Search

The basic physics and the chemical role of H in amorphisation and exfoliation of H-ion-implanted silicon were investigated by Raman scattering and transmission electron microscopy. We found that a continuous, thermally stable, 50-nm-thick amorphous layer formed upon implantation of 5-keV H ions at 150 K. Such amorphisation is contrary to earlier theoretical predictions but gives support to recent molecular dynamics calculations. After high temperature annealing, H-saturated mono- and di-vacancies became dominant in the Si-H Raman spectrum. (001) platelets and microcracks apppeared in the c-Si beyond the a-Si layer, leading to the exfoliation of a 90-nm thick layer. Deuterium implantation leads to the formation of a thicker (∼100 nm) amorphous layer but inhibits exfoliation. (© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

Despite decades of study, voids in silicon produced by implantation of H or He followed by annealing continue to be a topic of interest. There are two key applications: gettering of heavy metal impurities, and “ion cutting” used in silicon-on-insulator fabrication. Positron annihilation is one of the few techniques that can probe the vacancies and vacancy clusters that are the precursors to void formation. Data from recent studies will be discussed, including (I) isotopic substitution, in which comparisons of H vs. D implantation permit examination of the impact of primary point defects vs. chemical effects. Remarkable differences exist between H and D in blistering of silicon - ion doses 2–3 times higher are required for blistering with D than with H, despite a higher rate of primary defect production for D; (II) the effect of annealing temperature ramp-rate, in which we show that ramp-rate has a significant impact on residual defects, despite which it is so disregarded as to often be omitted from published reports; and (III) comparisons with electron microscopy which suggest that positron annihilation can be insensitive to large voids. In these studies, positron annihilation augments data from techniques including ion channelling, Raman scattering and electron microscopy; the suite of techniques allows elucidation of the interplay between implanted impurities and the vacancies and interstitials created by implantation.

Germanium was implanted with 5 keV H and D ions at −120 °C or room temperature and thermally annealed in several steps. The samples were analysed at various stages by atomic force microscopy, ion channeling and Raman spectroscopy of Ge-H/D local vibration modes. The results are discussed in comparison with those in the well studied silicon. In general, the evolution of the different types of defects, in germanium at a given temperature, tends to be similar to that of the corresponding defects in silicon at 100–300 °C higher temperature. However, the behaviour of the defects detected by ion channeling (interstitials, lattice distortions) often appears unrelated to the chemical evolution measured by Raman scattering and to the temperature and isotope dependence of blistering.

The giant isotope effect, observed when D ions were substituted for H ions in low energy blistering of silicon, challenged common assumptions based on ion irradiation theory and H–Si chemistry. We report a systematic investigation of the effect in several materials, in an attempt to identify its origin. Samples of Si, Ge, GaAs, 6H–SiC and SrTiO3 (STO) were implanted at 5 keV with various H or D doses (1 × 1016 to 2 × 1017 cm− 2) and rapid thermal annealed. The resulting surface morphology was examined by atomic force microscopy. On Si and GaAs, D-ion blistering requires 2 to 3 times higher doses than H-ion blistering. On SiC, no D-ion blistering whatsoever is observed. On the other hand, on Ge and STO, there is little difference between H and D-induced blistering. For Si, Raman spectroscopy of Si–H/D bonding in conjunction with lattice kinetic Monte-Carlo calculations led to an interpretation of the effect in terms of reactions between the mobile entities, i.e. H (or D) atoms, vacancies and interstitials. The new data will be discussed by reference to this scenario.

When silicon is implanted with a sufficient concentration of H ions, at low to moderate temperature, and subsequently annealed at high temperature, dome-shaped gas-filled blisters and/or craters of exploded blisters appear on the surface. Under particular conditions, blistering can be produced by plasma hydrogenation as well. The phenomenon is another facet of hydrogen behaviour in silicon, a question with both fundamental and applied implications. Blistering is at the origin of the "ion-cutting" process for the fabrication of silicon-on-insulator and other heterostructures; this process is particularly useful whenever atomically sharp interfaces between layers are required. The novelty and vast potential of this process has spurred since the mid-1990's a burst of experimental activity on blistering. The purposes of those works were either to improve or extend the ion-cut process, or to clarify its underlying mechanisms. In "mechanisms", the plural is used to convey the fact that it is a multi-step phenomenon. Because of this complexity, the theoretical work, in comparison, is far less abundant. Hydrogen blistering of silicon is qualitatively understood in broad terms: H being insoluble in Si, it tends to segregate into cavities which grow and coalesce at high temperature, and the H 2 pressure in the cavities finally deforms the surface. In fact, our understanding of the microscopic mechanisms has progressed much beyond that level thanks to the sophisticated work that has been carried out using techniques such as transmission electron microscopy, Ruther-ford backscattering in the channelling mode, infrared spectroscopy of local vibrational modes, stress and strain measurements, and others. The effects of n- or p-doping, He ion coimplantation, and isotope substitution have also greatly helped in discriminating between different hypotheses. After a review of the most relevant experimental facts, the blistering mechanisms that have been proposed in the literature will be discussed and their conformity with the data assessed. Finally an attempt will be made to identify the key questions and suggest a few avenues for future work.

We studied the thermal evolution of H/D-stabilized defects aiming at understanding the microscopic mechanisms leading to crystalline silicon surface blistering and exfoliation. The critical hydrogen-defect complexes involved in this phenomenon have been identified by Raman spectroscopy analysis. A plausible mechanism for the ion-cut process emerging from our analyses is proposed. Surprising but instructive effects observed under isotope substitution are also discussed.

We have investigated the patterning of silicon surfaces using ion blistering in conjunction with e-beam lithography. Variable width (150–5000 nm) trenches were first written in 500 nm thick PMMA resist spin coated on silicon, using an electron beam. Next, 10 keV H2+ ions were implanted to various fluences through the masks. The resist was then removed and the samples were rapidly thermally annealed at 900 °C. The resulting surface morphologies were investigated by atomic force microscopy. In the wider trenches, round blisters with 600–900 nm diameter are observed, which are similar to those observed on unmasked surfaces. In submicron trenches, there is a transition in morphology, caused by the proximity to the border. The blisters are smaller and they are densely aligned along the trench direction ('string of pearls' pattern). Unusual blister geometries are observed in the narrowest trenches (150 nm) at higher H doses (≥1 × 1017 H cm−2)—such as tubular blisters aligned along the trench. It was also found that for H doses of ≥6 × 1016 H cm−2 the surface swells uniformly, which has implications for the blistering mechanism. The prospects for accomplishing ion cutting, layer transfer and bonding of finely delineated patterns of silicon onto another material are discussed in the light of the above results.

Radioactive phosphorus (32P) can be implanted in biomedical devices using plasma-based ion implantation (PBII). A pulsed ECR (2.45 GHz) coaxial argon plasma reactor has been developed in which the radioisotopes are implanted into a negatively biased target. A 1 ms microwave pulse creates the plasma, which sputters material from a sputter target. A fraction of the radioisotopes is then ionized and the ions are implanted into a negatively biased cylindrical titanium implant. The theory underlying plasma-based radioisotope ion implantation is described in this article and the calculations are compared to experimental results for different plasma conditions and sputtering target positions. This study confirms the importance of radioactive source positioning and the impact of high plasma densities when using a PBII reactor. Implantation efficiency rates of about 1% are measured, which is about 100 times higher than conventional beam-line ion implantation.

A pulsed, co-axial electron cyclotron resonance (ECR, 2.45?GHz) plasma reactor was designed and tested to demonstrate the feasibility of plasma-based radioactive ion implantation (32P radioisotope). The geometry of the reactor was designed to produce an efficient implantation of cylindrical implants. Therefore, the reactor is cylindrical in shape, and is equipped with a cylindrical grid in a co-axial geometry. The plasma is created between the wall and the grid; the plasma surrounds the implant, allowing for a radial implantation. A 1?ms microwave pulse creates a plasma in argon, which sputters material from a radioactive cathode. A fraction of the radioisotopes is then ionized, and the ions are implanted into negatively biased metal samples. The plasma was characterized by means of electrostatic probes, giving spatial evaluations of the electron temperature, plasma potential and electron density. Titanium samples were implanted with 32P during a study that aimed at optimizing the position of the radioactive sputter cathode in the plasma. From an analysis of the distribution of the radioactive fragments, we deduce that the plasma potential has a marked effect on the ion trajectories. In particular, it provides a more uniform implantation distribution than one would otherwise expect. For plasma densities ~8 ? 1010?cm?3, implantation efficiencies as high as 1% are measured; this is about 100 times higher than conventional beam-line ion implantation.

The implementation at the sub-100-nm scale of ion cleaving requires ion beams of ∼5 keV/amu or less. The blistering efficiency in 5-keV H-ion-implanted and annealed Si has been found to peak and vanish in a narrow range of ion fluence of (1.5–3.5)×1016 H/cm2. In order to understand this effect, the defect profiles in 5-keV H-irradiated Si were studied by Rutherford backscattering/channelling, while the Si-H bonding configurations during annealing were investigated by Raman scattering spectroscopy. Three types of defects play major roles: the ‘broad-band’ monohydride multivacancy complexes, the fully or partially passivated monovacancy VHn, and H-terminated internal surfaces Si(100):H. Blister absence at high fluence is characterised by the persistence up to 550 °C of the Si(100):H structures, which are blister embryos that failed to coalesce and grow. Radiation-induced stresses and fracture toughening may play roles in inhibiting cleavage at high fluence; however, widening towards the surface of the zone of high H and defect concentration is the likely major factor.

Microwave-based oxygen plasma source ion implantation (OPSII) is an emerging technique to harden the surface of aluminum, increase its wear resistance and reduce its friction coefficient. In earlier work, evidence of radiation-enhanced diffusion (RED) was found. It implied that thicker treated layers could be obtained at lower voltage. In order to understand the RED phenomenon and take full advantage of it, 22-keV O2/Ar co-implantations were carried out in pure Al at different substrate temperatures and ion doses. The oxygen depth profiles were determined by X-ray photoelectron spectroscopy (XPS), the surface hardness by depth-sensing nanoindentation and the scratch resistance by atomic force microscopy (AFM). Optimisation of the tribological properties in terms of oxygen profiles and temperature was carried out. Enhancement of oxygen penetration by RED peaks around 225 °C, in qualitative agreement with theory, and it gives the highest tribological properties: the hardness reaches ∼4 GPa, twice that obtained without Ar and four times that of untreated Al. The scratch resistance is similarly improved.

The properties of a variety of plasma produced fluorocarbon (CFx) thin films are described. The project objectives were: (i) to explore the widest range of process parameters and properties, with an emphasis on comparative characterizations of wetting and mechanical properties. (ii) To develop effective interfacial bonding layers for films of 1 μm and more. (iii) To utilize safe, non-toxic, environmentally acceptable and relatively inexpensive precursor gases. Wetting and surface energy were studied by the dynamic advancing and receding contact angle method. These properties are interpreted in terms of the surface chemistry revealed by X-ray photoelectron spectroscopy and the morphology characterized by optical and atomic force microscopy. The mechanical properties were determined by micro- and nano-indentation and scratch testing. Finally, simple tests of resistance to pollution, handling and aging were carried out. The process–structure–properties relationships are discussed. Process parameters giving films displaying attractive combinations of wetting and mechanical properties are identified.

Selected aspects of ion implantation at low energy are reviewed. Some are common to any ion implantation experiment or process, whereas others are specific to plasma based ion implantation (PBII). First, the physics of ion–solid interactions are outlined in the perspective of when to expect characteristic low-energy effects. Next, two examples are given, where low-energy implantation effects indeed deviate significantly from extrapolations of the high energy data. These are ion channelling and ion blistering on differently oriented silicon crystals. Following this, an attempt is made at identifying aspects of low-energy implantation that are relevant for PBII, particularly the ion energy and impact angle distributions. Finally, suggestions for further work are proposed, especially in connection with the fabrication of nano-scale lithographically patterned devices.

The damage production rate is an important factor in the diffusion of different species during implantation, for instance O-ion PBII of Al and its alloys. In this experiment, we try to distinguish the effect of purely radiation-enhanced diffusion (RED) from the synergistic effect between RED and Mg segregation driven by oxidation observed in earlier work with Mg-containing Al alloys. We investigated the influence of the irradiation damage on pure Al for which no segregation effect can occur. We used different gas compositions [O 2 , Ar, Ar(10%O 2 ) and Ne(10%O 2 )] to provide simultaneous species implantation and a variable ratio of damage to active ion implantation rate. The effect of the damage rate on the O implanted profiles was measured by XPS. Following implantation with an Ar(10%O 2 ) plasma, a very significant inward displacement of the peak concentration was observed. Analysis of the profiles leads to the conclusion that a purely RED effect is involved in this O migration.

Hydrophobic, adherent, and strong coatings would be desirable for the glazed porcelain insulators and aluminum conductors of electric power grids. Micron-thick films consisting of fluorinated carbon (CFx) on top of adherence layers (a-C:H on Si or Cr) were deposited on insulator pieces and aluminum, using a benign Ar/C2H2/C3F8 chemistry in a dual magnetron sputtering/plasma-enhanced chemical vapor deposition system. A wide range of process parameters was explored: pressure from 1 to 67 Pa, substrate r.f.-induced d.c. offset from –50 to –1000 V, and different gas ratios. Advancing and receding contact angle measurements, X-ray photoelectron spectroscopy, optical and atomic force microscopy, micro- and nano-indentation, and scratch testing were performed. Contact angles as large as 130° were obtained. The film F content was

The blistering of Si wafers with (100), (110), and (111) orientations, induced by implantation of low energy (5–8 keV) H and/or He ions, was studied. Our earlier work had used either pure H implants at high dose (5×1016 cm−2) or low dose coimplantation (1×1016 H cm−2+1×1016 He cm−2). Here we study pure He implantation, and coimplantation with variable He/H dose ratio, in order to understand the roles of H and He, in the hope of improving the blistering efficiency (as used, e.g., in the Smart-cut® process). After room temperature implantation, the samples were vacuum annealed in three steps at 500, 550, and 600 °C. At each step, the development of blistering and exfoliation was quantified by scanning electron microscopy (“exfoliation” is here used to mean blister explosion). We found that the thermal budget required for blistering is higher than for comparable doses of high energy (>30 keV) ions. On the Si(110) surface, H blistering is very weak, He blistering is the strongest, but exfoliation absent. On S...

Implantation of 30 keV oxygen ions in some magnesium containing aluminum alloys was performed using an ECR plasma source (“OPSII” process). The mechanical implantation effects were studied by AFM-nanoindentation, and the chemical effects by high resolution XPS and SIMS depth profiling. Surface hardening is obtained, consistent with the formation of very fine oxide precipitates. Since the temperature rises to ∼400°C during oxygen plasma source ion implantation (OPSII), Mg surface segregation is observed as expected. However, O and Mg are also found at the ∼10 at.% level as far as 0.5 μm below the surface, ten times deeper than the O-ion range. The possibility of an artefact has been eliminated by various tests. XPS indicates that MgO dominates while Al remains mostly metallic.

Oxygen implantation (5×10 17 O cm −2 , 30 keV) of the Al-7075 alloy (Al-Mg-Zn-Cu-Cr) was performed using an electron cyclotron resonance plasma source. The chemical implantation effects were studied by X-ray photoelectron spectroscopy and secondary ion mass spectrometry depth profiling, and the mechanical effects by atomic force microscopy. Strong Mg segregation (up to ∼100% MgO) was found. For purely thermal near-surface (∼15 nm) segregation, in O 2 gas, our measured activation energy of 68 kJ mol −1 is consistent with other results. Following O-ion irradiation, however, MgO is surprisingly observed 10× deeper (e.g. 10% MgO at 500 nm) than the O-ion range (∼50 nm). Al remains practically metallic. The nanohardness and Young's modulus are enhanced 2–4× to 1.5 GPa and ∼300 GPa, respectively. An increase in rms roughness from 4.5 to 47 nm, likely due to MgO precipitates, does not modify the nanoscale friction coefficient (0.026), dominated by adhesive forces.

Thin ( 0.1 μm ) layer splitting by H and He co-implantation at low energy was investigated. Si(1 0 0), (1 1 0) and (1 1 1) wafers were implanted with 1×1016 5- keV H + / cm 2 , followed by 1×1016 8- keV He + / cm 2 , and annealed in three steps at 500–600°C for 1 or 2 h. Surface blistering was studied by scanning electron microscopy. The threshold for blistering is 500°C. On Si(1 0 0) and Si(1 1 1), the exfoliated area (as % of total) reaches 56–58%, but on Si(1 1 0) it does not exceed ∼30%. At 30 keV, in comparison, exfoliation only requires a 20 min 450°C anneal. This low-energy limitation is due in part to widening of the implant profile and lowering of the peak density by channeling, as evidenced by ERD-TOF measurements of the profiles. The possible role of the damage profile is also discussed.

Numerous collector samples have been exposed in the TdeV tokamak, either to plasma assisted chemical vapor boronization only, or to boronization plus tokamak power discharges. They have been analyzed by X-ray photoelectron spectroscopy (XPS) in order to characterize and better understand the process by which the boron-rich film getter ambient oxygen. It was found that the state of oxidation of the samples after boronization is the most reliable predictor of subsequent machine performance. The gettering capacity is high, on the order of 1021 O/m2, and affects a surprisingly thick layer (∼100 nm), but the oxide always remains substoichiometric (with a formula BOx, x

Pure aluminum was implanted with 30 keV oxygen ions by means of electron-cyclotron resonance plasma based ion implantation. Virgin and treated samples were investigated by Auger electron microscopy and x-ray photoelectron spectroscopy to determine the depth distribution and chemical interactions of oxygen near the surface. The results are consistent with precipitation of oxygen and the formation of particles of Al2O3. Several samples were subsequently annealed at temperatures of up to 550 °C for 30 min. As-implanted and annealed specimens were characterized using an ultralow load nanoindentation test system. The results indicated an increase in aluminum hardness by a factor of 2 to 3; annealed samples were found to have lower hardness values than the as-implanted samples. A decrease of the experimental hardness with depth correlates with the concentration of implanted oxygen. This is consistent with the assumption that the improvement in strength is due to the formation of nanoprecipitates whose density i...

The depth profiles of H and D implanted at low energy (0.7 and 1.5 keV) in Highly Oriented Pyrolytic Graphite (HOPG) ( a -axis) and vitreous carbon (v-C), and then partially released by pulsed laser induced desorption were investigated in order to understand the H behaviour in carbon at high temperature. During the laser desorption experiment, the re-emitted gas quantity was measured as a function of laser energy density by mass spectrometry. Depth profiles of H and D remaining after partial desorption were also obtained by means of the elastic recoil detection (ERD)- E × B technique. Firstly, a significant difference between the re-emitted H 2 or D 2 gas quantity and the released amount deduced from the ERD profiles is observed. This could be explained by atomic H release or by far in-depth diffusion of molecules. Secondly, it is seen from the depth profiles of hydrogen in HOPG that no profile modification is observed below 1500 K. The desorption process is compatible with a local molecular recombination model with fast diffusion of the molecules. At higher temperatures, significant diffusion of atomic hydrogen allows it to be retrapped in the vacancy profile created during the implantation.

We present measurements of the range profiles of 0.5 and 1 keV H and D ions implanted in amorphous Si (a-Si) and crystalline Si (c-Si) in random, 〈1 0 0〉, 〈1 1 0〉 and 〈1 1 1〉 directions. The profiles were obtained by the E × B filtered Elastic Recoil Detection (ERD) technique with a resolution of ∼10 nm. This is the first such complete set of low-energy data to our knowledge. The whole set of profiles could not be simulated self-consistently with standard versions of existing codes. After critically reviewing the codes, we conclude this is due to the neglect of the anisotropy of the valence electron density in covalent solids. The localization of these electrons in the walls of the channels causes a reduction in screening and prompter dechanneling in the narrower channels 〈1 0 0〉 and 〈1 1 1〉. Moreover, it reduces the inelastic losses near the channel axes. We implemented an appropriate adjustment in the binary collision code MARLOWE (M.T. Robinson, Phys. Rev. B 40 (1989) 10717). We then successfully fitted all the data using only three parameters. The first of these is essentially determined by the random profiles and indicates that the inelastic energy loss is ∼1.5 × larger than the well-known Lindhard–Scharff–Schiott (LSS) value; this agrees with high energy data. An Oen–Robinson (OR) type local energy loss is also required. The second parameter, introduced by us, serves to adjust the screening function in the interatomic potential to account for the anisotropy of the electron density around the Si atoms. The third parameter is connected with simultaneous interactions of the ion with several atoms due to the crystal symmetry; in particular, focussing of the ion by trios of atoms in the 〈1 1 0〉 channel is evidenced.

The conditions required for blistering after annealing in n-type GaAs(1 0 0) were investigated using low energy H, D or He ion implantations at room temperature for different fluences from 1 × 1016 to 1 × 1017 ion/cm2. The threshold fluence for H+ at 5 keV is ∼3 × 1016 H+/cm2 and blistering is still present for fluences as high as 1 × 1017 H+/cm2, in contrast with silicon where blisters due to low energy H+ implantation appear at comparable fluence (2 × 1016 H+/cm2) but disappear at slightly higher fluence (5 × 1016 H+/cm2). The critical annealing temperature in GaAs was found to be 175 ± 25 °C, which is much lower than in silicon (350 °C). Blistering is also obtained with He at low fluence, contrary to Si which is impervious to He blistering in this fluence range. We have discovered that the exfoliated crater depths depend strongly on the He fluence, varying from 75 ± 12 nm (for 1.6 × 1016 He+/cm2), consistent with the most probable ion projected range determined by simulation, to a saturation value of 155 ± 10 nm for doses >4 × 1016 He+/cm2. Our results suggest that exfoliation is triggered at a local He concentration of 1.5–3.0 at.% inside GaAs.

An infrared (IR) camera has been installed on TdeV (Tokamak de Varennes) to view both the inner and the outer divertor plates. With a finite element code (PDE Protran), the energy deposited on the plates is calculated from the surface temperature profile measured by the IR camera. The experimental results have revealed that the divertor plate temperature decreases with an increase of the main plasma density, suggesting that the plasma detaches from the divertor plates at higher plasma densities (above 4.0×1019 m−3). During divertor biasing, the heat flux increases on the plates of the `active' divertor as a result of the E×B flow. However, energy deposition can be reduced by increasing the plasma density. The neutral gas pressure in the divertor region increases strongly at higher plasma density, and appears to be correlated with the reduction in deposited energy.

Strong toroidal asymmetries in power deposition during lower hybrid current drive experiments have been observed on TdeV and Tore Supra. These asymmetries are characterized by high heat loads on components magnetically connected to the LHCD antenna. The experimental results indicate that fast electrons generated by the rf fields in the scrape-off layer in front of the antenna are responsible for the localized heat loads. The rf power lost in the edge though this channel increases with both the edge density and the rf power level. In TdeV, the fraction of the injected power lost can exceed 10% at high power and high density, while it stays under 2% in Tore Supra. A simple model describing the trajectories of electrons in front of the antenna suggests that the high N| components of the near field spectrum are responsible for the acceleration of the scrape-off layer electrons.

The main aspects of radiative plasma experiments carried out in the LH-heated TdeV tokamak are summarized. D 2 , N 2 and Ne puffing, and divertor biasing, are assessed in terms of their ability to increase edge and divertor radiative losses and reduce neutralizer plate heat loads. Detailed patterns of radiation and plate power deposition are compared. Combined N 2 and D 2 puffing gives the best results. Biasing enhances divertor radiation but also causes an increase in power conducted and convected to one of the divertor legs. The improvements obtained are still insufficient due to disruptive instabilities that appear at high impurity puffing rates. The instability is triggered at the edge with N 2 and in the core with Ne.

We have studied the retention of a recycling impurity, neon, and a wall-pumped impurity, nitrogen, in the closed and cryopumped divertor of the TdeV tokamak, under a variety of heating, puffing, pumping and biasing conditions. Retention times were deduced from either the decay time or the rise time of the plasma impurity content, measured spectroscopically, following either a short puff or a steady injection. For both neon and nitrogen, the compression ratio increases rapidly with the main plasma density. Retention is not degraded by plasma detachment. In density scans, other conditions being kept constant, the compression ratio for neon is found to grow with the divertor neutral pressure as ∼ P D 1.5 . However, by puffing, pumping or biasing, retention and P D can be varied quite independently of each other.

The authors present a comparison of the wall deuterium retention and plasma fueling requirements of three diverted tokamaks, DIII-D, TdeV, and ASDEX-Upgrade, with different fractions of graphite coverage of stainless steel or Inconel outer walls and different heating modes. Data from particle balance experiments on each tokamak demonstrate well-defined differences in wall retention of deuterium gas, even though all three tokamaks have complete graphite coverage of divertor components and all three are routinely boronized. This paper compares the evolution of the change in wall loading and net fueling efficiency for gas during dedicated experiments without Helium Glow Discharge Cleaning on the DIII-D and TdeV tokamaks. On the DIII-D tokamak, it was demonstrated that the wall loading could be increased by > 1,250 Torr-1 (equivalent to 150 {times} plasma particle content) plasma inventories resulting in an increase in fueling efficiency from 0.08 to 0.25, whereas the wall loading on the TdeV tokamak could only be increased by < 35 Torr-{ell} (equivalent to 50{times} plasma particle content) plasma inventories at a maximum fueling efficiency {approximately} 1. Data from the ASDEX-Upgrade tokamak suggests qualitative behavior of wall retention and fueling efficiency similar to DIII-D.

At high densities, the plasma detaches from the outboard divertor plates in TdeV. The signatures are a reduction of the ion flux to the divertor plate, movement of the radiating zone from the plate toward the X-point, a pressure gradient between an ionization front and the target plate, and strong cross-field transport in the divertor. A toroidally-viewing TV imaging system allows us to observe local interactions between the divertor plasma and the different divertor plates. As the plasma detaches, the gas pressure in the divertor continues to rise, and there is evidence for molecular processes in the cold plasma near the divertor plates. Auxiliary heating increases the power and particle flow across the separatrix; our results suggest that detachment depends on the energy transported per particle. Simulations using the B2/EIRENE and DIVIMP codes give reasonable agreement with the measurements for the attached phase.

Experimental data, analysis and simulations are used to describe the physics of divertor detachment and He exhaust under detached conditions in TdeV. With increasing density, the plasma is found to detach progressively from the outboard divertor plates with a marked reduction of the ion flux to the plates, the generation of a pressure gradient between an ionization front and the target plate, and strong cross-field transport in the divertor. Local interactions between the divertor plasma and the plates are described, with evidence for carbon sputtering and molecular processes near the divertor plates. Divertor exhaust and retention continue to increase through detachment and He exhaust is not affected although the divertor He enrichment remains low but constant at about 0.2. A moderate density of seems to be sufficient both for efficient peak power load reduction at the divertor plate and good He exhaust through the divertor. Simulations of the edge and divertor plasmas using the B2/EIRENE and DIVIMP codes give reasonable agreement with the measurements and indicate possible divertor geometry improvements

Laser desorption of deuterium implanted in hot‐pressed silicon carbide samples was performed. The samples had been implanted at ion energies ranging from 1 keV/D to 3 keV/D and at fluences from 0.6×1020 m−2 to 1.8×1021 m−2. For the high fluence samples (≥3×1020 m−2), comparison of the experimental data of desorption rate versus laser energy with the results of computer simulations reveals that second order detrapping is the limiting re‐emission process. The effective trap activation energy decreases from 0.9±0.1 eV/molecule at 1 keV/D to 0.5±0.1 eV/molecule at 3 keV/D. Blister formation and exfoliation or other forms of radiation damage could explain the energy dependence in the high fluence samples. At low fluence (

Laser desorption experiments were performed on deuterium implanted carbon. Ion energies ranged from 0.5 up to 6.0 keV and fluences from 0.6 × 1020 to 16 × 1020 m−2. The desorption results were compared to computer simulations using the DTRLAS code. Analysis indicated that the desorption process implies a second order detrapping and a diffusion limited transport. The best fitting parameters give a detrapping rate vV of (4 ± 1) × 10−21 exp(−(0.36 ± 0.02) eV/molecule/kT) m3/s and an effective diffusion coefficient D = (2.0 ± 0.5) × 10−7 exp(−(0.20 ± 0.02) eV/kT) m2/s. The amount of H2 and HD released simultaneously with D2 as well as the deuterated hydrocarbons were also measured. Indirect evidence for atomic D release at high temperature was found.

H-isotope trapping and release by plasma facing components (PFCs) control fuel retention and recycling in fusion devices, and melting can occur when PFCs are subjected to high heat loads. Pulsed laser melting and desorption were used to investigate the transport of deuterium (D) pre-implanted in beryllium (Be). A numerical model with a finite difference form was used to compute the space and time dependence of the temperature, the phase changes, and the D 2 evolution. The activation energies and kinetic factors characterizing detrapping, diffusion and molecular recombination are free parameters of the code which have been determined by fitting the data on desorbed amount versus laser energy. The D diffusivity in liquid Be is found to be: D 1 = (9.0 ± 0.7) × 10 −8 exp[(−0.24 ± 0.02)eV/ kT ] m 2 /s. Migration of near-surface deuterium or tritium into the bulk of molten Be is implied by these results.

We have used the codes MARLOWE, CRYSTAL-TRIM, TRIM-SP and TRIM-95, to simulate our experimental data on the penetration of 1-keV hydrogen ions in a-Si and in c-Si in “random” and channeled directions. The dominant energy loss mechanism is inelastic for most of the path length of the penetrating ion. The effect of using an impact parameter dependent inelastic energy loss using the Oen-Robinson (OR) model or a drag force as in the LSS and ENR theories was investigated. In the case of c-Si, the data can be reproduced with the OR model using a single set of physical parameters, whereas in the case of a-Si only partial agreement was found. Even though elastic energy loss is not dominant, simultaneous collisions have to be taken into account. Also, the effect of the computationally required maximum impact parameter (imposed by the binary collision approximation) was examined.

We have found that blisters due to low-keV H-ion implantation and annealing of Si appear at low fluence (e.g., 2×1016 H cm−2) but disappear at slightly higher fluence (3.5×1016 H cm−2); this fluence “window” widens at higher ion energy. For D-ion blistering the window is shifted upwards by a surprising factor of 2–3. Thermal desorption spectrometry suggests that D is somehow more stable than H in Si. Hypotheses to explain blistering disappearance and the giant isotope effect are discussed. This phenomenon has an impact for the ion-cut process at the sub 100 nm scale.

Effective control of the particle and power flux distributions between the upper and lower divertors has been demonstrated on TdeV, using biasing of the divertor plates, in ohmic discharges with a double null magnetic configuration. Negative biasing of the divertor plates with respect to the walls significantly increases the pressure and the emitted power from the plasma in the active divertor into which E*B drift is directed, but it has little influence on the opposite divertor. This enhanced power dissipation is the result of improved particle retention in the active divertor. The incident power on the plates in the active divertor also increases with biasing, but at a slower rate. The Franck-Condon and charge exchange neutrals contribute significantly to the power loss in the divertors of TdeV in the present operating mode. Positive biasing of the divertor plates has little influence on the total losses from the divertor regions

A new instrument (“Variable Geometry Sniffer Probe” or VGSP), allowing one to sample and mass analyse atoms, ions or molecules moving in different directions and at different locations at the periphery of a plasma, has been built and used in plasma edge studies in the TdeV tokamak. Three different regimes of particle sampling have been identified. First, the VGSP can measure the fluxes of hydrogen and impurity molecules issuing from the walls. Second, it has the capability of detecting low energy charge-exchange and Franck-Condon neutrals. Finally, there is a parallel ion flux sampling regime, for which it is shown that both the connection lengths to the divertor plates and the X × B plasma flows induced by edge electric fields play major roles.

Measurements in TdeV in the double null configuration show strong asymmetries between the upper and lower divertors. The divertor to which the ion ∇ B drift is directed has a higher density, which varies little with the central line-average density. The electron temperature also exhibits a poloidal variation which depends on the direction of the toroidal magnetic field. The divertor plasma shows a double peak structure which evolves as a function of the central density, with the outer peak varying more rapidly. We have measured flow reversal near the separatrix in the bottom divertor when the ion ∇ B drift is directed away from the X -point. Divertor plate biasing allows us to control the flux of plasma in the SOL; under negative biasing there is a strong increase of the electron density in the active divertor, the one favoured by the E × B flow. Calculations of the energy deposition derived by flush-mounted probes agree well with measurements of the heat load derived from the temperature increase of the divertor tiles.

The effect of lithium wall conditioning in a boronized tokamak was investigated in ohmic discharges in the TdeV. In the course of 14 discharges, lithium was spread on the walls to a coverage of about 10 21 atoms/m 2 by inserting a lithium-containing graphite crucible in the scrape-off-layer. Lithium caused an increase of some tens of percent in the wall pumping capacity during discharges, as seen from gas puffing and H α + D α measurements. Post-shot wall outgassing also increased such that the short term in-vessel retention was only enhanced by about 20%. This increased retention level was confirmed by subsequent analysis of collector probe samples positioned in the SOL near the wall and was lower than one H or D atom per Li atom. Evidence of oxygen gettering by lithium was also found.

We present the salient features and interpretations from a multi-technique study of the high temperature H isotope behaviour in Be, including liquid Be. The effects of the implantation of H or D ions in Be at 300 to 983 K as well as the effects of in situ annealing in the transmission electron microscope and annealing with a 30 ns laser pulse were investigated. Quantitative measurements of the gas release, ion-beam depth profiling, scanning electron and atomic force microscopy, and numerical modeling were used in the analysis. For H Be , the H trapping energy is 1.7 eV and diffusion out of the sample is effectively 10 times faster than anticipated from other work. The release is not connected with blister rupture. For H Be >- 14% , the trapping energy is only 1.0–1.2 eV, but explosive release can take place under some circumstances. The implications for Be plasma facing components in tokamaks are discussed.

Preliminary results concerning the influence of negative biasing of the divertor plates on particle recycling and on power loss distribution in single null discharges of TdeV during lower hybrid (LH) current drive and heating experiments are presented. The beneficial effects of negative biasing of the divertor plates, such as the ability to control power and particle fluxes in the SOL, remain effective in the presence of auxiliary heating and current drive. Up to 0.7 MW of auxiliary power were injected in these experiments. With a negative biasing of 150 V, and the E ⊗ b flow vector pointing towards the outer divertor chamber, a roughly 2 fold increase in the divertor pressure and the radiation from the divertor region is observed.

In this letter we report a giant isotope effect in silicon blistering induced by low-energy (5 keV) H or D ion implantation and annealing. Atomic force microscopy and thermal desorption spectroscopy revealed abundant blistering and exfoliation with a D fluence of 6×1016 D cm−2, whereas no blistering was observed with a H fluence up to 1×1017 H cm−2. This is accompanied by premature H2 release, and the evidence suggests that inefficient defect trapping is the cause. The blister density and the degree of exfoliation also depend strongly on the crystal orientation.

Segregation of surprising magnitude and depth was discovered in oxygen plasma-implanted Mg-containing Al alloys. In order to identify the respective roles of diffusional kinetics and oxidation energetics in the effect, pure Al and an Al–Mg–Zn–Cu–Cr alloy were implanted for various times, or oxidized in air, at 700 K. Examination of depth profiles and chemical states of O, Al, and Mg by x-ray photoelectron spectroscopy reveals a potent synergy between the chemical potential drive and the radiation-enhanced diffusion.

The isotope ratio RI=D/(H+D) was measured during ohmic discharges in the magnetically diverted tokamak TdeV, using spectroscopy (Halpha and Dalpha ), gas analysis (H2, HD and D2), mass analysis of charge exchange neutrals, and detection of H and D in a collector probe, in the course of four different experimental phases: (I) 'dynamical' isotope switchover of the fuelling gas during a discharge, (II) shot-to-shot evolution with pure D2 fuelling, (III) same with walls conditioned by D2 glow discharge, and (IV) 'dynamical' switchover combined with divertor biasing and pumping. The different values obtained for RI are compared and analysed in terms of the particle recycling properties of the device. The dynamical scenario (phase I) revealed a lack of equilibrium between the edge and core isotopic composition that is an indication of the degree to which the particle flux is amplified by recycling in the vicinity of surfaces. The phase II results were fitted to a model which yielded an estimate of the magnitude of the wall inventory available for recycling. In phase III, the effectiveness of conditioning was assessed. Phase IV revealed an even more pronounced disequilibrium than phase I, and demonstrated the power of divertor biasing and pumping in exhausting the wall inventory and controlling the isotope ratio.

Pulsed laser desorption was used to investigate the reemission of deuterium implanted in beryllium. The amount of D2 molecules released as a function of the laser energy density was measured on samples with different implantation energies and doses. Numerical modelling of our results indicates that detrapping and diffusion are the limiting processes. The data are best fitted with second order detrapping with a trap energy of 1.7 ± 0.1 eV and the effective diffusion coefficient is described by the equation D = (8 ± 1) × 10−8 exp(− 0.32 ± 0.02 eV/kT) m2 s−1.

Experiments in which the divertor plates are biased with respect to the grounded vacuum chamber were conducted in the Tokamak de Varennes, with ohmic discharges and boronized walls. A comprehensive set of diagnostics was used and those effects relating to plasma-surface interactions and their consequences, namely recycling and impurity production, are reported here. The main conclusion is that the bias voltage actively controls the balance between the particle fluxes to the walls and the diverters, with a weak influence on the confinement. The quantities associated with plasma surface interactions, i.e. the impurity influx, the plasma contamination and the global recycling coefficient, increase as a function of the voltage applied between the plates and the wall. Typically, the quantities of interest vary by a factor of two or more in the range from -240 to 290 V. The poloidal fluxes to the upper or lower divertor are enhanced by negative or positive biasing, respectively, in agreement with the sign of the E*B drift. For negative biasing the reduction in the wall flux and increase in the poloidal flux combine to improve greatly the divertor efficiency: the divertor pressure and the impurity retention time are increased by factors of up to 5 and 8, respectively. The above effects, together with the SOL density profiles and the biasing current-voltage characteristic, are consistent with a model in which the radial transport in the SOL is limited by the ion mobility; a mobility of 0.04 m2/V.s is found. Besides allowing active density and impurity control, it is suggested that biasing could facilitate the attainment of a radiative divertor by injection of well retained impurities in the divertor

Simple extrapolation of the Lindhard analytic theory of channeling predicts very wide channeling angles (∼ 10°) at low keV energies; however, the effect is highly sensitive to the exact atomic interaction potential chosen, as shown by computer simulations. Amorphous, polycrystalline, and single-crystalline 〈110〉, 〈111〉 and 〈100〉 silicon samples have been implanted with 1 keV deuterons in channeled and random directions. The ranges are measured by the ERD biE × biB (elastic recoil detection) ion beam technique and compared to simulations with the codes MARLOWE, TRIM and BABOUM. Some significant conclusions are that the channeling angle is approximately 10° in the 〈110〉 direction, and that polycrystalline material gives significant “accidental” channeling.

Electrically insulated divertor plates are used on TdeV (Tokamak de Varennes) [18th EPS Conference on Controlled Fusion and Plasma Physics Berlin (European Physical Society, Petit‐Lancy, 1991), Vol. 15C, Part I, pp. 1–141] to produce various biasing configurations, which can be decomposed into two basic modes. Plasma biasing, with a radial electric field Er in the scrape‐off layer (SOL), is most promising for divertor applications. The Er field is produced with a particular divertor plate geometry, causing a nonambipolar radial current and a particle flow in the Er×BT direction, toward one of the divertors (the active divertor). The pressure and impurity retention in the active divertor are shown, in the Ohmic regime, to be strongly increased by biasing. He exhaust through this divertor is increased by a factor of almost 3 with modest biasing voltages and currents scalable to larger devices. Biasing also modifies the power repartition between the divertors, with the active divertor also receiving a larger...