Twenty Five Years of Vibrational Kinetics and Negative Ion Production in H2 Plasmas: Modelling Aspects

Different approaches to study vibrational kinetics coupled to electron one for modeling different kinds of negative ion sources are presented. In particular two types of sources are investigated. The first one is a classical negative ion source in which the plasma is generated by thermoemitted electrons; in the second one, electrons already present in the mixture are accelerated by an RF field to sufficiently high energy to ionize the gas molecules. For the first kind of ion source a new computational scheme is presented to couple heavy particle and electron kinetics. Moreover models developed for an RF inductive discharge and for a parallel plate discharge are described. Introduction The development of high current negative ion sources for generating intense neutral beams represents an important topic in the thermonuclear research field. Negative hydrogen ions are very important for the generation of neutral beams for heating magnetically confined fusion plasmas. Negative ion based neutral beam injector (N-NBI) is required because negative ion beams present a higher neutralization efficiency respect to positive ion ones at high energies. Different kinds of negative ion sources are studied both from theoretical and experimental point of view, in order to optimize negative ion production. At present the volume plasma sources are widely used for H generation, even though the addition of alkaline metals increases the concentration of negative ions. Vibrational Kinetics in H2 Plasmas Vibrational kinetics of molecular hydrogen under plasma conditions started his long way many years ago in an attempt to understand alternative mechanisms in the dissociation of H2 under electrical discharges. Molinari and coworkers in an attempt to rationalize their experimental dissociation rates introduced the concept of a vibrational temperature higher than the translational one able to increase the dissociation of H2 by heavy particle collisions . This mechanism could prevail over the dissociation mechanism by electron impact in the cases in which this well known mechanism failed to describe the dissociation process. Capitelli et al. building up a vibrational kinetic model mutuated by the laser community first tried to implement the Molinari's ideas. The first model was very simple including the pumping of vibrational quanta through the e-V (electron-vibration energy exchange) processes and the redistribution of vibrational quanta on the vibrational ladder by V-V (vibration-vibration) and V-T (vibration-translation) energy transfer processes. V-V up pumping mechanism or Treanor's mechanism was able to bring the vibrational quanta up to the dissociation limit leading to dissociation rates higher than the dissociation rate induced by electron impact. This mechanism was called a laser type mechanism or pure vibrational mechanism (PVM). This kind of name was given because the vibrational distribution of molecular hydrogen presented a long plateau typically met in CO lasers. The enthusiasm in the new mechanism was strongly damped when becomes clear that atomic hydrogen should be considered a killer of vibrational level populations. V-T deactivation of vibrationally excited hydrogen molecules, not considered in Ref.3, presented indeed large rates which strongly destroy the vibrational quanta introduced by e-V processes. In a subsequent work the V-T term due to atomic hydrogen was inserted in the vibrational master equation having as result a strong decrease of the vibrational plateau and a consequent loss of importance of the pure vibrational mechanism in the dissociation process. Only at very high electron density i.e. at high vibrational quanta pumping rates the pure vibrational mechanism could be effective in dissociating H2 molecules. On the other hand large plateaux again arose as result of the recombination process which selectively pump vibrational quanta on the top of vibrational ladder. Despite the strong deactivating effect of atomic hydrogen on the vibrational distributions of H2 the research continued. In particular the authors of Ref.5 developed a joint vibrational dissociation mechanism (JVD) which included both the pure vibrational mechanism (PVM) and the electron impact dissociation (DEM) mechanism in the same model. Moreover the DEM model considered the dissociation transitions by electron impact from each vibrational level of vibrational ladder. The same authors realized the possibility of non-Maxwellian distribution functions for the electrons so that a Boltzmann equation for the electron energy distribution function (eedf) was solved to get the actual eedf. Soon after the importance of second kind collisions involving vibrationally excited H2 molecules and electrons in affecting eedf was realized as well as the dependence of eedf on the presence of atomic hydrogen. For long time only superelastic vibrational collisions were considered in the kinetics; recently however two groups of research emphasized the role of metastable electronic states in forming structures in the eedf of H2 plasmas. At the same time a large effort was done to calculate electron impact dissociation and ionization cross sections involving each vibrational level of the H2 vibrational manifold , a problematic which is still under study nowdays. Three independent codes were built up in Europe mainly in Bari (Gorse), Lisboa (Loureiro and Ferreira) and Saint Petersburg (Baksht). The research in the field shows a sharp increase when Bacal and Hamilton discovered the presence of large concentrations of negative ions (H) in multicusp magnetic plasmas. It was soon realized that the mechanism at the basis of the creation of negative ions in this kind of research was dissociative attachment from vibrationally excited molecules. The relevant cross sections infact were shown to dramatically depend on the vibrational quantum number. The development of negative ion sources for fusion applications gives new impetus to the research in the field of vibrational kinetics in H2 plasmas from both experimental and theoretical points of view. Different codes (Bari-Gorse, Palaiseau-Bacal, Livermore-Hiskes, Japan-Fukumasa) were built up to describe the complex phenomenology occurring in the plasma. In particular Hiskes discovered a new elementary process for exciting the whole vibrational manifold of H2 the so called E-V process which consists in the excitation of singlet electronically excited states of H2 followed by radiative decay on the ground state. Corresponding cross sections first calculated by Hiskes have been then refined by Celiberto et al.. Different diagnostics were used to monitor the quantities entering in the model i.e. concentration of H, H, Hn + species as well as electron number density and electron temperatures. At the same time sophisticated experiments were dedicated to measure the vibrational distribution (Essen-Dobele), the eedf (Hopkins-Graham-Dublin) and the ratio between cold and hot atomic hydrogen (Sultan-Orsay). Atomic and molecular physics methods were also implemented to shed light on different elementary processes important to understand the physics of these discharges. In particular Laganà et al. and Esposito et al. presented complete sets of H-H2(v) rate coefficients, while Billing and Cacciatore 23 start their pioneristic work on the interaction of vibrationally excited states with copper surfaces and on the recombination of atomic hydrogen on the same surface. The last process indeed is important also for pumping vibrational energy in the molecules formed during heterogeneous recombination and finally desorbed by the surface. Atomic hydrogen changes his role from being the killer of vibrationally excited molecules to being a source of them. On the other hand the experimental works of Hall et al. and of Eenshuistra et al. at the end of 1980s confirmed the production of vibrationally excited H2 molecules during atom recombination thus starting a topics of large actuality at the present. It should be noted that the synergy between atomic/molecular physics and vibrational plasma kinetics has contributed to the advancement of negative ion production reactors. We believe indeed, that without the work of Wadehra and Bardsley on the dependence of dissociative attachment cross sections on the vibrational quantum number and the work of Hiskes on the pumping of vibrational states by high energy electrons through the E-V process the concept of hybrid reactor for the formation of negative ions could not be probably developed. Nowadays another elementary process that one involving dissociative attachment on Rydberg states could push the negative ion community towards the development of other kind of reactors. Negative ion sources are still a fascinating topic in which vibrational kinetics constitutes one of the most important aspects. At the beginning of 90s the H2 plasma community was attracted to other kind of reactors those used in material science. In particular the microwave discharges used for the production of diamond films and parallel plate RF discharges for microelectronics acquired a noticeable technological importance. Also in this case two sophisticated codes were developed in Villateneuse and Bari for finding the optimum conditions in these reactors. At the same time a complete kinetic model was developed by Matveyev et al. for describing high energy H2 plasma expansion. In the last 5 years a new impetus on vibrational kinetics arose for handling problems met by fusion people in the divertor plasma. In particular the increase of ionic recombination assisted by vibrational kinetics at the edge of the divertor as well as the use of Monte Carlo simulation for Hion and neutral transport are topics of current interest. In particular our team is trying to improve the numerous input data entering in the kinetics describing the negative ion production in multipole, microwave and parallel plate reactors. Details of this modeling will be reported in the present lecture. Negative Ion Sources Different methods can be used to generate plasmas. In this section we consider two different kinds of plasma sources that can be used to generate negative hydrogen ions. In multicusp ion sources the plasma is produced by high energy electrons, emitted by hot filaments and accelerated by the negative voltage between the filaments and the source wall. The flux of accelerated electrons impinges on the H2 target and gives rise to molecular and atomic reactions such as ionization, excitation and dissociation. In particular, negative ion production in the plasma volume is driven by the dissociative attachment of low energy electrons to highly vibrationally excited molecules. On the other hand high energy electrons cause H destruction through electron detachment. This production mechanism evidences the necessity of dividing the source into two regions where electron energy distributions are optimized respectively for the plasma excitation and the dissociative attachment process. This separation is reached in multicusp ion sources by two kind of “filters”: the first one acts separating spatially electrons with different energies by means of a magnetic filter; the second one, realized pulsing the discharge (temporal filter), creates different electron energy distributions at different times. Multicusp ion sources present some technical limitations due to the damage of the emitting filaments and their evaporation that causes also a contamination and then a variation in the operating conditions. These problems are absent in a RF discharge: because of their low mass, electrons already present in the gas can be easily accelerated by the electric field to energies which are sufficient to ionize a gas molecule, then generating the plasma. Plasma Kinetic Modeling In order to model kinetically a plasma and in particular a negative ion source we need to solve the vibrational kinetics of H2(v), the dissociation kinetics of H2(v), the kinetics of electronically excited states of H2 and H, the ion kinetics. The time evolution of the heavy particle densities can be described by a set of nonlinear differential equations that reads as: € dNv dt       = dNv dt       e−V + dNv dt       E−V + dNv dt       V −V + dNv dt       V −T + dNv dt       e−D + dNv dt       e−I + dNv dt       e−da + dNv dt       e−E + dNv dt       wall (1) where each term on the right hand side represents the gain or the loss for the v-th species due to a specific reaction and is given by: € ± dNv dt       TOT = ki ⋅ N j j=1 react ∏ 