Refine
Year of publication
- 2009 (10) (remove)
Document Type
- Doctoral Thesis (7)
- Conference Proceeding (2)
- Article (1)
Language
- English (10) (remove)
Has Fulltext
- yes (10)
Is part of the Bibliography
- no (10)
Keywords
Institute
- Institut für Physik (10) (remove)
Publisher
- IOP Publishing (3)
Application of quantum cascade laser absorption spectroscopy to studies of fluorocarbon molecules
(2009)
The recent advent of quantum cascade lasers (QCLs) enables room-temperature mid-infrared spectrometer operation which is particularly favourable for industrial process monitoring and control, i.e. the detection of transient and stable molecular species. Conversely, fluorocarbon containing radio-frequency discharges are of special interest for plasma etching and deposition as well as for fundamental studies on gas phase and plasma surface reactions. The application of QCL absorption spectroscopy to such low pressure plasmas is typically hampered by non-linear effects connected with the pulsed mode of the lasers. Nevertheless, adequate calibration can eliminate such effects, especially in the case of complex spectra where single line parameters are not available. In order to facilitate measurements in fluorocarbon plasmas, studies on complex spectra of CF4 and C3F8 at 7.86 μm (1269 – 1275 cm-1) under low pressure conditions have been performed. The intra-pulse mode, i.e. pulses of up to 300 ns, was applied yielding highly resolved spectral scans of ∼ 1 cm-1 coverage. Effective absorption cross sections were determined and their temperature dependence was studied in the relevant range up to 400 K and found to be non-negligible.
Fluorocarbon containing capacitively coupled radio frequency (cc-rf) plasmas are widely used in technical applications and as model systems for fundamental investigations of complex plasmas. Absorption spectroscopy based on pulsed quantum cascade lasers (QCL) was applied in the mid-IR spectral range of 1269-1275 cm-1. Absolute densities of the precursor molecule CF4 and of the stable product C3F8 were measured with a time resolution of up to 1 ms in pulsed CF4/H2 asymmetrical cc-rf (13.56 MHz) discharges. For this purpose both the non-negligible temperature dependence of the absorption coefficients and the interference of the absorption features of CF4 and C3F8 had to be taken into account in the target spectral range. Therefore, at two different spectral positions composite absorption spectra were acquired under the same plasma conditions in order to discriminate between CF4 and C3F8 contributions. A total consumption of∼ 12 % was observed for CF4 during a 1 s plasma pulse, whereas C3F8 appeared to be produced mainly from amorphous fluorocarbon layers deposited at the reactor walls. A gas temperature increase by ∼ 100 K in the plasma pulse was estimated from the measurements. Additionally, not yet identified unresolved absorption (potentially from the excited CF4 molecule) was found during the àon-phase'.
An interesting aspect in the research of complex (dusty) plasmas is the experimental study of the interaction of micro-particles with the surrounding plasma for diagnostic purposes. Local electric fields can be determined from the behaviour of particles in the plasma, e.g. particles may serve as electrostatic probes. Since in many cases of applications in plasma technology it is of great interest to describe the electric field conditions in front of floating or biased surfaces, the confinement and behaviour of test particles is studied in front of floating walls inserted into a plasma as well as in front of additionally biased surfaces. For the latter case, the behaviour of particles in front of an adaptive electrode, which allows for an efficient confinement and manipulation of the grains, has been experimentally studied in terms of the dependence on the discharge parameters and on different bias conditions of the electrode. The effect of the partially biased surface (dc and rf) on the charged micro-particles has been investigated by particle falling experiments. In addition to the experiments, we also investigate the particle behaviour numerically by molecular dynamics, in combination with a fluid and particle-in-cell description of the plasma.
This thesis constitutes a computational study of charge and ion drag force on micron-sized dust particles immersed in rf discharges. Knowledge of dust parameters like dust charge, floating potential, shielding and ion drag force is very crucial for explaining complex laboratory dusty plasma phenomena, such as void formation in microgravity experiments and wakefield formation in the sheaths. Existing theoretical models assume standard distribution functions for plasma species and are applicable over a limited range of flow velocities and collisionality. Kinetic simulations are suitable tools for studying dust charging and drag force computation. The main aim of this thesis is to perform three dimensional simulations using a Particle-Particle-Particle-Mesh ($P^3M$) model to understand how the dust parameters vary for different positions of dust in rf discharges and how these parameters on a dust evolve in the presence of neighboring dust particles. At first, rf discharges in argon have been modelled using a three-dimensional PIC-MCC code for the discharge conditions relevant to the dusty plasma experiments. All necessary elastic and inelastic collisions have been considered. The plasma background is found collisional, charge-exchange collisions between ions and neutrals being dominant. Electron and ion distributions are non-Maxwellian. The dominant heating mechanism is Ohmic. Then, simulations have been done to compute the dust parameters for various sizes of dust located at different positions in the rf discharges. Dust charge and floating potential in the presheath are slightly larger than the values in the bulk due to the higher electron flux to the dust particle in the presheath. From presheath to the sheath the charge and floating potential values decrease due to the decrease of the electron current to the dust. A linear dependence of dust potential on dust size has been found, which results in a nonlinear dependence of the dust charge with the dust size when the particle is assumed to be a spherical capacitor. This has been verified by independently counting the charges collected by the dust. %where indeed it has been noted that the dust charge %scales nonlinearly with the dust size. The computed dust parameters are also compared with theoretical models. Simulated dust floating potentials are comparable to values obtained from Allen-Boyd-Reynolds (ABR) and Khrapak models, but much smaller than the values obtained from Orbit Motion Limited (OML) model. The dust potential distribution behaves Debye-H\"{u}ckel-like. The shielding lengths are in between ion and electron Debye lengths. % indicating shielding by both ions and electrons. Further, the orbital drag force is typically larger than the collection drag force. The total drag force for the collisional case is larger than for the collisionless case and it scales nonlinearly with the dust size. The collection drag values and size-scaling agrees with Zobnin's model. The charging and drag force computation is then extended to two and multiple static dust particles in the rf discharge to study the influence of neighboring dust particles on the dust parameters. Initially, the dust parameters on two dust particles are computed for various interparticle separation distances and for dust particles placed at different locations in the rf discharge. It is observed that for dust separations larger than the shielding length the dust parameters for the two dust particles match with the single dust particle values. As the dust separation is equal to or less than the shielding length the ion drag force increases due to the buildup of a parallel drag force component. However, the main dust properties like charge, potential, vertical component of ion drag are not affected considerably. This is attributed to the smaller collection impact parameter values compared to the dust separation. %This is because the %collection impact parameter values in the sheath and the presheath are smaller %than the smallest dust separation and in case of the dust in the bulk, the %collection impact parameter is comparable with the dust separation. Then the dust charges on multiple dust particles located at different positions in the discharge and arranged along the discharge axis are also computed. It is found that the charges of the multiple dust particles in the bulk or presheath do not differ much from the single particle values at that location. But the dust charges of multiple dust particles located in the sheath drastically differ from the single dust parameter values. Due to ion focusing from dust particles in the upper layers, the ion current increases to dust particles in the lower layers resulting in smaller charge values. This is as well the case where dust particles are vertically aligned as in the standard experiments of dusty plasmas. In conclusion, this work used a fully kinetic (PIC and MD or $P^3M$) model to study the physics of dust charging in rf plasmas. Our simulations revealed that the dust parameters vary considerably from the bulk to the sheath. The CX collisions increase flux to the dust thereby affecting the dust parameters and their scaling with dust size. Also, a dust particle affects the charging dynamics of its neighbor only when their separation is within the shielding length. In the plasma sheath, ion focussing can cause great reduction in dust charges.
Infrared laser absorption spectroscopy (IRLAS) employing both tuneable diode and quantum cascade lasers (TDLs, QCLs) has been applied with both high sensitivity and high time resolution to plasma diagnostics and trace gas measurements.
TDLAS combined with a conventional White type multiple pass cell was used to detect up to 13 constituent molecular species in low pressure Ar/H2/N2/O2 and Ar/CH4/N2/O2 microwave discharges, among them the main products such as H2O, NH3, NO and CO, HCN respectively. The hydroxyl radical has been measured in the mid infrared (MIR) spectral range in-situ in both plasmas yielding number densities of between 1011 ... 1012 cm-3. Strong indications of surface dominated formation of either NH3 or N2O and NO were found in the H2-N2-O2 system. In methane containing plasmas a transition between deposition and etching conditions and generally an incomplete oxidation of the precursor were observed.
The application of QCLs for IRLAS under low pressure conditions employing the most common tuning approaches has been investigated in detail. A new method of analysing absorption features quantitatively when the rapid passage effect is present is proposed. If power saturation is negligible, integrating the undisturbed half of the line profile yields accurate number densities without calibrating the system. By means of a time resolved analysis of individual chirped QCL pulses the main reasons for increased effective laser line widths could be identified. Apart from the well-known frequency down chirp non-linear absorption phenomena and bandwidth limitations of the detection system may significantly degrade the performance and accuracy of inter pulse spectrometers. The minimum analogue bandwidth of the entire system should normally not fall below 250 MHz.
QCLAS using pulsed lasers has been used for highly time resolved measurements in reactive plasmas for the first time enabling a time resolution down to about 100 ns to be achieved. A temperature increase of typically less than 50 K has been established for pulsed DC discharges containing Ar/N2 and traces of NO. The main NO production and depletion reactions have been identified from a comparison of model calculations and time resolved measurements in plasma pulses of up to 100 ms. Considerable NO struction is observed after 5 ... 10 ms due to the impact of N atoms.
Finally, thermoelectrically cooled pulsed and continuous wave (cw) QCLs have been employed for high finesse cavity absorption spectroscopy in the MIR. Cavity ring down spectroscopy (CRDS) has been performed with pulsed QCLs and was found to be limited by the intrinsic frequency chirp of the laser suppressing an efficient intensity build-up inside the cavity. Consequently the accuracy and advantage of an absolute internal absorption calibration is not achievable. A room temperature cw QCL was used in a complementary cavity enhanced absorption spectroscopy (CEAS) configuration which was equipped with different cavities of up to ~ 1.3 m length. This spectrometer yielded path lengths of up to 4 km and a noise equivalent absorption down to 4 x 10-8 cm-1Hz-1/2. The corresponding molecular concentration detection limit (e.g. for CH4, N2O and C2H2 at 1303 cm-1/7.66 μm) was generally below 1 x 1010 cm-3 for 1 s integration times and one order of magnitude less for 30 s integration times. The main limiting factor for achieving even higher sensitivity is the residual mode noise of the cavity. Employing a ~ 0.5 m long cavity the achieved sensitivity was good enough for the selective measurement of trace atmospheric constituents at 2.2 mbar.
(A paperback version is published by Logos under ISBN 978-3-8325-2345-9.)
Asymmetrical capacitively coupled RF discharges in oxygen, argon and hydrogen have been experimentally investigated with the innovative technique of the phase resolved optical emission spectroscopy. This diagnostic tool allows to measure spatio-temporally resolved emission intensities of electronically excited species with a high resolution. The spatial (axial) resolution was better than 1 mm and a temporal resolution of about 1.5 ns has been achieved. Therefore the plasma induced optical emission within the RF cycle (TRF = 73.75 ns) from the RF sheath region with a typical mean sheath thickness of about 5mm has been studied. Spatio-temporally resolved optical emission patterns of the following optical transitions have been measured for a total gas pressure in the range of 20 to 100 Pa and self-bias voltages between -50 and -550 V: Oxygen plasma Emission at 777.4 nm and 844.6 nm (atomic oxygen) Argon plasma Emission at about 751 nm and 841 nm (argon) Hydrogen plasma Emission at 656.3nm (atomic hydrogen, H alpha-line) These transitions are the most prominent ones of the investigated excited species in these plasmas as could be shown from overview spectra of the plasma induced optical emission in the range from 350 to 850 nm. For the first time such extensive PROES measurements in oxygen CCRF plasmas are presented in this work. The additional investigations of argon and hydrogen plasmas serve as a reference and for a direct comparison with results from the literature. The temporal behavior of the emission intensity is influenced by the effective lifetime of the emitting states which is on the order of the nanosecond time scale of the RF cycle. Therefore, it does not represent the real temporal behavior of the excitation. A simple method has been applied to calculate relative excitation rates from the measured emission intensities to distinguish different excitation mechanisms and their correct relative temporal behavior. In a close collaboration within the framework of the Sonderforschungsbereich Transregio 24 'Fundamentals of Complex Plasmas' a newly 1d3v PIC-MCC code for simulations of capacitive RF discharges in oxygen has been developed by Matyash et al. The very close coupling of experiment and modeling allowed a really detailed and microscopic understanding of the processes and dynamics from the sheath to the bulk plasma in CCRF discharges. The spatio-temporally resolved excitation rate profiles show four different excitation structures (I-IV). Excitation processes due to the following mechanisms in CCPs could be identified and characterized: I Electrons expelled from growing sheath II Electrons detached from negative ions (collisions with neutrals) + secondary electrons from the electrode surface (ion bombardment) III Field-reversal effect, reduced mobility of electrons (electron-neutral collisions) IV Heavy-particle collisions These excitation mechanisms are characterized by different temporal and spatial behaviors of the excitation rate within the RF cycle. Additionally it has been shown that the excitation by electron impact in the investigated oxygen plasmas results mainly from dissociative electron impact excitation (O2 + e -> O + O* + e) and not from direct electron impact excitation (O + e -> O* + e). Actinometry measurements show that the results are not really credible. Thus actinometry is not applicable on the investigated oxygen RF plasma. A challenge in interpretation is the observed excitation pattern IV. Pattern IV has to be caused in connection with heavy particle collisions nearby the electrode surface and could be observed in all the three plasmas oxygen, argon and hydrogen. It is located directly in front of the powered electrode and appears during almost the whole RF cycle. The temporal modulation is nearly sinusoidal and weak in comparison to the first three patterns. This is due to the weak RF modulation of the ion flux towards the electrode surface which has been proven by a PIC simulation. It could be shown that the modulation degree of pattern IV depends on the transition time of the corresponding positive ions through the RF sheath which is influenced by the ion mass. In oxygen as well as in argon CCRF plasmas pattern IV is less modulated than in hydrogen CCRF plasmas due to the heavier ions in oxygen and argon. Additionally the modulation degree increases with increasing pressure due to the more confined plasma at higher pressures which is yielding in a stronger modulated ion current towards the powered electrode.
In the present work, mass determinations of the eleven neutron-deficient nuclides (99-109)Cd, of ten neutron-rich silver nuclides (112-114,121,123)Ag, and seven neutron-rich cadmium nuclides (114,120,122-124,126,128)Cd are reported. Due to the clean production of the neutron-deficient nuclides it was possible to reduce the experimental uncertainties down to 2 keV, whereas the measurements of neutron-rich nuclides were hampered by the presence of contaminations from more stable In and Cs nuclides. In the case of 99Cd and 123Ag the masses were determined for the first time and for the other nuclides the mass uncertainties could be reduced by up to a factor of 50 as in the case of 100Cd. In the case of a potential isomeric mixture as for (115,117,119)Ag and 123Cd, where no assignment to either the ground state or the excited state was possible, the experimental results were adjusted accordingly. Afterwards all results were included in the framework of the atomic-mass evaluation and thus linked and compared with other experimental data. In the case of the neutron-deficient Cd nuclides a conflict between the mass values obtained in the present work and those published by the JYFLTRAP group could be solved by performing an atomic-mass evaluation. These mass measurements are an important step towards an understanding of the physics of the rp process that will enable a more reliable determination of the composition of the produced material at A = 99. It has been shown that the mass of 99Cd strongly affects the A = 99 production in an X-ray burst model, and that uncertainties have been significantly reduced from more than an order of magnitude to about a factor of 3. The dominant source of uncertainty is now the mass of 100In. In principle, other uncertainties will also contribute. These include those of masses of lighter Cd isotopes, where similar rp-process branchpoints occur and which might affect feeding into the 99Cd branchpoint. In addition, nuclear reaction rate uncertainties will also play a role. However, as reaction rates affect branchings in a linear fashion, while mass differences enter exponentially, mass uncertainties will tend to dominate. Also, which reaction rates are important depends largely on nuclear masses. For example, for low Sp(100In) a (p,γ)-(γ,p) equilibrium will be established between 99Cd and 100In and the 100In(p,γ) reaction rate would affect the A = 99 production, while for larger Sp(100In) the 99Cd(p,γ) reaction rate might be more relevant. Therefore, the mass uncertainties should be addressed first. The presented results are relevant for any rp-process scenario with a reaction flow through the 99Cd region. Here, an X-ray burst model has been used to investigate in detail the impact of the present measurements on such an rp process. The νp process in core collapse supernovae might be another possible scenario for an rp process in the 99Cd region. It it is planed to also explore whether in that case mass uncertainties have a similar impact on the final composition. On the neutron-rich side of the valley of stability for the Cd and Ag chains of nuclides, the r process has not yet been reached. Further technical development on suppression of contaminants are required. This includes improvements on the ISOLDE side, e.g., by improving the selectivity of the transfer line or on the ISOLTRAP setup by implementing an electrostatic ion beam trap for a fast and efficient isobaric selection. Nevertheless the obtained results contribute to the knowledge of nuclear structure. The trends in the two-neutron separation-energy S2n and the interaction between the last neutrons and last protons ΔVpn were corrected to more smooth evolutions, as already seen in other regions of the nuclear chart. The strongest corrections have been observed for even-N nuclides, were more new experimental data are available. Thus, new measurements on odd-N nuclides are suggested. This also is underlined by the trends observed in the Garvey-Kelson relations for the neutron-rich Cd nuclides. Furthermore, it has been shown, that the prominent structure of the ΔVpn for an entire chain of nuclides including inflexion points can be reproduced by using simple relations between quantum numbers of the occupied orbits. This approach connects ten values for each nuclide with only one adjusted parameter. This has been investigated for 63 ΔVpn values of even-even nuclides in the vicinity of Z = 50 and 50 ≤ N ≤ 82. The simple model works remarkably well for the elements Cd, Sn, and Te. Small deviation have been observed for the Xe and Pd nuclides which were explained with the limitations of the model to the vicinity of the close shells, where the nuclides have only few valence protons and neutrons.
The experimental determination of the electron energy distribution of a low pressure glow discharge in neon from emission spectroscopic data has been demonstrated. The method extends an approach by Fischer and Dose [5]. The spectral data were obtained with a simple overview spectrometer and analyzed using a strict probabilistic, Bayesian data analysis. It is this Integrated Data Analysis (IDA) approach, which allows the significant extraction of non-thermal properties of the electron energy distribution function (EEDF). The results bear potential as a non-invasive alternative to probe measurements. This allows the investigation of spatially inhomogeneous plasmas (gradient length smaller than typical probe sheath dimensions) and plasmas with reactive constituents. The diagnostic of reactive plasmas is an important practical application, needed e.g. for the monitoring and control of process plasmas. Moreover, the experimental validation of probe theories for magnetized plasmas as a long-standing topic in plasma diagnostics could be addressed by the spectroscopic method.
Turbulence is a state of a physical system characterized by a high degree of spatiotemporal disorder. Turbulent processes are driven by instabilities exhibiting complex nonlinear dynamics, which span over several spatial as well as temporal scales. Apart from fluids and gases, turbulence is observed in plasmas. While turbulent mixing of a system is sometimes a desired effect, often turbulence is an undesired state. In hot, magnetically confined plasmas, envisaged for energy generation by thermonuclear fusion, plasma turbulence is clearly a problem, since the magnetic confinement time is drastically deteriorated by turbulent transport. Hence, a control mechanism to influence and to suppress turbulence is of significance for future fusion power devices. An important area of plasma turbulence is drift wave turbulence. Drift waves are characterized by currents parallel to the ambient magnetic field, that are tightly coupled to a coherent mode structure rotating in the perpendicular plane. In the present work, the control of drift waves and drift wave turbulence is experimentally investigated in the linear magnetized helicon experiment VINETA. Two different open-loop control systems - electrostatic and electromagnetic - are used to drive dynamically parallel currents. It is observed that the dynamics of the drift waves can be significantly influenced by both control schemes. If the imposed mode number as well as the rotation direction match those of the drift waves, classical synchronization effects like, e.g., frequency locking, frequency pulling, and Arnold tongues are observed. These confirm the nonlinear interaction between the control signal and the drift wave dynamics. Finally, the broadband drift wave turbulence, and thereby turbulent transport, is considerably reduced if the applied control signal is sufficiently large in amplitude.
In this work, the investigation of dusty plasma by means of tunable diode laser spectroscopy was carried out. Special interest was focused on the interactions of dust particles and metastable atoms. At first, Al density and temperature in dc and pulsed magnetron discharges were measured. Measurements with argon as working gas show an expected behavior of the measured atom density and temperature. Decrease of absorption signal was observed in argon/oxygen and argon/methane mixtures. A small admixture of oxygen leads to a complete disappearance of the absorption signal indicating vanishing Al atom density. The effect is believed to be caused by the oxidation of the magnetron target. This decrease reveals typical hysteresis behavior caused by poisoning of the target. Significant difference between critical oxygen flow value in dc and pulsed modes was registered. Then dust formation and plasma behaviors in hydrocarbon containing plasmas were analysed. The dust growing plasmas (Ar/C2H2, Ar/CH4 and Ar/C3H6 rf plasmas) were characterized by laser transmission and scattering methods, ion energy distribution function and mass spectrum evolution by plasma processing monitor, and the spatial distribution in pristine plasma and the temporal behavior of the metastable atom density in processing plasma using TDLAS. Pristine plasma were then characterized in term of metastable density and temperature. The radial distribution of neon metastable atom density in capacitive coupled rf discharge can be approximated to a Gaussian profile with the width smaller than plasma chamber radius. The diffusion flow of metastable atoms deduced from their spatial density distribution gives the loss of metastable atom in the plasma sheath. Argon metastable density was measured in rf plasma and compared with a simple model for metastable density. The model explains well the trend of metastable density with respect to the change of plasma input power. Metastable density of dusty plasma with injected dust particles was measured and compared to that of pristine plasma. The particle heating by metastable atoms was strongly evidenced. The power absorbed by dust particles due to bombardment of metastable atoms onto a dust particle surface in our experiments is about 0.04 Wm-2 for the low dust density case and lower for higher dust density which is in the same order as the contributions of kinetic energy of ions and electrons and the energy released by their recombination on the grain surface. The influence of dust particle density and size on metastable density was studied. Through measuring metastable density, TDLAS can be used as a tool to study the dust growth process in processing plasma.