Search

Thrusters that exploit vacuum arc discharges to produce high-velocity plasma jets directly or as sources of plasma that is subsequently accelerated electrostatically have been proposed or are currently under development. Vacuum arc discharges exhibit certain regularities in their behavior that allow the performance of these thrusters to be described by simple semiempirical models. Empirical data on the current density distribution, charge state and velocity of ions created in vacuum are discharges, and the total cathode mass loss rate are used to develop expressions for the expected thrust and specific impulse as a function of thruster geometry. Thruster electrical efficiency and thrust-to-power ratio are calculated based on measurements of the burning voltage for given thruster operating parameters. Estimates of achievable thruster performance for a wide range of cathode materials are presented. This analysis suggests that thrusters using vacuum arc sources can be operated efficiently with a range of propellant options that gives great flexibility in specific impulse. In addition, the efficiency of plasma production in these devices appears to be largely independent of scale because the metal vapor is ionized within tens of micrometers of the cathode electron emission sites, so this approach is well suited for micropropulsion. Index Terms--Cathodes, plasma arc devices, space vehicle propulsion, vacuum arcs.

After pioneering work in the 1980s, plasma-based ion implantation (PBII) and plasma-based ion implantation and deposition (PBIID) can now be considered mature technologies for surface modification and thin film deposition. This review starts by looking at the historical development and recalling the basic ideas of PBII. Advantages and disadvantages are compared to conventional ion beam implantation and physical vapor deposition for PBII and PBIID, respectively, followed by a summary of the physics of sheath dynamics, plasma and pulse specifications, plasma diagnostics, and process modeling. The review moves on to technology considerations for plasma sources and process reactors. PBII surface modification and PBIID coatings are applied in a wide range of situations. They include the by-now traditional tribological applications of reducing wear and corrosion through the formation of hard, tough, smooth, low-friction, and chemically inert phases and coatings, e.g., for engine components. PBII has become viable for the formation of shallow junctions and other applications in microelectronics. More recently, the rapidly growing field of biomaterial synthesis makes use of PBII and PBIID to alter surfaces of or produce coatings on surgical implants and other biomedical devices. With limitations, also nonconducting materials such as plastic sheets can be treated. The major interest in PBII processing originates from its flexibility in ion energy (from a few electron volts up to about 100 keV), and the capability to efficiently treat, or deposit on, large areas, and (within limits) to process nonflat, three-dimensional workpieces, including forming and modifying metastable phases and nanostructures. Index Terms--Ion-assisted thin film deposition, plasma-based ion implantation, pulsed bias.

The ion current from different cathode materials was measured for 50-500 A of arc current. The ion current normalized by the arc current was found to depend on the cathode material, with values in the range from 5% to 19%. The normalized ion current was generally greater for elements of low cohesive energy. The ion erosion rates were determined from values of ion current and ion charge states, which were previously measured in the same ion source. The absolute ion erosion rates ranged from 16-173 [micro]g/C. Index Terms--Cathodic vacuum arcs, ion current, ion erosion rates, ion flux.

Cathode spot phenomena show many features of fractals, for example self-similar patterns in the emitted light and arc erosion traces. Although there have been hints on the fractal nature of cathode spots in the literature, the fractal approach to spot interpretation is underutilized. In this work, a brief review of spot properties is given, touching the differences between spot type 1 (on cathodes surfaces with dielectric layers) and spot type 2 (on metallic, clean surfaces), as well as the known spot fragment or cell structure. The basic properties of self-similarity, power laws, random colored noise, and fractals are introduced. Several points of evidence for the fractal nature of spots are provided. Specifically, power laws are identified as signature of fractal properties, such as spectral power of noisy arc parameters (ion current, arc voltage, etc) obtained by fast Fourier transform. It is shown that fractal properties can be observed down to the cutoff by measurement resolution or occurrence of elementary steps in physical processes. Random walk models of cathode spot motion are well established: they go asymptotically to Brownian motion for infinitesimal step width. The power spectrum of the arc voltage noise falls as 1 / [f.sup.2] where f is frequency, supporting a fractal spot model associated with Brownian motion. Index Terms--Cathode spot, fast Fourier transform, fractals, random colored noise, vacuum arcs.

Experimentally observed charge state distributions are known to be higher at the beginning of each arc discharge. Until know, this has been attributed to cathode surface effects in terms of changes in temperature, chemical composition, and spot mode. Here it is shown that the initial decay of charge states of cathodic arc plasmas may be at least in part due to charge exchange collisions of ions with neutrals. Neutrals gradually fill the discharge volume, and therefore, the effect of charge exchange shows delayed onset after arc initiation. Besides desorbed gases, sources of neutrals may include evaporated atoms from macroparticles and still-hot craters of previously active arc spots. More importantly, atoms are also produced by energetic condensation of the cathodic are plasma. Self-sputtering is significant, and additionally, ions have a low sticking probability when impacting at oblique angle of incidence. Estimates show that the characteristic time for filling the discharge volume agrees well with the charge state decay time, and the likelihood of charge exchange is reasonably large to be taken into account. Index Terms--Charge exchange collisions, ion charge states, vacuum arcs.

The effect of ion mass and charge state on plasma transport through a 90[degrees]-curved magnetic filter is experimentally investigated using a pulsed cathodic arc source. Graphite, copper, and tungsten were selected as test materials. The filter was a bent copper coil biased via the voltage drop across a low-ohm, 'self-bias' resistor. Ion transport is accomplished via a guiding electric field, whose potential forms a 'trough' shaped by the magnetic guiding field of the filter coil. Evaluation was done by measuring the filtered ion current and determination of the particle system coefficient, which can be defined as the ratio of filtered ion current, divided by the mean ion charge state, to the arc current. It was found that the ion current and particle system coefficient decreased as the mass-to-charge ratio of ions increased. This result can be qualitatively interpreted by a very simple model of ion transport that is based on compensation of the centrifugal force by the electric force associated with the guiding potential trough. Index Terms--Arc plasma, cathodic vacuum arc, ion charge states, macroparticle filter, plasma transport.

Continuous discharges could only be obtained after enduring energy sources became available, namely in the form of a battery of electrochemical calls, invented by Volta in late 1799. Humphry Davy is often credited with the discovery of the arc discharge, which later led to the development of the carbon arc lamp. Indeed, as early as 1800, he obtained short pulsed arcs with his Voltaic pile. Independently, and earlier than Davy in the sense of continuous discharges, the Russian Vasilii Petrov of St. Petersburg made carbon arcs in 1802. Petrov used a pile of 4200 electrochemical cells to drive what was the most powerful discharge at that time. Petrov's publication of 1803 appeared only in Russian, and his work was ignored and forgotten for over a century. Davy pursued highly successful electrochemical experimentS and was unaware of Petrov's work. He increased the size of his battery in several steps, which led to increasingly powerful discharges, most likely an undesired side effect. After 1808, using the new battery of the Royal Institution, Davy demonstrated continuous arc discharges. The exact dates and circumstances of early arc demonstrations around 1810 are still the subject of research, but later arc experiments such as those at the London Institution of 1821 are well documented. While Petrov could claim priority for continuous carbon arcs, it was Davy who made a lasting impact on further development. Index Terms--Arcs, carbon arc lamp, history, Humphry Davy, electrochemical cell, Vasilii Petrov, voltaic battery.

The early development of arc plasma physics is closely related to the development of suitable sources of electrical energy. The harnessing of electrostatic charge in Leyden jars (early capacitors) enabled the controlled production of sparks and pulsed arcs. A contemporary introduction to sparks and arcs gives the foundation for critically assessing the early development, when observation and discovery of phenomena were still far from understanding. The modern development of discharge physics can be traced back to the 18th century. Notably, and generally unnoticed by the scientific community until today, Joseph Priestley observed cathode erosion and film deposition of cathode material as early as 1766. Index Terms--Arc, Capacitor discharges, cathodic arc deposition, history, Priestley Rings, spark discharges.

Plasma immersion ion implantation (PIII) offers high throughput and efficiency in the synthesis of silicon-on-insulator (SOI) materials. In the separation by plasma implantation of oxygen (SPIMOX) process, the spatial and time variation of the sample temperature must be known and well controlled to ensure uniform buried oxide and silicon overlying layer thicknesses over the entire silicon wafer. In this paper, we describe a two-dimensional model and derive the temperature distribution on the silicon wafer with respect to time and other process parameters. Our results show laterally nonuniform heating by the incoming ions and the local temperature is influenced more by the sample voltage and thermal irradiation coefficient of the target than the pulse duration and plasma density. The model provides a simple and quick means to determine whether external heating will be needed to maintain the sample temperature at 600[degrees]C during the SPIMOX process. Index Terms--Modeling, plasma implantation, temperature.

The cathodic vacuum arc is a simple means of creating a highly ionized metal or carbon plasma, which can be used to produce a wide variety of metal and metal compound thin films. In order to produce smooth, defect-free films suitable for optical and electronic applications, the plasma needs to be filtered of micrometer-sized molten and solid cathode debris known as macroparticles. To date, most of the research effort on plasma filtering has been focused on development and optimization of magnetic filter designs based on curved solenoids. Ryabchikov et al. proposed and employed novel filter designs for a combined plasma and ion beam source. The new designs are based on a nested set of conical plates in one version and a set of angled plates ('venetian blind') in another. In the present work, we report on a detailed characterization of the operation of a filter employing a linear venetian blind geometry. The effects of parameters such as plate and arc currents, as well as plate bias on the throughput and deposition profile are reported. Index Terms - Coatings, plasma arc devices, plasmas, vacuum arcs.

Charge state distributions (CSD's) of Fe, Co, Zr, Nb, Hf, Ag, Pt, Au ions produced at vacuum arc cathode spots are investigated without often-used averaging procedures. Statistical results are used to evaluate the contribution of fluctuations to the broadening of averaged CSD's. Calculations of Saha equations in the Debye-Huckel approximation are often successful in replicating experimental CSD's even when using the simple model of 'instantaneous freezing.' This allows us to use charge state spectrometry as a plasma diagnostic tool for the dense plasma near the cathode spot. Typically, total heavy particle (atoms and ions) densities and electron temperatures of [10.sup.24]-[10.sup.25] [m.sup.-3] and 3-4 eV have been measured this way at the transition zone from equilibrium to nonequilibrium, the 'event horizon.' It was also found that CSD's for some elements such as Co, Ag, and Hf could not be accurately simulated with the assumption of instantaneous freezing. Here, a more gradual transition to nonequilibrium occurs. This transition can be described by the model of partial local Saha equilibrium. Finally, an axial magnetic field was found to soften the transition to nonequilibrium thus causing broadening of experimental CSD's.

Micron and submicron 'macroparticles' are produced along with the plasma at vacuum arc cathode spots. Published data refer to the size range 0.2-100 [[micro]meters]. The lower limit is determined by the resolution of the equipment used (usually scanning electron microscopes). In the present study we focus on the detection and characterization of nanosize macroparticles ('nanoparticles') using atomic force microscopy and field-emission scanning electron microscopy. New information is gathered on material-dependent size distribution functions as well as on the effectiveness of magnetic filtering for nanoparticles.

Ion charge state distributions (CSD's) of alloy-cathode vacuum arc plasmas have been calculated under the assumption that thermodynamic equilibrium is valid in the vicinity of the cathode spot, and equilibrium CSD's 'freeze' when the plasma is rapidly expanding. In this way, experimental data of titanium-hafnium alloy-cathode plasmas have been simulated using a system of Saha equations generalized for multiple elements. It was found that the CSD's of titanium and hafnium freeze approximately at the same plasma temperature and density. The freezing parameters depend slightly on the plasma composition. For the example considered, freezing occurs at temperature of T = 3.0-3.8 eV and a total heavy particle density of order [10.sup.25] [m.sup.-3].

A hydrodynamic model of free-boundary vacuum arc plasma jet flow through a curved magnetic field has been developed. The plasma flow was modeled by sourceless steady-state equations, and the free boundary of the plasma was determined as a part of the solution of the gas-dynamic equations. It was found that the maximum plasma density shifts with respect to the centerline. The shift is maximum at the 90 [degrees] guide exit plane and for the case of low magnetic field. The maximum increase in longitudinal plasma velocity along the duct was shown to be a factor of three. The plasma distribution within the duct was analyzed experimentally using a deposition probe technique. The plasma jet expansion predicted by the model was compared with the experiment and a good qualitative agreement was found. Index Terms - Curved magnetic field, macroparticle filter, plasma jet.

The plasma composition and ion charge-state distributions of pulsed vacuum arcs have been investigated for carbon, aluminum, silver, platinum, and tantalum cathodes using a time-of-flight (TOF) charge-to-mass spectrometer. With the exception of carbon, it was found that all results depend on the arc pulse repetition rate, a fact which, up to now, has not been reported in the literature. It is shown that adsorption of gas on the cathode between arc pulses leads to contamination of the metal plasma and to a reduction of metal ion charge states. These usually undesired effects can be avoided by operating at high arc pulse repetition rates of order 10 Hz or more. The results can be interpreted in terms of cathode spot type 1 (on contaminated cathode surfaces) and type 2 (on clean surfaces) which are well known from their different brightness and erosion behavior. The transition between these modes was found to be gradual.

The plasma-generating parts of ion sources including their power supplies are usually floated to high potential (ion extraction voltage), thus requiring great insulation efforts and causing high costs for high-energy ion beams. A new concept for pulsed ion sources is presented in which a single power supply is used to simultaneously produce the plasma and high extractor voltage via a modified Marx generator. Proof-of-principle experiments have been performed with high-current spark discharges in vacuum where multiply charged ions are produced with this Marx-generator-based ion source ('Magis'). Using 'Magis,' it has been demonstrated that pulsed ion beams of very high energies can be obtained with relatively low voltage. For copper, ions of charge states up to 7+ have been found whose energy was 112 keV for a charging voltage of only 10 kV. Index Terms - Ion-charge state distribution, ion sources, Marx generator, metal ion beams, plasma generation.

A new magnetic macroparticle filter design consisting of two 90 [degrees] filters forming an S shape is described. The transport properties of this S filter are investigated using Langmuir and deposition probes. It is shown that the filter efficiency is the product of the efficiencies of two 90 [degrees] filters, and the deposition rate is still acceptably high to perform thin-film deposition. Films of amorphous hard carbon have been deposited using a 90 [degrees] filter and the S filter, and the macroparticle contents of the films are compared. Index Terms - Films, filtering, plasma generation, vacuum arcs.

The transport of vacuum arc plasmas through a 90 [degrees] curved magnetic macroparticle filter was investigated using a high-current pulsed arc source with a carbon cathode. The peak arc current was in the kiloampiere range, exceeding considerably the level of what has been reported in the literature. The main question investigated was whether magnetic macroparticle filters could be scaled up while maintaining the transport efficiency of small filters. In front of the cathode, we found that arc current dependent total ion saturation currents were in the range from 10% to 23% of the arc current. The best relative transmission was 25% (time integrated output/time integrated input) at a duct wall bias of 12.5 V and at an axial magnetic field of about 100 mT. The measured relative transmission of the used high-current arrangement is comparable to what has been observed with other low-current filters. The absolute measurable ion saturation currents at the filter exit reached 70 A at an arc current of about 1000 A. Index Terms - Carbon films, deposition, filtering, hard coatings, high-current, microparticle, plasma, vacuum arc.

We have investigated the charge state distributions of metal ions produced in a high current vacuum arc plasma located in a strong magnetic field. The arc current was varied over the range 200 A to 4 kA and the magnetic field was from zero up to 10 kG. In general, the effect of both high arc current and high magnetic field is to push the distribution to higher charge states - the mean ion charge state is increased and new high charge states are formed. These effects are explained in terms of increased power input (higher plasma temperature) and delayed freezing of the charge state distribution during the plasma expansion process.

Self-sustained self-sputtering occurring during high current pseudospark operation ([approximately equal to][10.sup.4]A/[cm.sup.2],I > [10.sup.3] A) is shown to be a possible mechanism for the superdense glow. The mean-free-path for ionization of cathode material sputtered in the low-current hollow-cathode phase can be shorter than the cathode-anode gap distance, and ionized atoms can return to the cathode surface, self-sputtering with a yield greater than one. The self-sputtered cathode atoms become ionized in the beam of electrons accelerated in the cathode sheath. A large fraction of the discharge current at the cathode surface can be carried uniformly over the surface by ions, and a very high electron emission density is not required to maintain the high current.

The plasma in vacuum arc cathode spots is always accompanied by a production of metallic macroparticles of size 0.1-10 micrometers. We have studied the macroparticle contamination of vacuum arc deposited thin films generated by a plasma source with an optional axial magnetic field, and have concentrated our attention on the macroparticle flux near the discharge axis. We have varied the arc current, metal species, deposition system geometry and axial magnetic field strength. Distribution functions for macroparticles of Pb, Ag, Cu, Pt, W, and Ni have been determined, normalized to the film thickness deposited or the charge transferred. Application of the axial magnetic field led to a considerable reduction of the normalized macroparticle flux since the plasma was effectively focused by the field whereas the macroparticle production was not influenced. The macroparticle content normalized to the deposited film thickness was reduced to about 20-35% of that without additional magnetic field.

The ignition and arc phases of vacuum arcs were investigated using differential dye laser absorption photography with simultaneous high spatial (micrometer) and temporal (nanosecond) resolution. The discharge duration was 800 ns, the current 50-150 A, the electrode material copper, and the cathode-anode distance less than 50 [Mu]m. A 0.4 ns laser pulse (tunable, [Lambda] = 480-530 nm) was used to obtain momentary absorption photographs of the cathode region. During ignition, an optically thick anode plasma expanded toward the cathode, decaying within 25 ns after bridging the electrode gap. In the arc phase, a fragmentary structure of the cathode spots was observed in situ for the first time. The microspots have a characteristic size of 5-10 [Mu]m. They appear and disappear on a nanosecond time scale. The plasmi density of the microspots was estimated o be greater than (3-6) x [10.sup.26] [m.sup.-3]. The results are in agreement with the model of explosive electron emission. A current density in the cathode spots as high as [10.sup.12] A/[m.sup.2] can be deduced.

Ion charge-state distributions have been measured with high time resolution for short (20 [micro]s) and long (250 [micro]s) vacuum discharges of 300 A. Charge-state data for 3 [micro]s after discharge ignition and quasi-steady-state values are given for most conductive elements in a Periodic Table, including data for a few elements (rhodium, europium, and terbium) that were never before reported in the literature. Ion charge states are significantly higher at the beginning of the discharge and decay to their quasi-steady-state are plasma values. It was found that the mean ion charge states can be fitted by functions of the form [bar]Q = [[bar]Q.sub.t [infinity]] [1 + A exp(-t/[Tau])] where A is an enhancement function that depends on the power density. For the present conditions, A ~ 1 and r ~ 50 [micro]s. Index Terms--Enropium, ion charge states, rhodium, terbium, vacuum arcs.

Chopping of current and induction of high-voltage spikes has been observed at the rising edge of cathodic vacuum arcs. It is argued that current chopping at the rising edge is similar to the current chopping effect that is well known for the arc current approaching current-zero. Current chopping can generate high-voltage spikes, and thus, cathodic arc equipment should be protected against voltage spikes that exceed the system's nominal voltage by an order of magnitude. Index Terms--Chopping current, vacuum arcs, vacuum interrupters, voltage spikes.

Thallium arc plasma was investigated in a vacuum arc ion source. As expected from previous considerations of cathode materials in the Periodic Table of Elements, thallium plasma shows leadlike behavior. Its mean ion charge state exceeds 2.0 immediately after arc triggering and diminishes to the predicted 1.60 and 1.45 after about 100 and 150 [micro]s, respectively. The most likely ion velocity is initially 8000 m/s and decays to 6500 and 6200 m/s after 100 and 150 [micro]s, respectively. Both ion charge states and ion velocities decay further toward steady-state values, which are not reached within the 300-[micro]s pulses used here. It is argued that the exceptionally high vapor pressure and charge exchange reactions are associated with the establishment of steady-state ion values. Index Terms--Cathodic vacuum arcs, ion charge states, ion velocity, thallium.

Cathodic vacuum arc plasmas can be used to produce well-adherent, dense, high-quality films and coatings, provided macroparticles are removed from the plasma. Both plasma and macroparticles are produced at cathode spots, and in most cases both can easily be observed because they emit light. The nature of this light emission is briefly discussed. Using open, curved solenoids as magnetic filters, one can readily optically record the separation of plasma and macroparticles by the filter. Imaging can be used as a powerful tool to help developing macroparticle filters and optimizing their design. Index Terms--Cathodic vacuum arc, imaging, macroparticles, plasma filter.

The quasi-linear constricted plasma source is a downstream plasma source with ten linearly aligned discharge cells. Each cell operates on the basis of a constricted glow discharge. The plasma output can easily be monitored by the plasma-emitted light. The information is not only intuitive but can also be used to operate on-line feedback control of the plasma source which is important for large-area plasma processing of materials. Index Terms - Constricted plasma source, gas plasma flow, plasma diagnostics, plasma processing of materials.

Ion charge states of vacuum arcs are enhanced at the beginning of each arc pulse. The ion charge states drop with time and become steady-state after about 100 [[micro]seconds], with some dependence on the cathode material. A reduction of ion charge states has been observed at the end of each are pulse. In this note it is shown that the charge-state-dependent ion drift velocity in the plasma expansion zone contributes to the effect measured by the time-of-flight method, although it is not solely responsible for it. A quantitative evaluation as well as literature data of other measuring methods prove the physical significance of the effects observed. Index Terms - Ion charge state distribution, time-of-flight measurements, vacuum arcs.

The events of the VVIIth International Symposium on Discharges and Electrical Insulation in Vacuum held on Jul 20-25, 1996 in Berkeley, CA, are presented. Vacuum arc deposition of thin films and pseudospark discharges were discussed as the newest developments in vacuum arc applications and electrode phenomena. The symposium featured invited speakers, oral contributions, posters, and two minicourses. The awarding of two outstanding scientists in the field of electrical discharges and insulation in vacuum was also included in the symposium's agenda.

Ultrathin ( Index Terms--Carbon overcoats, cathodic vacuum arc, diamond-like carbon, macroparticle filtering, magnetic storage.