Table of contents

This is my last issue of Plasma Sources Science and Technology in my role as Editor-in-Chief. I would like to take this opportunity to describe the origins of what has been for me a 17-year adventure.

Maureen Clarke, then responsible for commissioning new journals at IOP, first conceived of a journal devoted to low-temperature plasmas. She contacted me, and, I imagine, others in the plasma community, with a set of questions about this possible new journal. Although I've lost that letter, I still have a copy of my e-mail response to her from 2 July 1990, from which the following extracts are taken:

Dear Ms. Clarke,

Thank you for an opportunity to comment on your new journal—Plasma Sources and Plasma Processing. I believe that there is a currently a place for a journal which is concerned with plasma source design and characteristics related to plasma processing and that this need is likely to continue for at least 10 years or more.

[ . . . ]

Right now there is considerable interest in the plasma processing community on the relative advantages of ECR and other microwave sources versus 13 MHz systems and a variety of different types of both sources have been invented and more seem to be coming along each day. Helicon sources are also starting to be interesting.

[ . . . ]

My view is that plasma processing includes all aspects of processes which employ charged particle plasmas in manufacturing processes. This runs from ion implantation out of very low pressure (<10–4 torr) plasmas to plasma sprays at atmospheric pressure. A journal which emphasizes the role of the charged particles and which covers the full spectrum of devices would be a welcome addition to other journals now available.

I am interested in the subject and I would be willing to serve on the Editorial Board.

At that time, the journal was tentatively named Plasma Sources and Plasma Processing. By 6 November 1990 she had offered me the position of Editor-in-Chief. I accepted, and by early 1991, IOP had approved the new journal. Soon afterwards, Maureen and I decided that the journal name should be Plasma Sources Science and Technology (PSST), and chose the original cover design (shown here).

I agreed to be Editor-in-Chief of PSST on the condition that I be responsible for choosing the referees. This differed from the policy of most other IOP journals, for which IOP staff made those decisions. I believed, and still believe, that the best way to build the journal and maintain its quality is to choose the best authorities in the field to be referees. This served to familiarize these scientists with PSST; eventually, they submitted their own papers to PSST, and sometimes even became members of the Editorial Board.

I contacted possible members of the Editorial Board, and most accepted. The Board consists of the leaders in the field and has guided the evolution of the journal over the years. Possibly the most important decision I made was to ask Daan Schram from Eindhoven University of Technology to serve as the Regional Editor for Europe. His experience and expertise were invaluable to the development and success of PSST. He was a key addition to the Editorial Board because he was very well established in the European plasma community and his plasma background complemented mine. Daan retired as Regional Editor in 2006, but still continues as a member of the Editorial Board.

The first issue of PSST appeared a year later, in March 1992. It was the only time that a printed issue of the journal did not appear in February, May, August, or November. It was 74 pages long. (To see the first table of contents, click here). The articles were arranged by the appropriate neutral pressures, a procedure that continued until the onset of electronic publication. (An alternative ordering of the papers according to neutral pressure was subsequently provided at the beginning of each issue.) In the first few issues, the papers covered the full range in neutral pressures, setting a pattern for the future.

In 1992, the first year of publication, we had 280 pages. This year there are over 900 regular issue pages. The latest impact factor of 2.346 is the highest ever, and evidence that PSST is now a well established journal. My comment in 1990, that the need for a journal which is concerned with plasma source design and characteristics related to plasma processing 'is likely to continue for at least 10 years or more', was clearly overly pessimistic. The field is continuing to grow. A variety of atmospheric pressure discharges that were hardly considered or unknown in 1990 are now hot topics, and sources and plasma characteristics at all pressures are still of interest.

PSST has not only evolved into the journal I envisioned in 1990, but has, in fact, exceeded my expectations. I feel privileged to have had the opportunity to start and develop a new journal, to have been able to get to know and work with many exceptional scientists, and to have been a part of IOP Publishing, a first-class organization. The IOP staff associated with the production of PSST have changed over the years but in all cases have been a pleasure to work with. IOP has led the way in electronic publication.

While I am retiring as Editor-in-Chief, I will still continue as an Associate Editor, so my adventure – albeit with reduced responsibilities – will continue.

Mark Kushner is taking over as Editor-in-Chief and I wish him all the best.

Noah Hershkovitz, University of Wisconsin-Madison, Editor-in-Chief

Recent developments in laboratory double layers from the late 1980s to the spring of 2007 are reviewed. The paper begins by a lead up to electric double layers in the laboratory. Then an overview of the main double layer devices and properties is presented with an emphasis on current-free double layers. Some of the double layer models and simulations are analysed before giving a more complete description of current-free double layers in radiofrequency plasmas expanding in a diverging magnetic field. Astrophysics double layers are briefly reported. Finally, applications of double layers to the field of plasma processing and electric propulsion are discussed.

The optimal plasma parameters for plasma processing, such as deposition of microcrystalline silicon from silane and hydrogen, are generally chosen in steady-state discharge conditions. However, this steady state must be reached in a short time after plasma ignition to avoid significant film deposition in non-optimal conditions during the plasma transient phase.

Simple and inexpensive time-resolved optical emission spectroscopy has been used to measure the plasma time evolution from ignition to steady-state conditions in a large area RF capacitive plasma reactor. Absolute values of silane and hydrogen molecular number densities, relative values of electron density, and qualitative information on electron temperature were obtained without the need for absolute intensity calibration. Apart from the experimental verification of constant electron temperature, the particular condition here is that the emission intensities should be followed from the instant of ignition, since the molecular densities are known at this instant.

A plasma model for the reactor, and a dispersive axial flow model for the pumping line, were used to show why the plasma chemistry in a well-designed large area reactor generally reaches steady-state conditions in less than one second. The optimal design for fast equilibration is a closed, directly-pumped showerhead reactor with a uniform plasma which fills the whole reactor volume.

A spatial and temporal numerical simulation has been carried out of a pulsed (100% modulated), rf inductively coupled plasma discharge in argon, connected to an additional (ballast) diffusion chamber of much larger volume. It is demonstrated that during the active phase, the presence of the large ballast volume has a small impact on the parameters of the plasma in the smaller discharge chamber. In this case the plasma parameters in the discharge chamber can be estimated separately from the diffusion chamber by a standard method using the characteristic ambipolar diffusion time (for example, using a global model). However, during the afterglow phase, the situation is changed significantly. In the afterglow, the densities of charged particles in the discharge chamber become lower than in the large ballast chamber due to more rapid diffusion loss. As a result, the reverse of the active phase situation occurs, namely, the plasma does not flow from the small to the large chamber, but in the opposite direction, from diffusive to discharge volume, and both the plasma density gradient and the self-consistent ambipolar electric field in the small chamber change directions. This phenomenon leads to new effects in the discharge volume, in particular a decreasing rate of decay of densities of charged particles and electron temperature. Thus, in the afterglow the presence of a large additional ballast volume has a significant impact on the plasma transport. In this case, a simple treatment of the plasma in the discharge chamber in the framework of a spatially averaged model (for example, the global model) is inadequate.

A pulsed corona facility for the treatment of gases at relatively high gas flow rates and powers is described. Data are presented, mainly for the example of toluene removal, which illustrate specific energy scaling of the system with power, gas flow rate, charging voltage of the pulsed power supply and inlet concentration. The effect of plasma reactor length and charging capacitance of the pulsed power supply on the system efficiency is also investigated. The implications of the scaling and efficiency in determining the size of systems and the relative operating and capital costs for practical applications are discussed briefly.

The formation of aluminum fluoride layers on the inner part of plasma reactor walls is known to be a serious issue for plasma etching processes: it causes process drifts and particle generation. AlF_x coatings are formed on the reactor walls as soon as the Al₂O₃ reactor walls are exposed to fluorine (F)-based plasmas. Since plasma reactors are always cleaned in fluorine chemistries, SF₆ for example, AlF_x formation is inevitable in the typical plasma clean conditions used in manufacturing. AlF_x material being extremely etch resistant, it cannot be removed from reactor walls: it accumulates until particles flake off, then imposing a wet clean of the plasma reactor. In this paper, we review the different methods to reduce or eliminate the formation of AlF_x and we report that SiCl_x radicals/ions etch AlF_x material by forming AlCl_x and SiF_x products. By using a dedicated technique based on x-ray photoelectron spectroscopy to analyze the reactor walls, we show that AlF_x (and also YF_x) coatings formed on the reactor walls are cleared in SiCl₄/Cl₂ plasmas, restoring clean Al₂O₃ (Y₂O₃) walls before processing the next wafer. As a result the wafer-to-wafer reproducibility and the mean time between reactor wet cleans are probably significantly improved. Furthermore, SiCl₄-based plasma chemistries efficiently remove other metal fluorides (such as HfF_x) and more generally metallic oxides (high-k) from the reactor walls.

The hairpin resonator probe has been developed in recent years into a sophisticated diagnostic technique capable of measuring spatially resolved electron number densities in sub-Torr discharges. In this paper, we extend the use of this technique to discharges at pressures greater than 1 Torr. In this regime, the effects of electron–neutral collisions become significant and a suitable correction is applied in conjunction with the sheath correction. We also describe elements of hairpin design and coupling that need to be more carefully controlled in order to maximize the range of electron densities that can be detected at higher pressures. Finally, we discuss limitations to the transmission-line model used routinely to interpret hairpin data as they apply to measurements in a nonuniform plasma.

Self-consistent two-dimensional modeling of a steady microwave discharge initiated at the end of the central electrode in nitrogen is presented. The discharge parameters are calculated at a gas pressure of 1 Torr and incident power of 30 W. The computational model includes the time-dependent Maxwell's equations, the balance equations describing the kinetics of charged and neutral plasma particles and the time-independent homogeneous Boltzmann equation for electrons. The processes involving vibrationally excited ground state molecular nitrogen are taken into account by the well-known analytic expression for the vibrational distribution of molecules in the diffusion approximation. It is shown that the spatially non-uniform microwave field causes the difference in plasma particles kinetics in different parts of the discharge. Results of numerical simulations for space distribution of electronically excited molecules have been found in good qualitative agreement with those taken from spectral measurements of first and second positive systems of nitrogen. Results confirm the concept according to which such a discharge comprises a self-sustained and a non-self-sustained discharge.

We investigated characteristics of a density peak observed in a magnetic field B₀ lower than 100 G in the case of using helicon plasma sources with particular wavelength. For B₀ > 30 G, the antenna launches an electromagnetic wave as a slow wave, the phase velocity of which becomes close to the electron thermal velocity under the density-peak condition for various gas pressures. The Landau damping frequency is higher than the electron–neutral and the electron–ion collision frequencies, which indicates that the wave produces the plasma via Landau damping at low B₀. The wavelengths estimated from the density n_e and B₀ for the density peaks agree with those of the electromagnetic field generated by helicon antennas of various lengths. The measured density is found to vary under the condition of the agreement between wavelengths of the propagating wave and the antenna-excited field during the density increase. One of the causes of the density peak appearing as a function of B₀ is considered to be the wavelength variation and the corresponding change of phase velocity of the slow wave which is enabled to propagate in the plasma by the introduction of B₀.

We present an analysis of laser-induced fluorescence and emission spectroscopy profile measurements of the excited states of neutral argon for two different helicon source pressures. Through use of a collisional–radiative model that calculates excited state profiles based on measured electron densities, electron temperatures, and the measured edge neutral pressure, the ground state neutral density profile is determined for each pressure case. The results demonstrate that it is possible to extract the ground state neutral density profile from such measurements and that the degree of ionization at the center (r = 0) of a helicon source plasma can vary significantly for a small change in source pressure.

Optical emission spectroscopy is used for the determination of hydrogen molecule rotational T_rot and vibrational T_vib and translational temperature T_tr. The experiment was carried out using copper or titanium hollow cathode discharges operated in pure hydrogen. The rotational temperature of excited electron energy levels is determined from Fulcher-α diagonal bands ( electronic transition, P-, Q- and R-branches), which were later used to derive the rovibronic temperature of the hydrogen molecule ground state, which is assumed equal to T_tr. The vibrational temperature T_vib is determined for the same electronic transition (Q-branch with ν' = 2,3; Δν = 0). For all temperatures the radial temperature distribution is reported.

In order to study the physical mechanism of an oscillation newly discovered by the Harbin Institute of Technology Plasma Propulsion Lab (HPPL) in the range of hundreds of kHz to several MHz, Hall thrusters with different magnetic coils are studied by changing one of the following three parameters: discharge voltage, anode flow and coil current, directly measuring the coil current and measuring plasma oscillations related to coil current oscillation with the Langmuir probe. Experimental results indicated that in the discharge process of a Hall thruster the broadband turbulence of the Hall current causes an unstable spatial magnetic field and this field causes the magnetic circuit to resonate as an equivalent high level resistance–inductance–capacitance (RLC) network. As the response of the network, the oscillation of the coil current has a large oscillating component at the natural frequencies of the network. Also, the oscillation of coil current has an effect on the discharge process at the same time, so that they reach a self-consistent equilibrium state. As a result of such a coupling, both coil current and the discharge current exhibit their oscillating component at the natural frequencies of the magnetic circuit. It is therefore concluded that the newly discovered oscillation is caused by the coupling between the magnetic circuit and the discharge circuit.

To understand the origin of the multiple arc–anode attachments in the anode region of high intensity arcs, a linear stability analysis of a non-uniform argon plasma consisting of electrons, argon neutrals and singly ionized argon ions and an arc–anode interface instability analysis are performed and presented. Plasma conditions typical of the anode region of high intensity argon arcs are emphasized. The calculations have shown that small-amplitude fluctuations presented in the plasma excite two different wave modes, space-charge relaxation mode and electron thermal mode, on a time scale shorter than the one in which the multiple-attachment mode forms. The space-charge relaxation mode has been found to be stable and not affected by the non-uniformity of the plasma. The electron thermal mode can be unstable, depending on the distributions of the current flow, the electric field, the electron temperature and the electron density. The analysis, together with a consideration of the evaporation–ionization instabilities at the arc–anode interface, has been applied to previous experimental data. The observed multiple-attachment mode, consisting of several constricted attachment spots in the arc fringes, can be explained by this analysis.

Hydrocarbon nanoparticles with diameters between 10 and 30 nanometres are created in a low pressure plasma combining capacitive and inductive power coupling. The particles are generated in the capacitive phase of the experiment and stay confined in the plasma in the inductive phase. The presence of these embedded particles induces a rotation of a particle-free region (void) around the symmetry axis of the reactor. The phenomenon is analysed using optical emission spectroscopy both line integrated and spatially resolved via an intensified charge coupled device camera. From these data, electron temperatures and densities are deduced. We find that the rotation of the void is driven by a tangential component of the ion drag force induced by an external static magnetic field. Two modes are observed: a fast rotation of the void in the direction opposite to that of the tangential component and a slow rotation in the same direction. The rotation speed decreases linearly with the size of the particles. In the fast mode the dependence on the applied magnetic field is weak and consequently the rotation speed can serve as a monitor to detect particle sizes in low temperature plasmas.

The UNU-ICTP plasma focus with a significantly longer than conventional anode can still be a reasonably good neon soft x-ray (SXR) source. The highest average neon SXR yield of 3.3 J was achieved at 3 mbar. The time difference between the two first peaks of the voltage probe signal at the radial collapse phase was found to be inversely related to the SXR yield, i.e. the smaller the time difference, the higher the yield and vice versa. The estimation of average current sheath speeds using the shadowgraphic method coupled with laser and focus peak timing signals showed that the average axial rundown speed is similar to the one obtained for the optimal anode length but the average radial compression speed is decreased significantly. The range of pressure for a good neon SXR yield, however, has become much narrower, making efficient plasma focus operation a very sensitive function of the filling gas pressure for longer than the optimal anode length.

The decomposition of toluene in a gliding arc discharge (glidarc) was performed and studied. Experimental results indicate that the glidarc technology can effectively decompose toluene molecules and has bright prospects of being applied as an alternative tool to decompose volatile organic compounds. It is found that a change in the electrode material had an insignificant effect on the toluene removal efficiency. The toluene removal efficiency increases with increasing inlet gas temperature. The water vapor present in the gas mixture has a favorable effect on the toluene decomposition in the plasma. The energy efficiency is 29.46 g (kWh⁻¹) at a relative humidity of 50% and a specific energy input of 0.26 kWh m⁻³, which is higher than other types of non-thermal plasmas. Too much or too little oxygen content does not favor toluene decomposition. The major gas phase products detected by FT-IR from the decomposition of toluene with air participation were CO, CO₂, H₂O and NO₂. Some brown depositions were found on the surface of the electrodes, which were polar oxygenous and nitrogenous compounds determined by the GC-MS analysis, such as benzaldehyde, benzoic acid, quinine and nitrophenol from the reaction of toluene with radicals. A possible mechanism for toluene destruction via glidarc technology is proposed and summarized.

An intense source of atomic oxygen is reported in this paper. Oxygen atoms were produced in an Ar/O₂, 2.45 GHz, microwave discharge at flow rates ranging from 10 to 150 sccm and pressures from 0.5 to 4 Torr. The absolute flow rate of atomic oxygen peaked at 70 sccm which corresponded to a fractional dissociation close to 0.5. The atomic oxygen flow rate was measured using a titration method with NO₂ gas. It was also shown that the emission ratio from the 3p³P state of an oxygen atom at 844.6 nm and from the 3p⁵4p state of an argon atom at 750.4 nm was strongly correlated with oxygen atom to argon concentrations at some experimental conditions, e.g. modest Ar/O₂ flow ratios.

An atmospheric argon plasma jet generated by an original dc double anode plasma torch has been investigated through its electrical and spectroscopic diagnostics. The arc instabilities and dynamic behavior of the argon plasma are analyzed using classical tools such as the statistical method, fast Fourier transform (FFT) and correlation function. The takeover mode is identified as the fluctuation characteristic of the double arc argon plasma jet in our experiment. The FFT and correlation analysis of electrical signals exhibit the only characteristic frequency of 150 Hz, which originates from the torch power and is independent of any change in the operating parameters. No high frequency fluctuations (1–15 kHz) are observed. This indicates that the nature of fluctuations in an argon plasma jet is induced mainly by the undulation of the tri-phase rectified power supply. It is found that each arc root attachment is diffused rather than located at a fixed position on the anode wall. Moreover, the emission spectroscopic technique is performed to determine the electron temperature and number density of the plasma jet inside and outside the arc chamber. Along the torch axis, the measured electron temperature and number density of the double arc argon plasma drop from 12 300 K and 7.6 × 10²² m⁻³ at the divergent part of the first anode nozzle, to 10 500 K and 3.1 × 10²² m⁻³ at the torch exit. In addition, the validity criteria of the local thermodynamic equilibrium (LTE) state in the plasma arc are examined. The results show that the measured electron densities are in good agreement with those calculated from the LTE model, which indicates that the double arc argon plasma at atmospheric pressure is close to the LTE state under our experimental conditions.

This paper examines the complex nature of highly polymerizing fluorocarbon plasmas. An inductively coupled modified GEC reference cell is used to look at process rates on SiO₂, p-Si and Si₃N₄ samples using various chamber geometries and gas chemistries. In an attempt to understand the process rates, a simple model based on the sticking and etch yield coefficients of radicals and ions is employed. Development of the model requires knowledge of radical flux, ion flux, ion energy and related process rates. These values are determined using in situ spectroscopic ellipsometry, in situ optical emission spectroscopy, in situ Fourier transform infrared spectroscopy and chuck self-bias measurements. Through the use of a variable electrode gap and changing feed gas chemistry, sticking radical densities are controlled almost independently of ions and etching radical densities. This control allows a partial deconvolution of the process rate equation. Estimated values for the upper bound sticking coefficients of fluorocarbon radicals are made. Additionally, values are reported for ion sticking coefficients and the fluorine etch yield coefficient. These values are then used in a basic low ion energy model to compare with experimental process rates.

In a pulsed dc discharge of an Ar–N₂ mixture containing 0.91% of NO the kinetics of the destruction of NO has been studied under static and flowing conditions, i.e. in a closed and open discharge tube (p = 266 Pa). For this purpose quantum cascade laser absorption spectroscopy (QCLAS) in the infrared spectral range has been applied as a new approach for fast in situ plasma diagnostics which is capable of achieving a time resolution below 100 ns. The time decay of the NO concentration was measured in single discharge pulses of 1 ms duration. Additionally, the temporal behaviour of the electric field and the applied power was followed during the pulse. The comparison of the time evolution of the NO concentration under static and flowing conditions and simplified model calculations enabled an analysis of the dynamics of the plasma heating to be made. The temperature increase during the pulse is below 40 K, but has a strong influence on the line strength of the NO absorption line. The apparent decrease in the NO concentration in a single pulse of about 20% is due to the heating of the gas which in turn makes the line strength vary while the concentration remains constant for several successive pulses. Therefore the QCLAS measurements combined with model calculations are a powerful non-invasive temperature probe with a remarkable time resolution approaching the sub-microsecond time scale.

We report on investigations of gas metal arc welding plasma operated in pure argon and in a mixture of argon and CO₂ at a dc current of 326 A. The spatially resolved electron densities and temperatures were directly obtained by measuring the Stark widths of the Ar I 695.5 nm and Fe I 538.3 nm spectral lines.

Our experimental results show a reduction of the plasma conductivity and transfer from spray arc to globular arc operation with increasing CO₂ concentration. Although the electron density n_e increases while approaching the core of the plasma in the spray-arc mode, a drop in the electron temperature T_e is observed. Moreover, the maximum T_e that we measure is about 13 000 K. Our experimental results differ from the Haidar model where T_e is always maximum on the arc axis and its values exceed 20 000 K. These discrepancies can be explained as a result of underestimation of the amount of metal vapours in the plasma core and of the assumption of local thermal equilibrium plasma in the model.

The behavior of dual frequency capacitively coupled plasma discharges (2f-CCP) is experimentally studied by Langmuir probe and rf current measurements and is compared with simulations from the literature. The driving frequency ratio, system pressure, high frequency (HF) power and low frequency (LF) power are varied in the experiments. An increase in LF power causes a moderate increase in electron density but a significant decrease in electron temperature. An increase in HF power causes a strong increase in electron density and populates the high energy part of the electron energy distribution function. These dependences can be explained on the basis of a global model. It is shown that the ratios of HF/LF power and driving frequency are the most important parameters. At integer frequency ratios a significant increase in electron density was found, which is explained by the indirect heating at the plasma series resonance. Several design guidelines are derived which address industrial applications and process stability.

The PDF file provided contains web links to all articles in this volume.