Design and preliminary tests of an Active Plasma Chamber for ECR Ion Sources

An innovative plasma chamber for Electron Cyclotron Resonance Ion Sources (ECRIS) has been developed at INFN and will soon be installed and tested with the AISHa (Advanced Ion Source for Hadrontherapy) ion source. It consists in inserting a particular liner into the existing chamber, which allows an electrical segmentation of the internal walls of the chamber. The purpose of this system is to reduce the ion losses induced by the anisotropic diffusion mechanism, to improve the plasma confinement and thus to increase the overall performance of the ion source. In fact, in ECRIS plasmas, electrons mostly diffuse along magnetic field lines while ions mostly leak across the same lines. In particular, the inner walls of the plasma chamber are covered with 30 tiles, each one polarized to a proper positive voltage. The tiles are made of Al-6082 and anodized except for the surface directly facing the plasma. The anodizing process makes each tile electrically insulated from the others and from the plasma chamber while preserving the correct operation of the cooling system. The tiles are wrapped by 2 half-cylinders made of Al-6082 acting as shells. Some tiles are equipped of a temperature sensor and machined to allow the wiring of the entire system. In this work the results of the preliminary tests of the thermal and electrical behaviour of the active chamber and the future perspectives are presented.


Introduction
Electron Cyclotron Resonance Ion Sources (ECRIS) are plasma-based highly performing ion sources able to produce high intensity beams of highly charged ions [1].Ions are extracted from a high-density, high-temperature plasma (ne∼ 10 17 -10 19 cm −3 , Te ∼ 0.1-100 keV) produced through electron cyclotron heating and confined by a magneto-hydro-dynamically stable B-min magnetic field.
The mean charge state <q> of the ions that composes the ECRIS plasma and the extracted beam is proportional to the ion lifetime within the plasmas,   .Indeed, generation of highly charged ion beam occurs only if plasma is well-confined.
Plasma confinement in ECRIS is guaranteed by the superposition of three different (but overlapping) physical phenomena: magnetic confinement, double-layer confinement and diffusional confinement (Fig. 1).
Figure 1: plasma confinement in an ECRIS chamber due to the overlap of three physical phenomena: magnetic confinement (through axial and hexapolar solenoids), double-layer confinement (through the electron cyclotron heating within the ECR layer, i.e. the plasmoid surface) and diffusional confinement (through mainly axial and radial charge losses, regulated by quasi-neutrality).
The magnetic confinement is the main responsible for the plasma confinement in ECRIS.Due to magnetic moment and energy conservation, most of the electrons are reflected at the electron cyclotron resonance surface.B-minimum magnetic configuration guarantees the confinement of particles by means of a magnetic force proportional to the magnetic field gradient.Roughly, the ions lifetime   ∝ ln(  /  ) where   and   are the maximum and minimum magnetic fields in the plasma chamber [1].More information about magnetic confinement and scaling laws in ECRIS con be found in Refs.[2][3][4].
The double-layer confinement is indirectly caused by the electron cyclotron heating within the ECR layer.High energy electrons populate just a thin slab overlapping the ECR layer, while their density drops down of more than one order of magnitude outside.Ions, instead, diffuse across the electron layer due to their high collisionality.This physical condition establishes a double-layer (DL) configuration which self-consistently originates a potential barrier.This "barrier" confines the ions inside the plasma core [5][6][7].Double layer confinement explains why closed ECR surfaces are required to improve plasma confinement and ion lifetime.
Plasma diffusion plays a relevant role in establishing ion lifetime too.Due to the presence of an external magnetic field, diffusion in ECRIS is not ambipolar [8].Simon diffusion is the typical diffusion mechanism in ECRIS [9,10].Strongly magnetized electrons diffuse along magnetic field lines.Unmagnetized ions diffuse isotropically.Due to the plasma chamber geometry, a net negative current flow along the axis; a net positive current is lost along the radial direction (Fig. 1).Due to plasma quasineutrality, positive and negative loss fluxes must compensate.The plasma potential adjusts itself to equalize electron and ion losses.Moreover, in order to maintain quasi-neutrality, if loss currents of positive (negative) charges reduce, loss currents of negative (positive) charges reduce too.
Over the past 30 years, R&D on ECRIS has enabled numerous technological and experimental advances that have optimized the plasma confinement induced by the magnetic field and by the double layer generation.For example, the application of polarized electrodes to reduce plasma losses in ECRIS dates back to the 1990s.The so-called Bias-disk, firstly tested at KVI in '91 [11], has become a standard for routine ECRIS operations.This electrode is placed along the plasma chamber axis.Once negatively biased, it reduces electron losses in the axial direction with a strong improvement in the overall ECRIS performances [12,13].
Furthermore, from the early 2000s, in order to control and reduce ion losses due to diffusion phenomena several attempts have been made to design a set of polarized electrodes positioned on the lateral surface of the plasma chamber [14,15].Although the operating principle has been fully demonstrated by several experiments [15], no set-up able to ensure long term operations has yet been proposed.The main problems that limit the reliability of polarized radial electrodes are difficulties in cooling, wiring and electrically insulating them.A certain margin for improvement is therefore expected in this direction.
To this end, the INFN's ECRIS R&D activity is engaged in the development of an active plasma chamber that should be able to guarantee the bias of radial electrodes and to ensure cooling, wiring and insulation.The main characteristic of the active plasma chamber, preliminary tests and its perspectives will be presented below.

The active plasma chamber
The active plasma chamber has been designed to be housed within the Advanced Ion Source for Hadrontherapy (AISHa).AISHa is a 18 GHz ECRIS, developed at INFN with the aim of producing high intensity and low emittance highly charged ion beams for hadrontherapy and research purposes [13,16].
In AISHa the plasma chamber consists of an AISI 316L cylinder with 92 mm internal diameter and 340 mm length, equipped with a water cooling circuit.The active chamber will cover the inner walls of the plasma chamber by 30 Al-6082 alloy tiles, 6 radial tiles for each of 5 axial sectors.Tiles will be grouped into 2 sub-sets, 15 tiles each, that will be screwed to 2 half-cylinders made of Al-6082 alloy, acting as two shells for the tiles.The design of a shell and the entire set of tiles are shown in Fig. 2. The tiles have a thickness of 4 mm.They are anodized on the rear face, in contact with the shell.The anodization process makes each tile electrically isolated from the others and from the plasma chamber, through a very thin oxide layer (50 microns) which allows good thermal properties to be preserved and to guarantee effective system cooling.
Each tile is machined on the back to be equipped with a temperature sensor and to allow the electrical wiring of the entire system.Each tile will be biased to a specific voltage between 0 and 100 V. To match the plasma region with the central axial sector, the tiles closest to the AISHa injection are 100.5 mm long while the others have a length of 53.5mm.The tiles are fixed to the shells by 8 half rings made of Al-6082 and completely anodized.
In order to improve the mechanical contact with the cooled walls of the plasma chamber, and thus obtain a good heat transfer, the 2 shells will work will be assembled using 4 springs to keep them apart.

Test-bench in a dummy chamber
Before mounting the active chamber on AISHa, a vacuum test of the system was performed using a dummy chamber.In particular, the behavior of the active chamber under the effect of heating was verified, using a tubular halogen lamp (5 cm long) which emulates the radiation generated by the plasma in AISHa (Fig. 3).No external cooling was applied to the dummy chamber.
Prior to assembly, all the components of the active chamber (tiles, shells, wires, thermocouples, etc.) were washed with acetone in ultrasound, rinsed in ethyl alcohol and dried in air, without drying in the oven to study the worst conditions for achieving the basic chamber vacuum.During heating, temperature monitoring on the tiles was achieved using K-type thermocouples (diameter 0.5 mm) inserted into 6 tiles positioned in different regions of the chamber.
The pumping system consisted of a turbo with a pumping speed of up to 350 l/s for N 2 , on a DN100 flange, mounted in axis with the chamber, with nominal final pressure < 110 -7 mbar.The setup also included a residual gas analysis device (<200 amu), in order to observe any anomalous degassing (of unwanted substances) of the materials under heating.
Figure 3: (left) view of the tiles mounted in a dummy chamber, with a halogen lamp positioned on the central sector of the tiles; (right) thermal behavior of 6 thermocouples (inserted in 6 tiles) positioned in different regions of the chamber, with a heating power of 56 W (left peaks) and up to 77 W (right peaks), applied for approximately 10 minutes, with no cooling applied to the chamber.

Preliminary tests and results.
The tests performed concerned the behavior in vacuum, with leakage tests of the dummy chamber (using helium), degassing of the dummy chamber (using an external heating band), analysis of the residual gas spectrum, direct heating of the tiles using the halogen lamp.
Degassing of the dummy chamber was accelerated using an external heating band that brought the external temperature to approximately 100°C for 15 minutes, followed by free cooling.After this phase and after approximately 17 hours of pumping in total, the minimum pressure reached is 7.410 -6 mbar.Subsequently, the halogen lamp was turned on to emulate the thermal conditioning (cleaning) phase which is normally performed in AISHa after closing the chamber.
Figure 3 shows the temperature curves for two heating ramps of the halogen lamp: the first up to a power of 56 W and the second up to 77 W, each lasting approximately 10 minutes and followed by a complete shutdown phase.With the first ramp the tile closest to the lamp reached 102°C, while in the second ramp the same tile reached 212°C.During these heating ramps, the drop in temperature between the hottest tile and the coldest one, i.e. the one furthest from the lamp, is approximately 50%, confirming the good heat conduction of the system.This is also confirmed when the lamp is turned off, where the temperature of the tiles drops rapidly, showing the low thermal inertia of the tiles and the entire system.At the end of these heating tests, the vacuum reached by the system after cooling, after a further 3 hours of pumping, is 410 -6 mbar.

Conclusions and perspectives
Preliminary tests of the active plasma chamber prototype have shown that it is possible to reach decent vacuum levels in reasonable pumping times, without particular preliminary degassing operations of the components.It is clear that to obtain a better degassing of the components it is necessary to increase the duration of the thermal conditioning, which is moreover compatible with the 2-3 days of conditioning that are normally applied in AISHa.
Apparently the thermal heat conduction behavior of the active plasma chamber is satisfactory and should not compromise the heat extraction produced by the plasma in AISHa.For further confirmation, it is planned to equip the dummy chamber with an external cooling system, with a cooling capacity equivalent to that of AISHa.
Complete tests of the active plasma chamber's vacuum behavior and physics tests are expected on AISHa in 2024.They will also provide confirmation (or not) on the overall performance improvement of the ECRIS machine.
Among the future perspectives is that of applying this model of active plasma chamber on the (bigger) ECRIS machine of the PANDORA experiment, under construction at the LNS laboratories in Catania.

Figure 2 :
Figure 2: (left) design of a shell with 15 tiles, connected to the AISHa injection side; (right) the entire set of 30 tiles is shown, together with 8 fixing half-rings; only the tiles' back side is anodized (dark grey) while the surface directly facing the plasma is not (light grey).