Temperature dependence of sputtering behavior of Cu-Li alloys

Three different Cu-Li alloys (4, 10, and 16% Li) have been sputtered by 1 keV and 100 eV D⁺ ions. The Cu sputtering yield and the total weight loss was measured as a function of the target temperature between 25 and 700°C. The yield was measured by the catcher foil technique and by the weight loss method in order to differentiate between the total weight loss and the Cu sputtering yield (i.e., Rutherford backscattering analysis (RBS) of the catcher foil). Targets with lower Li concentration (4 and 10%) did not show a significant change of the Cu sputtering yield [1] as found by other authors [2] and this is probably due to the higher current density (10¹⁵ cm⁻² s⁻¹) in this experiment. An increase in weight loss at temepratures above 550°C was caused by Li evaporation. The target with high Li concentration (16% Li) showed a reduction of the Cu sputtering yield by more than a factor of 50 for both the 1 keV and the 100 eV D⁺ ions. This reduction occurs in a small temperature range around 550°C, which coincides with the transition on the Cu-Li phase diagram from the a-phase to α + liquid. For temperatures above 550°C the sputtering yield increases again, most probaby due to an enhanced evaporation of Li. At optimum temperature conditions, the evaporation rate of Li for the Cu-Li alloy is many orders of magnitude lower than the rate for pure Li. According to the phase diagram, the Cu-Li alloy with even higher Li concentration could reach optimum conditions at lower temperatures and, therefore, would be a promising first-wall candidate.     

Modeling of plasma/lithium-surface interactions in NSTX: Status and key issues

We are studying lithium sputtering, evaporation, transport, material mixing, and surface evolution for the National Spherical Torus Experiment (NSTX) for various surfaces and plasma conditions. Lithium modeling is complex, particularly for NSTX short pulse, multiple material, variable plasma conditions. Cases examined include: (1) liquid lithium divertor (LLD) with planned high heating power/low-D-recycle plasma, (2) non-pumping/high-recycle solid or liquid divertor surface, (3) Li and C impingement on a molybdenum surface. An impurity erosion/redeposition code package is the overall integration tool, with sputter yield and velocity distributions from binary collision mixed-material codes, sheath code input for NSTX boundary conditions, and inputs of plasma edge solutions from external data-calibrated plasma fluid codes. Analysis predictions are generally favorable, showing non-runaway lithium self-sputtering, acceptable net erosion (～5 nm/s), and moderate edge (～10%) and core plasma (～0.1–1%) Li contamination, for the cases studied. A Mo divertor surface is significantly affected by C and Li impingement but with low core plasma contamination predicted for a high-recycle edge plasma.     

Observations of hydrocarbon film deposition in the MAST tokamak

Two divertor tiles from the Mega-Ampere Spherical Tokamak (MAST) were analysed using interference fringe analysis, nuclear reaction analysis and Rutherford backscattering analysis. The analysis allowed the quantification of the co-deposited layers thickness of the deuterium and carbon mixture and the detection of the presence of other impurities such as oxygen and iron. A layer thickness varying from 10 to 270 nm was measured with a typical D/(D+C) ratio of ≈0.36. During the same campaign in which the analysed divertor tiles were exposed to the plasma, the ion flux to the divertor plates has been measured by a set of Langmuir probes mounted on the divertor ribs. This paper shows that the sputtering co-efficient that can be inferred from the ion fluences measured by the probes and by the redeposition of C in those amorphous layers is consistent with physical and chemical sputtering yields in conventional tokamaks. This paper reports the beginning and the intended developments of a project undertaken in MAST to study and quantify the phenomena of plasma surface interaction both at the walls and at the divertor of the machine in order to contribute to the material choice and design of the plasma facing components of a next step device such as ITER.     

Behavior of carbon readsorbed on tungsten during low energy Ar ion irradiation at elevated temperatures

A study of the behavior of carbon sputtered and readsorbed after scattering collisions with particles of surrounding gas on the tungsten surface affected by Ar ion irradiation with the flux equal to 2 × 10¹⁶ cm⁻² s⁻¹ extracted from plasma under 300 V negative bias voltage in the temperature range 370-870 K was performed. The dependence of the W sample weight change on the working gas pressure in the range 0.1-10 Pa was registered and the information was deduced about prevailing sputtering-redeposition processes. The depth profiles of carbon at the tungsten surface were measured. We found that carbon distribution profiles in tungsten depend on the C redeposition rate for fixed ion irradiation parameters. Three regimes have been distinguished: (i) at working gas pressure equal to 5 Pa and more, the C redeposition rate prevails the sample surface erosion rate and the W surface is covered by continuous amorphous carbon film (the C film growth regime), (ii) at working gas pressure equal to about 1 Pa, the C redepostion rate is approximately equal to the erosion rate and the W surface is partially covered by redeposited carbon, and (iii) at working gas pressure less than 0.2 Pa, the erosion rate prevails the C redeposition rate (the W surface erosion regime). In the regime of balanced redeposition and erosion deep C penetration depth into nanocrystalline W was registered.It is suggested that under simultaneous C adsorption and ion irradiation at elevated temperature C adatoms are driven from the W surface into grain boundaries and into the bulk by the difference in chemical potentials between the activated W surface and grain boundaries. As the W surface is covered by amorphous C film, the grain boundaries are blocked and the efficiency of carbon transport decreases.     

Atomistic self-sputtering mechanisms of rf breakdown in high-gradient linacs

Molecular dynamics (MD) models of sputtering solid and liquid surfaces - including the surfaces charged by interaction with plasma, Coulomb explosion, and Taylor cone formation - were developed. MD simulations of self-sputtering of a crystalline (1 0 0) copper surface by Cu⁺ ions in a wide range of ion energies (50 eV-50 keV) were performed. In order to accommodate energetic ion impacts on a target, a computational model was developed that utilizes MD to simulate rapid atomic collisions in the central impact zone, and a finite-difference method to absorb the energy and shock wave for the collisional processes occurring at a longer time scales. The sputtering yield increases if the surface temperature rises and the surface melts as a result of heat from plasma. Electrostatic charging of the surface under bombardment with plasma ions is another mechanism that can dramatically increase the sputtering yield because it reduces the surface binding energy and the surface tension. An MD model of Taylor cone formation at a sharp tip placed in a high electric field was developed, and the model was used to simulate Taylor cone formation for the first time. Good agreement was obtained between the calculated Taylor cone angle (104.3°) and the experimental one (98.6°). A Coulomb explosion (CE) was proposed as the main surface failure mechanism triggering breakdown, and the dynamics of CE was studied by MD.     

Studies of Li sputtering and particle recycling in glow discharges and in TJ-II

Lithium sputtering is studied at TJ-II with lithiated, boron coated walls. As reported previously, the sputtering yield measured in H plasmas was found to be significantly lower than expected, based on laboratory experiments and TRIM code calculations. The edge temperature dependence of the yield could be consistent with either an energy threshold far exceeding the corresponding pure lithium threshold, or with a strong perturbation of the impinging ion energies. The material mixing effect, one of the candidates for explaining the observed behavior, was studied by depositing boron on the Li layer. Also, particle retention and release was investigated in H/He plasmas in this very low recycling scenario. The release of either species in the opposite plasma was studied in different plasma conditions as well as in glow discharge (GD) plasmas. In He GD plasmas, the I/V characteristics of lithiated stainless steel electrodes were found to be anomalous, in spite of the fact that the dependence of sputtering on the incident particle energy agreed reasonably with expectations. The possible implications of these phenomena for the interaction of reactor plasmas with lithium elements are addressed.     

Assessment of approximations for efficient topography simulation of ion beam processes: 10 keV Ar on Si

The main assumption of existing efficient topography simulations is that sputtering is a local process that depends only on the angle of incidence and not on the detailed shape of the surface. If redeposition is considered, sputtered atoms are redeposited and cause no further sputtering when they hit another part of the surface. Furthermore the angular distribution of sputtered atoms follows a cosine law. If ion reflection is considered, ions do not lose energy during backscattering. Using binary collision simulations (IMSIL) and comparing them with results obtained by a topography simulator (IonShaper^®) we show that all these assumptions need refinement for the simulation of nanostructures except the neglect of sputtering by sputtered atoms. In addition we show that a nonlocal model is essential for ion beam induced deposition of narrow structures.     

Nanostructured carbide surfaces prepared by surfactant sputtering

Nanostructured surface layers of titanium carbide and tungsten carbide were prepared on tetrahedral amorphous carbon (ta-C) films using the surfactant sputtering technique. Surfactant sputtering is a novel ion beam erosion technique, which utilizes the steady state coverage of a substrate surface with foreign atoms simultaneously during sputter erosion by combined ion irradiation and atom deposition. These foreign atoms act as surfactants, which strongly modify the substrate sputtering yield on atomic to macroscopic length scales. The novel technique allows smoothing of surfaces, the generation of novel surface patterns and nanostructures, controlled shaping of surfaces on the nanometer scale and the formation of ultra-thin compound surface layers. We have sputter eroded ta-C films using 5 keV Xe ions under continuous deposition of either tungsten or titanium surfactants. This leads to the steady state formation of a W_xC or a TiC/a-C nanocomposite surface layer of few nm thickness. Depending on the ion angle of incidence the layer is either smooth or nanostructured with a ripple- or dot-like surface topography. We have analyzed the surface topography, the composition and microstructure of the metal-carbon nanocomposites, and compare coverage dependent sputtering yields with SRIM and TRIDYN simulations.     

Evaluation of energy and particle impact on the plasma facing components in DEMO

We analyze the first wall blanket W/EUROFER configuration for DEMO under steady-state normal operation and off-normal conditions, such as vertical displacement events (VDE) and runaway electrons (RE). The main issue is to find the optimal thickness of the W armor which will prevent tungsten surface from evaporation and melting and, on the other hand, will keep EUROFER below the critical thermal stresses. Under steady-state operation heat transfer into the coolant must remain below the critical heat flux (CHF) to avoid the possible severe degradation of the coolant heat removal capability. From the plasma side it is particularly demanding to keep the bulk plasma contamination during the reactor long operational discharges below the fatal level. The possible damage of the FW materials due to the plasma sputtering erosion is estimated. The minimum thickness of the tungsten amour about 3 mm for W/EUROFER sandwich structure will keep the maximum EUROFER temperature below the critical limit for EUROFER steel under steady-state operation and ITER like cooling conditions.     

Chemical effects of thermonuclear plasma interactions with insulator and metal surfaces

The interactions of chemically reactive ions with insulator and metal surfaces result in specific chemical effects which must be considered in sputtering, blistering, trapping and re-emission processes involving these projectiles. Chemical sputtering and chemical trapping of ion beam flux are examples of chemical interaction effects which are discussed. Bombardment of a niobium surface by oxygen and nitrogen ions results in the formation of sputtered niobium oxide and nitride molecules. The molecular species are identified and characterized by means of their vibrational spectra using isotopic substitution and matrix isolation techniques. Results using matrix isolation spectroscopy for both physical and chemical sputtering studies on materials of CTR interest are presented. Chemical trapping is evaluated as a method of chemically pumping the major fraction of ion beam flux in the divertor of next generation tokamaks.     

Study of ion beam induced mixing during sputter depth profiling of thin films by LEIS

Ion beam induced mixing during sputter depth profiling was studied for tantalum and lead marker layers in silicon with 5 keV neon by low energy ion scattering spectroscopy (LEIS). The diffusion approximation was used to calculate mixing efficiency values (D/JF_d) from decay length measurements. The mixing efficiency values are shown to be sensitive to the preferential sputtering which takes place during ion bombardment. TRIM simulations for Ta/Si are shown to agree with the experimentally determined value for preferential sputtering. Depth profiling at high temperature is shown to separate some of the interrelated mixing mechanisms of radiation enhanced diffusion (RED) and radiation induced segregation (RIS). For the Ta/Si system the mixing efficiency value is observed to remain constant regardless of the 5 keV inert gas ion beam used for depth profiling.     

强流脉冲离子束辐照对钛合金显微组织的影响

为研究强流辐照过程中钛合金微观组织结构和化学成分的变化,利用束流成分70%H+和30%C+的混合强脉冲离子束对钛合金进行了表面轰击,对离子束诱发的显微组织形貌和化学成分在扫描电子显微镜(SEM)上进行分析。结果表明,钛合金内部组织由α和β两相构成,为α+β型两相钛合金。试样不同区域组织类型存在明显差异,大部分区域组织类型一致,局部区域属于典型的魏氏组织。强流辐照钛合金表层组织和微观结构发生一定的变化,边缘β相结构遭到一定程度的破坏。显微组织结构的变化是由于表层温度梯度引起的热应力及其在靶材体内的传播造成的。     

Effect of Frequency and Power of Bias Applied to Substrate on Plasma Property of Very-High-Frequency Magnetron Sputtering

The effect of the frequency and power of the bias applied to the substrate on plasma properties in 60 MHz(VHF) magnetron sputtering was investigated.The plasma properties include the ion velocity distribution function(IVDF),electron energy probability function(EEPF),electron density n_e,ion flux Γ_i,and effective electron temperature T_(eff).These parameters were measured by a retarding field energy analyzer and a Langmuir probe in the 60 MHz magnetron sputtering,assisted with 13.56 MHz or 27.12 MHz substrate bias.The 13.56 MHz substrate bias led to broadening and multi-peaks IVDFs,Maxwellian EEPFs,as well as high electron density,ion flux,and low electron temperature.The 27.12 MHz substrate bias led to a further increase of electron density and ion flux,but made the IVDFs narrow.Therefore,the frequency of the substrate bias was a possible way to control the plasma properties in VHF magnetron sputtering.     

Modeling of sputtering erosion/redeposition—status and implications for fusion design

This paper reviews sputtering erosion/redeposition modeling for plasma facing surfaces. Basics of the WBC Monte Carlo impurity transport code are described. An example analysis is shown for erosion of a liquid tin fusion reactor divertor. Multiyear erosion studies of candidate divertor and first wall coating materials (Li, Be, C, W, etc.) are summarized—showing generally serious erosion concerns for low-Z solid materials, and encouraging results for high-Z materials and liquid divertor surfaces. Future goals are discussed, e.g. for supercomputer modeling.     

Growth of carbon impurity density and adequate divertor wall temperature for fusion experimental reactors

Sudden increase of carbon impurity called carbon bloom has terminated the energy breakeven condition in the present large tokamak. In order to lengthen the burning plasma state in the next device, carbon bloom has to be well suppressed. The temporal evolution of carbon impurity density is analytically examined by using a simple one-point kinetic or zero-dimensional model including the effects of graphite erosions due to oxygen and ion, and gettering for oxygen due to boron or beryllium. The growth of carbon bloom due to radiation-enhanced sublimation is discussed based on the effective self-sputtering of carbon. Even when the self-sputtering yield is less than unity, carbon density is observed to continuously increase with the discharge time if the oxygen gettering action is not perfectly conducted. From the present analysis and data on the erosion of carbon materials, and the evaporation of gettering materials, it is suggested that the divertor wall temperature has to be kept less than approximately 900–1000°C to avoid the continuous growth of the carbon density.     

Argon irradiation damage at a Cu/Al2O3 interface for different alumina structures

The evolution of damages at a Cu/Al₂O₃ device interface after Ar⁺ irradiation, depending on alumina structure, and the effect of surface roughness on sputtering have been studied. A polycrystalline Cu/Al₂O₃ bilayer and polycrystalline Cu on amorphous alumina were irradiated with 400 keV Ar⁺ ion beam at doses ranging from 5 × 10¹⁶ to 10¹⁷ Ar⁺/cm² at room temperature. The copper layer thicknesses were between 100 and 200 nm. RBS analysis was used to characterize the interface modification and to deduce the sputtering yield of copper. The SEM technique was used to control the surface topography. A RBS computer simulation program was used to reproduce experimental spectra and to follow the concentration profile evolutions of different elements before and after ion irradiation. A modified TRIM calculation program which takes into account the sputtering yield evolution as well as the concentration variation versus dose gives a satisfactory reproduction of the experimental argon distribution. The surface roughness effect on sputtering and the alumina structure influence at the interface on mixing mechanisms are discussed.     

Erosion yield of Ti from TiC-deposited graphite by hydrogen ion bombardment at high temperatures

Erosion yields of Ti atoms from a TiC-deposited graphite by bombardment with 1 keV hydrogen ion beam of various current densities at 900°C have been investigated by means of the Rutherford backscattering (RBS) technique. It has been observed that the sputtering yields for Ti atoms at 900°C are almost zero below a critical ion flux of 1 × 10¹⁵/cm²·s, compared with the sputtering yield of Ti atoms at room temperature which has been measured to be 1 × 10⁻²atoms/ion. No sputtering of Ti atoms observed at 900°C is explained in terms of self-sustaining coating of the TiC surface with segregated carbon layer. The condition for the self-sustaining coating is discussed.     

An ion optical decel system for oblique angular impact of low-energy ions on targets

A beam retardation system is presented for performing sputtering investigations with oblique angles of incidence of the ion beam. The system is designed for ion energies from 90 eV to 3 keV. For oblique ion impact it is necessary to have a field-free space in front of the target to avoid deflection of the beam in front of the tilted target due to an otherwise distorted electric field. A double-lens system yielding a parallel ion beam can meet this requirement if the lenses are movable in the beam direction. Such a system is described for a two-dimensional field. The system was tested with a tungsten ion beam in a tungsten self-sputtering investigation at an energy of 350 eV. For this ion-target combination a large angular dependence of the sputtering yield is expected. The results are compared with Monte Carlo calculations.     

The effect of bulk carbon impurity on the radiation-induced surface composition modification of V-Cr alloys at elevated temperature

The effect of 1 to 3 keV argon ion sputtering on the surface composition of a V/Cr alloy containing 31.6 at% Cr and carbon as the dominant impurity has been investigated at 300 and 800 K. The results obtained closely parallel those reported previously for V/Cr alloys which contained bulk oxygen impurity. Sputtering at 300 K depletes the surface in Cr whereas at 800 K sputtering enhances the surface Cr concentration. Annealing at 800 K in the absence of radiation causes the surface to become enriched in both vanadium and carbon, which co-segregate. The mechanism for Cr enrichment is thus co-segregation of V and C, only possible at high temperature, and a consequent depletion in the surface in vanadium on sputtering. Sequential cycles of 800 K annealing without sputtering and 300 K sputtering, produce the same result, Cr surface enrichment which occurs during the 300 K sputtering segment of the cycle. This shows that impurity co-segregation with V and sputter-depletion of V rather than radiation induced segregation of Cr is the dominant mechanism of surface composition modification.     

Modeling of chemical erosion of graphite due to hydrogen by inclusion of chemical reactions in SDTrimSP

Erosion of carbon-based materials plays a very crucial role in determining its candidature as a potential plasma facing material in fusion devices. Carbon erosion yield shows strong dependence on energy, flux of the incoming ions and the target temperature. Various experiments and theoretical models have made successful attempts to understand some aspects of the erosion mechanism of carbon bombarded by hydrogen ions. They are valid for a particular parameter range. None of the models alone can explain the temperature, energy and flux dependence of erosion simultaneously. In the present work different aspects of erosion mechanism are parameterized. A new model which takes into account the basic merits of different theoretical models and the various parameterized quantities is embedded in the SDTrimSP [1], [2] program. This upgraded SDTrimSP enables us to calculate the physical and chemical sputtering in a single model for various parameter ranges. The results obtained by the upgraded SDTrimSP are in good agreement with experimental data.