PRISMAP Deliverable D10.4 - Report on ion source efficiency for a 11C PET ISOL beam

Deliverable D10.4 Report on ion source efficiency for a 11C PET ISOL beam This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 101008571 (PRISMAP). This document reflects only the view of the author(s). The Agency is not responsible for any use that may be made of the information it contains. Deliverable D10.4 ii Project Acronym PRISMAP Project Title The European medical isotope programme: Production of high purity isotopes by mass separation Grant Agreement No. 101008571 Topic INFRAIA-02-2020: Integrating Activities for Starting Communities Project start date 01 May 2021 Nature Report Dissemination level Public Due date M52 Date of delivery M52 Lead partner MedAustron Contributing partner CERN Authors N. Gambino, M. Orsolic, L. Penescu, C. Schmitzer (MedAustron) Reviewers T.E. Cocolios (KULeuven), S. Rothe (CERN), M. Manzolaro (INFN) Point of Contact C. SCHMITZER Institution MedAustron E-mail [email protected] Phone +4366480878839 © PRISMAP 2021. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. Deliverable D10.4 iii Revision History Version Date Author Comment 0.1 09.07.2025 M. Orsolic, L. Penescu, C. Schmitzer First draft 0.2 01.08.2025 M. Orsolic, L. Penescu, C. Schmitzer Final draft for review 0.3 14.08.2025 S. Rothe, M. Manzolaro, T.E. Cocolios Internal review 0.4 21.08.2025 M. Orsolic, L. Penescu, C. Schmitzer Revised version 0.9 25.08.2025 V. Gobry, K. Leufgen Review and formatting 0.91 28.08.2025 M. Manzolaro, K. Leufgen Version for final review and approval by the PRISMAP general assembly (GA) 1.0 29.08.2025 K. Leufgen Final version, approved by the GA Deliverable D10.4 iv Contents Abbreviations, Participant short names v Abbreviations v Participant short names v List of Figures vii List of Tables vii Summary 1 1. Context 2 2. Motivation 2 2.1 Limitations in current clinical setup 2 2.2 Expected target output 3 2.3 Potential overall integration 3 3. Supernanogan test results 4 3.1 Conclusion and possible use of Supernanogan 6 4. VADIS MK7 test results 6 4.1 Source testing with noble gas mixture 7 4.2 Source tuning for operation with [C-12]CO2 8 4.3 Efficiency measurement with [C-13]CO2 11 5. COMIC test results 11 6. Conclusion and outlook 14 References 15 Deliverable D10.4 v Abbreviations, Participant short names Abbreviations COMIC Compact Microwave and Coaxial EBIS Electron Beam Ion Source ECR Electron Cyclotron Resonance FEBIAD Forced Electron Beam Induced Arc Discharge ISOLDE Isotope Separator Online Device LPSC Laboratoire de Physique Subatomique et de Cosmologie PET-CT Positron Emission Tomography – Computed Tomography RF radiofrequency SPS Spark Plasma Sintering TESIS Tubular Electron String Ion Source VADIS Versatile Arc Discharge Ion Source Participant short names CERN European organisation for nuclear research NPL National Physical Laboratory PSI Paul Scherrer Institut CEA Commissariat à l’énergie atomique et aux énergies alternatives IST-ID Associação do Instituto Superior Técnico para a IST-ID Investigação e Desenvolvimento DTU Danmarks Tekniske Universitet CHUV Centre hospitalier universitaire vaudois GANIL Grand Accélérateur National d’Ions Lourds SCK CEN Studiecentrum voor Kernenergie / Centre d'étude de l'énergie nucléaire ARRONAX Groupement d’intérêt public ARRONAX ESS European spallation source ERIC TUM Klinikum rechts der Isar der technischen Universität München KULeuven Katholieke Universiteit Leuven MedAustron Entwicklungsund Betriebsgesellschaft MedAustron GmbH SCIPROM SCIPROM Sàrl MUI Medizinische Universität Innsbruck ILL Institut Max von Laue - Paul Langevin Deliverable D10.4 vi JRC JRC -Joint Research CentreEuropean Commission NCBJ Narodowe Centrum Badań Jądrowych GSI GSI Helmholtzzentrum fr Schwerionenforschung GmbH LU Latvijas Universitāte INFN Istituto Nazionale di Fisica Nucleare UiO Universitetet i Oslo Deliverable D10.4 vii List of Figures Figure 1: CO2 to C4+ ionisation efficiency with respect to CO2/He gas injection 5 Figure 2: CO2 to C4+ ionisation efficiency with respect to CO2/Ar gas injection 5 Figure 3: Effect of turbo pump power on the CO₂ to C4⁺ ionisation efficiency, for the best ratio of He to CO2 identified in Figure 1. The other source settings are the ones given in Table 2. 6 Figure 4: Ionisation efficiency for noble gas mixture (helium, neon, argon, krypton, xenon). Operating parameters: cathode temperature: 2020 °C, ion source magnet: 27.5 mT, anode voltage: 250 V 7 Figure 5: Total ion current as a function of anode voltage at different gas injection values with noble gas mixture. Note: Although the feed injection comprises five noble gases in equal proportions (20 % each), this ratio shifts inside the ion source because the low leak rate allows lighter atoms to enter more easily than heavier ones. The actual composition is helium 46 %, neon 21 %, argon 15 %, krypton 10 %, and xenon 8 %. Operational parameters: cathode temperature: 2020 °C, ion source magnet: 30 mT 8 Figure 6: CO2 to CO+ ionisation efficiency with CO2 gas as a function of the cathode temperature. Operating parameters: CO2 gas injection of 5.45 x 1014 pps, ion source magnet: 30 mT, anode voltage: 250 V 9 Figure 7: CO+ current as a function of the ion source magnetic field strength. Operating parameters: CO2 gas injection of 2.79 x 1014 pps, cathode temperature: 2107 °C, anode voltage: 200 V 9 Figure 8: Ionisation efficiency ratio for CO+/CO2+, C+/CO2+ and C+/CO+ as a function CO2 gas injection. Operational parameters: cathode temperature: 2020 °C, ion source magnet: 30 mT, anode voltage: 250 V 10 Figure 9: Ionisation efficiency ratio of CO+/CO2+, C+/CO2+ and C+/CO+ of the COMIC ion source as a function of forward power. CO2 gas injection of 2.24 x 1016 pps 12 Figure 10: CO+, CO2+ and C+ current as a function of time of a 200 ms microwave pulse. CO2 gas injection of 2.24 x 1016 pps 12 Figure 11: CO+ current as a function buffer gas injection for (a) argon and (b) helium. CO2 gas injection of 3.32 x 1016 pps, forward power 20 W 13 List of Tables Table 1: Expected gas output from the BN target when operated at 1500 °C with a controlled oxygen leak, and beam on target as available from the 18 MeV Cyclone KIUBE 3 Table 2: Optimised ion source setting for C4+ operation 4 Table 3: Ionisation efficiencies of CO+, CO2+ and C+ for CO2 gas injection at different leak pressure values, uncorrected for background carbon coming from the ion source materials. Operating parameters: cathode temperature: 2020 °C, ion source magnet: 30 mT, anode voltage: 250 V 10 Table 4: The measured values for CO+ ionisation efficiencies. No background was observed at mass 29. A transport efficiency of 80% is considered from the ion source to the Faraday cup located after the separator magnet. 11 Summary The goal of this research activity task was to identify potential solutions to establish radioactive ion beams in the form of C-11 in a clinical ion beam therapy accelerator complex. Previous studies [1],[2] provided prototypes of C-11 production targets and concepts for overall integration. A crucial component is the ion source to ionize that scarcely produced radioactive gas. Three ion source options have been tested with CO2 gas to assess their performance and gas efficiency in an attempt to mimic similar conditions as provided by the production target prototype. A 14.5 GHz Supernanogan ECR source at MedAustron, a FEBIAD VADIS MK7 at CERN and a 2.4 GHz COMIC ECR source at The University of Manchester were studied under lab conditions by MedAustron and CERN staff. The Supernanogan ion source is a commercial compact Electron Cyclotron Resonance (ECR) ion source, whereas the FEBIAD and COMIC sources originate from the ISOL domain, in particular relying on the extensive experience accumulated over the years at the most representative ISOL facility, ISOLDE. Measurements have confirmed that the Supernanogan ion source demonstrates relatively low gas efficiency (observed maximum 0.6%). While it did not meet the particle output necessary for therapeutic applications, it still presents a viable option due to its operational simplicity. One key advantage is its compatibility with direct injection into an existing linear accelerator, eliminating the need for a mass separator or charge breeder—components that typically lead to additional particle losses and complexity. Despite falling short of the requirements for full clinical use, the current setup is still suitable for generating diagnostic carbon beams at low intensity (estimated 4x107 particles). These can be valuable for applications such as physiological and range verification in treatment planning. However, while academically interesting, its practical use in a clinical setting is doubtful. The primary concern is its negative impact on workflow efficiency, particularly the increased in-room time required for patients during irradiation procedures. To improve the Supernanogan’s performance, a possible upgrade involves integrating a cryogenic trap, similar to the one used in the Krion electron string ion source [9]. This would allow for a more concentrated release of radioactive carbon monoxide through pulsed sublimation into the plasma chamber, potentially increasing ion production efficiency. However, further studies are needed to assess how significantly this setup could enhance the yield of C4+ ions. In addition, issues like beam reproducibility and long-term stability throughout treatment sessions must be carefully evaluated. While observed COMIC efficiencies and intensities were too low to prove useful in the envisaged setup (observed maximum 0.35%), the FEBIAD ion source demonstrated more favourable gas efficiencies. The FEBIAD source, in particular, performed well at low gas pressures (observed maximum 5.4%), making it a strong candidate for use with the proposed boron nitride (BN) target. However, a limitation of this source is its inability to efficiently produce high charge states. As a result, its implementation would necessitate both a mass separator and a charge breeder. While the required mass resolution is relatively modest and thus technically achievable, the charge breeder itself would need to be highly capable—able to deliver around 1x1011 particles per pulse. This would likely require enhancements to existing EBIS (Electron Beam Ion Source) designs to meet those output levels. An alternative and potentially more robust solution could be the use of a cryogenic trap in combination with a high-performance electron string ion source, such as the TESIS system. TESIS has shown promising results for producing carbon beams at the required charge states for therapeutic purposes. Since similar sources have already been successfully integrated with cryotraps, this configuration could meet both the intensity and quality demands of carbon ion therapy. In conclusion, while the Supernanogan ion source in its current form is not suitable for therapeutic applications due to low particle yield, its simplicity and potential for upgrades may justify further research. Meanwhile, sources like FEBIAD and advanced systems like TESIS offer more immediate pathways toward achieving the necessary beam quality and quantity, particularly when paired with appropriate separation and breeding technologies. Ongoing development and refinement of these systems are critical to ensuring practical, high-efficiency solutions for clinical C-11 ion beam therapy. Deliverable D10.4 2 1. Context One prominent quality assurance limitation of particle-based radiation therapy is to verify the 3-dimensional location of the Bragg peak and thus the deposited dose within the patient. While the ion beam position and incoming energy can be well determined, the patient’s physiology will change over the course of the treatment which usually lasts several weeks. As the ion beam dose deposition is effectively an implantation process, the use of radioactive isotopes as primary ion beam could enable in-vivo monitoring of the dose location. Positron emitters like C-11 would thus enable the use of standard PET-CTs for direct correlation between the PET signal and the deployed dose distribution. While this is also possible using stable C-12, using an active positron emitter increases the signal by an order of magnitude and results in a better signal to noise ratio, thus improving imaging quality. As this method can be employed during normal treatment, no extra machine time or other dedicated imaging CT sessions are required. Every treatment session can be used to gather new quality assurance data and trigger a necessary update of the treatment plan to take physiology changes into account. Additional online monitoring will also help to reduce range uncertainties and help guarantee the required safety margins. While the first application would be post-treatment imaging to provide more data on the patient’s treatment response, in the future, in-vivo online dose verification will improve tumour targeting and support 4dimensional treatment planning. In order to provide a C-11 beam, high yield production targets have to be used [1] and very efficient radioactive ion beam lines are required. Current accelerator transport efficiency demands around 1x1011 C4+ or C6+ particles per second at the source level to provide sufficient intensity for proficient treatment. The scarce amount of [C-11]CO gas results in tight requirements of ion source gas efficiency. An extensive analysis of the possible technical solutions for the implementation of C-11 radioisotopes in an accelerator-based particle therapy centre was done under the MEDICIS-PROMED Horizon 2020 project [2]. In this report we will present the work conducted at MedAustron, CERN and The University of Manchester on carbon ion source characterisation in view of gas efficiency for potential C-11 radioactive ion beam lines. 2. Motivation 2.1 Limitations in current clinical setup C-12 beams are generated in particle therapy centres using high-performance ECR ion sources. These ion sources generally deliver currents of over 100 microamperes (µA) in continuous mode, or, more recently, also in a long-pulse (>100 ms) mode. In both cases, however, almost all of the generated beam is discarded, as the ion beam can only be injected into the synchrotron within a time window of a few tens of ms. As stable C-12 is available in sufficient quantities, these losses do not pose a problem. Radioactive particle beams present a different situation. Medical synchrotrons are usually filled once every 1-10 seconds. The beam is subsequently accelerated and then extracted throughout an extraction spill which lasts up to 8 seconds before the next filling cycle starts. Radioisotope production is a complicated and costly process which therefore calls for efficient use of the short-lived radioisotopes. The time between spills should therefore be used to accumulate the particles to ensure the intensity required for irradiation. Each step must achieve high efficiency to ensure that the largest possible proportion of the produced particles is accelerated. However, the ion sources currently used in particle therapy centres are not designed for high gas efficiencies and thus not intrinsically optimised for the low gas flow rates typically expected from target materials. Furthermore, these ion sources have no storage capacity. An EBIS-type charge breeding system was therefore proposed to meet the requirements. Deliverable D10.4 9 Figure 6: CO2 to CO+ ionisation efficiency with CO2 gas as a function of the cathode temperature. Operating parameters: CO2 gas injection of 5.45 x 1014 pps, ion source magnet: 30 mT, anode voltage: 250 V Increasing the cathode temperature from approximately 1,900 °C to 2,100 °C causes thermionic electron emission to increase, as illustrated in Figure 6, in agreement with the Richardson–Dushman equation 𝑗 ∝ 𝑇2𝑒𝛼/𝑇. As the anode drain current (which is a measure of the thermionic emission current) increases, the CO2 to CO+ efficiency also improves. The rise of the electron density, however, deepens the electrostatic potential well in the anode chamber, described in Error! Reference source not found. as the inactive volume. This is the space in which ions remain trapped in the potential funnel and can no longer be extracted. As a result, the proportion of the active volume from which CO⁺ ions can actually escape decreases, and the CO⁺ efficiency reaches a plateau despite further increase of the electron emission. Figure 7: CO+ current as a function of the ion source magnetic field strength. Operating parameters: CO2 gas injection of 2.79 x 1014 pps, cathode temperature: 2107 °C, anode voltage: 200 V The VADIS operates within a weak axial magnetic field generated by a solenoid. As the current in the ion source magnet increases, a shift in CO₂ to CO⁺ efficiency can be observed (Figure 7), which can be explained by the presence of two competing mechanisms [5]. On the one hand, the magnetic field forces the electrons Deliverable D10.4 10 into spiral orbits, thus extending their residence time in the anode cavity, which significantly improves the probability of ionisation. At the same time, the increased electron accumulation in the anode centre leads to the formation and deepening of the previously mentioned inactive volume. The produced CO⁺ ions remain trapped and can no longer be extracted. The best ionisation efficiency was measured at higher magnetic field strengths. At higher gas injection rates, the extracted current saturates due to space charge limitations according to the Child-Langmuir equation. This effect reduces the overall ionisation efficiency, as shown in Table 3. Table 3: Ionisation efficiencies of CO+, CO2+ and C+ for CO2 gas injection at different leak pressure values, uncorrected for background carbon coming from the ion source materials. Operating parameters: cathode temperature: 2020 °C, ion source magnet: 30 mT, anode voltage: 250 V Leak pressure [mbar] Injected C02 gas (pps) CO+ [%] CO2+ [%] C+ [%] 254 2.76 x 1014 6.5 1.56 0.68 456 4.95 x 1014 4.7 1.74 0.50 952 1.03 x 1015 3.4 1.66 0.34 1454 1.58 x 1015 2.5 1.30 0.25 The data from Table 3 is presented in Figure 8 as ratios between the different carbon compounds, to illustrate how the chemical equilibrium evolves inside the ion source depending on the injected gas amount. It is observed that the improvement of the CO+ ionisation efficiency at lower values of the injected gas is not only due to an overall improvement of the ionisation efficiencies, but also to an improvement of the ratio CO/CO2. The ratio C/CO is unaffected by the amount of injected gas. Figure 8: Ionisation efficiency ratio for CO+/CO2+, C+/CO2+ and C+/CO+ as a function CO2 gas injection. Operational parameters: cathode temperature: 2020 °C, ion source magnet: 30 mT, anode voltage: 250 V These measurements led to the following conclusions:  The carbon compound with the best ionisation efficiency is CO+ Deliverable D10.4 11  The ion source has better ionisation efficiencies when less gas is injected  The best settings to use for the ion source are:  Line temperature between 2050°C and 2100°C  Anode voltage at 250 V  Solenoid magnet at 6A (which generates a magnetic field estimated at 30 mT)Table 4 4.3 Efficiency measurement with [C-13]CO2 The goal of the test with [C-13]CO2 was to obtain a reliable measurement of the ionisation efficiency for the carbon molecules, because the VADIS has inherently a background of C-12 ions when operated at high temperatures. After performing anode voltage, magnet and line heating scans, the settings identified in section 4.2 were confirmed as optimal and yielded the following results for C-13 (see Table 4), which are expected to also represent the C-11 efficiencies. Table 4: The measured values for CO+ ionisation efficiencies. No background was observed at mass 29. A transport efficiency of 80% is considered from the ion source to the Faraday cup located after the separator magnet. Measurements at mass 29 are not affected by any background as there was no contamination observed at this mass during the tests with noble gases and with C-12. An important contamination (on the order of µA) was still observed at mass 28, for temperatures above 1900 °C. This contamination was increasing during the operation with [C-12]CO2. During the tests with noble gases, the contamination was found to be negligible (on the order of a few nanoamperes). 5. COMIC test results The 2.45 GHz COMIC (Compact Microwave and Coaxial) ion source was designed as an ultra-compact ECR source at the LPSC in Grenoble in close collaboration with CERN ISOLDE. The following section provides an overview of the tests conducted at The University of Manchester with regard to C-11 compatibility. Leak pressure [mbar] Injected 13C02 gas (pps) 13CO+ [A] 13CO+ [%] 250 2.71 x 1014 1.87 5.4 500 5.43 x 1014 2.64 3.8 Deliverable D10.4 12 Figure 9: Ionisation efficiency ratio of CO+/CO2+, C+/CO2+ and C+/CO+ of the COMIC ion source as a function of forward power. CO2 gas injection of 2.24 x 1016 pps Figure 9 shows that the CO/CO₂ ratio increases with higher microwave power. This is possibly due to the increase in electron density, which is important for the neutral dissociation of CO₂. Initially, electrons split CO₂ predominantly into CO and O (neutral dissociation energy 5.5 eV), with CO subsequently being more easily ionised to CO⁺ (ionisation energy 13.8 eV) [6]. In addition, the proportion of hot electrons also increases with higher microwave power, which favours the direct ionisation of CO₂ to CO⁺ (dissociative ionisation energy 19.5 eV) [7]. Therefore, the CO/CO₂ ratio increases until equilibrium is reached at higher powers, as almost all available CO₂ molecules have already been converted. The formation of C⁺, on the other hand, remains very low, because even higher electron energies are required. As a result, both C/CO₂ and C/CO reach only low values. Figure 10: CO+, CO2+ and C+ current as a function of time of a 200 ms microwave pulse. CO2 gas injection of 2.24 x 1016 pps (a) Deliverable D10.4 13 Figure 10 shows the CO₂⁺, CO+ and C+ ion current of a 200 ms microwave pulse. For the CO2+ current, a rapid increase and overshoot can be seen, followed by a significant drop to a steady plateau. After the initial depletion of CO2 and the direct production of CO2+ stabilises, the stepwise ionisation of other species and dissociative products takes over [8]. This results in a short-term current peak of CO₂⁺ ions. Shortly after, the CO₂⁺ current drops again which may have several causes that occur simultaneously. First, some of the CO₂ is not only ionised by electron collisions, but also dissociated (dissociation energy 5.5 eV), producing CO and atomic oxygen. This reduces the CO₂ concentration in the plasma, leaving fewer molecules available for direct ionisation to CO₂⁺. Second, CO₂⁺ ions that have already been produced are further converted into CO⁺ or other ions by charge transfer and collision processes, which further reduces the CO₂⁺ current. CO+ and C+ currents show a rapid increase up to a plateau, in agreement with expectations. Figure 11: CO+ current as a function buffer gas injection for (a) argon and (b) helium. CO2 gas injection of 3.32 x 1016 pps, forward power 20 W Different buffer gases have been injected in order to attempt an efficiency boost of CO and CO2 ion production. Results of argon and helium tests are illustrated in Figure 11. The CO⁺ current decreases in both investigated cases with increasing injection of buffer gases. When argon is used, the plasma is only stable (a) (b) Deliverable D10.4 14 within a very narrow range, whereas helium allows for a significantly broader stable operating window. Compared to operating the source exclusively with CO₂, the addition of buffer gases has an overall suppressing effect on CO⁺ production, as the achievable ion current in both cases remains below the values for pure CO₂ operation. The highest measured CO⁺ current was approximately 6 µA with a CO₂ gas injection of 1.34 × 10¹⁷ pps and a forward power of 50 W. Despite the overall lower CO⁺ current yield, however, the buffer gases enable a stable plasma operation at lower CO₂ gas injection rates while still maintaining good CO⁺ currents. The best measured CO₂ to CO⁺ efficiency was 0.35% with an argon injection of 2.24 × 10¹⁶ pps, a CO₂ injection of 4.93 × 10¹⁵ pps, and a forward power of 20 W. 6. Conclusion and outlook Conducted measurements have confirmed a rather low gas efficiency of the Supernanogan ion source. Although it was not possible to achieve the treatment requirements in terms of number of particles, it remains a viable setup due to its advantage of simplicity. The ECR ion source can directly inject into an existing linear accelerator and does not require any mass separator or charge breeder which by themselves yet reduce the transmitted particle count. Current results will only suffice for diagnostic carbon beams which can be used for physiology and range verifications. Although interesting from an academic point of view it is very unlikely that such a setup would be applied for clinical use sinceit will increase overall in-room time of the patient in the irradiation room. In order to boost the overall performance, the Supernanogan could be equipped with a cryogenic trap as implemented in the Krion [9] electron string ion source. This would enable a higher concentration of [C-11]CO followed by a pulsed sublimation release [9] into the plasma chamber. It remains to be studied how much the C4+ production can be boosted with such a setup. Additional points to address would be reproducibility and stability throughout the duration of a treatment. The observed gas efficiencies of the other tested and optimised sources, especially the FEBIAD, show promising performances. The FEBIAD also shows a good performance at low gas pressures which makes it a suitable candidate for use with the envisaged BN target. Unfortunately, only low charge states can be generated efficiently which results in the need of a dedicated charge breeder and another mass separator. The efficiency of this process was evaluated in [2], where it was found that for a combination of a trap and EBIS charge breeder, a maximum efficiency of 8% for the conversion of CO1+ to C6+ can be achieved. Yet this efficiency was decreasing towards 1% with increasing number of C6+ ions per pulse, due to space charge limitations and to the high collection time on the trap compared to the standard operation (on the order of 100 ms, instead of few µs). If we assume an average efficiency of 5% for this charge breeding step, together with the VADIS ionisation efficiency of about 5%, we obtain a total efficiency of 0.25%, which is below the measured ionisation efficiency for Supernanogan, which does not require these additional steps. Another potential solution could be a combination of a cryogenic trap with a high-performance electron string ion source called TESIS [11] which promises suitable beam currents of the required carbon charge states for ion beam therapy. As similar ion sources have already been equipped and tested with such cryotraps, such a setup promises the required beam intensities for treatment. Deliverable D10.4 15 References [1] S. Stegemann et al., “A porous hexagonal boron nitride powder compact for the production and release of radioactive 11C,” Journal of the European Ceramic Society, vol. 41, no. 7, pp. 4086–4097, July 2021, doi: 10.1016/j.jeurceramsoc.2020.12.029. [2] L. Penescu, T. Stora, S. Stegemann, J. Pitters, E. Fiorina, R. Augusto, C. Schmitzer, F. Wenander et al., "Technical Design Report for a Carbon-11 Treatment Facility", Frontiers in Medicine, vol. 8, 2021, doi:10.3389/fmed.2021.697235 [3] P. Jardin et al., “Atom-to-ion transformation time in singly charged ECRIS”, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, vol. 225, no. 3, pp. 374–382, 2004, doi: 10.1016/j.nimb.2004.06.004. [4] L. Penescu, R. Catherall, J. Lettry, and T. Stora, “Development of high efficiency Versatile Arc Discharge Ion Source at CERN ISOLDE”, Review of Scientific Instruments, vol. 81, no. 2, 2010, doi: 10.1063/1.3271245. [5] F. Maldonado Millan, T. Day Goodacre, and A. Gottberg, “Multiphysics simulation of a FEBIAD ion source”, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, vol. 463, pp. 302–304, Jan. 2020, doi: 10.1016/j.nimb.2019.04.078. [6] M. Tsuji, T. Tanoue, K. Nakano, and Y. Nishimura, “Decomposition of CO2 into CO and O in a Microwave-Excited Discharge Flow of CO2/He or CO2/Ar Mixtures”, Chemistry Letters, vol. 30, no. 1, pp. 22–23, 2001, doi: 10.1246/cl.2001.22. [7] G. Juve, “Lawrence Berkeley National Laborator, yBerkeley, California”, XRDS, vol. 17, no. 4, pp. 58– 59, 2011, doi: 10.1145/1961678.1961690. [8] M. Salahshoor and M. Aslaninejad, “Resonance surface, microwave power absorption, and plasma density distribution in an electron cyclotron resonance ion source”, Physical Review Accelerators and Beams, vol. 22, no. 4, p. 043402, 2019, doi: 10.1103/PhysRevAccelBeams.22.043402. [9] D.E. Donets, E.D. Donets, E.E. Donets, V.V. Salnikov and V.B. Shutov, "Production of highly charged ion beams Kr32+, Xe44+, Au54+ with Electron String Ion Source (ESIS) Krion-2 and corresponding basic and applied studies", Journal of Instrumentation, vol. 5, 2010, doi: 10.1088/17480221/5/09/C09001 [10] D.E. Donets, E.D. Donets, E.E. Donets, V.V. Salnikov, V.B. Shutov, E.M. Syresin, "Formation of high intensive radioactive carbon ion beams in electron string ion source", Proceedings of HIAT09, Venice, Italy, 2009, doi: 10.1088/1748-0221/5/09/C09001 [11] D.E. Donets et al. "Design of a novel tubular electron string ion source (TESIS)", Proceedings of EPAC08, Genoa, Italy, 2008