Search

Beam tomography is a non-invasive diagnostic that allows us to reconstruct the beam emission profile by measuring the light emitted by the beam particles interacting with the background gas, along an elevated number of lines of sight, which is related to the beam density by assuming a uniform background gas. In the framework of the heating and current drive of future nuclear fusion reactors, negative ion beams of hydrogen and deuterium are required for neutral beam injectors (NBIs) due to their elevated neutralization efficiency at high energy (in the MeV range). Beside the beam energy, beam divergence and homogeneity are two critical aspects in the design of future NBIs. In this paper, the characterization of the negative ion beam of the negative ion source NIO1 (a small-sized radio-frequency driven negative ion source, with 130 mA of total extracted H - current and 60 kV of maximum acceleration) using the tomographic system composed of two visible cameras is presented. The Simultaneous Algebraic Reconstruction Technique (SART) is used as an inversion technique to reconstruct the 3 × 3 matrix of the extracted beamlets, and the beam divergence and homogeneity are studied. The results are compared with the measurements of the other diagnostics and correlated with the source physics. The suitability of visible cameras as a diagnostics system for the characterization of the NIO1 negative ion beam is a small-scale experimental demonstration of the possibility to reconstruct more complicated multi-beamlet profiles, resulting in a powerful diagnostic for large NBIs.

A systematic analysis of the charge state distribution (CSD) of each element (or isotope) X extracted from a highly charged ion source as a function of control variables or time needs properly automated tools for the large amount of data involved and the possibility of peak superposition, due to the mass separator properties. Some key features of identification of ion currents with the new program DAT2PEAK are discussed here. A preprocessor LECR2DAT reads files of experimental scans. To obtain a reliable identification of element X and charges i several filtering and calibration routines were necessary; principles of their robust algorithms are discussed. Moreover, the concept of constrained fit allows one to use information both from superposed peak and single ion peak and to select different fit laws of CSD for each element X. A robust rule to distribute the current of superposed peaks to related elements is also presented. For each measurement processed, DAT2PEAK produces a listing of peaks and CSDs and updates the history files for the convenience of following statistical analysis.

The ELTRAP device installed at the Department of Physics of the University of Milan is a Malmberg–Penning trap, with a magnetic field up to 0.2 T, equipped with charge coupled device optical diagnostics. It is intended to be a small scale facility for electron plasma and beam dynamics experiments, and in particular for the study of collective effects, equilibrium states, and the formation of coherent structures in these systems. The device features a relatively long solenoid, corrected by 4 shims and 16 dipole coils, in order to obtain a large uniform magnetic field region. The modular electrode design allows several variations of the experimental configuration. The first experiments which assess the operation of the facility are described. Plasma confinement times up to several minutes have been obtained and an electron temperature of 4–8 eV has been measured.

Charge breeding inside electron cyclotron resonance (ECR) ion sources is a promising way to produce radioactive ion beams with a high charge state, based on deceleration and capture of the primary ion beam by the ECR plasma, trapping, multi-ionization, diffusion, and extraction. A detailed analysis of these processes could help to improve the efficiency (reducing ion loss to walls) and the charge breeding time. An estimate of the deceleration length is given. Collisions, external magnetic field, and ambipolar potential are the fundamental parameters for the ion motion inside the trap, which is simulated by a program, including statistical averaging over many (400–3200) particle histories. Characteristics of trapped and extracted ions are shown with simulations taking in account the design of the PHOENIX ion source. Since particle losses are found to increase rapidly with the radius of injection, the injection system should form an image at the beam stopping plane. Analysis of additional source parameters is discussed.Charge breeding inside electron cyclotron resonance (ECR) ion sources is a promising way to produce radioactive ion beams with a high charge state, based on deceleration and capture of the primary ion beam by the ECR plasma, trapping, multi-ionization, diffusion, and extraction. A detailed analysis of these processes could help to improve the efficiency (reducing ion loss to walls) and the charge breeding time. An estimate of the deceleration length is given. Collisions, external magnetic field, and ambipolar potential are the fundamental parameters for the ion motion inside the trap, which is simulated by a program, including statistical averaging over many (400–3200) particle histories. Characteristics of trapped and extracted ions are shown with simulations taking in account the design of the PHOENIX ion source. Since particle losses are found to increase rapidly with the radius of injection, the injection system should form an image at the beam stopping plane. Analysis of additional source parameters ...

A sharp transition to a space-charge dominated regime is induced in a low-energy electron beam produced in a Malmberg–Penning trap by increasing the emission current of the source. The transition is characterized by the appearance of a region, around the axis of the beam, not accessible to beam electrons, and by the fast development of coherent structures in the remaining electron plasma, due to the sharp increase of local vorticity. The results are interpreted in the framework of a cold fluid drift–Poisson model, and using a three-dimensional particle-in-cell simulation code.

The ion cooling exerted by lighter ions inside an ECR plasma is often credited as an explanation of the well-known gas mixing effect, since the ion temperature is about a constant Tp for all the ion species in the plasma. Here the energy balance equations are written in detail for any ion species, including the improved evaporation cooling in the collisional confinement regime. A formula for Tp and decoupled equations for small corrections around Tp are given, accounting for the colder gas input. The charge state distribution are written in closed form. Gas flow between plasma center and periphery is described. Some equilibria for most common gas mixing cases are compared, showing in particular the efficacy of neon.

The electron cyclotron resonance ion source Alice is a small 14.4 GHz ion source intended to be a general purpose ion source for the ALPI ion linac. To increase the speed of emitted ions to more than 0.009 c it was installed on a high voltage platform. Noteworthy details of the platform complex such as the 400 kV insulation transformer with the power line connecting it to the platform and the acceleration tube are reviewed. The whole complex can operate up to 350 kV and up to 36 μA of extracted current, the current dispersed is small and upgrading to 400 kV looks possible. Details of the beam current, ion source optimization, source emittance and beam transport are discussed and show good agreement between experiments and calculated beam sizes.

Kinetic equilibria implications on the ionization efficiency ηi of electron cyclotron resonance (ECR) ion sources when used as a charge breeder are discussed. The ion density distribution n(i) where i is the ion charge state is here expressed by a properly factorized formula, which allows analytic consideration. Agreement with some numerical simulation is excellent. In the case of low ion recombination rates, a detailed solution is given, and some relevant approximation of ionization rate Si and ionization energies are proposed. A useful approximation log(Si)=a−(cs−k3)i is found, with a and cs mainly related to ionization energies and macroscopic plasma parameters, such as the hot electron density nh and temperature Th, while k3 is related to the ambipolar electrostatic potential. Formulas for the charge state at which efficiency is maximum and for the maximum value are written in detail.

NIO1 (Negative Ion Optimization 1) is a versatile ion source designed to study the physics of production and acceleration of H- beams up to 60 keV. In ion sources, the gas is steadily injected in the plasma source to sustain the discharge, while high vacuum is maintained by a dedicated pumping system located in the vessel. In this paper, the three dimensional gas flow in NIO1 is studied in the molecular flow regime by the Avocado code. The analysis of the gas density profile along the accelerator considers the influence of effective gas temperature in the source, of the gas temperature accommodation by collisions at walls, and of the gas particle mass. The calculated source and vessel pressures are compared with experimental measurements in NIO1 during steady gas injection.

Electromagnetic traps are a flexible and powerful method of controlling particle beams, possibly of exotic nuclei, with cooling (of energy spread and transverse oscillations) provided by collisions with light gases as in the Radio Frequency Quadrupole Cooler (RFQC). A RFQC prototype can be placed inside the existing Eltrap solenoid, capable of providing a magnetic flux density component B(z) up to 0.2 T, where z is the solenoid axis. Confinement in the transverse plane is provided both by B(z) and the rf voltage V(rf) (up to 1 kV at few MHz). Transport is provided by a static electric field E(z) (order of 100 V/m), while gas collisions (say He at 1 Pa, to be maintained by differential pumping) provide cooling or heating depending on V(rf). The beamline design and the major parameters V(rf), B(z) (which affect the beam transmission optimization) are here reported, with a brief description of the experimental setup.

The operation of the electron cyclotron resonance (ECR) ion source Alice is described, with attention to the improved ion analysis and the new diagnostic equipment. Ion current peaks as small as tens of nA can be measured; resolving power is about 100. Routine operation is still restricted to gases, but prototypes of ovens have passed the first tests. Since currents are still rather limited (about 400 nA for 132Xe18+), a new removable plasma chamber with increased inner diameter (63 mm), made in aluminum, is under construction. Preliminary analysis of charge state distribution of xenon isotopes shows good fits with an approximately thermodynamical distribution; a sort of effective chemical potential μc is defined from fit parameters; in the particular case shown here μc≅390 eV. A decision to install a high voltage insulation transformer to allow platform operation up to Vp=350 kV was made, and an adequate location is being found, with proper wall feedthroughs in design.

A metal vapor vacuum arc (Mevva) ion source was finally installed and operated inside the electron cyclotron resonance (ECR) ion source Alice; Mevva cathode was made of Al, while ECR buffer gas was nitrogen, and among residuals, argon. Installation, hardware control, and improvements of diagnostic are discussed. Magnetic analysis of the ECR emitted current shows the presence of an Al3+ peak of about 5–10 μA following Mevva pulse end after 15 μs, against a steady state current of 10 nA. Peak width is about 5 μs, which is the limit of our current amplifier. Peak remains when ECR is turned off. Application of a bias potential of ±46 V to the Mevva anode shows fairly small effects on time of flight, which may indicate a large kinetic energy of Mevva ions (280 eV) at the end of Mevva shot. Bias voltage and ECR operation proved compatible only at low power, in present installation; then, some or most of Mevva ions are not trapped inside the ECR.

It is a known fact that in electron cyclotron resonance ion sources, plasma impurities play a role in the formation of the ion charge state distribution (CSD). Writing the balance equations for the ion densities in a matrix form allows us to calculate CSD easily for plasmas containing an arbitrary number of elements, taking into account electron ionization, radiative recombination, charge exchange between ions, and ion flowing out of the magnetic trap. Moreover, by defining two global quantities—S0 and Sc—representative of charge exchange, the stationary case can be reduced to the resolution of only one nonlinear equation for S0 and Sc. Therefore plasma equilibria even with several species of impurity ions (carbon, oxygen, hydrogen, etc.) can be easily described and solved in this model. The application to the special case of no ion flow is discussed, to study a case similar to thermodynamic equilibrium. An additional factor exp{μ i4/3}, where i is the charge state, is shown to be necessary to fit the CS...

A miniaturized Mevva source, designed to be inserted in an electron cyclotron resonance (ECR) source in order to sputter metals, was proposed to supplement our ECR source Alice. The major problem was to control the Mevva arc at the low level of current (20–30 A for Al, Cu), which are required to limit metal consumption. Installation of our Mevva in a separate test stand was completed. Extensive electrical tests are here reported, discussing components that were rebuilt and optimized during experiments. The improvement in trigger voltage (10 kV) and rise time (3 μs) obtained by an air transformer, with little help from a ferrite core, allowed a reliable arc start‐up. Increasing the main voltage was also important, as was evident from the new ferrite transformer (giving 3.5 kV instead of 2 kV) and from the detailed shape of arc voltage in some shots, where wide fluctuation appeared. A neutral beam spectrometer for the MetAlice experiment is being developed. Perspective of schematic improvement and of installation in Alice are discussed.

The propagation of the transverse microwave Ex Ey into electron cyclotron resonance (ECR) plasmas is studied (with simplified plasma dynamics), emphasizing that plasma dimensions and wavelength λ are comparable, which is the case with most ion sources. The coupling to the Ez field is included. A simple system of coupled equations is found and a closed fourth‐order differential equation is obtained. Nonpropagating modes are easily solved. Numerical solutions are shown for propagating modes applied to the ECR Alice dimension (radius 3 cm, λ≂2 cm). A simplified formula for the cavity admittance is given, involving the matching of the guide to cavity, where each mode contribution is weighted; weights are computed for the circular waveguide case.

The particular microwave off‐axis injection into the electron cyclotron resonance (ECR) ion source Alice, the reflected power and the occasionally observed waveguide arcs, required a careful study both of the waveguide termination and the microwave transmission line. The high‐voltage microwave window was redesigned, improving the choke flange and showing the opportunity of treating the teflon insulator to obtain a surface resistivity in the range of 107 Ω. A simple plasma model, unidimensional and based on the Krook model for electron motion, is here considered for the ECR plasma; effects of rapid spatial variation of the external magnetic field are naturally included. Profiles of the waves in the ECR cavity are given, as well as an expression for the impedance of the plasma; no substantial reflections are evidenced for a typical plasma case (ωpl/ω=0.7).

Two linear trap devices for particle beam manipulation (including emittance reduction, cooling, control of instabilities, dust dynamics, and non-neutral plasmas) are here presented, namely, a radiofrequency quadrupole (RFQ) beam cooler and a compact Penning trap with a dust injector. Both beam dynamics studies by means of dedicated codes including the interaction of the ions with a buffer gas (up to 3 Pa pressure), and the electromagnetic design of the RFQ beam cooler are reported. The compact multipurpose Penning trap is aimed to the study of multispecies charged particle samples, primarily electron beams interacting with a background gas and/or a micrometric dust contaminant. Using a 0.9 T solenoid and an electrode stack where both static and RF electric fields can be applied, both beam transport and confinement operations will be available. The design of the apparatus is presented.

Selfconsistent laminar flow models, which enable to predict the optimal cathode and anode geometry in simple diodes, must be modified to account for the anode aperture and the effect of other electrodes. An equation for charge coupled to arbitrary laminar flows is here first presented and its numerical solutions are obtained with a new method, based on mesh transformations. It is found that a close match to theoretical flows requires an increase of the simple diode voltage v0 by an amount vδ, which, for a typical case designed for zero exit angle condition, are v0 = 0.7465 and vδ = 0.0294 in adimensional units. States "in" and "out" for the anode lens are also shown, where "out" is a new and nonlinear solution for the beam expansion in a drift tube.

Flow of particles in diodes is solved selfconsistently assuming an approximated system of flow lines, that can be easily represented by an analytic transformation in a complex plane, with assumed uniformity in the third spatial direction. Beam current compression is tunable by an angle parameter α(0); transformed coordinate lines are circular arcs, exactly matching to the curved cathode usually considered by rectilinear converging flows. The curvature of flow lines allows to partly balance the transverse effect of space charge. A self-contained discussion of the whole theory is reported, ranging from analytical solution for selfconsistent potential to electrode drawing to precise numerical simulation, which serves as a verification and as an illustration of typical electrode shapes. Motion and Poisson equation are written in a curved flow line system and their approximate consistency is shown to imply an ordinary differential equation for the beam edge potential. Transformations of this equation and their series solutions are given and discussed, showing that beam edge potential has a maximum, so supporting both diode (with α(0) =/~ π∕3) and triode design. Numerical simulations confirm the consistency of these solution. Geometrical details of diode design are discussed: the condition of a zero divergence beam, with the necessary anode lens effect included, is written and solved, as a function of beam compression; accurate relations for diode parameters and perveance are given. Weakly relativistic effects including self-magnetic field are finally discussed as a refinement.

Radiofrequency heating for ICP (inductively coupled plasma) ion sources depends on the source operating pressure, the presence or absence of a Faraday shield, the driver coil geometry, the frequency used, and the magnetic field configuration: in negative ion source a magnetic filter seems necessary for H(-) survival. The result of single particle simulations showing the possibility of electron acceleration in the preglow regime and for reasonable driver chamber radius (15 cm) is reported, also as a function of the static external magnetic field. An effective plasma conductivity, depending not only from electron density, temperature, and rf field but also on static magnetic field is here presented and compared to previous models. Use of this conductivity and of multiphysics tools for a plasma transport and heating model is shown and discussed for a small source.

The propagation of a charged particle beam is accompanied by the production of secondary particles created in the interaction of the beam itself with the background gas flowing in the accelerator tube. In the drift region, where the electric field of the electrodes is negligible, this production results in an accumulation of space charge and a consequent creation of a plasma shielding the self induced potential of the charged beam. This phenomenon, known as space charge compensation (SCC) is a typical issue in accelerator physics, where it is usually addressed by means of 1D radial transport codes [1] or Monte Carlo codes [2]. The present paper describes some theoretical studies on this phenomenon, also presenting a PIC-Monte Carlo code investigating the space charge neutralization of a beam in a 2D domain on the zx plane, z being the beam axis. With respect to the classical 2D xy transversal simulations, the particular geometry chosen allows to investigate the beam parallel dynamics of plasma particles. A two time scale approach also gives the possibility to analyze the transient phase of the SCC as well as the mechanisms regulating the phenomenon at regime. These peculiarities of the model, offering a new insight on the problem, are described and the results discussed. REFERENCES 1. N. Chauvin et al. "Beam dynamic simulation of the low energy beam transport line for IFMIF/EVEDA". In Proceeding of LINAC Conference 2008. 2. J. L. Vay and C. Deutsch. "Charge compensated ion beam propagation in a reactor sized chamber". Physics of Plasma, 5, 1998.

The resistive wall instability of the l=2 diocotron mode in a pure electron plasma has been investigated with a systematic variation of the parameters of the external impedance connected to a pair of sectored electrodes. The measured growth rate is well described by a linear perturbation theory of the two-dimensional drift-Poisson system.

Modeling of negative ion sources proposed for spallation sources or neutral beam injectors must include several ion production mechanisms, in particular, for wall production, the initially backward ion motion must be reversed in the presheath to obtain extraction. An adequate analytic simplification of the full Vlasov-Boltzmann equations to a one space dimension model for the investigation of the transport effects of the presheath voltage is described here. Integrodifferential collision dominated transport equations are written and solved for typical profiles of the presheath voltage. The results for space charge density improve previous estimates. A transverse magnetic field can be included in the formalism to treat electron transport: principles of its solution are summarized and numerical examples are given; plasma quasineutrality is approximately verified.

The effect of an externally imposed time-varying quadrupole electrostatic perturbation on a pure electron plasma trapped in a Malmberg-Penning trap is investigated experimentally. A resonance occurs when the frequency of the drive matches the frequency of the azimuthal l=2, kz=0 mode. Very good agreement is found between the experimentally determined resonance frequency and the real part of the frequency calculated from a two-dimensional linear model for a top-hat density profile, while nonlinear deviations for the mode amplitude are found especially close to the resonance.

The injection phase in a Malmberg-Penning trap is investigated both experimentally in the ELTRAP [M. Amoretti et al., Rev. Sci. Instrum. 74, 3991 (2003)] device, and numerically. The resulting plasma density distribution is studied by varying the source parameters, the external magnetic field strength, and the axial position of the external potential barrier. Space charge phenomena dominate the dynamics of the system; formation of hollow plasma columns and three-dimensional structures are observed. The processes are interpreted using a three-dimensional particle-in-cell code which solves the drift-Poisson system.

The off-axis bulk rotation (l=1 diocotron mode) of an electron plasma column confined in a Malmberg-Penning trap is strongly destabilized by a small population of positive ions formed by energetic electron-neutral collisions. The instability, known as ion resonance instability, drives the plasma against the wall, destroying the confinement. A new experimental technique based on the static or time dependent application of low voltages to the inner conductors of the trap is shown to be effective in controlling the instability. The efficiency of the control technique is experimentally investigated by a systematic variation of the amplitudes, time duration, and periodicity of the additional potentials.

In electron cyclotron resonance ion sources (ECRISs) the neutral gas density ng is typically smaller than the plasma density, which is the opposite of many other source cases, for example, radio-frequency ion source for H (plus or minus) production. For the latter, a global model of equilibrium between ambipolar diffusion and ionization gives a firm prediction of the electron temperature Th. With the purpose of obtaining firm predictions also for ECRIS plasmas, the difficulty of a highly charged ion distribution and of a more complicated model of ambipolar potential ϕ is here taken into account. Using an approximate solution for the highly charged ion density, the global balance equations are written and discussed as a function of ϕ, of the gas density ng, and of the electron temperature Th. Some examples of the application of equilibrium balance laws to a typical ECRIS are given.