Some interruption criteria for short high-frequency vacuum arcs

If the contacts of a vacuum interrupter open shortly before a current zero, the transient recovery voltage (TRV) can cause a reignition and reestablish the arc. When the current in a diffuse vacuum arc passes through zero, there is a distinct pause before the TRV builds up (approximately 40 ns for copper). During this pause the gap carries conduction current only with an ion component which depends on dI /dt, varying between 3 A for dI/dt=60 A/μs and 60 A for dI/dt=1235 A/μs. The ion current subsequently decays in tens or hundreds of nanoseconds. It can be distinguished from the displacement current at this time by varying dV/dt, keeping the other parameters constant. Among the interruption criteria for short high-frequency vacuum arcs, dI /dt prior to current zero and initial dV/dt are the most important. High values of dI/dt are more likely to precipitate reignitions, but breakdowns can occur after lower dI/dt's if the gap has been subjected to a high current for a relatively long time (>100 μs)     

The short vacuum arc. I. Experimental investigations

High-current vacuum arcs drawn between commercial radial-magnetic field, chromium-copper contacts were studied by high-speed photography. The aim of the investigation was to study phenomena of relevance to high-current interruption, such as arc constriction and arc motion. The arcs were drawn at contact opening speeds typical of commercial devices, are duration being 10 ms or less. The arcs were `short' for much of their lifetime, and strong electrode-discharge coupling could be expected. Such arcs are also of principal interest. Arc behavior was found to be strongly influenced by the gap length d and the arc current I. No diffuse mode was observed at d less the d_min≈4 mm and at instantaneous currents I above I_max±25 kA. The diffuse discharge mode was assumed more readily when d was large. At d below 2 mm, the arcs could not be moved by a magnetic field. Increasing both I and d resulted in an increase of the probability and duration of arc motion and of the arc path length. Arc speed was often irregular, showing that arc motion is also affected by parameters other than the purely electrodynamic ones     

The role of dust particles with large gyroradii in the `2/3'fall-down process

The authors consider the Alfven-Arrhenius fall-down process and propose a mechanism whereby the Rosseland electric field (the field needed to maintain quasineutrality) may be responsible for the capture and confinement of large-gyroradius dust particles within a plasma shell stratified along the direction of the magnetic-field lines. For these particles, the effect of the magnetic force is rather weak, and they move with a constant z component of the angular momentum in a one-dimensional equivalent potential (gravitational plus centrifugal). This has a maximum at the equator and a minimum at the `2/3' points, i.e. the points where the field-aligned components of the gravitational and centrifugal forces balance. It is shown that under suitable initial conditions these are points of maximum dust density and minimum plasma density. The plasma-planetisemal transition is therefore expected to take place at the `2/3' points in accordance with the Alfven-Arrhenius mechanism. It is also shown that the fraction of infalling dust particles that can accrete onto the equatorial plane by the Alfven-Arrhenius and Rosseland mechanisms is rather small (~(L/ R_e)≪1), L being the thickness of the plasma shell, and R_e, a characteristic length scale of the field line     

Theory of the cathode sheath in a vacuum arc

A previous theory of the plasma sheath transition starting from the charge exchange model for ion collisions is extended to account for ionization and recombination. It is applied to the quasi-neutral boundary layer (presheath) in front of the cathode sheath of a vacuum arc. An essential potential and density difference between the sheath edge and cathodic plasma ball is found. This difference is accounted for in a unified theory of the arc cathode based on G. Ecker's (1971) existence diagram method, which indicates possible areas of arc operation in the T_cj plane, where T_c is the spot temperature and j is the current density. A numerical evaluation for Cu gives the results which are qualitatively similar to Ecker's theory. The existence areas are quantitatively enlarged and shifted to lower current densities     

A revised theoretical model of vacuum arc spot plasmas

The quasi-stationary hemispherical expansion of the cathodic plasma in vacuum arcs can be modeled with hydrodynamic two-fluid equations. In any case, the state of the plasma is determined by the only variable (I/r)^2/5 (with current I , distance r). In order to avoid some deficiencies of the model (as published) and to investigate more carefully the dependence of the plasma parameters on the arc current, the known analytic solution to the problem is improved by taking into consideration the variability of the Coulomb logarithm and the dependence of the boundary conditions on I. These effects are treated separately. Examples are used to illustrate the new results, with particular emphasis on ion acceleration. The influence of the above factors turns out to be rather unimportant. Quantitatively, they cause some shifts, but no qualitative change of the basic behavior of the plasma is seen     

Pressure limits for the vacuum arc deposition technique

Observations of the cathodic copper plasma expansion at low pressures of He, Ar, and SF₆ showed that, for background gas mass densities of ρ_g=1 to 4×10^-4 kg/m ³ and higher, the plasma and gas are separated into two volumes. A shock wave acts as a boundary between the two volumes. The boundary attains a stationary position once its expansion velocity decreases to the velocity of sound in the background gas. This position corresponds to a distance R_c to the cathode that agrees with a snowplow expansion model, giving R_c ^βf=E_r, where f is a function of the arc current and background gas characteristics, E _r is the erosion rate of the cathode, and β varies between 2.1 and 2.5. The interaction model is based on kinetic energy exchanges between two gas-like volumes without other energy losses. A maximum pressure limit for vacuum arc deposition is set for ρ_g /I=2 to 9×10^-6 kg/m³ A     

Ion current collected at various distances and argon backgroundpressures in a copper vacuum arc

The ion current collected by a probe biased at the cathode potential and located behind an annular anode of a vacuum arc is measured as a function of distance to the cathode and background argon pressure. The arc is formed between a circular Cu cathode and an annular anode. Arc current is 170 A, and the arc duration is 0.9 s. The arc is ignited by momentary contact of a movable W trigger rod (held at anode potential) with the cathode. Arc voltage, arc current, and ion current are measured using an analog data acquisition card and a personal computer. Arc voltage and arc current values are stable during the arc and their normalized standard deviation is less than 0.07. Ion current is noisy and fluctuates during the arc with a normalized standard deviation that varies from 0.5 at p<0.1 torr up to more than 1.5 at p>1 torr     

A hydrodynamic model of vacuum arc plasmas

The properties of plasmas expanding from cathode spots of vacuum arcs are calculated with a one-dimensional two-fluid model. The system of simplified hydrodynamic equations can be solved under stationary conditions using asymptotic power series. Although necessarily only an approximation, such analytical solutions prove to be advantageous compared with numerical integrations. All the plasma parameters are functions of (I/r)^2/5 (current, I: distance, r). The three forces accelerating the ions to high kinetic energies are quantitatively calculable: the electric field, the ion pressure gradient, and the electron-ion friction. The potential is decreasing towards the anode, and the residence of the plasma is negative. The ion temperature reaches only about 35% of the electron temperature. Although only asymptotic, the solution is suited to describe the arc plasma in a sufficient manner all over the expansion region     

Interrupting capacity of vacuum interrupters depending on thefrequency of current

Results of arc reignition voltages during current interruption of frequencies from 5.9 to 60 kHz by a short vacuum gap are presented. Measured arc reignition voltages depend on current amplitude and frequency, the values of preliminary arc current at the moment of switching on the HF current, and the discharge mode in the preceding current semiwave. Threshold amplitudes of the first semiwave of currents l_m1l and I_m1h as a function of frequency are determined. I_m1l and I_m1h divide current into three ranges to which different kinds of arc reignition voltage distribution correspond. Particularly large dispersion of reignition voltages takes place in the current range from I_m1l to I_m1h. The threshold current I_m1h is inversely proportional to frequency in the range from about 10 to 60 kHz, which is in agreement with the elaborated mathematical model     

Finite conductivity and the interchange instability in railgun arcs

A first-order perturbation expansion of the MHD (magnetohydrodynamic) equation is used to describe the onset of the interchange instability due to the high accelerations in railgun plasma-arc armatures. J.D. Powell (1986) considered the problem of perturbing an initial isothermal equilibrium with appropriate distributed current and density profiles, treating the perturbation equations with infinite conductivity. Here, the authors model the arc the same way, but they include the effects of finite conductivity σ. A fourth-order mode equation is derived and solved numerically. The authors find continuous spectra of unstable modes for a nonzero plasma acceleration g, whose growth rates are mostly greater than √kg, depending on the values of k and σ. The resistivity always raises growth rates higher than Powell's especially for large k and large resistivity. The resulting growth rates in typical railgun situations are large enough to permit full development of the instability     

Development and test of an L-band magnetron

An L-band relativistic magnetron has been designed, built, and tested at approximately 1.1 GHz. The authors have obtained a peak RMS power of 2.4 GW in one waveguide, and a total of 3.6 GW from two waveguides. The overall efficiency reached 9%, compared with the input power. In contrast with previous S- and X-band magnetron results, the L-band magnetron impedance does not collapse at the end of the pulse, and microwave generation tends to last as long as high voltage is applied. This implies that large total energies can be produced by applying long high-voltage pulses     

Townsend-type breakdown as a criterion for high-frequency vacuumarc interruption at submillimeter gaps

In vacuum circuit breakers, multiple reignitions give rise to HF current arcing (≈500 A; ≈200 kHz). Due to the small contact distance and the very large current gradient, at every arc-current zero pressures of several tens of millibars can be expected. Very soon thereafter (≈30 ns) this gap is dielectrically stressed by the first component of the restriking voltage (≈10 MHz), originating from parasitic impedances. The combination of the associated high electric field and the relatively high neutral density may cause Townsend-type breakdown, leading to another half-sine of continued arcing. Both dielectric stress and residual neutral density are expressed as a function of di/dt, yielding values of interruptible di/dt as a function of the contact distance with the Townsend breakdown criterion. Comparison with experiments shows fair agreement in the range of di/dt of 100-1000 A/μs and distances of 0.1-0.5 mm for two different circuits     

Initial stage of discharge current growth in a triggered vacuum gap

The dynamics of light emission accompanying the initial stage of electric discharge in a substantially undervoltaged vacuum gap was studied with a knock-down model using high-speed photorecording. Voltage across the gap was maintained within the range of 0.5-5 kV, which corresponded to the minimum operating voltage of vacuum-triggered gaps. It was found that front layers of a plasma cloud near a cathode, formed by a firing pulse, scattered at a speed of (5-8)×10⁶ cm/s. During firing, a channel directed to the opposite electrode was formed from the plasma cloud near the cathode. It was found that the average switch-on delay time of the triggered vacuum gap is ~d(1+h/d) I_f^-α, where d is the interelectrode gap length, h is the trigger assembly penetration height, and I_f is the firing current     

A model of the cathode spot glow discharge for a cathode withinclusions of low work function material

A model that describes the self-maintained glow discharge for a composite cathode in which cathode spots are formed on low work function material inclusions is presented. Based on this model, expressions are obtained for the radius of the spot, the cathode voltage, and the spot current as functions of Pd (where P is the gas pressure and d is the thickness of the cathode layer), the parameters of the inclusions, and the basic material matrix are obtained. Numerical results for the case of a glow discharge in 75 torr neon gas with a copper cathode matrix and inclusions of LaB₆ are presented. The average radii of inclusions were 0.05 and 0.015 mm. Conditions under which the glow discharge migrates from the inclusions and begins to spread onto the surface of the basic matrix are obtained. The transition of the glow discharge into a contracted state, depending on the structure of the composite material, is also considered     

Electrical characteristics of pulsed glow discharges

The i-v characteristics, energy partitioning, and time evolution of the discharge current and reduced field (E/N) for a nitrogen discharge are simulated using a self-consistent calculation of the electron energy distribution function and the vibrational level populations. The model includes diffusion losses and takes account of the external circuit parameters. The results discussed are for pressures of 1-100 torr, discharge currents in the range of 10^-3-5.0 A, and a reduced field (E/N) in the range of 150-250 Td. For a typical discharge in a tube of 2-cm diam. and a current of a few amperes, the results show that the energy stored in the vibrational manifold saturates a few milliseconds after the initiation of the discharge     

Computation of the electric and magnetic fields induced in a plasmacreated by ionization lasting a finite interval of time

S.C. Wilks et al. (1988) showed that when an infinite expanse of gas, carrying a linearly polarized electromagnetic wave, is instantly ionized, the initial wave is frequency upshifted. This phenomenon of frequency upconversion through flash ionization gives rise to steady-state transmitted and reflected electromagnetic waves and to a time-independent magnetic field. The case in which the final state of ionization is achieved not instantly but in a finite turn-on time, 0⩽t⩽t 0, which is followed by the steady state, is studied. It is shown that the electric field is obtained from the one-dimensional Helmholtz equation, d²F(t)/dt² +ω₀²g(t)F( t)=0, if electrons are born at rest when they are created during ionization. As a result, the instantaneous frequency of the upshifted radiation is ω(t)=√g(t). The electric field can be solved exactly for specific choices of g(t). It is solved using WKB approximations for arbitrary g(t). The magnetic field is then found by integrating Faraday's law. It is found that the steady-state electric field amplitude depends on the steady-state value o f g(t) but does not depend on the ionization time t₀. Conversely, the static magnetic field amplitude decreases with increasing turn-on time     

The type and extent of non-LTE in argon arcs at 0.1-10 bar

A 30-A, 3-mm-diameter, wall-stabilized argon arc with 1% hydrogen is examined spectrometrically at pressures of 0.1-10 bar. Values of T_g, and T_exa diverge as pressure decreases below 5 bar (n_e⩽1×10¹⁷ cm^-3) at r=0. T_exβ is 20-40% larger than the other temperatures. The results are dependent on the transition probability scale used     

Electrical discharge initiation and a macroscopic model forformative time lags

Formative times in electrical discharges in overvoltaged gaps are analyzed with a model having no spatial dependence and with simple assumptions about discharge channel temperature T and discharge voltage. The model treats the early temporal evolution of the discharge. Specifically, the dissipative voltage drop, V*, across the discharge is taken to be a step function of T. Thermal quasi-equilibrium is assumed in the discharge medium, and it is shown that d(In t_d)/_d(In &thetas;)=-1, i.e., &thetas;t_d=constant, where &thetas; is the fractional overvoltage and t_d is the formative time lag, in agreement with measured values of t_d for much of the experimentally explored range of &thetas;. Highly-time-resolved (~92 ps) experimental data are presented for the first 10 ns of electrical discharge initiation; these data suggest that the authors' model should provide a reasonable representation of t _d when t_d>10-100 ns     

Acoustic double layers in multispecies plasma

Particle simulation in a one-dimensional bounded system is used to examine the formation of acoustic double layers in the presence of two ion species. Double-layer formation depends critically on the details of the distribution functions of the supporting ion populations, and their relative drifts with respect to the electrons. The effect of having two ion components, an H⁺ and an O⁺ beam, on double-layer evolution from ion acoustic turbulence driven by an electron drift relative to the H⁺ beam of ≈0.5u _e, where uu_e is the electron thermal speed, is examined. The ratio of ion drifts is taken to be consistent with acceleration by a quasi-static auroral potential drop (i.e. V _H/V_O=√M_O/ M_H=4.0). Acoustic double layers form in either ion species on the time scale τ≈100ω_ps^-1, where ω_ps is the ion plasma frequency for species `s' and s=H or O, and for drifts relative to the electrons lower than that required for double layer formation in simulations of single ion component plasma