Research and development activities to increase the performance of the CAPRICE ECRIS at GSI

At GSI the CAPRICE ECRIS is in operation to deliver high charge state ion beams from gaseous and metallic elements to the accelerator facility. A test campaign has been carried out at the ECR test bench to fulfill the demand for higher intensity and stability of high charge state ions and for mixed ion beams. The ion beam stability has been monitored by an Optical Emission Spectrometer (OES), which has been already used to check the plasma and the temperature of the resistively heated oven during metal ion beam operation, particularly for Ca ion beams. During the test campaign, it was investigated that the OES can be used for monitoring the stability of ion beams from gaseous elements and mixed ion beams, and the main achieved results are reported. The ion beams extracted from the ECRIS have been simulated with a particle tracking code in order to study and improve the beam matching into the RFQ. The preliminary results of the study together with possible modifications of the extraction column are presented.


Introduction
The CAPRICE ECRIS is in operation at the high charge states injector of GSI to deliver high charge state ion beams from gaseous and metallic elements to the accelerator facility [1].An ECR test bench, where the same ECRIS and similar LEBT are installed, is used for testing new ion beams and carrying out research and development activities to fulfill the demand of the experimentalist in terms of ion source performance.The last test campaign has been carried out to improve the stability of ion beams from gaseous elements by using an Optical Emission Spectrometer (OES) as a monitoring tool.The use of an OES has been proven to be effective for the investigation of the properties of ECRIS plasma [2,3] and to check the temperature of the resistively heated oven during metal ion beam operation, particularly for Ca ion beams [4].Mixed carbon-helium ion beams have been proposed for simultaneous carbon ion beam therapy and helium radiography [5][6][7].In preparation for a corresponding fundamental experiment, those mixed ion beams have been established and investigated at the test bench in terms of intensity and C-to-He-ratio by measuring ion beam currents, mass spectra, and optical emission lines.The ion beams extracted from the CAPRICE ECRIS have been simulated with a particle tracking code in order to study and improve the beam matching into the RFQ.The main achieved results are reported.

Optical Emission Spectroscopy Measurements
A measurement campaign has been carried out with the CAPRICE ECR Ion Source (ECRIS) at the EIS (ECRIS Injector Setup) test bench of GSI to determine whether a diagnostic tool based on optical emission spectroscopy could be used for the early detection of instabilities during the production of ions from gaseous elements and for monitoring mixed ion beams.To monitor the internal plasma condition, the spectral content in the visible wavelength range has been measured with an optical emission spectrometer, looking through the ECRIS extraction aperture.The description and the data sheet of the spectrometer QE-Pro used for the measurements are given in [8] and the main features of the experimental setup are given in [9].The main parameters and settings of the ECRIS during operation have been continuously recorded: the valve setting of the main gas, the reflected microwave power level, the ECRIS ion source pressure, the ion beam current measured after the dipole magnet and the drain current of the extraction power supply.The analysis of the time variations of spectral lines together with the ECRIS settings has been related to the intensity of the extracted ion beam to monitor the ion source stability.The experimental campaign has been performed with noble gases and with oxygen as a support gas.For each combination of working and support gas, the ECRIS was in operation for at least 8 hours.Since no plasma instabilities were observed with the experimental set-up previously described during this experimental campaign, it was decided to stimulate their appearance by actively changing the source parameters.Therefore, the present analysis focuses mainly on the periods when the source optimization was performed to evaluate the effect of the parameters change on the ECRIS and the visible spectral content.It should be pointed out that the optical emission spectra were continuously recorded once per five minutes in the present analysis.Decreasing the time between individual optical emission measurements could help to gain information on the upcoming long-or short-term instabilities.As soon as the ECRIS was optimized for the desired charge states, the optical emitted spectra in the visible wavelength range were saved.The optical emission lines of each element (neutral and 1+ ion) were identified using the NIST database [10] and the time variation of the lines was monitored and saved.Fig. 1 shows the OES measurement for xenon (left) and for krypton (right) when the ECRIS was optimized for Xe 18+ , Xe 23+, and Kr 13+ , respectively.As observed for xenon, a direct dependence of the intensity of the emission lines with gas pressure and microwave power coupling efficiency has been measured.The analysis of the data obtained for the noble gases together with the intensity of the measured optical emission lines allows us to conclude that the ion source operation and the optical lines remain stable if the ECRIS settings are not adjusted.However, according to this feasibility study, it is possible to conclude that the optical emission spectrometer can be used as a tool for the long-term instability detection of gaseous ion beams as already pointed out for the metallic ones [3].The reaction time of the ECRIS plasma to a parameter change can be observed with the spectrometer after a shorter time in comparison with the metallic ion beam production because the solid material has to be evaporated with a resistively heated oven first.
In the context of fundamental research in carbon ion beam therapy a mixed carbon/helium ion beam has to be provided to the users.The objective from the ion source's perspective is to set up a steady carbon ion beam of approximately 150 eµA ( 12 C 3+ or 12 C 4+ ) containing a helium particle fraction of about 10 %, i.e. approx.5 eµA ( 4 He + or 3 He + ) in front of the subsequent linac-synchrotron accelerator chain.By this, the demands in terms of intensity are entirely met.The associated measurements were performed at the EIS test bench.The results of a mixed 12 C 4+ / 3 He + ion beam out of CH4/ 3 He are reported here.The approach is to set up the C 4+ ion beam followed by stepwise adding helium to the plasma while recording the optical emission lines and the corresponding mass spectra.In the latter, there is no distinction between 12 C 4+ and 3 He + , while the optical emission lines of carbon (wavelength 465 nm) and helium (728 nm) may allow for an estimate of the C-to-He ratio.Fig. 3 (left, black graph) shows the He I peak at 728 nm wavelength which is increasing with stepwise opening of the He-valve (Fig. 3, right).At the same time, the combined 12 C 4+ / 3 He + peak and 3 He 2+ are increasing in the mass spectrum (Fig. 4, left).Thus, it is possible to establish a relation between increasing current in the combined peak and the optical emission lines (Fig. 4, left), but not yet giving the absolute number of particles.

Particle Tracking Simulations
In order to improve the understanding of the effect on the ion beam dynamics of the main components of the extraction system of the CAPRICE ECRIS, particle tracking simulations with CST Design Studio® [11] were performed.The software has been already used to analyze the Uranium ion trajectories downstream of the extraction system and through the post-acceleration system of the high current test injector (HoSTI) of GSI [12].For this purpose, the mechanical 3D model of the ion source is imported in the simulation software and the entire geometry is simulated by using the imported drawing.The length of the layout is 800 mm and to model the smaller geometries, e.g. the 10 mm extraction holes, a fine and adapting mesh with more than 49 million cells is used.The non-magnetic metal parts are modeled as Perfect Electric Conductors (PEC) and the electromagnetic properties of the other materials are included.To study the extracted ion beams, an electrostatic and magnetostatic model is required.The ion beam is extracted with an "acceleratingdecelerating" triode system to match the injection energy of 2.5 keV/u of the RFQ.The details of the HLI RFQ together with the beam dynamic design parameters are available at [13].A fixed potential is applied at the electrodes: 15 kV at the plasma electrode, -2 kV at the screening electrode, and 0 kV at the puller electrode.The magnetostatic model includes the longitudinal magnetic field generated by two coils with a maximum field of 1.2 T and the radial confining field produced by a Halbach-type 12segment hexapole made of permanent magnets with 1 T remanence field.The potential map and the magnetostatic field obtained with the electromagnetic simulations are used for tracking the extracted charged particles.Approximately 100,000 argon macro particles with a Maxwell-Boltzmann initial temperature distribution were simulated to be emitted from an equipotential surface in the middle of the plasma chamber.This plasma sheath emission model includes a plasma core at a fixed potential, e.g. 10 V higher than the plasma electrode potential.As an example, the trajectory plots of the Ar 4+ ions trapped inside the plasma chamber and extracted by the ECRIS are shown in Fig. 5 where the ion energy is indicated in eV in the right scales.The main characteristics of extracted Ar 4+ , Ar 8+ and Ar 12+ ion beam, e.g.real space, emittances and brilliance are investigated by tuning the plasma electrode diameter from 8 to 14 mm in steps of 2 mm.In order to analyze the envelope of the real spaces and the emittances, different beam cross sections were selected and the real space profile and the vertical emittance at a cross-plane 36 mm downstream of the puller electrode are shown in Fig. 6.The simulation results are consistent with the ones obtained for the same design with other software [14,15].Furthermore, the particle tracking results concerning the real space profile are in good agreement with the images recorded with viewing target detectors [16].The main characteristics of the extracted ion beam, e.g.intensity, RMS emittances, and brilliance, are investigated for the different diameters of the plasma electrode.The simulation results in terms of RMS emittance and brilliance at 36 mm downstream of the puller electrode by increasing the plasma electrode diameter for different argon charge states are shown in Fig. 7.The small decrease of the emittance for large plasma electrode diameter, i.e. 14 mm, can be referred to the beam cut taking place since the aperture of screening and the ground electrodes was not modified.The actual diameter of the plasma electrode placed at the CAPRICE ECRIS extraction column is 10 mm and according to the simulation results a reduction to 8 mm will provide an extracted beam with higher quality for a better matching into the RFQ.The modified plasma electrode will be machined at GSI workshop and an experimental verification of the simulation results will be performed.The particle tracking simulations with CST Design Studio® provided a good matching with the emittance and real space measurements, for this complex and large geometry, with a moderate computational time (<8 hours with a workstation equipped with a 128 GB RAM).Nevertheless, it would be necessary, for accurate particle dynamics simulations, to take into account the space charge effect not included in the actual investigation.

Figure 1 :
Figure 1: OES measurements for xenon (left) when the ECRIS is optimized for Xe 18+ , Xe 23+ and for krypton (right) for and Kr 13 optimization.Oxygen used as support gas.

Fig. 2
Fig.2shows the time variations of the xenon spectral lines at 559 and 829 nm together with the ECRIS settings (valve setting of the main gas, reflected microwave power level, ECRIS pressure, ion beam current measured after the dipole magnet, and the drain current of the extraction power supply) during the operation with Oxygen as support gas.Two ion source settings were selected to maximize the intensities of two different charge states: Xe 18+ and Xe 23+ beams.The setting change can be noted at 16:25 in Figure2(left).The analysis of this figure shows that the intensity of the measured emission lines increases with the gas pressure thus affecting the microwave power to the plasma coupling, as well.

Figure 2 :
Figure 2: Time [hh:mm] variations of xenon spectral lines at 559 and 829 nm (left) and of krypton spectral lines at 526 and 761 nm (right) together with the ECRIS settings.Please note that the line corresponding to the gas valve refers to the normalized voltage of the board controlling the gas opening.The long-term stability of the xenon beam operation has been monitored with the OES within the first three hours of operation.The time variations of the krypton spectral lines

Figure 3 :Figure 4 :
Figure 3: OES number of counts for CH4+ 3 He (left) and at 728 nm for He valve stepwise opening (right)

Figure 5 :
Figure 5: Ar 4+ ions at different planes in the plasma chamber (left) in the vertical cross-section (right) of CAPRICE ECRIS

Figure 6 :
Figure 6: Real space profile (top figures) and vertical emittance (bottom figures) @ 36 mm from the puller electrode by increasing the plasma electrode diameter for different argon charge states

Figure 7 :
Figure 7: RMS emittance (left) and brilliance (right) @ 36 mm from the puller electrode by increasing the plasma electrode diameter for different argon charge states