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

Re . Sci. Ins um. 92, 043554 (2021); h ps://doi.o g/10.1063/5.0043768 92, 043554 © 2021 Au ho (s). Design and simula ion o an imaging neu al pa icle analyze o he ASDEX Upg ade okamak Ci e as: Re . Sci. Ins um. 92, 043554 (2021); h ps://doi.o g/10.1063/5.0043768 Submi ed: 11 Janua y 2021 • Accep ed: 25 Ma ch 2021 • Published Online: 26 Ap il 2021 J. Rueda-Rueda, M. Ga cía-Muñoz, E. Viezze , e al. COLLECTIONS Pape published as pa o he special opic on P oceedings o he 23 d Topical Con e ence on High-Tempe a u e Plasma Diagnos ics ARTICLES YOU MAY BE INTERESTED IN Implemen a ion o syn he ic as -ion loss de ec o and imaging hea y ion beam p obe diagnos ics in he 3D hyb id kine ic-MHD code MEGA Re iew o Scien i ic Ins umen s 92, 043558 (2021); h ps://doi.o g/10.1063/5.0043757 Sel -adap i e diagnos ic o adial as -ion loss measu emen s on he ASDEX Upg ade okamak (in i ed) Re iew o Scien i ic Ins umen s 92, 053538 (2021); h ps://doi.o g/10.1063/5.0043756 Upg ade and absolu e calib a ion o he JET scin illa o -based as -ion loss de ec o Re iew o Scien i ic Ins umen s 92, 043553 (2021); h ps://doi.o g/10.1063/5.0043760 Re iew o Scien i ic Ins umen s ARTICLE sci a ion.o g/jou nal/ si Design and simula ion o an imaging neu al pa icle analyze o he ASDEX Upg ade okamak Ci e as: Re . Sci. Ins um. 92, 043554 (2021); doi: 10.1063/5.0043768 Submi ed: 11 Janua y 2021 •Accep ed: 25 Ma ch 2021 • Published Online: 26 Ap il 2021 J. Rueda-Rueda,1,a) M. Ga cía-Muñoz,1,2 E. Viezze ,1,2 P. A. Schneide ,3J. Ga cía-Domínguez,2 J. Ayllon-Gue ola,2J. Galdón-Qui oga,1,3 A. He mann,3X. D. Du,4M. A. Van Zeeland,4P. Oyola,1 M. Rod iguez-Ramos,5and ASDEX Upg ade Teamb) AFFILIATIONS 1Depa men o A omic, Molecula and Nuclea Physics, A . Reina Me cedes s/n, 41012 Se ille, Spain 2Cen o Nacional de Acele ado es (CNA) CSIC, 41092 Se ille, Spain 3Max Planck Ins i u e o Plasma Physics, Bol zmanns . 2, 85748 Ga ching, Ge many 4Gene al A omics, P.O. Box 85608, San Diego, Cali o nia 92186-5608, USA 5Labo a o y o Ion Beam In e ac ions, Ru-- de Boˇ sko i´ c Ins i u e, 10000 Zag eb, C oa ia No e: Pape published as pa o he Special Topic on P oceedings o he 23 d Topical Con e ence on High-Tempe a u e Plasma Diagnos ics. a)Au ho o whom co espondence should be add essed: [email p o ec ed] b)See he au ho lis o H. Meye e al., Nucl. Fusion 59, 112014 (2019). ABSTRACT An Imaging Neu al Pa icle Analyze (INPA) diagnos ic has been designed o he ASDEX Upg ade (AUG) okamak. The AUG INPA diagnos ic will measu e as neu als escaping he plasma a e cha ge exchange eac ions. The neu als will be ionized by a 20nm ca bon oil and de lec ed owa d a scin illa o by he local magne ic ield. The use o a neu al beam injec o (NBI) as an ac i e sou ce o neu als will p o ide adially esol ed measu emen s, while he use o a scin illa o as an ac i e componen will allow us o co e he whole plasma along he NBI line wi h unp eceden ed phase-space esolu ion (<12 keV and 8cm) and a as empo al esponse (up o 1 kHz wi h he high esolu ion acquisi ion sys em and abo e 100 kHz wi h he low esolu ion one), making i sui able o s udy localized as -ion edis ibu ions in phase space. Published unde license by AIP Publishing. h ps://doi.o g/10.1063/5.0043768 I. INTRODUCTION A de ailed unde s anding o he as -ion (FI) beha io in he p esence o magne ohyd odynamic (MHD) luc ua ions is manda o y o achie ing a good as -ion con inemen in u u e usion de ices.1Fo his pu pose, no el diagnos ic ech- niques o measu e he FI dis ibu ion in phase space wi h Al énic empo al esolu ion a e cu en ly being de eloped. The Imaging Neu al Pa icle Analyze (INPA), al eady ins alled a he DIII-D okamak,2,3 which is able o measu e he adial posi ion and ene gy o he con ined FI popula ion wi h a as empo al esponse, is one example among hese no el diagnos ics. This pape is s uc u ed as ollows: he INPA wo king p inciple is explained in Sec. II. Sec ion III p esen s he syn he ic diagnos- ic, while Sec. IV explains he in luence o he di e en geome i- cal pa ame e s on he de ec o pe o mance. Sec ion Vp esen s he esponse o he INPA o FI edis ibu ions due o magne ohyd ody- namic (MHD) ac i i y. II. WORKING PRINCIPLE The INPA combines he al eady wo king p inciples o neu al pa icle analyze s (NPAs)4and as -ion loss de ec o s (FILDs)5 o p o ide he ene gy and loca ion o he con ined as -ion popula ion Re . Sci. Ins um. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-1 Published unde license by AIP Publishing Re iew o Scien i ic Ins umen s ARTICLE sci a ion.o g/jou nal/ si FIG.1.Schemeo heINPAwo kingp inciple.Inblueis he ajec o yo heneu al pa icle, while in ed is he ajec o y o he ion. (a) and (b) show he side and op iews, espec i ely, while (c) shows a 3D ske ch. wi h a as empo al esponse. Simila o NPA sys ems, he INPA analyzes neu al pa icles p oduced in cha ge exchange (CX) eac- ions. These pa icles a e no con ined by he magne ic ield and e ain all he in o ma ion o as ions, as no signi ican exchange o ene gy o momen um occu s du ing he CX eac ion.6A e being collima ed, neu als a e ionized by an ul a- hin [20nm, o he ASDEX Upg ade (AUG) se up] ca bon oil and de lec ed owa d a scin illa o pla e, as can be seen in Fig. 1. The s ike posi ion o he pa icles in he scin illa o will be de e mined by hei ene gy and pi ch (λ=− ∥/ , whe e ∥is he eloci y along he magne ic ield) and he local magne ic ield o he okamak. The use o a neu al beam injec o (NBI) as an ac i e sou ce o neu als gi es he possi- bili y o ela ing, ia modeling, he measu ed pi ch o he neu als o hei adial bi h posi ion and hence o he loca ion o he con- ined FI. This is possible hanks o he geome ic ela ion be ween he eloci y o ien a ion o he measu ed neu al and i s adial bi h posi ion. Figu e 2 shows an example o his ela ion o he case o AUG.The in- esselposi ioning o he diagnos icopens upa b oade lexibili yin selec ing he lines-o -sigh (LOSs)compa ed o he con- en ional elec omagne ic NPA ins alled ou - essel and using hei own elec omagne ic ields.4The use o a scin illa o will enable he simul aneous explo a ion o a wide egion o he plasma, which FIG.2.Rela ionbe ween heline-o -sigh (LOS)and adialbi hposi ion.Asneu al pa icles a el along a s aigh line, and assuming ha hey a e all bo n along he NBI line, he e is a one- o-one ela ion be ween he LOS ( eloci y di ec ion o he neu al) and i s adial bi h posi ion. In blue is he NBI line, in g een a e some o he explo ed LOSs, and in ed is he loca ion o he INPA head. would no be co e ed by indi idual LOSs. Ligh emi ed by he scin- illa o is collec ed by a se ies o lenses and guided o a Phan om came a and a se o pho omul iplie s (PMTs). The came a allows us o ob ain high esolu ion images o he scin illa o , while he PMTs sac i ice his esolu ion o gain a as e empo al esponse. III. SYNTHETIC DIAGNOSTIC A syn he ic INPA diagnos ic has been de eloped o s udy he easibili y o he ins alla ion o an INPA a AUG and o op imize i s design. The syn he ic INPA is based on FIDASIM7,8 ou pu and he INPASIM code, which has been de eloped du ing his wo k. Gi en he magne ic equilib ium, plasma p o iles, he as -ion dis ibu ion unc ion, and he de ec o geome y, FIDASIM calcula es he lux o neu al pa icles, coming om CX eac ions, wi h a Mon e Ca lo (MC) app oach. To his end, a collisional- adia i e model is sol ed.7 FIDASIM has been ex ensi ely e i ied agains expe imen al da a a he DIII-D,9AUG,10 and TCV11 okamaks, he LHD12 s ella a- o , and o he de ices.8Using he FIDASIM ou pu ( eloci y-space esol ed neu al densi ies) as inpu , he INPASIM code calcula es he syn he ic signal, he esolu ion, and he ins umen unc ion o pe o m omog aphic econs uc ions as o FILD.13 INPASIM is di ided in o wo independen sec ions: calcula ion o he sig- nal/ins umen unc ion and de e mina ion o he scin illa o s ike map. In he o me , he code acks he FIDASIM ma ke s inside hediagnos ic head un il hey collide wi h hecollima o o scin illa- o , bo h conside ed as 3D elemen s. In he la e , MC ma ke s wi h gi en ene gy and pi ch a e launched a he pinhole in o de o c ea e amap,which ela es he(R,E)-space o he s ikeposi iono he pa - icle in he scin illa o . Fo bo h sec ions, sca e ing and ene gy loss in he ca bon oil a e modeled ollowing SRIM16 simula ions, he ioniza ion yield in he ca bon oil ollows DIII-D INPA modeling,17 and he scin illa o yield ollows Bi k’s model as applied in absolu e measu emen so as -ionlossesa AUG.18,19 This modelp edic s he numbe o pho onsemi ed by hescin illa o pe inciden ion. A e he scin illa o emission is calcula ed, INPASIM o ms he came a image conside ing he ansmission ac o o he op ical sys em and in oducing a 2D Gaussian unc ion o mimic he ini e ocusing o he op ics, bo h based on Zemax simula ions. IV. DESIGN OF THE INPA DIAGNOSTIC A. Selec ion o he posi ion o place he INPA a AUG The selec ion o he posi ion o he INPA diagnos ic is a com- p omise be ween ou ac o s: phase space co e age, signal le el (a enua ion), a ailable space inside AUG, and esolu ion. The o i- en a ion o he diagnos ic will de e mine which eloci y di ec ions can be measu ed and he e o e which egion o he FI phase space canbep obed.Specialca ehasbeen akeninma ching hepi chp o- ile o he slowing-down as -ion dis ibu ion c ea ed by NBI #8 and #6o AUG,which a e ypicalexamples o he on- ando -axis FIdis- ibu ions achie able in AUG ia NBI hea ing. A compa ison o he pi ch p o iles o hese dis ibu ions and he one INPA will explo e can be ound in Fig. 3. He e, he pi ch p o ile explo ed by he INPA diagnos icis highligh ed wi h he whi e dashedline and he apped- passing bounda y wi h he o ange line. No ice ha he INPA will be mainly sensi i e o passing ions. Re . Sci. Ins um. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-2 Published unde license by AIP Publishing Re iew o Scien i ic Ins umen s ARTICLE sci a ion.o g/jou nal/ si FIG. 3. Pi ch p o ile explo ed by he INPA diagnos ic. (a) P ojec ion in he (R,λ)- spaceo he as -iondis ibu ion,calcula edwi h heTRANSP14,15 code,usingNBI #6 and sho #19913. The do ed (squa es) whi e line co esponds o a geome ic calcula ion o he explo ed pi ch p o ile based on he NBI geome y and he mag- ne ic equilib ium, and he o ange line (ci cles) co esponds o he apped-passing bounda y a he heigh o he magne ic axis. (b) Equi alen o NBI #8 and sho #30585. The accep ance o INPA is a ound ±0.1 on he pi ch p o ile shown. The pinhole posi ion o achie e his good o e lapping is ound in sec o 16 o AUG, he sec o closes o NBI #3, which p o ides he ac i e sou ce o neu als. The e o e, he pa h ha CX neu als should a el inside he plasma and he co esponding losses due o e-ioniza ion a e minimized. The ansmission ac o o a high densi y case [ne(0)=6.5 ⋅1019 m−3] can be ound in Fig. 4. FIG. 4. F ac ion o CX neu als c ea ed in eac ions wi h he ac i e sou ce, which each hede ec o pinhole,asa unc iono hei adialbi hposi ion.I co esponds o a densi y o ne(0)=6.5 ⋅1019 m−3. Calcula ed wi h FIDASIM. B. Collima o and ene gy esolu ion The basic shape o he INPA collima o is ske ched in Fig. 1. The collima o leng h, l, is ixed by he a ailable dis ance o he i s wall, bu all he o he pa ame e s a e ee o be modi ied. The ape - u e angles o he collima o , αland α , con ol i s accep ance in he di ec ion o he NBI, as can be seen in Fig. 5. They a e 30○and 40○, espec i ely. These alues enable a co e age o he egion om R=1.35m up o he ou e sepa a ix, app oxima ely a R =2.16m. The accep ance o he diagnos ic in he pe pendicula di ec ion, deno ed by βin Fig. 6(a), domina es he ene gy esolu ion. This accep ance is con olled by he pinhole size and collima o leng h and heigh , h. Dec easing he pinhole adius o hwill imp o e he ene gy esolu ion bu will educe he signal le el. As dec easing he pinhole adius educes he neu al in lux quad a ically, his he cho- sen ac o o pu sui he desi ed alue o ene gy esolu ion, as i only a ec s he signal le el linea ly. A de ailed compa ison o he ene gy esolu ion o di e en alues o hcan be ound in Fig. 6(b). Val- ues below h=3 mm a e no conside ed in o de o main ain a lux high enough o each he Al énic ime scale wi h he PMTs, and alues abo e h=4 mm a e a oided o keep a good ene gy esolu- ion. The alue h=3 mm was selec ed as he inal choice, looking o he bes ene gy esolu ion. Possible sca e ing caused by he ca - bon oil will also de e io a e he esolu ion, bu he small hickness o heca bon oil(20nm)makes hissca e ingunimpo an o NBI injec ion ene gies a AUG. C. Radial esolu ion Two ac o s domina e he INPA adial esolu ion: he NBI sou ce diame e , which ac s as an ac i e neu al sou ce, and he diame e o he pinhole. Syn he ic signals calcula ed by he INPASIM code ha e been used o es ima e he adial esolu ion. To his end, he ac ual bi h posi ion o he ma ke s has been compa ed wi h he posi ion gi en by he s ike map. The ull wid h hal maxi- mum (FWHM) o he s uc u es esul ing om his compa ison can be ound in Fig. 6(c), whe e a pa abolic i has only been included as a guide o he eye. A p io i, he e is no model ha jus i ies ha he adial esolu ion mus ha e a pa abolic dependence wi h he majo adius. D. Tempo al esolu ion The signal- o-noise a io (SNR) a he pho o senso o he cam- e a and PMT was es ima ed as desc ibed in Re . 20. The SNR, o he pho on lux p edic ed by INPASIM, is plo ed in Fig. 6(d). Tak- ing a SNR o 10 up o 1 kHz o bandwid h could be achie ed using FIG. 5. E ec o he collima o angles in he INPA iew. In ed is a 3D ske ch o he INPA collima o , in shaded g ay is he solid angle co e ed by he INPA, while in blue is he NBI line. Re . Sci. Ins um. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-3 Published unde license by AIP Publishing Re iew o Scien i ic Ins umen s ARTICLE sci a ion.o g/jou nal/ si FIG. 6. INPA esolu ion. (a) in luence o he collima o heigh on he ene gy eso- lu ion; se e al possible ajec o ies o pa icles wi h he same ene gy a e plo ed. The la ge he heigh o he ca bon oil, he wide he dis ibu ion o impac s on he scin illa o . (b) Ene gy esolu ion (FWHM) o INPA o di e en alues o he collima o heigh . He e, he magne ic ield is 2.5T on he axis.(c) Radial esolu ion o INPA. Calcula ed o sho #30585. The magne ic axisposi ion is indica edwi ha solid black line while he ou e sepa a ix posi ion wi h a do ed one. The high ield side (HFS) and low ield side (LFS) a e also indica ed. (d) Signal- o-noise a io o heacquisi ionsys ems.Thecalcula ion hasbeenpe o med o acasewi haco e plasma densi y o ne(0)=6.5 ⋅1019 m−3. Calcula ed wi h INPASIM. FIG. 7. Sec ion o he CAD design o he AUG INPA diagnos ic. (a) In- essel componen s and (b) ou - essel componen s. he Phan om came a (which enables he ene gy and adial esolu- ions p esen ed abo e) and abo e 100 kHz could be achie ed wi h he PMTs. E. Mechanical design An o e iew o he mechanical design o he AUG INPA diag- nos icisshowninFig.7.Ascanbeseen,inside hede ec o head,no only he i s op ical elemen s and he scin illa o bu also a calib a- ion lamp is loca ed, which allows us o check he alignmen o he op ical componen s be ween sho days. The head is connec ed wi h he po window wi h a pe iscope. This pe iscope no only allows us o mee he spa ial bounda y condi ions bu also p o ides he sys- em wi h he necessa y deg ees o eedom o accommoda e small de ia ions be ween compu e aided design (CAD) and eali y du - ing ins alla ion. In he ou - essel egion, a beam spli e di ides he pho on lux and edi ec s i owa d he Phan om came a and he PMTs. V. RESPONSE OF THE INPA TO MHD FLUCTUATIONS A comple e e iew o he INPA esponse o di e en plasma scena ios and FI edis ibu ions is ou o he scope o his a icle and will be p esen ed in a ollow-up pape . He e, we p esen only one example o show he capabili ies o he diagnos ic. In his case, we simula e he esponse o he INPA o a localized edis ibu ion. An ad hoc anomalous di usion coe icien has been inse ed in o TRANSP o mimic he e ec o a localized MHD luc ua ion in he FI dis ibu ion. A maximum alue o he anomalous di usion coe - icien o 0.5 m2/s21,22 has been se , wi h an ex ension o 0.3 in he ρ space (ρ is he no malized o oidal lux adius), which is in con- co dance wi h he size o he di e en poloidal modes o a To oidal Al én Eigenmode (TAE) measu ed and simula ed a DIII-D.23,24 250ms was simula ed o gi e enough ime o he FI o slow down. Re . Sci. Ins um. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-4 Published unde license by AIP Publishing Re iew o Scien i ic Ins umen s ARTICLE sci a ion.o g/jou nal/ si FIG. 8. (a) Di e ences in he FI dis ibu ion due o he assumed anomalous di usion. Only he mid-plane egion (z =0) has been plo ed, which is calcula ed wi h TRANSP. (b) Rela i e changes in he scin illa o yield. To a oid simula ion noise, only pixels ha ha e mo e han 5% signal a e plo ed. (b) Cu a 72 keV. Shaded a eas ep esen a s anda d de ia ion o 5%, he expec ed uppe limi o measu emen noise. The g ay dashed ames indica e whe e he anomalous di usion was applied. As men ioned in Re . 25, his simula ion scheme canno ep oduce he p ecise in e ac ion be ween he MHD ins abili ies and as ions bu gi es an es ima e o he pa icle anspo . The anomalous di - usion has been selec ed o be cons an in he in e al ρ ∈(0,0.3) and ze o o he wise. In ene gy, he coe icien akes a Gaussian shape cen e ed a 72 keV, wi h a FWHM o 10 keV, o mimic a na ow esonance. Anomalous di usion has only been applied o passing pa icles. The ela i e di e ence in he dis ibu ion unc ion can be seen in Fig. 8(a). No ice ha in he egion be ween R =1.55 and 1.85m, he FI densi y is smalle when he di usion is ac i a ed, as expec ed ( he ange whe e he di usion was applied is highligh ed in g ay). The di e ences in he scin illa o signal can be seen in Fig. 8(b). No e how i ag ees wi h he di e ences in he FI dis i- bu ion, inside he esolu ion o he diagnos ic. In Fig. 8(c), he di - e ence along he cons an line o 72 keV can be seen. Bo h cu es will be dis inguishable, e en wi h he assumed 5% o noise. Du ing hiscompa ison, heplasma p o iles ha ebeen conside ed o emain cons an . The INPA diagnos ic is, in i s cu en con igu a ion, no expec ed o be sensible o a edis ibu ion o apped pa icles because he explo ed pi ch p o ile is basically always in he passing egion o he phase space (see Fig. 3). Only o egions close o he sepa a ix, he apped-passing bounda y en e s he egion explo ed by he INPA and he diagnos ic becomes sensible o apped pa i- cles. The e o e, in gene al, only passing ions o apped ions, which become passing and en e he INPA ield o iew a e he di usion, will be measu able. VI. CONCLUSIONS An imaging neu al pa icle analyze has been designed and op imized o he ASDEX Upg ade okamak. This diagnos ic will allow measu emen s o he dis ibu ion o sup a- he mal pa icles in ene gy and adius wi h good esolu ion and a as empo al esponse simul aneously, complemen ing he AUG sui e o as -ion diagnos ics o ob ain a comple e unde s anding o he dynamics and anspo o sup a- he mal pa icles. The inal design o he diagnos ic ea u es an ene gy esolu ion o 12 keV o 100 keV ions and a adial esolu ion below 8cm a he low ield side o AUG, wi h a empo al esponse o 1 kHz. I he as acquisi ion sys em (wi h low spa ial esolu ion) is used, he esponse is inc eased o abo e 100 kHz. ACKNOWLEDGMENTS This p ojec ecei ed unding om he Eu opean Resea ch Council (ERC) unde he Eu opean Union’s Ho izon 2020 esea ch and inno a ion p og amme (G an Ag eemen No. 805162) and he Spanish Minis e io de Ciencia, Inno ación y Uni e sidades (G an No. FPU19/02486). DATA AVAILABILITY The da a ha suppo he indings o his s udy a e a ailable om he co esponding au ho upon easonable eques . REFERENCES 1A. Fasoli e al.,Nucl. Fusion 47, S264 (2007). 2X. D. Du e al.,Nucl. Fusion 58, 082006 (2018). 3M. A. Van Zeeland e al.,J. Ins um. 14, C09027 (2019). 4S. S. Medley e al.,Re . Sci. Ins um. 79, 011101 (2008). 5M. Ga cía-Muñoz e al.,Re . Sci. Ins um. 80, 053503 (2009). 6I. H. Hu chinson, P inciples o Plasma Diagnos ics, 2nd ed. (Camb idge Uni e si y P ess, 2002), p. 322. 7W. W. Heidb ink e al.,Commun. Compu . Phys. 10, 716–741 (2011). 8B. Geige e al.,Plasma Phys. Con olled Fusion 62, 105008 (2020). 9Y. Luo e al.,Phys. Plasmas 14, 112503 (2007). 10P. A. Schneide e al.,Re . Sci. Ins um. 86, 073508 (2015). 11B. Geige e al.,Plasma Phys. Con olled Fusion 59, 115002 (2017). 12Y. Fujiwa a e al.,Nucl. Fusion 60, 112014 (2020). 13J. Galdon-Qui oga e al.,Plasma Phys. Con olled Fusion 60, 105005 (2018). 14R. J. Haw yluk, Phys. Plasmas Close The monucl. Cond. 1, 19–46 (1981). 15J. B eslau e al., T ansp, Compu e So wa e h ps://doi.o g/10.11578/ dc.20180627.4, 2018. 16J. F. Ziegle , Nucl. Ins um. Me hods Phys. Res., Sec . B 219-220, 1027 (2004). 17X. D. Du e al.,Nucl. Fusion 60, 112001 (2020). 18M. Rod íguez, Ph.D. hesis, Uni e si y o Se ille, Se ille, 2017. 19J. Galdon-Qui oga e al.,Nucl. Fusion 58, 036005 (2018). 20Hamama su-Pho onics, Pho omul iplie Tubes, Wo d Technology ed. (Hama- ma su Pho onics, 2007), p. 73. 21M. A. Van Zeeland e al.,Phys. Plasmas 18, 056114 (2011). 22W. W. Heidb ink e al.,Nucl. Fusion 48, 084001 (2008). 23R. B. Whi e e al.,Phys. Plasmas 17, 056107 (2010). 24M. A. Van Zeeland e al.,Phys. Re . Le . 97, 135001 (2006). 25W. W. Heidb ink e al.,Phys. Plasmas 24, 056109 (2017). Re . Sci. Ins um. 92, 043554 (2021); doi: 10.1063/5.0043768 92, 043554-5 Published unde license by AIP Publishing