[en] The understanding of turbulence in the edge region of toroidal magnetised plasmas is one of the key issues in modern fusion research. Great efforts are spend, experimentally and theoretically, to quantify and understand particle, energy and momentum transport. Blob-like transport was observed first by Zweben in 1985 during studies of edge density turbulence in the Caltech research tokamak. Zweben described coherent magnetic field-aligned structures of higher or lower plasma density and called them consistently blobs and holes. These long-lived filaments are born in the edge shear layer, where zonal flows shear off meso-scale coherent structures. The occurrence of localized filamentary blobs and holes have recently found renewed interest with the advent of fast camera diagnostics. It is now well known that Edge Localized Modes (ELMs) lead to the ejection of a number of filamentary structures into the scrape off layer (SOL). ELMs thus generate structures with excess energy and density and it can be conjectured that they leave corresponding holes behind. In contrast to blobs/filaments, holes are usually quickly filled by the edge plasma along the magnetic field and therefore exhibit a restricted lifetime and consequently they are more difficult to observe. If such a hole, however, is able to reach a resonant surface with low rational q, due to its size or low shear, it closes on itself and increases its lifetime significantly. We believe that a signature of such an event can be found in the JET tokamak, called Palm Tree Mode (PTM). The PTM is a special MHD phenomenon, which was only detected in JET type-I ELMy H-mode plasmas as long as the rational q =3 surface is in the ELM perturbed region. Fast sampling magnetic pickup coils were used to study onset, decay, dynamics and the helical structure of the mode. Comparisons with charge exchange recombination spectroscopy show that the PTM is co-rotating with the edge plasma after recovery of ELM induced momentum losses. The decay rates of the current give insight in the closing phase and the filling mechanisms, which determine the lifetime of the PTM. Based on relative signal strengths and decay rates a new method has been developed to locate the mode. Fluctuations of PTMs can also be found in the electron cyclotron emission diagnostic. The effect of the mode on the diagnostic is twofold. One is the increase in temperature measured inside the filament. The other is the change in temperature perceived through the change in geometry by the magnetic perturbation of the PTM. These two contributions have been separated for the first time with the method of empirical mode decomposition. Forward modelling suggests that PTMs are closed unipolar current filaments with negative current. The size of the current hole is of the order of the edge current density. It is demonstrated that a rotating filament can produce the same characteristic spectra like observed in FFTs of ECE and magnetics data. The indications, especially found from forward modelling, support the hypothesis. The concept of blobs and holes therefore gives insight in the genesis and nature of the PTM. The replacement of a wave-like description of the PTM by a filament or quasiparticle picture appears to be possible and also contributes to explain other edge phenomena like outer modes or edge snakes. (author)

[en] High power impulse magnetron sputtering (HiPIMS) plasmas exhibit a high ionization fraction of the sputtered material and ions with high kinetic energies, which produce thin films with superior quality. These ion energy distribution functions (IEDF) contain energetic peaks, which are believed to be linked to a distinct electrical potential hump $\mathrm{\Delta <!--\; ??\; -->}{\mathrm{\Phi <!--\; ??\; -->}}_{\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{z}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{z}\mathrm{o}\mathrm{n}\mathrm{e}}$ inside rotating localized ionization zones, so called spokes, at target power densities above 1 kW cm⁻². Any direct measurement of this electrical potential structure is, however, very difficult due to the dynamic nature of the spokes and the very high local power density, which hampers the use of conventional emissive probes. Instead, we use a careful analysis of the IEDFs for singly and doubly charged titanium ions from a HiPIMS plasma at varying target power density. The energy peaks in the IEDFs measured at the substrate depend on the point of ionization and any charge exchange collisions on the path between ionization and impact at the substrate. Thereby, the IEDFs contain a convoluted information about the electrical potential structure inside the plasma. The analysis of these IEDFs reveal that higher ionization states originate at high target power densities from the central part of the plasma spoke, whereas singly charged ions originate from the perimeter of the plasma spoke. Consequently, we observe different absolute ion energies with the energy of Ti²⁺ being slightly higher than two times the energy of Ti⁺. Additional peaks are observed in the IEDFs of Ti⁺ originating from charge exchange reactions from Ti²⁺ and Ti³⁺ with titanium neutrals. Based on this analysis of the IEDFs, the structure of the electrical potential inside a spoke is inferred yielding $\mathrm{\Delta <!--\; ??\; -->}{\mathrm{\Phi <!--\; ??\; -->}}_{\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{z}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{z}\mathrm{o}\mathrm{n}\mathrm{e}}$ = 25 V above the plasma potential, irrespective of target power density. (paper)

[en] Reactive high power impulse magnetron sputtering (HiPIMS) of metals is of paramount importance for the deposition of various oxides, nitrides and carbides. The addition of a reactive gas such as nitrogen to an argon HiPIMS plasma with a metal target allows the formation of the corresponding metal nitride on the substrate. The addition of a reactive gas introduces new dynamics into the plasma process, such as hysteresis, target poisoning and the rarefaction of two different plasma gases. We investigate the dynamics for the deposition of chromium nitride by a reactive HiPIMS plasma using energy- and time-resolved ion mass spectrometry, fast camera measurements and temporal and spatially resolved optical emission spectroscopy. It is shown that the addition of nitrogen to the argon plasma gas significantly changes the appearance of the localized ionization zones, the so-called spokes, in HiPIMS plasmas. In addition, a very strong modulation of the metal ion flux within each HiPIMS pulse is observed, with the metal ion flux being strongly suppressed and the nitrogen molecular ion flux being strongly enhanced in the high current phase of the pulse. This behavior is explained by a stronger return effect of the sputtered metal ions in the dense plasma above the racetrack. This is best observed in a pure nitrogen plasma, because the ionization zones are mostly confined, implying a very high local plasma density and consequently also an efficient scattering process. (paper)

[en] High power impulse magnetron sputtering (HiPIMS) plasmas generate energetic metal ions at the substrate as a major difference to conventional direct current magnetron sputtering (dcMS). The origin of these very energetic ions in HiPIMS is still an open issue, which is unravelled using two fast diagnostics: time-resolved mass spectrometry with a temporal resolution of 2 µs and phase resolved optical emission spectroscopy with a temporal resolution of 1 µs. A power scan from dcMS-like to HiPIMS plasmas was performed, with a 2 inch magnetron and a titanium target as sputter source and argon as working gas. Clear differences in the transport as well as the energetic properties of Ar⁺, Ar²⁺, Ti⁺ and Ti²⁺ were observed. For discharges with highest peak power densities a high energetic group of Ti⁺ and Ti²⁺ could be identified with energies of approximately 25 eV and of 50 eV, respectively. A cold group of ions was always present. It is found that hot ions are observed only when the plasma enters the spokes regime, which can be monitored by oscillations in the IV characteristics in the MHz range that are picked up by the used VI probes. These oscillations are correlated with the spokes phenomenon and are explained as an amplification of the Hall current inside the spokes as hot ionization zones. To explain the presence of energetic ions, we propose a double layer (DL) confining the hot plasma inside a spoke: if an atom becomes ionized inside the spokes region it is accelerated because of the DL to higher energies whereas its energy remains unchanged if it is ionized outside. In applying this DL model to our measurements the observed phenomena as well as several measurements from other groups can be explained. Only if spokes and a DL are present can the confined particles gain enough energy to leave the magnetic trap. We conclude from our findings that the spoke phenomenon represents the essence of HiPIMS plasmas, explaining their good performance for material synthesis applications. (paper)

[en] Full text: Instabilities and turbulence, which occur in a tokamak, are of high interest in fusion physics and a detailed examination is essential for a successful operation of upcoming fusion experiments. Two dimensional turbulent mixing leads to particle and energy transport perpendicular to the magnetic field and especially the outer parts of a tokamak are subject to various turbulence events, such as Edge Localized Modes (ELMs). In order to reveal potential similarities or special features, both classical analysis methods and wavelet methods have been applied on ELM Type-I Langmuir probe measurements at ASDEX Upgrade. Moreover, simulation data obtained with a gyrofluid code have been investigated for comparison reasons. Wavelet techniques allow a deeper insight into the characteristics of fluctuation data by means of a time-scale representation. Consequently, they provide a reasonable addition to common statistical tools like probability density and correlation functions or Fourier spectrograms, particularly in the case of highly non stationary ELM data. (author)

[en] The rotation of localised ionisation zones, i.e. spokes, in magnetron discharge are frequently observed. The spokes are investigated by measuring floating potential oscillations with 12 flat probes placed azimuthally around a planar circular magnetron. The 12-probe setup provides sufficient temporal and spatial resolution to observe the properties of various spokes, such as rotation direction, mode number and angular velocity. The spokes are investigated as a function of discharge current, ranging from 10 mA (current density 0.5 mA cm⁻²) to 140 A (7 A cm⁻²). In the range from 10 mA to 600 mA the plasma was sustained in DC mode, and in the range from 1 A to 140 A the plasma was pulsed in high-power impulse magnetron sputtering mode. The presence of spokes throughout the complete discharge current range indicates that the spokes are an intrinsic property of a magnetron sputtering plasma discharge. The spokes may disappear at discharge currents above 80 A for Cr, as the plasma becomes homogeneously distributed over the racetrack. Up to discharge currents of several amperes (the exact value depends on the target material), the spokes rotate in a retrograde $\mathbf{E}\times <!--\; ??\; -->\mathbf{B}$ direction with angular velocity in the range of 0.2–4 km s⁻¹. Beyond a discharge current of several amperes, the spokes rotate in a $\mathbf{E}\times <!--\; ??\; -->\mathbf{B}$ direction with angular velocity in the range of 5–15 km s⁻¹. The spoke rotation reversal is explained by a transition from Ar-dominated to metal-dominated sputtering that shifts the plasma emission zone closer to the target. The spoke itself corresponds to a region of high electron density and therefore to a hump in the electrical potential. The electric field around the spoke dominates the spoke rotation direction. At low power, the plasma is further away from the target and it is dominated by the electric field to the anode, thus retrograde $\mathbf{E}\times <!--\; ??\; -->\mathbf{B}$ rotation. At high power, the plasma is closer to the target and it is dominated by the electric field pointing to the target, thus $\mathbf{E}\times <!--\; ??\; -->\mathbf{B}$ rotation. (paper)

[en] The temporal distribution of the incident fluxes of argon and titanium ions on the substrate during an argon HiPIMS pulse to sputter titanium with pulse lengths between 50 to 400 µs and peak powers up to 6 kW are measured by energy-resolved ion mass spectrometry with a temporal resolution of 2 µs. The data are correlated with time-resolved growth rates and with phase-resolved optical emission spectra. Four ion contributions impinging on the substrate at different times and energies are identified: (i) an initial argon ion burst after ignition, (ii) a titanium and argon ion flux in phase with the plasma current due to ionized neutrals in front of the target, (iii) a small energetic burst of ions after plasma shut off, and (iv) cold ions impinging on the substrate in the late afterglow showing a pronounced maximum in current. The last contribution originates from ions generated during the plasma current maximum at 50 µs after ignition in the magnetic trap in front of the target. They require long transport times of a few 100 µs to reach the substrate. All energy distributions can be very well fitted with a shifted Maxwellian indicating an efficient thermalization of the energetic species on their travel from target to substrate. The energy of titanium is higher than that of argon, because they originate from energetic neutrals of the sputter process. The determination of the temporal sequence of species, energies and fluxes in HiPIMS may lead to design rules for the targeted generation of these discharges and for synchronized biasing concepts to further improve the capabilities of high-power impulse magnetron sputtering (HiPIMS) processes. (paper)

[en] The growth rate during reactive high power pulsed magnetron sputtering (HIPIMS) of titanium nitride is an inherent time-dependent process. By using a rotating shutter setup it is possible to gain an insight into its variation with a temporal resolution of up to 25 µs. In this apparatus a 200 µm slit is rotated in front of the substrate synchronous with the HIPIMS pulses. This ensures that the incoming growth flux is laterally distributed over the substrate. By measuring the resulting deposition profile with profilometry and x-ray photoelectron spectroscopy, the temporal variation of the titanium and nitrogen growth flux per pulse is deduced. The analysis reveals that film growth occurs mainly during an HIPIMS pulse, with the growth rate following the HIPIMS phases ignition, current rise, gas rarefaction, plateau and afterglow. The growth fluxes of titanium and nitrogen follow slightly different behaviours with titanium dominating at the beginning of the HIPIMS pulse and nitrogen at the end of the pulse. This is explained by the gas rarefaction effect resulting in a dense initial metal plasma and metal films which are subsequently nitrified. (paper)

[en] The growth rate during high-power pulsed magnetron sputtering (HIPIMS) of titanium is measured with a temporal resolution of up to 25 µs using a rotating shutter concept. According to that concept a 200 µm slit is rotated in front of the substrate synchronous with the HIPIMS pulses. Thereby, the growth flux is laterally distributed over the substrate. By measuring the resulting deposition profile with profilometry, the temporal variation of the growth flux per pulse is deduced. The time-resolved growth rates are measured for 0.25, 0.5 and 1 Pa with pulse lengths of 50, 200 and 400 µs for an average power of 100 W. We can clearly identify, the individual phases of a HIPIMS pulse consisting of ignition, current rise, gas rarefaction, plateau/self-sputtering, and afterglow as described in the literature. In addition, the maximum film growth is only reached after gas rarefaction, indicating a dynamic change in local transport properties. After the end of the HIPIMS pulse, the growth rate decays following two time constants of 100 µs and of ∼ms, respectively. The first is consistent with the decay of the ion flux in the afterglow; the second with a decay of reactive neutrals. The absolute comparison of growth rates indicates that a reduction of the efficiency to 30% for very short pulses is typical for a true HIPIMS plasma. (paper)

[en] High power magnetron sputtering (HiPIMS) discharges generate ions with high kinetic energies in comparison to conventional dc magnetron sputtering. The peculiar shape of the ion energy distribution function (IEDF) is correlated to the formation of localized ionization zones (IZ) in the racetrack of a HiPIMS discharge, so called spokes. This is explained by a local maximum of the electrical potential inside these localized IZ. By using ion energy mass spectrometry, probe experiments and plasma spectroscopy the connection between IZ and IEDFs is evaluated with high temporal resolution. The data of a floating probe next to the target is used to directly monitor the movement of the spokes in the $\text{E}\times <!--\; ??\; -->\text{B}$ direction. Chromium is used as target material, because the plasma undergoes a sequence from stochastic spoke formation, to regular spoke pattern rotating in the $\text{E}\times <!--\; ??\; -->\text{B}$ direction to a homogeneous plasma torus with increasing plasma power. In particular, the analysis of the transition from the regular spoke pattern to the homogeneous plasma torus at very high plasma powers shows that the high energy part of the IEDF is not affected and only the low energy part is modified. Consequently, one could consider the homogenous plasma torus at very high plasma powers as a a single ionization zone localized over the complete torus, which is formed by merging individual spokes with increasing power. Details and consequences of that model are discussed. (paper)