Target Erosion Research Articles

Impact of rigid angular particles with fully-plastic targets Part II: Parametric study of erosion phenomena

In the accompanying paper (Papini M, Spelt JK. Impact of rigid angular particles with fully-plastic targets. Part I: Analysis. International Journal of Mechanical Sciences 2000;42(5):991–1006), the erosion of targets of arbitrary dynamic hardness and friction coefficient due to the impact of individual angular particles was analyzed with the purpose of predicting crater size, shape, and rebound parameters as a function of incident velocity, angle, particle orientation, and shape. A rigid-plastic model of impact was utilized, and the specific case of symmetric angular particles was studied in detail. In this paper, the model is used to predict crater size in a parametric study of the effect of particle size, shape, incident velocity and orientation, as well as target dynamic hardness and friction coefficient. A dimensional analysis revealed some useful relevant dimensionless parameters. The conditions which maximize the crater size are identified for both constant friction and frictionless surfaces. A parametric study is presented to identify the effects of the most important parameters and can be used as a tool to guide further experimentation.

Nb oxide thin film resistors

Recently Kerr and Pan have proposed development of integrated image separating mixer circuits. To realize the 50 and 100 ohm terminations required by these circuits, a 4.2 K thin film resistor technology is required. It is also desirable to implement each resistor with one or two "squares" of material so that they are electrically short. Existing commonly used Ti-Au based resistor technologies are limited to values well below 50 /spl Omega///spl square/. We are investigating a new Nb-oxide technology where the resistivity is controlled by varying the oxygen content of the sputtered films. Initial experiments using a DC magnetron gun with a Nb target proved difficult to control due to the changing deposition rate and electrical characteristics of the gun with target erosion. Greatly improved control was obtained with the use of a RF diode sputter arrangement and a 5" diameter target. As target voltage is decreased or oxygen partial pressure is increased, the deposited films become less metallic and more resistive. Films with ohms//spl square/ values ranging from five to 1,000 have been obtained with this technique while a partial pressure of oxygen on the order of 10/sup -3/ Pa is needed for 50 /spl Omega///spl square/ material. A valuable attribute of these resistive films is that the measured room and liquid He temperature resistivities are the same.

The effect of velocity on the solid particle erosion rate of alloys

Velocity is a critical test variable in erosion, and can easily overshadow changes in other variables, such as target material, impact angle, etc. The effect of velocity on erosion rate was studied in 70–30 brass (cold worked and annealed) and Fe–C martensite (as-quenched and tempered) and it was found, as previously shown, that erosion rate is dependent on velocity by a power law, given by ER= kV n . However, the velocity exponent n was found to be target material independent and is governed by test conditions, including particle characteristics and the erosion test apparatus. In addition, n is not dependent on the erosion mechanism. Results from tested as-quenched martensite and tempered martensite showed that the exponent is approximately 2.9 for both materials, even though martensite eroded by a brittle cracking mechanism, while tempered martensite eroded by a plastic deformation mechanism. No difference in the erosion rate relationship was found between the fully annealed and the 70% cold worked brass. The exponent n was found to change over time with nominally the same erosion test conditions, indicating that n is very sensitive to slight changes in erodent particles and/or the test apparatus, and that it must be measured periodically if erosion results generated at different times are to be compared. 1 Prime Novelty: The velocity exponent n is independent of both target material and erosion mechanism and is highly dependent on erosion test conditions for the materials tested. 1

Titanium carbide film deposition by DC magnetron reactive sputtering using a solid carbon source

Titanium carbide films with various C/Ti ratios have been deposited by DC magnetron sputtering using carbon sheets on the Ti target erosion area as a solid carbon source. By changing the number of carbon sheets (5 × 5 × 1 mm 3) from 0 to 24 pieces, the C/Ti compositional ratio of the films was changed. The composition, structure, and hardness of the deposited films were estimated as a function of the ratio of the source carbon area to the titanium target erosion area. The results of X-ray photoelectron spectroscopy showed that the C contents in the films increased to Ti: C = 60: 40 as the source C/Ti areal ratio increased. By X-ray diffraction, the films obtained for C/Ti areal ratios above 0.1 were found to possess the face-centered cubic structure and that the d-value of TiC (111) increased monotonically from 0.238 to 0.249 nm as the C/Ti areal ratio increased. The hardness of the films also increased monotonically as the C/Ti areal ratio increased, yielding a maximum of 11 GPa.

Erosion of beryllium and deposition of carbon and oxygen due to bombardment with C + and CO + ions

The bombardment of clean, metallic beryllium with carbon and the co-bombardment with carbon and oxygen is investigated experimentally and by computer simulation. The experiments use C + and CO + ions with energies between 3 and 12 keV at normal incidence. The simulation is carried out with the Monte Carlo code TRIDYN. The results of carbon bombardment from experiment and simulation agree well and show the buildup of a carbon adlayer on the metal surface. After the carbon adlayer formation, no more Be is eroded. For C and O co-bombardment, carried out by CO + ion bombardment, simulation and experiment agree for low fluences. At higher fluences, another erosion channel besides kinematic sputtering is opened by the chemical interactions of C and O. This leads to a deposition/erosion equilibrium which is not predicted by the kinematic TRIDYN code. The formation of a ternary mixture layer on the Be target leads to a continuous erosion of the beryllium target.

Studies on the optimisation of unbalanced magnetron sputtering cathodes

The optimisation is reported on the design of unbalanced magnetron (UBM) sputtering cathodes. For the study, a planar circular cathode backed by a double-coil electromagnet (compatible for a 100 mm diameter target) was developed. The variation of the structure and strength of the magnetic field in front of the target was investigated for different current combinations in the electromagnetic coils, and its effect on the sputtering process was analysed. The observations on the magnetic field geometry revealed some interesting features, such as the balancing point of the fields along the axis (null-point), and the zero axial region over the target surface ( B z =0 ring). The positions of both could be controlled by adjusting the ratio of the electric current in the coils. The magnetic field null-point could be used as a reference for the region of homogeneous film growth. The B z =0 ring was the location where the glow discharge concentrated (or where the maximum target erosion occurred). The diameter of the ring determined the area covered by the discharge and thus the sputtering efficiency. The optimum substrate position can be fixed according to the position of the null-point and optimisation of sputtering can be achieved by adjusting the diameter of the B z =0 ring. The results of this study should be helpful in the designing of an ideal UBM using permanent magnets as well as electromagnets.

Rectangular magnetron with full target erosion

The article reports on the operation of a rectangular magnetron with three and four closed discharges simultaneously formed on the surface of sputtered target. The closed discharges, elliptically shaped with the longer axis perpendicular to the long side of the rectangular magnetron, are formed using magnetic circuits placed behind the sputtered target. By moving these magnetic circuits along the long axis of a rectangular target, it is possible to achieve almost full target erosion and therefore significantly increase the usage lifetime of the target. This principle can be used for rectangular magnetrons with arbitrary length of the target. The longer the target, the greater the number of closed discharges that are required. Experimentally, it was found that neighboring closed discharges can interact which may significantly modify the sputtering characteristics of the magnetron cathode. This interaction is described and the physical requirements to avoid its occurrence are given. The I–V characteristics of a rectangular magnetron with one, two, and three closed discharges are given.

Plasma chemistry model of DC magnetron reactive sputtering in Ar[sbnd]O2 gas mixtures

A space averaged model of reactive sputtering which includes plasma chemistry for the first time in DC magnetron discharge has been developed by Ershov and Pekker [A. Ershov. L. Pekker, Thin Solid Films, 289 (1996) 140.]. Argon and oxygen have been chosen in this model as buffer and reactive gases, respectively, and silicon as the target material. The present work is a further development of this model. The present model includes (1) The contribution of secondary electron emission induced by ion bombardment to the total current density and (2) A new plasma chemical reaction of dissociation of molecular oxygen due to its collisions with metastable argon. Thus, the present model has a more advanced plasma chemistry and also allows to take into account the secondary electron emission to investigate reactive sputtering and plasma chemistry. In the model, we show that the consideration of the secondary electron emission can significantly decrease the target erosion rate and shift the hysterisis curves. Numerical calculations also shows that the dissociation of molecular oxygen by metastable argon can significantly change the plasma chemistry equilibrium of the discharge. Generally, deposition of any oxide film can be studied with this model.

Simulation of sputter deposition at high pressures

Sputtering, although often optimized by trial and error, increasingly needs sound modeling techniques because of the high capital investment. Usually the simulation problem is broken down into three stages: emission at the target, transport through the host gas, and deposition at the substrate. It turns out that these stages are necessarily linked and must be properly decoupled. Here we are mainly interested in the transport through the vapor phase, which may be simulated by our Monte Carlo model, SIMSPUD. However, at high operating pressures, significant deposition takes place onto the target. To modify our Monte Carlo routine, rather than emitting particles in proportion to the erosion profile of the target and waiting for some of these to completely escape, all particles which land on the target are re-emitted until they ultimately escape to the substrate or chamber walls. This modification makes the net emission rate equal to the target erosion rate. At sufficiently high pressures, the Monte Carlo routine becomes very slow. In fact, the problem then becomes a diffusion problem best handled by numerical solution of the Laplace equation. At any rate, it is shown that when SIMSPUD has been modified in this way, good agreement with experiment is obtained over a wide range of working pressures.

Time dependence and spatial distribution of the deposition rate of YBa2Cu3O7 thin films in 90° off-axis sputtering

We have examined the time dependence and spatial distribution of the deposition rate of YBa2Cu3O7 (YBCO) thin films in 90° off-axis radio frequency planar magnetron sputtering throughout the life of a target. We have found that although the amount of YBCO sputtered from the target is constant, the amount of YBCO deposited on the substrate decreases linearly over time. Furthermore, the time dependence of the deposition rate changes at different positions on the substrate. The changes in the deposition rate at any point along the substrate as a function of time can be explained by the change in the angular distribution of the sputtered particles that is caused by changes in the target erosion groove geometry, and the decrease in the kinetic energy of sputtered particles that is indicated by the decrease in the self-bias voltage over time.

Plasma/Surface Interaction in ITER Tokamak Disruption Simulation Experiments

In evaluating the lifetime of plasma-facing components for the International Thermonuclear Experimental Reactor (ITER) against nonnormal high heat loads, credit is taken from the existence of a plasma shield that protects the target from excessive evaporation. Formation and physical properties of plasma shields are studied at the dual plasma gun facility, 2MK-200, under conditions simulating ITER hard disruptions and edge-localized modes (ELMs). The experimental results are used for validation of the theoretical modeling of the plasma/surface interaction. The important features of the non-local thermodynamic equilibrium plasma shield, such as temperature and density distribution, its evolution, the conversion efficiency of the energy of the plasma stream into total and soft X-ray radiation from highly ionized evaporated target material, and the energy balance in the plasma shield, are reproduced quite well. Thus, realistic modeling of ITER disruptive plasma/wall interaction is now possible. Because of the rather small target erosion in the simulation experiments, material erosion for ITER typical disruptions and ELMs cannot be evaluated from these simulation experiments. This requires additional simulation experiments with hot plasma streams of longer pulse duration and a separate numerical analysis, which can now be performed with validated theoretical models.

Study of impurity radiation in ASDEX Upgrade with a spatially scanning spectrometer for the VUV and visible ranges

A combined visible and vacuum ultraviolet (VUV) scanning spectrometer for the investigation of the boundary layer of ASDEX Upgrade has been developed. A rotating mirror is used for poloidal scanning. This mirror can be tilted perpendicular to its rotation axis, enabling observation in the toroidal direction too. The broad spectral range opens a variety of investigation possibilities. Here we present poloidal radiation profiles of different ion species in the VUV. Regions of special interest such as the divertor for target erosion studies and the upper boundary for the estimation of ion penetration depths were investigated with high spatial resolution. Additionally, the application of this instrument is shown for measuring the spatial distributions of impurities injected for radiation cooling.

Improvement of Sputtering Target Utilization Using Dynamic Plasma Processing

A newly proposed design of the magnetron sputtering cathode shows to vastly improve target utilization efficiency for depositing films. In order to investigate this design's primary intent to obtain widely diffused target erosion areas, the finite element method (FEM) was employed to study the magnetic field distributions on the target surface. Favorable results from the study compelled the fabrication of this magnetron sputtering cathode design. This cathode creates widely diffused erosion areas by moving the position of the intense plasma, where the target becomes deeply sputtered, on the target surface by controlling the magnetic field with solenoid coils mounted on the inside of the sputtering cathode. This sputtering cathode design proved to achieve target utilization efficiencies as high as 40% for a 1150 ×150 ×6t mm indium tin oxide (ITO) target, in addition to satisfying the properties of deposited ITO films for use in transparent electrodes.

Composition change in magnetron-sputtered Fe-Zr-N film with erosion of Fe-Zr alloy target

The composition of magnetron-sputtered Fe-Zr-N films is found to change with the erosion of the high-permeability Fe-Zr alloy target. Zr is preferentially sputtered parallel to the target surface, and this tendency is more pronounced as erosion increases, Incident ion energy decreases with erosion because the leakage magnetic flux intensity increases, and the plasma density increases as a result of enhanced ionization frequency. This composition change has been attributed to a reflective collision process becoming dominant as the incident ion energy decreases with increasing erosion. A novel mosaic-type target, consisting of both alloy and single-element phases, has been developed, and the composition change is successfully suppressed and compositional uniformity is markedly improved.

Film thickness variation in a cylindrical magnetron deposition device

The 2D thickness distributions of Cu and Al planar thin films prepared by a home-made cylindrical magnetron with rectangular exit window were examined. The experimental results were compared with several models of sputtered particle transport. It has been shown that in the azimuthal direction the best agreement with experimental results was achieved with the free flight model. Due to an uneven target erosion, the agreement with the model calculations proved to be somewhat less satisfactory in the axial direction. An overall similarity of thickness distributions in the case of Al deposition at P Ar≈0.7 Pa and Cu deposition at P Ar≈2.25 Pa working gas pressure has been found. However, it has been found that the portion of backscattered flux was greater in the case of Al than it was for Cu sputtering.

Erosion rate prediction and correlation technique for ceramic surfaces exposed to high speed flows of abrasive suspensions

We describe a simple method to predict (based on available erosion yield data) erosion rates for cylindrical ceramic targets (e.g. circular tube coatings, leading edges of turbine blades, or the ceramic-lined target zone inside CFBC-cyclones) exposed to high-speed abrasive particle-laden streams. Use is made of a convenient parameterization/extrapolation of published laboratory results giving, in effect, average erosion yields per particle impact (ϵp) for particular planar ceramic target and projectile materials over a range of impact velocities, Vp, incidence angles, θj, and particle sizes, νp. For a given target/flow geometry we reduce the engineering problem of predicting absolute target erosion rates to that of multiplying a readily calculated characteristic erosion rate by the universal dimensionless erosion rate functions explicitly approximated here in the limit of impacts by particles large enough to be undeflected or slowed down by the local target gas flow. Our characteristic erosion rate is that which would be associated with the mainstream abrasive particle volume flux if all particles struck at normal incidence with the mainstream velocity, U. Dimensionless erosion rate results are cast in terms of the following four dimensionless parameters characterizing the erodent/ceramic target system of interest: sensitivity (exponent n) of erosion yield to projectile incident velocity; sensitivity (exponent m appearing in (costm(θi)) of erosion yield to angle of incidence θi: sensitivity (exponent l) of erosion yield to projectile particle size (volume, νp); and the reference erosion yield, ϵp,ref (here, ϵp is evaluated at Vp = 100 ms−1, θt = o, and νp corresponding to dp = 100 μm). Based on our preliminary survey of available erosion yield experimental data, we provide a table givin “best-fit” values of the four parameters: l, m, n, and ϵp,ref required to complete a prediction of local and spatially-averaged erosion rates according to our present formalism. For the latter, useful closed-form approximations are provided for convex or concave cylindrical target geometries in the high Stokes number limit. Moreover, convenient correction factors are developed to account for a (Rosin-Rammler) particle size distribution in the erodent mainstream, and mainstreams not perpendicular to the cylinder axis. The more general case of arbitrary (finite) Stokes numbers is outlined. Using two numerical examples (convex leading edge coating on a turbine stator blade, and concave sector target zone in a CFBC-cyclone), we demonstrate that casting required erosion yield data in this suggested format greatly facilitates erosion design calculations for ceramic targets (or coatings) exposed to high-speed abrasive particle suspensions. Organizing empirical data in this manner will also facilitate the longer range goal of correlating each of the above-mentioned parameters with independently measurable physical properties of the participating materials.

Monte Carlo numerical analysis of target erosion and film growth in a three-dimensional sputtering chamber

A combination of two mathematical models for three-dimensional Monte Carlo particle simulation of a low pressure sputtering environment is proposed. One is intended for the simulation of a discharge gas flow, and the other for the sputtered atom transport. The combination is used to characterize target erosion and film growth. The models are refined with recourse to experimental measurements made in a practical sputtering apparatus (SPF-210 AS, ANELVA Ltd.) over the range of operating pressures and flow rates of 0.3–10 Pa and 0.5–5 sccm, respectively. A considerable number of numerical analyses are done to find possible reasons for the measured nonuniformity of target erosion and film growth rate distributions. Simulation results show that under the operating and design conditions treated here the nonuniformity of gas flow field appears to be too weak to explain the experimental data. Film growth rates simulated for measured erosion rates show a good agreement with the experimental data for the various operating pressures above. Important features of the present simulation are the inclination of the angular distribution of sputtered atoms, and taking into consideration reemission of sputtered atoms deposited onto the target. The inclination of angular distribution results from the oblique ion incidence onto the target surface. The existence of highly eroded etch track along the target edge is shown to be correlated with the greater sputtering yield for the oblique ion incidence near the edge. Axial asymmetry of the erosion and growth rate distributions may be due to the asymmetry of the electric field lines at the periphery of the target, which leads to a nonuniform incidence angle along the edge. In contrast to erosion and growth rates, the redeposition rate of sputtered atoms onto the target is found to increase considerably with pressure.

Model of d.c. magnetron reactive sputtering in ArO 2 gas mixtures

A model of reactive sputtering, which includes plasma chemistry for the first time in d.c. magnetron discharges, has been developed. Argon and oxygen have been chosen as buffer and reactive gases, respectively, and silicon as the target material. We have developed a space averaged model of the plasma chemistry in a d.c. magnetron reactive sputtering process. We have calculated the rate coefficients of plasma chemical reactions and solved the plasma chemistry equations together with sputtering deposition equations. We have assumed different sticking probabilities of atomic oxygen created in plasma as compared with molecular oxygen. The model allows the investigation of the sputtering process and the plasma chemistry of the discharge for an arbitrary ratio of partial pressures of reactive and buffer gases. This makes it possible to simulate reactive sputtering for large erosion rates of target material. It is obtained that the dissociation of molecular oxygen in the discharge is so significant that the atomic oxygen is the main oxidizer for the target and substrate for almost all current densities. Numerical calculations also show that sputtering by oxygen ions can be quite significant, causing up to 1/4 of the total target erosion rate. The model describes well-known experimental results, for example, the hysteresis for small pumping speeds. Generally, deposition of any oxide film can be studied with this model.

Investigation of the candidate divertor materials erosion at the powerful electron beam

The powerful electron beam is a rather convenient energy source for the investigation of the material erosion because its parameters at the target can be determined with sufficient accuracy and can be controlled in a wide range. The electron beam facility ELDIS being used for investigation of the candidate divertor materials erosion has the following parameters: electron energy 60–140 keV, pulse duration 10–60μs, power density at the target up to 5 MW/cm 2, minimum beam diameter at the target 1.2 cm, magnetic field at the target up to 4 T. The maximum energy deposition density had approximately the same value for the materials investigated (C, Cu, W) equal to 10–12 kJ/g. In the course of experiments the target erosion and target plasma parameters were determined. Shielding effect is negligible and the erosion is mainly determined by the energy deposition of the unattenuated electron beam. The erosion of the graphite targets cannot be explained by surface evaporation or volumetric sublimation. It seems likely that it is connected with intensive evaporation inside the pores causing the material destruction.

Kinetic modeling of scrape-off layer plasmas

Electron transport along open field lines in the diverted scrape-off layer of a tokamak is studied numerically via a kinetic Fokker–Planck approach. The method allows calculation of the distribution function in a situation where large parallel temperature gradients are maintained by collisional relaxation and, at the same time, superthermal electrons stream freely from the midplane of the plasma to the target/sheath boundary. The method also allows calculation of the self-consistent electrostatic field associated with parallel gradients in the distribution function, as well as the potential drop across the target/sheath boundary, where the latter is calculated to enforce appropriate boundary conditions at the target, although the sheath itself is not resolved. The kinetic results are compared to classical fluid results for the case of a simple (nonradiative) divertor. The kinetic solutions exhibit an enhanced superthermal electron population in the vicinity of the target, which results in a larger sheath energy transmission factor, a lower bulk electron temperature, and a smaller sheath potential drop. The sheath potential largely determines the energy with which ions impact the target, thereby affecting the rate of target erosion. Ionization rates and radiation rates from impurities in the vicinity of the target also depend strongly on the local electron temperature and can be sensitive to superthermal tails.