Design and simulation of an imaging neutral particle analyzer for the ASDEX Upgrade tokamak

Rev. Sci. Instrum. 92, 043554 (2021); https://doi.org/10.1063/5.0043768 92, 043554 © 2021 Author(s). Design and simulation of an imaging neutral particle analyzer for the ASDEX Upgrade tokamak Cite as: Rev. Sci. Instrum. 92, 043554 (2021); https://doi.org/10.1063/5.0043768 Submitted: 11 January 2021 • Accepted: 25 March 2021 • Published Online: 26 April 2021 J. Rueda-Rueda, M. García-Muñoz, E. Viezzer, et al. COLLECTIONS Paper published as part of the special topic on Proceedings of the 23rd Topical Conference on High-Temperature Plasma Diagnostics ARTICLES YOU MAY BE INTERESTED IN Implementation of synthetic fast-ion loss detector and imaging heavy ion beam probe diagnostics in the 3D hybrid kinetic-MHD code MEGA Review of Scientific Instruments 92, 043558 (2021); https://doi.org/10.1063/5.0043757 Self-adaptive diagnostic of radial fast-ion loss measurements on the ASDEX Upgrade tokamak (invited) Review of Scientific Instruments 92, 053538 (2021); https://doi.org/10.1063/5.0043756 Upgrade and absolute calibration of the JET scintillator-based fast-ion loss detector Review of Scientific Instruments 92, 043553 (2021); https://doi.org/10.1063/5.0043760 Review of Scientific Instruments ARTICLE scitation.org/journal/rsi Design and simulation of an imaging neutral particle analyzer for the ASDEX Upgrade tokamak Cite as: Rev. Sci. Instrum. 92, 043554 (2021); doi: 10.1063/5.0043768 Submitted: 11 January 2021 •Accepted: 25 March 2021 • Published Online: 26 April 2021 J. Rueda-Rueda,1,a) M. García-Muñoz,1,2 E. Viezzer,1,2 P. A. Schneider,3J. García-Domínguez,2 J. Ayllon-Guerola,2J. Galdón-Quiroga,1,3 A. Herrmann,3X. D. Du,4M. A. Van Zeeland,4P. Oyola,1 M. Rodriguez-Ramos,5and ASDEX Upgrade Teamb) AFFILIATIONS 1Department of Atomic, Molecular and Nuclear Physics, Av. Reina Mercedes s/n, 41012 Seville, Spain 2Centro Nacional de Aceleradores (CNA) CSIC, 41092 Seville, Spain 3Max Planck Institute for Plasma Physics, Boltzmannstr. 2, 85748 Garching, Germany 4General Atomics, P.O. Box 85608, San Diego, California 92186-5608, USA 5Laboratory for Ion Beam Interactions, Ru-- der Boˇ skovi´ c Institute, 10000 Zagreb, Croatia Note: Paper published as part of the Special Topic on Proceedings of the 23rd Topical Conference on High-Temperature Plasma Diagnostics. a)Author to whom correspondence should be addressed: [email protected] b)See the author list of H. Meyer et al., Nucl. Fusion 59, 112014 (2019). ABSTRACT An Imaging Neutral Particle Analyzer (INPA) diagnostic has been designed for the ASDEX Upgrade (AUG) tokamak. The AUG INPA diagnostic will measure fast neutrals escaping the plasma after charge exchange reactions. The neutrals will be ionized by a 20nm carbon foil and deflected toward a scintillator by the local magnetic field. The use of a neutral beam injector (NBI) as an active source of neutrals will provide radially resolved measurements, while the use of a scintillator as an active component will allow us to cover the whole plasma along the NBI line with unprecedented phase-space resolution (<12 keV and 8cm) and a fast temporal response (up to 1 kHz with the high resolution acquisition system and above 100 kHz with the low resolution one), making it suitable to study localized fast-ion redistributions in phase space. Published under license by AIP Publishing. https://doi.org/10.1063/5.0043768 I. INTRODUCTION A detailed understanding of the fast-ion (FI) behavior in the presence of magnetohydrodynamic (MHD) fluctuations is mandatory for achieving a good fast-ion confinement in future fusion devices.1For this purpose, novel diagnostic techniques to measure the FI distribution in phase space with Alfvénic temporal resolution are currently being developed. The Imaging Neutral Particle Analyzer (INPA), already installed at the DIII-D tokamak,2,3 which is able to measure the radial position and energy of the confined FI population with a fast temporal response, is one example among these novel diagnostics. This paper is structured as follows: the INPA working principle is explained in Sec. II. Section III presents the synthetic diagnostic, while Sec. IV explains the influence of the different geometrical parameters on the detector performance. Section Vpresents the response of the INPA to FI redistributions due to magnetohydrodynamic (MHD) activity. II. WORKING PRINCIPLE The INPA combines the already working principles of neutral particle analyzers (NPAs)4and fast-ion loss detectors (FILDs)5to provide the energy and location of the confined fast-ion population Rev. Sci. Instrum. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-1 Published under license by AIP Publishing Review of Scientific Instruments ARTICLE scitation.org/journal/rsi FIG.1.SchemeoftheINPAworkingprinciple.Inblueisthetrajectoryoftheneutral particle, while in red is the trajectory of the ion. (a) and (b) show the side and top views, respectively, while (c) shows a 3D sketch. with a fast temporal response. Similar to NPA systems, the INPA analyzes neutral particles produced in charge exchange (CX) reactions. These particles are not confined by the magnetic field and retain all the information of fast ions, as no significant exchange of energy or momentum occurs during the CX reaction.6After being collimated, neutrals are ionized by an ultra-thin [20nm, for the ASDEX Upgrade (AUG) setup] carbon foil and deflected toward a scintillator plate, as can be seen in Fig. 1. The strike position of the particles in the scintillator will be determined by their energy and pitch (λ=−v∥/v, where v∥is the velocity along the magnetic field) and the local magnetic field of the tokamak. The use of a neutral beam injector (NBI) as an active source of neutrals gives the possibility of relating, via modeling, the measured pitch of the neutrals to their radial birth position and hence to the location of the confined FI. This is possible thanks to the geometric relation between the velocity orientation of the measured neutral and its radial birth position. Figure 2 shows an example of this relation for the case of AUG.The in-vesselpositioning ofthe diagnosticopens upa broader flexibilityin selectingthe lines-of-sight (LOSs)compared tothe conventional electromagnetic NPA installed out-vessel and using their own electromagnetic fields.4The use of a scintillator will enable the simultaneous exploration of a wide region of the plasma, which FIG.2.Relationbetweentheline-of-sight(LOS)andradialbirthposition.Asneutral particles travel along a straight line, and assuming that they are all born along the NBI line, there is a one-to-one relation between the LOS (velocity direction of the neutral) and its radial birth position. In blue is the NBI line, in green are some of the explored LOSs, and in red is the location of the INPA head. would not be covered by individual LOSs. Light emitted by the scintillator is collected by a series of lenses and guided to a Phantom camera and a set of photomultipliers (PMTs). The camera allows us to obtain high resolution images of the scintillator, while the PMTs sacrifice this resolution to gain a faster temporal response. III. SYNTHETIC DIAGNOSTIC A synthetic INPA diagnostic has been developed to study the feasibility for the installation of an INPA at AUG and to optimize its design. The synthetic INPA is based on FIDASIM7,8 output and the INPASIM code, which has been developed during this work. Given the magnetic equilibrium, plasma profiles, the fast-ion distribution function, and the detector geometry, FIDASIM calculates the flux of neutral particles, coming from CX reactions, with a Monte Carlo (MC) approach. To this end, a collisional-radiative model is solved.7 FIDASIM has been extensively verified against experimental data at the DIII-D,9AUG,10 and TCV11 tokamaks, the LHD12 stellarator, and other devices.8Using the FIDASIM output (velocity-space resolved neutral densities) as input, the INPASIM code calculates the synthetic signal, the resolution, and the instrument function to perform tomographic reconstructions as for FILD.13 INPASIM is divided into two independent sections: calculation of the signal/instrument function and determination of the scintillator strike map. In the former, the code tracks the FIDASIM markers inside thediagnostic head until they collide with thecollimator or scintillator, both considered as 3D elements. In the latter, MC markers with given energy and pitch are launched at the pinhole in order to create amap,whichrelates the(R,E)-spacetothe strikepositionofthe particle in the scintillator. For both sections, scattering and energy loss in the carbon foil are modeled following SRIM16 simulations, the ionization yield in the carbon foil follows DIII-D INPA modeling,17 and the scintillator yield follows Birk’s model as applied in absolute measurementsoffast-ionlossesatAUG.18,19 This modelpredictsthe numberof photonsemitted by thescintillator per incident ion. After the scintillator emission is calculated, INPASIM forms the camera image considering the transmission factor of the optical system and introducing a 2D Gaussian function to mimic the finite focusing of the optics, both based on Zemax simulations. IV. DESIGN OF THE INPA DIAGNOSTIC A. Selection of the position to place the INPA at AUG The selection of the position for the INPA diagnostic is a compromise between four factors: phase space coverage, signal level (attenuation), available space inside AUG, and resolution. The orientation of the diagnostic will determine which velocity directions can be measured and therefore which region of the FI phase space canbeprobed.Specialcarehasbeentakeninmatchingthepitchprofile of the slowing-down fast-ion distribution created by NBI #8 and #6of AUG,which are typicalexamples ofthe onandoff-axis FIdistributions achievable in AUG via NBI heating. A comparison of the pitch profiles of these distributions and the one INPA will explore can be found in Fig. 3. Here, the pitch profile explored by the INPA diagnosticis highlighted withthe white dashedline and thetrappedpassing boundary with the orange line. Notice that the INPA will be mainly sensitive to passing ions. Rev. Sci. Instrum. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-2 Published under license by AIP Publishing Review of Scientific Instruments ARTICLE scitation.org/journal/rsi FIG. 3. Pitch profile explored by the INPA diagnostic. (a) Projection in the (R,λ)- spaceofthefast-iondistribution,calculatedwiththeTRANSP14,15 code,usingNBI #6 and shot #19913. The dotted (squares) white line corresponds to a geometric calculation of the explored pitch profile based on the NBI geometry and the magnetic equilibrium, and the orange line (circles) corresponds to the trapped-passing boundary at the height of the magnetic axis. (b) Equivalent for NBI #8 and shot #30585. The acceptance of INPA is around ±0.1 on the pitch profile shown. The pinhole position to achieve this good overlapping is found in sector 16 of AUG, the sector closest to NBI #3, which provides the active source of neutrals. Therefore, the path that CX neutrals should travel inside the plasma and the corresponding losses due to re-ionization are minimized. The transmission factor for a high density case [ne(0)=6.5 ⋅1019 m−3] can be found in Fig. 4. FIG. 4. Fraction of CX neutrals created in reactions with the active source, which reachthedetectorpinhole,asafunctionoftheirradialbirthposition.Itcorresponds to a density of ne(0)=6.5 ⋅1019 m−3. Calculated with FIDASIM. B. Collimator and energy resolution The basic shape of the INPA collimator is sketched in Fig. 1. The collimator length, l, is fixed by the available distance to the first wall, but all the other parameters are free to be modified. The aperture angles of the collimator, αland αr, control its acceptance in the direction of the NBI, as can be seen in Fig. 5. They are 30○and 40○, respectively. These values enable a coverage of the region from R=1.35m up to the outer separatrix, approximately at R =2.16m. The acceptance of the diagnostic in the perpendicular direction, denoted by βin Fig. 6(a), dominates the energy resolution. This acceptance is controlled by the pinhole size and collimator length and height, h. Decreasing the pinhole radius or hwill improve the energy resolution but will reduce the signal level. As decreasing the pinhole radius reduces the neutral influx quadratically, his the chosen factor to pursuit the desired value of energy resolution, as it only affects the signal level linearly. A detailed comparison of the energy resolution for different values of hcan be found in Fig. 6(b). Values below h=3 mm are not considered in order to maintain a flux high enough to reach the Alfvénic time scale with the PMTs, and values above h=4 mm are avoided to keep a good energy resolution. The value h=3 mm was selected as the final choice, looking for the best energy resolution. Possible scattering caused by the carbon foil will also deteriorate the resolution, but the small thickness ofthecarbonfoil(20nm)makesthisscatteringunimportantforNBI injection energies at AUG. C. Radial resolution Two factors dominate the INPA radial resolution: the NBI source diameter, which acts as an active neutral source, and the diameter of the pinhole. Synthetic signals calculated by the INPASIM code have been used to estimate the radial resolution. To this end, the actual birth position of the markers has been compared with the position given by the strike map. The full width half maximum (FWHM) of the structures resulting from this comparison can be found in Fig. 6(c), where a parabolic fit has only been included as a guide to the eye. A priori, there is no model that justifies that the radial resolution must have a parabolic dependence with the major radius. D. Temporal resolution The signal-to-noise ratio (SNR) at the photo sensor of the camera and PMT was estimated as described in Ref. 20. The SNR, for the photon flux predicted by INPASIM, is plotted in Fig. 6(d). Taking a SNR of 10 up to 1 kHz of bandwidth could be achieved using FIG. 5. Effect of the collimator angles in the INPA view. In red is a 3D sketch of the INPA collimator, in shaded gray is the solid angle covered by the INPA, while in blue is the NBI line. Rev. Sci. Instrum. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-3 Published under license by AIP Publishing Review of Scientific Instruments ARTICLE scitation.org/journal/rsi FIG. 6. INPA resolution. (a) influence of the collimator height on the energy resolution; several possible trajectories of particles with the same energy are plotted. The larger the height of the carbon foil, the wider the distribution of impacts on the scintillator. (b) Energy resolution (FWHM) of INPA for different values of the collimatorheight. Here, the magneticfield is 2.5T on the axis.(c) Radial resolution ofINPA. Calculated forshot#30585. The magnetic axisposition is indicatedwitha solid black line while the outer separatrix position with a dotted one. The high field side (HFS) and low field side (LFS) are also indicated. (d) Signal-to-noise ratio of theacquisitionsystems.Thecalculation hasbeenperformedfor acasewithacore plasma density of ne(0)=6.5 ⋅1019 m−3. Calculated with INPASIM. FIG. 7. Section of the CAD design of the AUG INPA diagnostic. (a) In-vessel components and (b) out-vessel components. the Phantom camera (which enables the energy and radial resolutions presented above) and above 100 kHz could be achieved with the PMTs. E. Mechanical design An overview of the mechanical design of the AUG INPA diagnosticisshowninFig.7.Ascanbeseen,insidethedetectorhead,not only the first optical elements and the scintillator but also a calibration lamp is located, which allows us to check the alignment of the optical components between shot days. The head is connected with the port window with a periscope. This periscope not only allows us to meet the spatial boundary conditions but also provides the system with the necessary degrees of freedom to accommodate small deviations between computer aided design (CAD) and reality during installation. In the out-vessel region, a beam splitter divides the photon flux and redirects it toward the Phantom camera and the PMTs. V. RESPONSE OF THE INPA TO MHD FLUCTUATIONS A complete review of the INPA response to different plasma scenarios and FI redistributions is out of the scope of this article and will be presented in a follow-up paper. Here, we present only one example to show the capabilities of the diagnostic. In this case, we simulate the response of the INPA to a localized redistribution. An ad hoc anomalous diffusion coefficient has been inserted into TRANSP to mimic the effect of a localized MHD fluctuation in the FI distribution. A maximum value of the anomalous diffusion coefficient of 0.5 m2/s21,22 has been set, with an extension of 0.3 in the ρtspace (ρtis the normalized toroidal flux radius), which is in concordance with the size of the different poloidal modes of a Toroidal Alfvén Eigenmode (TAE) measured and simulated at DIII-D.23,24 250ms was simulated to give enough time for the FI to slow down. Rev. Sci. Instrum. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-4 Published under license by AIP Publishing Review of Scientific Instruments ARTICLE scitation.org/journal/rsi FIG. 8. (a) Differences in the FI distribution due to the assumed anomalous diffusion. Only the mid-plane region (z =0) has been plotted, which is calculated with TRANSP. (b) Relative changes in the scintillator yield. To avoid simulation noise, only pixels that have more than 5% signal are plotted. (b) Cut at 72 keV. Shaded areas represent a standard deviation of 5%, the expected upper limit of measurement noise. The gray dashed frames indicate where the anomalous diffusion was applied. As mentioned in Ref. 25, this simulation scheme cannot reproduce the precise interaction between the MHD instabilities and fast ions but gives an estimate of the particle transport. The anomalous diffusion has been selected to be constant in the interval ρt∈(0,0.3) and zero otherwise. In energy, the coefficient takes a Gaussian shape centered at 72 keV, with a FWHM of 10 keV, to mimic a narrow resonance. Anomalous diffusion has only been applied to passing particles. The relative difference in the distribution function can be seen in Fig. 8(a). Notice that in the region between R =1.55 and 1.85m, the FI density is smaller when the diffusion is activated, as expected (the range where the diffusion was applied is highlighted in gray). The differences in the scintillator signal can be seen in Fig. 8(b). Note how it agrees with the differences in the FI distribution, inside the resolution of the diagnostic. In Fig. 8(c), the difference along the constant line of 72 keV can be seen. Both curves will be distinguishable, even with the assumed 5% of noise. During thiscomparison, theplasma profiles havebeen consideredto remain constant. The INPA diagnostic is, in its current configuration, not expected to be sensible to a redistribution of trapped particles because the explored pitch profile is basically always in the passing region of the phase space (see Fig. 3). Only for regions close to the separatrix, the trapped-passing boundary enters the region explored by the INPA and the diagnostic becomes sensible to trapped particles. Therefore, in general, only passing ions or trapped ions, which become passing and enter the INPA field of view after the diffusion, will be measurable. VI. CONCLUSIONS An imaging neutral particle analyzer has been designed and optimized for the ASDEX Upgrade tokamak. This diagnostic will allow measurements of the distribution of supra-thermal particles in energy and radius with good resolution and a fast temporal response simultaneously, complementing the AUG suite of fast-ion diagnostics to obtain a complete understanding of the dynamics and transport of supra-thermal particles. The final design of the diagnostic features an energy resolution of 12 keV for 100 keV ions and a radial resolution below 8cm at the low field side of AUG, with a temporal response of 1 kHz. If the fast acquisition system (with low spatial resolution) is used, the response is increased to above 100 kHz. ACKNOWLEDGMENTS This project received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant Agreement No. 805162) and the Spanish Ministerio de Ciencia, Innovación y Universidades (Grant No. FPU19/02486). DATA AVAILABILITY The data that support the findings of this study are available from the corresponding author upon reasonable request. REFERENCES 1A. Fasoli et al.,Nucl. Fusion 47, S264 (2007). 2X. D. Du et al.,Nucl. Fusion 58, 082006 (2018). 3M. A. Van Zeeland et al.,J. Instrum. 14, C09027 (2019). 4S. S. Medley et al.,Rev. Sci. Instrum. 79, 011101 (2008). 5M. García-Muñoz et al.,Rev. Sci. Instrum. 80, 053503 (2009). 6I. H. Hutchinson, Principles of Plasma Diagnostics, 2nd ed. (Cambridge University Press, 2002), p. 322. 7W. W. Heidbrink et al.,Commun. Comput. Phys. 10, 716–741 (2011). 8B. Geiger et al.,Plasma Phys. Controlled Fusion 62, 105008 (2020). 9Y. Luo et al.,Phys. Plasmas 14, 112503 (2007). 10P. A. Schneider et al.,Rev. Sci. Instrum. 86, 073508 (2015). 11B. Geiger et al.,Plasma Phys. Controlled Fusion 59, 115002 (2017). 12Y. Fujiwara et al.,Nucl. Fusion 60, 112014 (2020). 13J. Galdon-Quiroga et al.,Plasma Phys. Controlled Fusion 60, 105005 (2018). 14R. J. Hawryluk, Phys. Plasmas Close Thermonucl. Cond. 1, 19–46 (1981). 15J. Breslau et al., Transp, Computer Software https://doi.org/10.11578/ dc.20180627.4, 2018. 16J. F. Ziegler, Nucl. Instrum. Methods Phys. Res., Sect. B 219-220, 1027 (2004). 17X. D. Du et al.,Nucl. Fusion 60, 112001 (2020). 18M. Rodríguez, Ph.D. thesis, University of Seville, Seville, 2017. 19J. Galdon-Quiroga et al.,Nucl. Fusion 58, 036005 (2018). 20Hamamatsu-Photonics, Photomultiplier Tubes, Word Technology ed. (Hamamatsu Photonics, 2007), p. 73. 21M. A. Van Zeeland et al.,Phys. Plasmas 18, 056114 (2011). 22W. W. Heidbrink et al.,Nucl. Fusion 48, 084001 (2008). 23R. B. White et al.,Phys. Plasmas 17, 056107 (2010). 24M. A. Van Zeeland et al.,Phys. Rev. Lett. 97, 135001 (2006). 25W. W. Heidbrink et al.,Phys. Plasmas 24, 056109 (2017). Rev. Sci. Instrum. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-5 Published under license by AIP Publishing