Michael Keidar

Recent progress in filtered vacuum arc deposition

Abstract During this decade significant advances have been made both in the understanding and implementation of filtered vacuum are deposition. Rigid rotor models have been analyzed statistically, and new models which treat the mutual influence of the electrons and ions on each other self-consistently, take into account the centrifugal force on the ions, and take into consideration collisions, have been formulated. It was shown that the plasma transport efficiency is limited by drifts caused by the centrifugal force and by the electric field generated by charge separation in the plasma. For a range of magnetic fields strengths for which the ions are not magnetized, i.e., confined to a Larmor radius less than the duct radius, the transport efficiency for Cu plasma is about 10%, and depends only weakly on the magnetic field strength. Increased transmission is found when the ions are magnetized, reaching about 50% for a 36–60 mT field in typical configurations. The plasma transport efficiency and spatial distribution has been measured over a large parameter range, and correlated with the various theories. The plasma beam may be approximated as a Gaussian distribution which is displaced in the B × G direction, where G is in the direction of the centrifugal force, while a displacement in the plane of symmetry is surprisingly found in the − G direction. The total convected ion current decreases exponentially with distance from the toroidal filter entrance. Macroparticle transport within the magnetic filter has been analyzed, and it has been shown that electrostatic reflection from the walls can occur if the magnetic field is weak. Filtered arc sources with improved throughput performance and novel geometries have been built, and are now available commercially. The range of coatings deposited with FVAD has been expanded to include metals, oxides, and nitrides, as well as diamond-like carbon. In several cases, coatings having the highest quality reported in the literature have been fabricated with the FVAD technique, and one commercial application has been reported.

2D expansion of the low-density interelectrode vacuum arc plasma jet in an axial magnetic field

The two-dimensional expansion of a current carrying plasma jet in the interelectrode gap of a vacuum arc with an axial magnetic field is analysed by finding the steady state solution of the fully ionized plasma in the hydrodynamic approximation. Two models are presented: (1) expansion into a duct with known geometry and (2) free jet expansion. The first approach models the plasma jet expansion with a conical shape. In the second model the geometric position of the free boundary was determined by the free hydrodynamic jet expansion into vacuum without and with the influence of a magnetic field. In the case of plasma expanding into a conical guide, it was found that the flow field in the near-axis region does not depend on the cone angle for cone angles . The radial velocity becomes comparable to the axial velocity due to the expansion, depending on the cone angle and the initial axial velocity. A model of the free boundary plasma expansion was developed, based on the jet-like (i.e. axial velocity larger than the radial velocity) plasma flow in the vacuum arc near the cathode spot. The free jet boundary was calculated by solving the equations for the normal and tangential velocity components at the free boundary. It was found that the plasma jet had a conical shape, and for axial distances 3 - 4 times greater than the initial jet radius, the radial velocity becomes comparable with the axial velocity if no magnetic field is imposed. Imposition of a magnetic field reduces the radial component of the plasma velocity. The streamline angle is about for a 0.001 T magnetic field and about for a 0.01 T magnetic field. The plasma remains quasi-neutral in all regions except in the space charge boundary layer, where an outward directed electric field appears for low magnetic fields, and an inward directed field is present for strong magnetic fields.

Theoretical study of plasma expansion in a magnetic field in a disk anode vacuum arc

The low-density plasma flow in an axial magnetic field to a disk-shaped anode in a vacuum arc was studied theoretically using a two-dimensional model. The plasma expansion was modeled using the sourceless steady-state hydrodynamic equations, where the free boundary of the plasma was determined by a self-consistent solution of the gas-dynamic and electrical current equations. The anode was modeled as a current and plasma collector, which does not influence the plasma flow field. Magnetic forces from both the azimuthal self-magnetic field, and the imposed axial magnetic field were taken into account. It was found that the self-magnetic field does not substantially influence either the plasma jet shape, density, velocity, or the current density distribution for arc currents I⩽200 A. On the other hand, the plasma jet angle (α0) at the starting plane and the radial plasma density gradient force in the expansion region do have a strong influence on the plasma and current flow. The mass and current flow in a 500...

Sheath and presheath structure in the plasma–wall transition layer in an oblique magnetic field

In this paper we present a hydrodynamic study of a current carrying transition layer that separates the plasma from electrically conducting or insulating walls. The plasma is placed in a magnetic field that could be parallel to, or intersects obliquely with the walls. The self-consistent model of the smooth presheath–sheath transition of the near-wall plasma layer developed here for a finite Debye radius to ion Larmor radius ratio (Ψ) is based on a previously developed model of the quasineutral plasma presheath [Phys. Plasmas 4, 3461 (1997)]. The potential distribution in the presheath is found to have a positive maximum with respect to the plasma–presheath interface. In the case of a wall with a floating potential, the value of the maximum decreases with the incidence angle θ of the magnetic line force, and approaches zero when θ=2°. The presheath thickness generally increases with the incidence angle, from about the electron Larmor radius up to the ion Larmor radius, and depends on the electron to ion c...

Nonstationary macroparticle charging in an arc plasma jet

The charging of liquid metal macroparticles in the rarified part of a vacuum arc plasma jet is studied. The sheath in the vicinity of the macroparticle is collisionless and the problem with different Debye length to macroparticle radius ratios is analyzed. Maxwellian velocity distribution functions with different temperatures for the electrons and ions in an arbitrary ratio are allowed in the model. By solving the equation for the electric field together with the equation for ion and electron flux, the charging time and the near electric field of the macroparticles were calculated. The kinetics of the macroparticle charging are controlled by the ion and electron flux to the macroparticle, which depend on the potential distribution in the sheath. The potential falls off slower than 1/r/sup 2/ in the case of the large Debye length to macroparticle radius ratio, and falls off more rapidly than 1/r/sup 2/ in the other case. The charge which accumulates on a macroparticle at distances of about 10 cm from a 100-/spl Aring/ cathode is about 10/sup -16/ C and the charging time is about 10/sup -5/ s. The influence of the plasma drift velocity on the macroparticle charging is small. The model presented here agrees well with an experimental study of macroparticle repulsion from biased substrates.

Transport of macroparticles in magnetized plasma ducts

The cathode spot of a vacuum are produces a highly ionized energetic plasma jet of vaporized cathode material which may be directed to a substrate to form a high-quality coating or thin film, and a spray of molten droplets, referred to as macroparticles (MPs). The plasma flux can be concentrated by magnetic collimation, while the MP spray can be filtered from the plasma jet by guiding the plasma around an obstacle using a magnetic field. In the present work, the motion of the individual electrically charged MP in the straight and quarter torus plasma guides is studied, taking into account MP charging by interaction with the plasma. The influence of the electric field in the plasma, which depends on the magnetic field, on the charged MP motion is calculated. The fraction of the MPs transmitted through the toroidal plasma guide was calculated as a function of the wall potential and MP velocity for different minor to major torus radius ratios r/R. The guide wall potential has a strong effect on the transmission of MPs having velocities in the 25-100 m/s range. In the case in which r/R=0.1, with the duct at floating potential, the fraction of the MPs transmitted through the torus approaches 100% for 0.1 /spl mu/m Ti MPs having an initial velocity parallel to the duct wall. The main mechanism of MP transmission through curved ducts is repeated electrostatic reflection of the charged MP from the wall. The calculation of the MP transmission through the magnetized straight duct was compared with experiment, where a significant reduction of the MPs was obtained with an increasing of the axial magnetic field. The calculated MP transmission was close to that previously measured for 0.1 /spl mu/m MPs with 70 m/s and for 0.5 /spl mu/m MPs with 20 m/s directed velocities.

Influence of an electrical field on the macroparticle size distribution in a vacuum arc

The results of experimental study of macroparticle distribution in a vacuum arc presented for Cu, Ti, Zr, and Cr cathodes. We have studied the macroparticle contamination of the films deposited on the substrate having floating potential and biased up to −1000 V with respect to the anode. It has been found that the macroparticle number significantly decreases (by a factor of 3–4) with substrate biasing for all examined cathode materials. A model of macroparticle motion in the quasineutral plasma and near-substrate sheath was proposed. The model bases on analyses of the macroparticle charging and motion in the quasineutral plasma and near substrate sheath. The model can qualitatively explain the macroparticle reduction in the coating due to negatively charged macroparticle reflection in an electric field.

Filtered vacuum arc deposition of semiconductor thin films

The cathode spot vacuum arc produces a jet of highly ionized plasma plus a spray of liquid droplets, both consisting of cathode material. The droplets are filtered from the plasma by passing the plasma through a curved, magnetized duct. A radial magnetic field may be applied to the face of the cathode to rotate and distribute the cathode spots in order to obtain even erosion and avoid local overheating. The choice of axial magnetic field strength in the vicinity of the cathode is a compromise between a relatively high field desired to collimate a large fraction of the plasma flux, and the need to collect a substantial fraction of the plasma at the anode in order to reduce arc voltage and insure arc stability. The transmission of the filter duct increases with magnetic field strength until a saturation value is reached. Entrainment of the droplets in the plasma jet can decrease the effectiveness of the filter at high plasma flux. Semiconducting thin films of amorphous silicon were prepared using cathodes of heavily B-doped Si. Arcs of 35-A current produced a deposition rate of 10 /spl Aring//s. The electrical conductivity of the films was similar to conventional a-Si:H films deposited by conventional Silane based PACVD at high temperatures, but had a higher room-temperature conductivity. Transparent conducting films of Sn-O were deposited at rates of up to 100 /spl Aring//s using 160-A arcs on a Sn cathode while injecting O/sub 2/ gas in the vicinity of the substrate. Adjustment of the O content is critical for optimizing conductivity, and complicated by pumping effects of the arc. Optimal conductivity was achieved at an oxygen pressure of 6 mtorr. Conductivities equal to the best reported to date were achieved by subjecting the room-temperature deposited films to a 30-s rapid thermal annealing at 350/spl deg/C. Both the deposited and annealed films are amorphous. The deposition rates achieved by the filtered vacuum arc technique for these semiconductor films are an order of magnitude greater than achieved with conventional methods, while the conductivities are equivalent or better.

Interelectrode plasma parameters and plasma deposition in a hot refractory anode vacuum arc

The new mode of Vacuum arc-Hot Refractory Anode Vacuum Arc-was studied experimentally using a Langmuir probe, two types of thermal probes, and film collection substrates. The plasma density, electron temperature, plasma energy flux, cathode erosion, mass deposition rate on a substrate, and macroparticle contamination in the deposited films were measured. The arc initially operated as a usual vacuum arc sustained by cathode spots, i.e., and the vapor and plasma source located at the cathode spot. At a later stage the anode heated up and metal vapor originating at the cathode was re-evaporated from the nonconsumable hot graphite anode. Initially, plasma density was about (3–4)⋅1020 m−3 but it increased with time, reaching about 2⋅1021 m−3 after 60 s in a 340 A arc. The electron temperature initially was about 1.6 eV and decreased with time to a steady-state value of about 1.1 eV after 20 s. The radial plasma energy flux generated by 175 and 340 A arcs was about 1 and 2 MW/m2, respectively, at 1.6 cm from th...

Macroparticle interaction with a substrate in cathodic vacuum arc deposition

Abstract The presence of macroparticles (MP) in the vacuum arc cathodic jet is the critical problem of cathodic vacuum arc deposition technology. The MP distribution on the substrate depends on MP generation, MP transport to the substrate, and the MP—substrate interaction. In the present work different aspects of the interaction of the MP with a substrate are studied theoretically. The MP temperature evolution during transport to the substrate was studied and it was found that the MP is liquid in the case of low melting point cathode materials. It was found that the MP temperature does not depend on the initial MP temperature in the near cathode region. Possible MP acceleration in the substrate direction resulting from ion friction was calculated by solving the equation of motion for the MP taking into account the plasma density and the velocity distribution obtained in previous analyses of the 2-D plasma expansion. The voltage drop in the substrate sheath as a function of the background pressure was calculated assuming charge exchange between ions and gas atoms. It was found that the voltage drop in this sheath increases with increasing background pressure pg by factor of 2 for pg ∼ 1 Pa. The MP motion in this sheath was calculated taking into account MP charging and reflection from the substrate. The probability of reflection decreases with MP size, velocity, angle of the MP velocity vector with respect to the substrate, and increases with the background pressure.