measured fraction of the H2 + ions produced by the PIG discharge which were dissociated was 0.0011±0.0002 at a field of 200 kV/cm. Numerous experimental checks demonstrated that the dissociation fraction could not be attributed to effects other than electric fielddissociation. The fraction is several times greater than that measured by other investigators. This difference is attributed to the method used in the preparation of the H2 + ions.
35(1964); http://dx.doi.org/10.1063/1.1713120View Description Hide Description
The contact potential difference in a low‐pressure thermionic energy converter can be written as , where and are averaged work functions of the emitter and collector, respectively. The estimation of and from existing work‐function data is discussed. As an example, values are estimated for well‐aged polycrystallinetungsten wire and those estimates are compared with experimental values.
35(1964); http://dx.doi.org/10.1063/1.1713121View Description Hide Description
The internal modulation of lasers by a variation of cavity losses is analyzed in terms of the normal modes of the system. Time‐dependent perturbation theory is used to describe the results of resistive mode coupling when the losses are modulated with small signals. For the single‐mode oscillator, an expression is obtained for the amplitude distortion in the modulation index of the light produced by a coupling‐type internal modulator when the losses are modulated at a frequency commensurate with the separation between longitudinal modes of the laser cavity. Low distortion is obtained when the modulating frequency is noncommensurate with the longitudinal‐mode frequency separation. The internal coupling‐type modulator in its present form can provide small amounts of modulation over bandwidths limited to the separation between adjacent interferometercavitynormal modes.
For a multimode oscillator, modulation of the internal losses at a frequency equal to the separation between adjacent longitudinal modes produces a pulse‐modulated output wave. The average intensity is unchanged and the peak intensity is increased over the intensity of the unmodulated laser by a factor equal to the number of oscillating modes.
35(1964); http://dx.doi.org/10.1063/1.1713122View Description Hide Description
The origin of moving striations has been investigated by measurement of striation propagation at a sharp change in tube diameter, by study of the relationship between anode spot oscillations and moving striations, by observation of anode spot oscillations at the cutoff current for moving striations, by use of an auxiliary discharge to suppress anode potential oscillations, and by study of a striation system isolated from electrodes. The experiments indicate that: (a) the striation frequency, velocity, and wavelength all change discontinuously at a change in tube diameter; (b) visible spot oscillations are accompanied by oscillations in discharge potential, and exist unchanged above the critical current for disappearance of moving striations; (c) anode oscillations and anode spots may be suppressed by operation of an auxiliary discharge to the anode; (d) the striation frequency synchronizes to an applied oscillation at the anode, provided the applied frequency is close to a small integral multiple of the natural striation frequency; and (e) a striation system can be maintained in a vessel separated from the electrodes by narrow tubes operated above the local critical current for moving striations. No oscillations can be detected in the isolating regions either photoelectrically or with electrostatic probes.
It is concluded that: (1) anode oscillations and moving striations are independent phenomena; and (2) moving striations do not constitute the response of the discharge to an external perturbation, but appear to result from a local instability in the positive column.
35(1964); http://dx.doi.org/10.1063/1.1713123View Description Hide Description
Thermionic diodes containing xenon at 1 Torr and cesium at 10−6−10−3 Torr can lead to higher diode currents than those with either gas alone at the given pressures.Negative resistance thermionic diodes result from xenon in the appropriate pressure range plus cesium.
35(1964); http://dx.doi.org/10.1063/1.1713124View Description Hide Description
An experimental tube to obtain accurate and reproducible data relating to the efficiency of the hollow cathodedischarge had a cathode of two plane‐parallel plates with variable separation, and a movable anode to eliminate the anode fall region. The cathode temperature was kept between 25° and 35°C by employing pulse techniques. Preliminary studies performed in helium and neon at pressures between 2 and 15 Torr gave the following results: (a) The transition from the plane cathodedischarge into the hollow cathodedischarge is a smooth function of the cathodic plate separation. (b) For each tube current and pressure there exist two or more optimum values of the cathodic plate separation for which the tube voltage has a minimum. The optimum distances depend on type of gas, pressure, and tube current. (c) For fixed tube current, the relative maximum decrease in tube voltage with respect to the plane‐cathode discharge voltage is, in most cases, larger in helium than in neon. (d) The variations of the discharge voltage and current with pressure are consistent with the Townsend scaling law for certain values of cathodic plate separation.
35(1964); http://dx.doi.org/10.1063/1.1713125View Description Hide Description
The velocity of 180° domain walls in BaTiO3 crystals with saltwater electrodes has been measured from 3° to 75°C at applied fields from 0.2 to 200 kV/cm by repeated partial switching and etching. Heating the crystal was found to lower the field required to attain any given wall velocity by a factor which is substantially the same for all wall velocities (or for all applied fields) and is determined wholly by the temperature change. The factor is about 3 for a change from 3° to 75°C. This behavior agrees with the Miller‐Weinreich mechanism of thermally activated wall movement. Domain shapes are found to be very rounded at 50° and 75°C, becoming almost exactly circular at 75°C around 1 kV/cm, in marked contrast to the well‐known square domains at 25°C and below.