Search

Erbium oxide (Er 2 O 3 ) coatings were deposited using filtered vacuum arc deposition (FVAD) and their structure and thermal stability were studied as a function of fabrication parameters. The coatings were deposited on silicon wafer and tantalum substrates with an arc current of 50 A and a deposition rate of 1.6 ± 0.4 nm/s. The arc was sustained on truncated cone Er cathodes. The influence of oxygen pressure ( P = 0.40-0.93 Pa), bias voltage ( V b = -20, -40 or grounded) and substrate temperature (room temperature (RT) or 673K) on film properties was studied before and after post deposition annealing (1273K for 1 hour, at P ~ 1.33 Pa). The coatings were characterized using X-ray diffraction (XRD), optical microscopy, scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS) and Knoop Hardness. Optical microscope images indicated that the coatings had very low macroparticle concentration on their surface. The macroparticle diameters were less than 2.5 I¼m. The coatings were composed of only Er 2 O 3 without any metallic phase under all deposition parameters tested. The coatings deposited on RT substrates were XRD amorphous and had a featureless cross-section microstructure. However, the coatings deposited on 673K heated substrates had a C -Er 2 O 3 structure with (222) preferred orientation and weak columnar microstructure. The coating hardness varied with deposition pressure and substrate bias, and reached a maximum value of 10 GPa at P = 0.4 Pa and V b = -40 V. The post-deposition annealing caused crystallization, and the coatings hardness dropped to 4 GPa with thermal treatment. However, after post-deposition annealing, no peeling or cracking appeared at the coating surface or the interface with the substrate.

The effect of oxygen and argon partial pressures (PO2, PAr) in a Zr vacuum arc on plasma ion current density Jp, arc voltage Varc, deposition rate vd, and selected coating properties was determined. A d.c. arc current of Iarc = 100 A was initiated between a Zr cathode and a grounded anode. Cathode spots produced a plasma jet, which entered a 1/8 torus macroparticle (MP) filter. The plasma was guided by a d.c. magnetic field through an aperture to a glass substrate or a flat disk probe, mounted on a rotatable holder. Jp was measured with the probe, negatively biased to Vb = − 60 V. Coating thickness was measured using a profilometer, and coating properties were investigated using optical microscopy, energy dispersive X-ray spectroscopy (EDS), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), nano-indentation and optical analysis. The discharge electrical characteristics and the coating deposition rate were found to be significantly influenced by PO2 and PAr. Jp and vd increased with PAr until a maximum at PAr = 0.27 Pa and decreased with PO2. Varc decreased with both PAr and PO2. The changes in Jp, Varc, and vd, with PAr were larger at larger PO2. The Jp, Varc, and vd dependencies suggest that addition of Argon increased the Zr ion emission from the cathode, possibly because Ar ion bombardment reduced Zr surface oxidation and improved plasma conductivity. Zirconium Oxide (ZrO2) coatings were transparent and had colored interference rings. Well adhered, MP-free ZrO2 coatings were deposited with PO2 ≥ 1.07 Pa. Coatings deposited with PO2 = 1.07 Pa + PAr = 0 were amorphous, whereas those deposited with PO2 = 1.07 Pa + PAr = 0.27 Pa had some degree of a monoclinic phase. Furthermore, the refractive index (n) and extinction coefficient (k) slightly decreased, from 2.22 to 2.17, and from 0.03 to 0.01, respectively and coating hardness (H) and Young's Modulus (E) decreased from ~ 12.9 to ~ 11.6 GPa and from ~ 153 to ~ 136 GPa respectively when PAr = 0.27 Pa was added to a PO2 = 1.07 Pa environment.

Nano-multilayered Zr–O/Al–O coatings with alternating Zr–O and Al–O layers having a bi-layer period of 6–7 nm and total coating thickness of 1.0–1.2 μm were deposited using a cathodic vacuum arc plasma process on rotating Si substrates. Plasmas generated from two cathodes, Zr and Al, were deposited simultaneously in a mixture of Ar and O 2 background gases. The Zr–O/Al–O coatings, as well as bulk ZrO 2 and Al 2 O 3 reference samples, were studied using X-ray photoelectron spectroscopy (XPS). The XPS spectra were analyzed on the surface and after sputtering with a 4 kV Ar + ion gun. High resolution angle resolved spectra were obtained at three take-off angles: 15°, 45° and 75° relative to the sample surface. It was shown that preferential sputtering of oxygen took place during XPS of bulk reference ZrO 2 samples, producing ZrO and free Zr along with ZrO 2 in the XPS spectra. In contrast, no preferential sputtering was observed with Al 2 O 3 reference samples. The Zr–O/Al–O coatings contained a large amount of free metals along with their oxides. Free Zr and Al were observed in the coating spectra both before and after sputtering, and thus cannot be due solely to preferential sputtering. Transmission electron microscopy revealed that the Zr–O/Al–O coatings had a nano-multilayered structure with well distinguished alternating layers. However, both of the alternating layers of the coating contained of a mixture of aluminum and zirconium oxides and free Al and Zr metals. The concentration of Zr and Al changed periodically with distance normal to the coating surface: the Zr maximum coincided with the Al minimum and vice versa . However the concentration of Zr in both alternating layers was significantly larger than that of Al. Despite the large free metal concentration, the Knoop hardness, 21.5 GPa, was relatively high, which might be attributed to super-lattice formation or formation of a metal–oxide nanocomposite within the layers.

Al2O3–ZrO2 coatings were deposited using a vacuum arc deposition system equipped with two co-planar cathodes. The plasma was injected into a cylindrical magnetic duct through annular anode apertures toward a substrate or an electrostatic ion current probe positioned on the duct axis, in vacuum and in a low-pressure oxygen or argon + oxygen background. Ion current and arc voltage measurements and visual observation of the cathode spots were used to find stable arcing conditions, using a straight plasma duct configuration. The cathode spot operation and transport of the plasma beam in the duct were studied as a function of arc current (I arc = 25–200 A) and oxygen or oxygen + argon pressures (P = 0.1–1.5 Pa). Coatings were fabricated by exposing Si or WC–Co substrates simultaneously to Al and Zr plasmas using a 1/8 torus filter configuration in O2 + Ar pressures. The coating composition, structure, microhardness, adhesion, and wear behavior were studied as functions of the deposition parameters. Favorable conditions for stable arcing were obtained with I arc = 75 and 100 A for Al and Zr plasmas, respectively. The ion current decreased, and the arc voltage increased with the oxygen pressure. Behavior of the ion current and arc voltage suggested that cathode poisoning started at P = 0.5 Pa. Deposition rates were 0.3-0.6 μm/min, depending on the substrate position. All coatings were “Zr rich”, i.e., the Zr:Al ratio was in the range of 1.2–5.6 depending on the substrate position and deposition conditions. The coatings with higher ZrO2 concentration were harder and had better resistance to wear. The coating’s hardness reached a maximum of ~22–24 GPa at a deposition temperature of 500 °C or a negative bias voltage of 75–100 V.

Thin films of Zr–O/Al–O were deposited on SKD 11 tool steel substrate using Zr and Al cathodes in a cathodic arc plasma deposition system. The substrates were mounted on a rotating holder which alternatively exposed them to plasma from the two cathodes. The influence of the Zr and Al cathode arc currents and the substrate bias on the mechanical and the structural properties of the films were investigated. Films with a nano-layered structure of alternating Al-rich and Zr-rich layers were obtained. The Zr layers contained nano-crystallites of (101) oriented t-ZrO structure. Crystallites with α-Al 2 O 3 structure were observed only when the substrate was negatively biased in the 100–150 V range. The hardness of the film decreased with the increase of Zr cathode current from 60 to 80 A, increased when the Al cathode current increased from 25 to 30 A, and decreased when the Al cathode current increased from 30 to 35 A. The hardness of the film increased with the increase of bias voltage up to − 150 V and then decreased with further increase of the negative bias. The film structure was elucidated by HRTEM microscopy. Good correlation between the residual stress and the hardness enhancement of the films was observed.

Er plasma was produced by a vacuum arc source with a truncated cone-shaped Er cathode and an annular copper anode. The plasma beam propagated into a cylindrical duct through the annular anode aperture and flowed in vacuum or in a low-pressure oxygen background along an axial magnetic field toward a substrate or an electrostatic ion current probe positioned on the duct axis. The cathode spot location and motion, the cathode erosion, the transport of the plasma beam in the duct, and the coating deposition rate were studied as a function of arc current, axial magnetic field, and oxygen pressure. Favorable conditions for stable arcs were obtained with relatively low arc current (I arc = 30 - 50 A) and an axial magnetic field of 12 mT, which confined the spot motion on the cathode apex. Under such conditions, the rates of the cathode erosion and coating deposition were ~235 mug/C and 2-5 mum/min, respectively. With higher I arc (150-175 A), the cathode spots tended to move on the circumference of the large cone base, eroding a deep groove, which significantly decreased the cathode life.

A novel filtered cathodic vacuum arc source of carbon plasma, in which plasma ions are reflected from a Hall stratum formed in the inter-electrode gap, was studied. The plasma source consisted of a rectangular carbon cathode, positioned between the poles of a permanent magnet, a copper anode, and a triggering mechanism. An arc was operated with current pulses having a duration of 7 ms and a peak current of 200 A. It was shown that the cathode spot plasma jets were reflected from the Hall stratum formed in the arched magnetic field, and bent through 180° towards substrates positioned alongside of the cathode, while macroparticles flew from the cathode spots in straight trajectories. The calculated thickness of the Hall stratum ranged from less than 1 mm to ≥ 10 mm. The ratio of the substrate ion current to the arc current was ∼ 7.5%, which is much larger than that of most known filtered systems, and approaches the plasma transfer efficiency of unfiltered vacuum arc plasma sources.

Binary Nb–N coatings, ternary Ti–Nb–N and Zr–Nb–N, and multi-layer TiN/NbN coatings consisting of up to 100 alternating TiN and NbN layers, were deposited onto WC–Co substrates, using two different vacuum arc deposition (VAD) systems: with and without magnetic guiding of the metal plasma flow. Binary Nb–N coatings were fabricated by deposition of metal plasma produced by a Nb cathode in a background of reactive nitrogen gas at different pressures, P . Ternary coatings were fabricated at co-deposition of plasmas originating from two different cathode materials. Multilayer coatings were fabricated by alternatively depositing plasmas of Ti and Nb in reactive nitrogen gas. The crystalline coating structure, phase composition, hardness and critical load for coating failure were studied. For binary Nb–N coatings fabricated using both deposition systems, the phase composition, the Vickers hardness, HV, and the critical load strongly depended on the deposition pressure. Using VAD with magnetic plasma guiding, the highest HV of ∼ 42 GPa was measured for coatings deposited at low nitrogen pressure. These coatings contained a hexagonal β-Nb 2 N phase and had a relatively low critical load. The highest critical load and HV ∼ 38 GPa were obtained for coatings consisted of a single phase NaCl-type cubic δ-NbN structure, deposited at a higher nitrogen pressure. The structure and properties of Nb–N coatings deposited using VAD without magnetic plasma guiding had a similar correlation with the deposition pressure, however, their hardness values were lower. Ternary Ti–Nb–N and Zr–Nb–N coatings fabricated by both deposition processes had a single phase cubic NaCl-type structure and the hardness higher than that of the binary nitrides TiN, ZrN and NbN. The hardest coatings, HV ∼ 51.5 Pa, deposited with magnetic plasma guiding had a single-phase cubic δ-(Ti,Nb)N structure and a Ti:Nb ratio of ∼ 50:50 (at.%). Multilayer coatings TiN/NbN consisting of 20–40 alternating TiN and NbN layers with total thickness of 4–5 μm increased the life time of cemented carbide cutting inserts at turning tough Ni-base alloys by 2–7 times relative to uncoated cutting tools, while conventional vacuum arc deposited TiN coatings were not effective in machining of these alloys.

Transparent conducting ZnO:Al and ZnO films of 380-800 nm thickness were deposited on glass substrates by filtered vacuum arc deposition (FVAD), using a cylindrical Zn cathode doped with 5-6 at.% Al or a pure Zn cathode in oxygen background gas with pressure P = 0.4-0.93 Pa. The crystalline structure, composition and electrical and optical properties of the films were studied as functions of P. The films were stored under ambient air conditions and the variation of their resistance as function of storage time was monitored over a period of several months. The Al concentration in the film was found to be 0.006-0.008 at.%, i.e., a few orders of magnitude lower than that in the cathode material. However, this low Al content influenced the film resistivity, ?, and its stability. The resistivity of as-deposited ZnO:Al films, ? = (6-8) × 10- 3 ? cm, was independent of P and lower by a factor of 2 in comparison to that of the ZnO films deposited by the same FVAD system. The ? of ZnO films 60 days after deposition increased by a factor of ~ 7 with respect to as-deposited films. The ZnO:Al films deposited with P = 0.47-0.6 Pa were more stable, their ? first slowly increased during the storage time (1.1-1.4 times with respect to as-deposited films), and then stabilized after 30-45 days. © 2006 Elsevier B.V. All rights reserved.

ZnO thin films, 100-250 nm thick, were deposited on microscope glass slides and UV fused silica (UVFS) substrates using a filtered vacuum arc deposition (FVAD) system, operating at room temperature (RT) and 200 A for 60 s. The cathode was prepared from 99.9% pure Zn metal and the oxygen background pressure during deposition was in the range 0.67-0.93 Pa. The films were characterized by x-ray diffraction (XRD), x-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), optical transmission and spectroscopic ellipsometry. As-deposited ZnO films were found to be polycrystalline with c-axis orientation. The atomic concentration ratio of Zn to O in the film as determined by XPS analysis was stoichiometric. Film transmission in the visible was 70-90%. The maximum and minimum values of the refractive index n and the extinction coefficient k in the visible, for all samples, were 2.23 to 1.90 and 0.6 to approximately 0, respectively. The type of inter-band electron transition was found to be direct transition with optical band gap in the range of 3.25-3.42 eV. © 2006 IOP Publishing Ltd.

Transparent conducting SnO 2 coatings of 0.16–1.2 μm thickness were deposited on polymer substrates using filtered vacuum arc deposition. Mechanical and electro-optical properties of the coatings were investigated. Surface topography, nano-indentation hardness and nano-scratch resistance were studied using an atomic force microscope equipped with a diamond pyramid tip. Electrical sheet conductivity and optical transmission were studied as a function of the coating thickness. The surface nanohardness of SnO 2 coated polymer samples was larger by at least an order of magnitude than that of the bare polymer surfaces. Transparent and conducting SnO 2 coatings reduced the light-load scratch wear rate of the polymer substrates by two to three orders of magnitude, while providing optical transmission and electrical conductivity. This combination of properties would be useful for fabricating transparent scratch resistant coatings for vehicle and aircraft windows.

The use of magnetic fields to control cathode spot location and motion and to collimate and direct the plasma flow in vacuum arc deposition apparatus is reviewed. Retrograde and acute angle motion are used to control the location and motion of cathode spots, in order to confine them to the cathode surface facing the substrates, to prevent local overheating, and to reduce macroparticle production. Axial fields are use to collimate cathode spot produced plasma jets and to bend them around macroparticle-occluding obstacles in filtered vacuum arc deposition. Advances have been made in understanding these phenomena, but theoretical models have not yet been formulated which can predict plasma behavior sufficiently well for apparatus design.

A rectangular filtered vacuum arc plasma deposition system was used to deposit SnO2 coatings on stationary or moving 400×420 mm substrates. The system consisted of a rectangular plasma gun, a rectangular macroparticle filter, a vacuum deposition chamber, a substrate carriage and a transport mechanism. An arc discharge was ignited in the plasma gun between a rectangular cathode and an anode with a rectangular aperture, both 66×440 mm. A Sn plasma jet produced by cathode spots passed through the anode aperture into the rectangular macroparticle filter, which had two 45° bends. A magnetic coil located in the vicinity of the cathode produced a magnetic field that confined the cathode spot motion to a defined trajectory on the cathode surface. Three magnetic coils, located around straight sections of the filter, produced a magnetic field, which guided the plasma through the filter to the substrate positioned on the substrate carriage in the deposition chamber. A stationary multi-element Langmuir probe situated opposite the filter outlet was used to measure the saturated ion current density distribution, as a function of the arc current, the magnetic field strength, and the oxygen pressure. The SnO2 coatings were fabricated by deposition of the Sn plasma in low-pressure (≤0.73 Pa) oxygen background gas, either on stationary substrates, to determine the coating thickness distribution, or on moving substrates to determine dynamic deposition rate (DDR). The ion current at first increased with the arc current, Iarc, in the range of 300–450 A and then saturated with further increasing of Iarc up to 700 A. For coatings deposited on stationary substrates at Iarc=400 A and oxygen pressure of 0.7 Pa, the coating thickness distribution could be fit by a parabolic function in the direction of the long aperture axis, and was almost triangular in the perpendicular direction, with a maximum deposition rate of 37 nm/s in the beam center. The coatings deposited on moving substrates with Iarc=350 A had a DDR of ∼100 nm-m/min, 250–450 nm thickness, a relative visual light transmission of 83% or larger, and a sheet resistivity of 100–170 Ω/□. Higher current experiments suggest that a DDR of 190 nm-m/min or larger could be achieved.

The transport of a vacuum-arc produced plasma beam along a magnetized cylindrical duct was studied experimentally. The plasma source consisted of a Sn or an Al cathode and a 17-mm internal diameter annular copper anode through which the plasma beam entered into the 160-mm diameter and 500-mm length cylindrical duct. The arc current I/sub arc/ was in the range of 30-100 A. Three magnetic coils positioned coaxially with the duct axis produced an approximately axial guiding magnetic field, B/sub g//spl les/20 mT in the duct. A 130-mm-diameter movable planar disk probe, positioned normal to the duct axis, was used to measure the ion saturation current I/sub probe/ along the duct. The ion current to the duct wall, I/sub duct/, and the probe and duct floating potentials, /spl phi//sub probe/ and /spl phi//sub duct/, respectively, were measured as functions of B/sub g/ and the axial distance of the probe from the anode, L. Generally, I/sub probe/ decreased while I/sub duct/ increased with L, and the sum I/sub probe/+I/sub duct/ was approximately independent of L. For an Al arc, I/sub probe/ and I/sub duct/ initially increased as a function of I/sub arc/, reached a maximum at I/sub arc/=40--45 A, and then decreased by a factor of 2-2.5 relative to their maximal values. Both /spl phi//sub probe/ and /spl phi//sub duct/ were negative relative to the grounded anode. /spl phi//sub probe/ became significantly more negative as L or B/sub g/ increased, while /spl phi//sub duct/ depended only weakly on L and B/sub g/.

The effect of an obstacle positioned normal to a plasma jet produced by a vacuum arc plasma source on the radial distribution of ion flux in the vicinity of the obstacle was studied. This study was motivated by interest in the mutual influence of tightly packed substrates on coatings in industrial vacuum arc deposition systems. The experimental system consisted of a vacuum arc plasma source, a straight plasma duct, and a multi-probe consisting of a removable disc obstacle and a set of ring probes for measuring the radial ion flux. A dc arc discharge was ignited in vacuum between a truncated cone-shaped Cu cathode and an annular anode. The plasma jet produced by cathode spots passed through the anode aperture into the straight plasma duct. An axial magnetic field guided the plasma jet in the duct. The multi-probe consisted of a removable disc obstacle and a set of five ring probes for measuring the radial plasma flux as a function of distance from the disc obstacle. The rings and the disc probes were coaxially arranged on the multi-probe assembly and positioned so that plasma from the source passed through the ring probes and then encountered the disc. The influence of the obstacle was determined by measuring the ring ion currents, both in the presence of the obstacle, and when the disc obstacle was removed. The difference between the measured ion currents with and without the obstacle was interpreted to be the contribution of reflected or sputtered particles from the obstacle to the radial ion flux. The ring probes were biased by ?60?V with respect to the grounded anode, to collect the saturated ion current. The multi-probe was connected to a movable stem, and positioned at different distances from the plasma source.A plasma density of ~6 ? 1017 m?3 was estimated in this study based on the ion current to the obstacle. The radial ion flux collected by the ring probes increased by 20?25% due to the presence of the obstacle. As the calculated mean free path for collisions between the plasma ions and sputtered or reflected particles is much larger than the plasma path length, the plasma jet should not be affected by returning particles. Hence, the increase of the radial flux is most probably due to ion reflection from the obstacle and ionized neutral atoms sputtered from the obstacle.

A vacuum arc carbon plasma source is described, in which an arc was ignited between a cathode and an anode having In aperture, by bringing the two electrodes into contact, and parting them while current was flowing. The inter-electrode gap length was varied. A focusing magnetic field was applied in the inter-electrode gap, and a toroidal magnetic field was applied to guide plasma through a toroidal duct. The influence of the arc current (I/sub arc/), gap length (L), and focusing magnetic field (B/sub f/) on the cathode spot motion, plasma behavior, and output ion current was studied. It was shown that with I/sub arc/=150-250 A and relatively weak B/sub f/, the spots moved on the cathode periphery, causing noisy arcing and fluctuation of the output ion current, while at a lower I/sub arc/ (50-100 A) and strong B/sub f/ (/spl ges/28 mT), there were fewer spots on the cathode surface, they tended to move toward the cathode center, the arc noise decreased and the average output ion current increased. With increasing L from 2 to 18 mm, the total ion current at first rapidly increased, and then saturated at L/spl ges/10 mm. Best performance of the carbon plasma source was obtained with an arc current chosen so that with a given field configuration, the cathode spots locate in the center of the cathode surface, and with strong magnetic fields in both the electrode and toroidal filter regions.

Transverse current injection (TCI) is a technique in which an electrical current is imposed parallel to the surface of a workpiece exposed to processing or to a service environment. In this work, TCI was applied during vacuum arc deposition of TiN film on WC cemented carbide substrates (90% WC, 8% Co, 2% TaNbC). Bar samples with dimensions 5×6.5×20 mm were mounted on a holder that provided electrical contacts at their ends, and a thermocouple to measure sample temperature. Prior to deposition, the substrates were heated in vacuum up to 150°C by a current of 100 A. The TiN coatings were obtained by vacuum arc deposition of Ti plasma in a 0.67-Pa nitrogen background. Films with different thickness were obtained by controlling of the plasma flow density. The arc current was 250 A, and substrate temperature during deposition was in the range 110–130°C. Two deposition rates,: 2.3–3 and 20–23 nm/s were used. The injected d.c. transverse current was in the range 0–40 A. Surface microhardness was measured by Vickers microindentation using a 25-g load. Film structure was examined by X-ray diffractometry (XRD), scanning electron microscopy (SEM) using scattered electrons (SE), back scattered electrons (BSE), and energy dispersive spectroscopy (EDS). EDS spectra indicated that approximately the same ratio of Ti and N atoms existed both in thick (4–4.5 μm) and in thinner films (∼0.4 μm). XRD patterns indicated a cubic NaCl-type TiN phase in the thick films, while this phase was absent in the thinner films. In comparison with films deposited without TCI, the maximum microhardness of thick TiN films (∼4 μm) was larger by a factor of 1.3 with 10 A of transverse current, whereas the maximum microhardness of thinner films (∼0.4 μm) was up to a factor of 1.7 larger with 20 A of transverse current. TiN coatings deposited at 110–130°C with TCI were uniform and did not contain delaminating fragments, whereas coating deposited using the same deposition parameters but without TCI contained delaminating fragments.

TiN coatings were deposited on WC–Co bar substrates using a vacuum arc plasma gun connected to a cylindrical plasma duct in which an axial magnetic field was imposed. During deposition, the cathode arc current was 200 A, nitrogen pressure was 0.67 Pa, and the substrate temperature was 420°C. Substrate bias voltage (Vbias) was varied in the range of −40 to −600 V. The coating structure and properties were studied both on the sample face surface, i.e. normal to the plasma flux, and on its side surfaces. The structure and phase composition were studied using scanning electron microscopy (SEM) and X-ray diffraction (XRD). Microhardness and scratch critical load were studied using Vickers micro-indentation and scratch tests, respectively. The TiN coatings had a single-phase cubic δ-TiN structure and consisted of oriented columnar grains. No difference was observed in the preferred grain orientation and grain size at the substrate–coating interface for the coatings deposited on the face and the substrate side surface for every studied Vbias. The deposition rate decreased both on the face and the side surfaces, while the ratio between deposition rates on the face and the side surfaces increased from three to seven times when |Vbias| increased from 40 to 600 V. With increasing |Vbias|, the preferred orientation of the columnar grains changed from a mixture of (200) and (111) at −40 V to a strong (111) at −200 and −400 V. At −600 V, (111) remained dominant, while the (220) orientation also appeared. Increasing |Vbias| increased the grain size on the coating surface. A zone of equiaxed grains was observed near the substrate–coating interface, whose thickness increased with increasing |Vbias|. Possibly, the grain size growth was a thermal effect due to an increase in ion beam heating with increased |Vbias|. The grain size on the side surfaces was smaller than that on the face. The coating surface roughness and friction coefficient were smaller on the side surfaces than those on the face, while no differences in microhardness were observed.

A novel thermal spray technology — an electro-thermal chemical spray (ETCS) for producing hard coatings is presented. The experimental coating apparatus consists of a machine gun barrel, a cartridge containing the coating material in powder form, a solid propellant, and a plasma ignition system. The plasma ignition system produces plasma in pulsed mode to ignite the solid propellant. On ignition, the drag force exerted by the combustion gases accelerates the powder particles towards the substrate. Using the ETCS technique, the process of single-shot WC–Co coating deposition on stainless steel substrate was studied. The influence of process parameters (plasma energy, mass of the solid propellant and the coated powder, distance between the gun muzzle and the substrate) on the coating structure and some of its properties were investigated. It was shown that ECTS technique effectively deposited the WC–Co coating with deposition thicknesses of 100–200 μm per shot, while deposition yield of ∼70% was attained. The WC–Co coatings consisted of carbide particles distributed in amorphous matrix. The powder particle velocity was found to depend on the solid propellant mass and was weakly dependent on the plasma energy, while the particle processing temperature was strongly dependent on the plasma energy and almost independent of the solid propellant mass. Whilst increasing the solid propellant mass from 5 to 7 g, the deposition rate and yield correspondingly increased. When increasing the plasma energy, the temperature of the powder particles increased, the average carbide particle size decreased and their shape became more rounded. The deposition yield and microhardness at first increased and then achieved saturation by increasing the plasma energy.

Ternary transition-metal nitride coatings of (Ti,Zr)N, (Ti,Nb)N and (Zr,Nb)N were deposited using a triple cathode vacuum arc deposition system by simultaneously operating two cathodes of the appropriate metals in a low-pressure nitrogen background. The magnetically collimated plasma flux distribution was measured using a 13-element ion saturation current probe. The microstructure, composition, and hardness of the coatings were determined using X-ray diffraction, Auger electron spectroscopy, and Vickers indentation, respectively. The plasma flux distribution had two peaks whose relative positions were related to the two cathodes. Coatings applied to extended substrates had a compositional gradient. In almost all cases, the ternary nitrides had a single-phase solid-solution microstructure, whose lattice constants had values intermediate between those of the parent binary nitrides. The ternary coatings were considerably harder than either of the parent binary nitrides. (Ti,Nb)N had the highest microhardness: ∼50 GPa.

Single-layer coatings of TiN and NbN, and multi-layer coatings of TiN/NbN, were deposited onto WC–Co substrates using a triple-cathode vacuum arc plasma gun connected to a cylindrical plasma duct onto which an axial magnetic field was imposed. Additional magnetic fields were applied by two beam steering coils orientated normal to the duct axis. The magnetic field produced by the steering coils directed the plasma beam onto a substrate placed on the system axis, increasing the plasma flux to the sample. The single-layer coatings were produced by generating Ti or Nb plasmas in a nitrogen background at a pressure P in the range of 0.67 to 2.67 Pa. Multi-layer coatings with 20, 50 and 100 layers were deposited by alternately switching the arcs on Ti and Nb cathodes. Coating structure and composition were studied by means of scanning electron microscopy (SEM) and X-ray diffraction (XRD). Microhardness and adhesion to the substrate were studied by Vickers' micro-indentation and scratch tests, respectively. It was shown that the phase composition of the NbN coatings depended on the deposition rate and P . The coatings deposited at low deposition rate (i.e. without the beam steering field) exhibited a single-phase cubic δ-NbN structure at P ≥0.67 Pa, whereas the coatings deposited with application of the beam steering field, at P =0.67 and 1.33 Pa, were composed of a mixture of cubic δ-NbN and hexagonal NbN 0.95 , while at P =2 and 2.67 Pa, the hexagonal phase was not found. The phase composition of the TiN coatings was independent of the deposition rate and P in the range 0.67–2.67 Pa. The highest microhardness (up to 38 GPa) and scratch critical load (80–95 N) were obtained for single-phase δ-NbN coatings deposited at P =0.67–1.33 Pa and at low deposition rate. The microhardness of multi-layer TiN/NbN coatings of 3.2–3.6 μm total thickness increased with increasing number of alternating layers, but did not exceed that of the pure δ-NbN observed in this study.

Thin conductive wear resistant metal/ceramic coatings were deposited onto polysulfone S2010 substrates using a triple-cathode vacuum arc plasma source connected to a magnetized plasma duct. Various combinations of two-layer coatings having an inner metallic conductive layer of Ti, Zr, Cu, Ag or Al attached to the plastic, and an external wear resistant layer of TiN or ZrN, were studied and evaluated. It was shown that bi-layer Ti/TiN and Ti/ZrN coatings had good adhesion to the substrate and low wear rate, but their electrical surface resistivity was high (3–6.6 Ω/□). Cu, Ag and Al metal sublayers had a superior adhesion to the polymer substrate and low surface resistivity (less then 0.1 Ω/□). However, the deposition of a TiN layer on Ag caused immediate coating delamination from the substrate. In the case of Ag/ZrN coatings, which initially possessed good adhesion to the substrate, cracking and delamination of ZrN at the Ag/ZrN interface was observed after some time (1–7 days), while the Ag layer remained adherent to the substrate. In the cases of Cu/TiN and Cu/ZrN coatings, the nitride layers had poor adhesion to the Cu layer. The best results were obtained for Al/ZrN coatings, which had low surface resistivity (0.08–0.11 Ω/□) with good adhesion of the Al layer to the substrate and of the nitride layer to Al.

The distribution of plasma beams produced by Ti, Nb and Zr cathodes in a triple-cathode vacuum arc deposition system was studied. The system consisted of a triple-cathode plasma gun, straight plasma duct, sample chamber, vacuum system and computerized control system. Three cathodes were located on a circle centred on the system axis. An arc was ignited between the cathodes and an anode with three apertures, each located opposite one of the cathodes. Each cathode had a separate trigger ignition system and the cathodes could be operated simultaneously or separately. The plasma produced by the cathode spots passed through the anode aperture to a 160 mm diameter straight duct. Four magnetic field coils coaxial with the duct axis produced a magnetic field to collimate the plasma flow, and two beam steering coils whose axes were normal to the duct axis produced a magnetic field that deflected the plasma beam in the x and y directions. The saturated ion current distribution in the plasma was measured by a 13-segment multi-probe which was positioned in the sample chamber at a distance of 560 mm from the cathode plane. The measurements were carried out for each cathode operated separately in vacuum and in a background nitrogen gas at pressures of 0.4-1.33 Pa. The ion currents collected by the individual elements of the multi-probe were fitted to a two-dimensional Gaussian function. It was shown that the beam steering coils displaced the plasma beam centres in a wide range in the substrate plane. Changing the current of the X or Y steering coils displaced the beam centre linearly in the x or y directions respectively, but had only a weak effect in the y or x (i.e. orthogonal) directions respectively. The background nitrogen pressure did not influence the displacements of the plasma beam centres; however increasing nitrogen pressure decreased the distribution width. The results show that the pair of beam steering coils may direct the plasma beams to any position in the substrate plane without affecting the other beam distribution parameters.

Thin metal/ceramic coatings were deposited onto polysulfone S2010 substrates using a triple-cathode vacuum arc plasma source connected to a magnetized plasma duct in order to improve the tribological properties of the surface. Various combinations of multi-layer coatings having Ti, Zr, or Nb sub-layers, and nitrides of Ti, Zr, and multi-component (Ti,Zr)N as wear-resistant layers, were deposited and evaluated. The deposition parameters (arc current, magnetic field strengths, deposition time) were optimized (1) to obtain the required deposition rate and coating thickness, while preventing substrate damage under the high-energy ion flux exposure, and (2) to obtain good adhesion of the coating to the substrate at low substrate temperatures. The structure and composition of the coatings were studied using XRD, AES, and SEM. Scratch tests were used to evaluate the adhesive strength between the substrate and the coating, and reciprocating wear tests against a steel ball were used to study the friction and wear rates of the coated samples. The wear tracks were examined by SEM. It was shown that TiN layers possessed a nanocrystalline structure or a mixture of an amorphous and a nanocrystalline structure with random orientation, whereas the ZrN and (Ti,Zr)N had a more defined crystalline nature. The lowest wear rate was observed for Ti/TiN bilayer coatings. The Zr/ZrN coatings failed completely because of poor adhesion to the substrate. It was observed that carbide formed at the interface of the Ti coating and the substrate. In contrast, Zr intermediate layers did not form a carbide, and the coatings had poor adhesion and wear resistance. The results suggest that the formation of a carbide interface improves the coating adhesion.

Thin NbN coatings were deposited using a vacuum arc plasma gun connected to a straight plasma duct, with an imposed axial magnetic field. The substrates were cemented carbide bars having a composition of 90% WC, 1.8% TaC, 0.2% NbC, and 8% Co. The influence of the nitrogen pressure in the deposition system, which was in the range of 0.13–2 Pa, on the structure, phase composition, microhardness, and scratch critical load of the coatings was studied. It was shown that for nitrogen in the pressures range of 0.13–0.4 Pa the coating is composed of a mixture of two phases: hexagonal β-Nb2N and cubic δ-NbN, whereas at pressures of 0.67 Pa and above single-phase δ-NbN with a NaCl type structure was obtained. In most cases the coatings consisted of randomly oriented equiaxial grains. A maximum microhardness of 42 GPa was obtained for the two-phase coatings deposited at a nitrogen pressure of 0.4 Pa. However the maximal critical load of 95 N was obtained with the homogeneous δ-NbN coatings, deposited at a nitrogen pressure of 0.93 Pa, while the coatings with the hexagonal β-Nb2N phase had a much lower critical load (50 N). The NbN coatings with the highest critical load exhibited an average Vickers microhardness of 38 GPa.

Thin coatings of ZrN, and bilayer coatings TiN/ZrN and ZrN/TiN of up to 3 μm thickness were deposited using a triple-cathode vacuum arc plasma gun connected to a straight plasma duct, where an axial magnetic field was imposed. The substrates were cemented carbide bars, having a composition of 90% WC, 1.8% TaC, 0.2% NbC, and 8% Co. The coatings were deposited with an arc current of 200 A, background nitrogen pressure of 0.4–2 Pa, substrate temperatures of 200 to 600 °C, and substrate bias voltages in the range of 0 to −200 V. The magnetic field in the duct was in the range of 1 to 10 mT. The structure and composition of the coating and interface morphology were studied by means of X-ray diffraction, Auger electron spectroscopy, transmission electron microscopy and scanning electron microscopy combined with energy dispersive spectroscopy analyses. It was shown that for nitrogen pressures higher than 0.4 Pa a single-phase ZrN coating with a NaCl-type structure was obtained. The microstructure of the ZrN coatings and ZrN and TiN layers of the bilayer coatings was found to be composed of (111) oriented columnar grains, although near the coating-substrate interface randomly oriented grains were also observed. In the bilayer coatings a sharp interface without intermixing between the TiN and ZrN layers was observed. The preferred grain orientation was independent of the substrate bias voltage and temperature. However, the coating grain size increased with the substrate temperature and decreased with the substrate bias voltage. It was shown that Co diffused from the cemented carbide substrate to the free surface of the coating, and its concentration there increased with the deposition temperature.

Vacuum arcs were established on a 90-mm-diameter Ti cathode in a deposition apparatus consisting of a spacer, 122 mm-diameter annular anode, quarter-torus magnetic macroparticle filter, and a deposition chamber. A toroidal magnetic field generally parallel to the torus walls of up to 20 mT was applied. The ion current in various cross-sections of the toroidal duct was measured using: 1) a disc probe of 130-mm diameter, oriented normal to the torus axis used to measure the transmitted ion current, and 2) a hollow cylindrical probe of 135-mm diameter and 25-mm height, whose axis coincided with the torus axis, used to measure ion current losses to the duct wall. The distribution of ion current loss was studied using an 8-segment hollow cylindrical multiprobe, where the individual probes were equally distributed on the circumference of a 130-mm-diameter circle. It was shown that: 1) the ratio of ion currents collected on the cylindrical and disc probes at first decreases with increasing the toroidal field, and then becomes approximately constant; 2) the presence of the large-diameter disc probe does not influence the value of the ion current on the cylindrical probe; and 3) the maximum ion current density near the torus walls is located in the +g direction and displaces in the -(B/spl times/g) direction with increasing the toroidal field, where g and B are the vectors of the centrifugal acceleration and the magnetic field, respectively.

A filtered vacuum arc deposition system consisted of a cathode, an annular anode, a quarter torus duct macroparticle filter, and a deposition chamber. Arcs were sustained on a Ti cathode in the presence of noble background gases helium and argon, and reactive gases nitrogen and oxygen. The gas pressure was continuously varied from 3 × 10 −5 Torr (4 mPa) to 0.1 Torr (13.3 Pa). A toroidal magnetic field of up to 20 mT, and a straight field in the deposition chamber of up to 10 mT were imposed for plasma guiding. The total saturation ion current was measured with a 130 mm diameter probe. The ion current density distribution was measured with a nine-segment multi-probe, and the individual probe element currents were fitted to a two-dimensional Gaussian distribution. It was shown that in the presence of a noble gas the total saturation ion current at first increases with increasing the background gas pressure, then achieves a maximum at a pressure of 10 mTorr (1.3 Pa) for He, and at 2 mTorr (0.26 Pa) for Ar, where its value is 1.4–2 times greater than in vacuum. With further increase in the pressure the ion current strongly decreases. In the presence of reactive gases this maximum is not observed, and the total ion current strongly decreases at pressures greater than 2–3 mTorr (0.26–0.4 Pa). In contrast to this, a maximum is observed in the ion current collected on small diameter individual probes of multi-probe, positioned towards the direction of the plasma beam displacement. With increasing gas pressure, the distribution width decreases, and a displacement of the beam center is observed both in the − g direction and in the ( B × g ) direction, where B and g are vectors of the toroidal field and centrifugal acceleration, respectively. The present results show that proper substrate positioning in the deposition chamber must take into account the beam displacement due to the background gas.

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.

An analytical solution to the problem of plasma beam transport in a toroidal magnetic filter for unmagnetized ions is derived. A two-fluid model taking into account electromagnetic and pressure forces, electron-ion collisions, magnetic force line curvature, and radial dependence of centrifugal force is used. From comparison with experimental data it is shown that the obtained solution describes well the main properties of plasma beam behavior in the filter, e.g. (1) the relative value of the ion current along the torus decreases exponentially, (2) the deflection of the plasma beam from the center of the torus correlates with the centrifugal drift of the plasma beam across a magnetic field, and (3) experiment and theory agree well on the weak correlation between magnetic field strength and filter efficiency.

The ion current distribution produced by a filtered vacuum arc deposition system consisting of a 90-mm diameter cathode, 122-mm internal diameter anode, and a 240-mm major radius / 80-mm minor radius quarter-torus duct macroparticle filter was measured with a nine-element multi-probe, and the total ion current was measured with a 130-mm diameter probe. Arcs were sustained on Ti and Sn cathodes, with currents of 300–340 and 160–175 A, respectively. Radial magnetic fields of up to 3 mT were imposed on the cathode surface for cathode spot rotation, and fields of up to 20 mT generally parallel to the duct were imposed for plasma guiding. In general, the plasma distribution function exiting from the duct was approximately Gaussian, but with an offset from the center of the duct. At the torus exit, the Ti plasma beam was offset 12 mm towards the outside of the torus at B = 4 mT, while with increasing toroidal field the beam center moved towards the center of the duct, and reached the center at B = 8 mT. The beam was offset towards the inside of the torus for B > 8 mT, reaching a displacement of up to 5 mm for B = 12 mT. The Sn beam, by contrast, was shifted towards the inside of the torus, and the displacement from the center was as much as 15 mm. In addition, for both cases an offset in the (B × g ) direction was observed, where B is the duct magnetic field vector and g is a vector pointing outward from the toroidal center of curvature. The displacements from the center were stronger for Sn plasma (up to 16 mm) than for Ti plasma (up to 7 mm). The displacement from the center in both directions grew with the distance from the exit of the torus. A beam steering coil, whose axis was normal to the axis of plasma duct, was used to shift the plasma beam center towards the chamber axis.

Sn−O transparent, conductive thin films were produced in a filtered vacuum-arc deposition system, consisting of a 90 mm diameter cathode, 122 mm diameter annular anode, and a magnetic quarter torus macroparticle filter. Arc current was in the range 30–200 A and the toroidal magnetic field ( B T ) was in the range 0–25 mT. A radial magnetic field ( B r ) at the cathode surface was produced by a coil placed inside a cavity in the cathode. Cathode spot motion, plasma jet motion through the toroidal field, and the characteristics of the deposited Sn−O films were studied as a function of arc current, applied magnetic fields and oxygen pressure. Cathode spot motion on the cathode surface was controlled by B T . Random spot motion on the entire cathode surface was obtained in vacuum when Br was less than 2 mT. The spots were forced to move on a circular path with average radius of 25 mm when B r was larger than 2 mT. Cathode surface oxidation occurred when oxygen was introduced into the system, and its rate increased with the gas pressure. The extent of the oxidized area on the cathode surface was affected not only by gas pressure, but also by the magnitude of arc current and the magnetic fields B r and B T . With B r smaller than 2 mT, oxidized surface regions appear on the cathode, and cathode spot activity on them is disturbed. Random cathode spot motion continues undisturbed on non-contaminated regions of the cathode surface. Cathode spots were hardly observed on the oxidized regions, yet visual examination of the cathode surface used in the arc revealed the existence of cathode spots erosion tracks on them. Transparent and conductive tin oxide films were produced at oxygen pressures of 0.8 Pa when arc operation and the plasma jet in the filter were stable. The stability of the arc and the plasma jet were found to depend on the combination of gas pressure, arc current and magnetic fields. When the plasma jet was not stable poorly conductive films were produced, containing the non-conductive oxide SnO.

Arc discharges were established on a Ti cathode in vacuum in an arc deposition system with a toroidal macroparticle filter. A magnetic field with a radial component of up to 3 mT on the cathode surface at the distance of 20 mm from the cathode center was applied in order to drive the cathode spots in an azimuthal motion on the front surface of the cathode, and an axial component of the field of up to 30 mT was applied parallel to the walls of the plasma ducts leading from the cathode region to the substrate in order to collimate the plasma beam. Cathode spot motion was observed by means of a television camera and VCR via a window installed at the substrate holder flange and a mirror located in the quarter torus. Ion current convected by the plasma beam was measured with a negatively biased probe. It was shown that the magnetic field of the coils located on the plasma duct has a strong influence on cathode spot behavior. If they produce a field stronger than 4 mT at the cathode surface, the spots move off ...

Arcs were established on a Ti cathode in a filtered vacuum arc deposition system in vacuum and in a low pressure (0.1–0.67 Pa) nitrogen atmosphere. External magnetic fields of up to 25 mT were applied with a radical component in the vicinity of the arc cathode in order to drive the cathode spots in an azimuthal motion on the front surface of the cathode, and with an axial component parallel to the walls of the plasma ducts leading from the cathode region to the substrate in order to collimate the plasma beam. Cathode spot motion was observed by means of a television camera and VCR via a window installed at the substrate holder flange, and a mirror located in the quarter-torus. Ion current convected by the plasma beam was measured with a negatively biased probe. It was shown that the magnetic field coils located on the plasma ducts have a strong influence on cathode spot behaviour. If their field is stronger than 4 mT at the cathode, the spots move off the cathode surface, and the probability of their return is slight. The arc voltage increases, and the arc becomes unstable, and tends to extinguish. Location of the cathode spots on the side of the cathode, which becomes increasingly likely with increasing duct coil current, results in a decrease in the plasma flux through the plasma duct, leading to an optimal field current for optimizing the plasma output flux.

An Er plasma was produced by a vacuum arc source with a truncated cone-shaped Er cathode and an annular copper anode. The plasma beam was injected into a cylindrical magnetic duct through the annular anode aperture and flowed towards the substrate or an electrostatic ion current probe positioned on the duct axis, in vacuum or in a low-pressure oxygen background. The cathode spot operation, transport of the plasma beam in the duct, and coating deposition rate were studied as a function of arc current, axial magnetic field and oxygen pressure. Favorable conditions for arcing were obtained with relatively low arc current (Iarc=30-50 A) and an axial magnetic field of 12 mT, which confined the spot motion on the cathode apex. Under such conditions, the arc operated stably, and the cathode erosion (~235 mug/C) and coating deposition rates (2-5 mum/min) were very high. With higher Iarc (150-175 A), the cathode spots tended to move on the circumference of the large cone base, eroding a deep groove, which significantly decreased the cathode life.

Transparent and conducting SnO2 films of 57-200nm thickness were deposited on microscope glass slide substrates, using a rectangular filtered vacuum arc deposition system. The 40 glass slides were equally distributed on a 400 × 420mm substrate carriage, and were exposed to a Sn plasma beam, produced by a rectangular vacuum arc plasma gun with a Sn cathode, and passed through a rectangular magnetic macroparticle filter towards the substrates. The carriage with the substrates was transported past the 94 × 494mm filter outlet. The SnO2 films were fabricated on the glass substrates at room temperature by maintaining the chamber oxygen background pressure at 0.52Pa. The film composition, and electrical and optical properties were studied as a function of the film thickness. The films were stored under ambient air conditions, and their electrical resistance was measured as a function of storage time over a period of several months. The average resistivity of films was 10-17m? cm for films with thickness (t) less than 100nm, but that of t > 100nm it was 5-9m? cm. The resistivity of the films with t > 100nm did not change significantly after 8months of storage in ambient air. The optical transmittance of the films in the visible spectrum was in the range of 75-90%. The optical constants, i.e., the refractive index and the extinction coefficient of the films at wavelength ? = 550nm were in the range of 2.02-2.09 and 0.013-0.023, respectively, and the optical band gap energy was 4.15-4.21eV. Unlike the electrical resistivity, the optical parameters weakly depended on t. © 2008 Elsevier B.V. All rights reserved. Ministry of Education, Science and Technology The authors gratefully acknowledge financial support from the Israeli Ministry of Science and Technology, Mr. M. Govberg for his engineering assistance, Mr. D. Abunie and Mr. Y. Epstein for their assistance in conducting the experiments, Mr. O. Margulis for his assistance in figure preparation, Dr. L. Burshtein for XPS studies, and Dr. Y. Rosenberg for the XRD analysis.

The plasma beam produced by a vacuum arc plasma source was injected into a cylindrical duct through an annular anode aperture. The plasma source consisted of a frustum cone-shaped Cu cathode, and either a 20-mm-thick annular Cu anode with aperture diameter D of 10, 17, 30, 40, or 50 mm, or 35 mm thick and D=40 or 50 mm. Magnetic coils positioned coaxially with the duct axis produced an approximately axial magnetic field guiding the plasma in the duct. The arc current, I/sub arc/, was in the range of 30-100 A. A 130-mm-diameter negatively biased planar disk probe, positioned normal to the duct axis at a distance of 150 mm from the anode exit, was used to measure ion saturation current I/sub p/. I/sub p/, as well as the ion saturation current to the duct wall I/sub d/, the arc voltage V/sub arc/, and the probe and duct floating potentials with respect to the anode /spl phi//sub p/ and /spl phi//sub d/, were measured as functions of D, I/sub arc/, and the axial magnetic field B. Generally, I/sub p/ and I/sub d/ increased with D and I/sub arc/. For D=10 mm, I/sub p/ was /spl sim/0.4% of I/sub arc/, while with D=50 mm, I/sub p/ reached /spl sim/9.5% of I/sub arc/. However, with large D, the probability of the arc extinguishing increased. Both /spl phi//sub p/ and /spl phi//sub d/ were negative relative to the anode. /spl phi//sub p/ became increasingly negative with increasing D, and approximately linearly depended on D. With a small D, I/sub p/ and I/sub d/ increased almost linearly with B, while /spl phi//sub d/ was almost independent of B. However, for a large D, I/sub p/ and I/sub d/ were only slightly affected by B, and /spl phi//sub d/ became less negative with increasing B.

Arc operation was studied in a vacuum arc deposition apparatus consisting of a Ti cathode, spacer, annular anode, straight duct, quarter torus macroparticle magnetic filter, and a deposition chamber. Superposition of fields from different magnetic coils allowed the formation of different field configurations in the vicinity of the cathode and in the cathode-anode gap. The analytical study of these fields together with the observation of cathode spot motion, made it possible to demonstrate the action of two mechanisms causing unstable arcing: (1) cathode spot movement off of the cathode surface to the side in the direction of the opening of the acute angle formed by the intersection of the field lines with the cathode surface, and (2) cutoff from the anode of the magnetized electron flow by the field lines. With an optimal field configuration both of the above mechanisms were avoided. The field configuration included (1) an arched field on the cathode surface, which rotated the cathode spots on the cathode surface, and (2) connection of a sufficient portion of the cathode surface to the anode with magnetic field lines, so that the random electron current in the vicinity of the anode can supply the required arc current.

On the basis of self-consistent two fluid hydrodynamic model taking into account electromagnetic and pressure forces, electron-ion collisions, magnetic force line curvature and radial dependence of centrifugal force the analytical solution of the problem of vacuum arc plasma beam transport in a quarter-torus magnetic filter is obtained. From comparison with experimental data it is shown that the solution obtained describes well the main properties of plasma beam behavior in the filter, e.g., (1) the relative value of the ion current along the torus decreases exponentially, (2) the deflection of the plasma beam from the center of the torus correlates with the centrifugal drift of the plasma beam, and (3) experiment and theory agree well on the weak correlation between magnetic field strength and filter efficiency.

Two plasma beams of different materials were produced from Ti, Zr, or Nb cathodes in a triple-cathode vacuum arc deposition apparatus. The cathodes were arranged in a circle centered on the system axis. The plasma produced by the cathode spots was transported through a straight plasma duct with an axial magnetic field, into a sample chamber, in which a single Langmuir probe, an array of probes, or a substrate could be mounted 560 mm from the cathode plane. The saturation ion current produced by one cathode, or by the simultaneous operation of two different cathodes, was measured in vacuum and in a 0.13-2.7 Pa nitrogen background. The spatial distribution of the composition of coatings deposited during simultaneous operation of different cathodes in nitrogen was also studied. It was shown that the ion current produced by each single cathode decreased by a factor of 20-100, and by a factor of ten during simultaneous operation of two cathodes, when the nitrogen pressure was increased from vacuum to P=2.7 Pa. During simultaneous operation of two cathodes, the ion saturation current was less than the sum of the ion currents produced by each cathode individually when P

A vacuum arc carbon plasma source is described, in which an arc was ignited between a cathode and an anode having an aperture, by bringing the two electrodes into contact, and parting them while current was flowing. The inter-electrode gap length was varied. A focusing magnetic field was applied in the inter-electrode gap, and a toroidal magnetic field was applied to guide plasma through a toroidal duct. The influence of the arc current (I/sub arc/), gap length (L), and focusing magnetic field (B/sub f/) on the cathode spot motion, plasma behavior and output ion current was studied. It was shown that with I/sub arc/=150-250 A and relatively weak B/sub f/, the spots moved on the cathode periphery, causing noisy arcing and fluctuation of the output ion current, while at lower I/sub arc/=50-100 A and strong B/sub f//spl ges/28 mT the number of the spots on the cathode surface decreased, and they tended to move towards the cathode center, decreasing the arc noise and increasing the average output ion current. With increasing L from 2 to 18 mm, the total ion current at first rapidly increased, and then saturated at L/spl ges/10 mm.

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.

The vacuum arc is a high-current, low-voltage electrical discharge which produces a plasma consisting of vaporized and ionized electrode material. In the most common cathodic arc deposition systems, the arc concentrates at minute cathode spots on the cathode surface and the plasma is emitted as a hypersonic jet, with some degree of contamination by molten droplets [known as macroparticles (MPs)] of the cathode material. In vacuum arc deposition systems, the location and motion of the cathode spots are confined to desired surfaces by an applied magnetic field and shields around undesired surfaces. Substrates are mounted on a holder so that they intercept some portion of the plasma jet. The substrate often provides for negative bias to control the energy of depositing ions and heating or cooling to control the substrate temperature. In some systems, a magnetic field is used to guide the plasma around an obstacle which blocks the MPs. These elements are integrated with a deposition chamber, cooling, vacuum gauges and pumps, and power supplies to produce a vacuum arc deposition system.

Thin coatings of the pure metals Ti, Zr, Cu, and multi-layer metal-nitride coatings Ti/TiN and Ti/ZrN of 0.2–1 μm thickness were deposited onto polysulfone S2010 substrate using a multi-cathode vacuum arc plasma gun connected to a magnetized straight plasma duct. A few Ti coatings were also deposited with the aid of transverse current injection (TCI). The structure and composition of the coatings were studied by means of X-ray diffraction, Auger electron spectroscopy and scanning electron microscopy (SEM) analysis. Scratch tests were used for evaluation of the adhesive strength between the substrate and the coating, and reciprocating wear tests against a steel ball were used to study the friction and wear rates of the coated samples. The wear tracks were examined by SEM. The nitride coatings possessed a nanocrystalline structure or a mixture of amorphous and nanocrystalline structures. The TiN and the ZrN coatings were stoichiometric. The adhesive strength was found to be dependent on the pre-coating treatment and the nature of the metallic bond layer. The Ti/ZrN coating exhibited improved tribological properties as compared to uncoated polysulfone. For this coating a negligible wear rate and a coefficient of friction of 0.1 against a steel ball (lower than the uncoated substrate) was obtained. In addition, TCI significantly reduced the wear rate of the Ti films by a factor of 5 to 6.5.