41CA measurement with low energy accelerator mass spectrometry (leams) at the Centro Nacional de Aceleradores.

UNIVERSIDAD DE SEVILLA DEPARTAMENTO DE F´ ISICA APLICADA I 41Ca measurement with Low Energy Accelerator Mass Spectrometry (LEAMS) at the Centro Nacional de Aceleradores A thesis submitted for the degree of DOCTOR IN PHYSICAL SCIENCES AND TECHOLOGIES PRESENTED BY CARLOS VIVO VILCHES SUPERVISORS Dr. Jos´ e Mar ´ ıa L´ opez Guti´ errez Dr. Christof Vockenhuber 2018 Jos´ e Mar´ıa L´ opez Guti´ errez, Profesor Titular del Departamento de F ´ ısica Aplicada I de la Universidad de Sevilla; y Christof Vockenhuber, Cient ´ ıfico Titular del Laboratorio de F ´ ısica de Haces de Iones de la Escuela Polit´ ecnica Federal de Z´ urich CERTIFICAN: Que la presente memoria ”41Ca measurement with Low Energy Accelerator Mass Spectrometry (LEAMS) at the Centro Nacional de Aceleradores”ha sido realiza bajo su direcci´ on conjunta en el Centro Nacional de Aceleradores por Carlos Vivo Vilches para optar al grado de Doctor en Ciencias y Tecnolog ´ ıas F ´ ısicas. Para que as ´ ı conste, y en cumplimiento con la legislaci´ on vigente, firman el presente documento. Sevilla, a X de junio de 2018 Jos´ e Mar ´ ıa L´ opez Guti´ errez Christof Vockenhuber A mi familia, con todo mi amor In this age of specialization men who thoroughly know one field are often incompetent to discuss another. The great problems of the relations between one and another aspect of human activity have for this reason been discussed less and less in public. When we look at the past great debates on these subjects we feel jealous of those times, for we should have liked the excitement of such argument. The old problems, such as the relation of science and religion, are still with us, and I believe present as difficult dilemmas as ever, but they are not often publicly discussed because of the limitations of specialization. Richard Feynman Agradecimientos Danksagung Ringraziamenti ¿Qu´ e queda por decir, adem´ as de alea iacta est? Pues falta agradecer a toda la gente que me ha acompa˜ nado durante este doctorado, que no es poca. Voy a intentar ir dejando los distintos agradecimientos en el orden m´ as est´ andar, ya que en esto de los agradecimientos de tesis ya hay cierta jurisprudencia. ¨ Uber diese Teile dass ich auf Deutsch schreibe, muss ich sache entschuldigung, weil ich mein Deutsch vergessen habe, und vielleicht etwas is nicht v¨ ollig richtig. Aunque en el t ´ ıtulo tambi´ en a˜ nado la palabra ringraziamenti, no me atrever´ e a escribir m´ as cosas en italiano que alg´ un grazie suelto. Suficiente italiano inventado han soportado los destinatarios durante estos a˜ nos. Y como sigo un orden est´ andar, empezar´ e estos agradicimientos por mis directores. Agradecer a Jos´ e Mar ´ ıa su constante supervisi´ on, su disponibilidad y su buen rollo. Si alguien se merece el reconocimiento principal por la direcci´ on de esta tesis y decir el d ´ ıa de ma˜ nana ”a ese chico le dirig ´ ı yo la tesis”, es sin duda ´ el. Ich muss also mein co-Betreuer, Christof, danken. Obwohl er noch nicht mein Betreuer w¨ arend meines 4-Monate Besuch nach ETH Zurich war, hat er wie mein Betreuer mit mir benommen. Seine Erfahrung in 41Ca AMS und in der berichtigen der Kaliuminterferenz waren notwendig f¨ ur die Messunge in Seville. Y debo tambi´ en dar las gracias a Manolo, quien me dio esta oportunidad y ha estado siempre atento a todos los problemas burocr´ aticos a pesar de estar siempre hasta arriba de trabajo. Que alguien con tanto jaleo est´ e siempre que puede disponible es muy de agradecer, y es una pena que al final no conste como mi director. Tambi´ en tengo que agradecer al resto del grupo de AMS del CNA, porque creo que es dif ´ ıcil encontrar un grupo con el mismo buen rollo, e imposible con mejor: a Javi, por sus tardes de juegos y su sentido del humor; a Elena, porque su inigualable conocimiento de SARA ha sido una ayuda constante, amable y a menudo requerida durante estos cuatro a˜ nos; a Juan I CONTENTS B.2 The experiment at SARA ................................ 112 C FORTRAN programs for 41Ca measurement analysis 115 C.1 Input file ........................................ 115 C.2 Output files ....................................... 117 C.2.1 out-K-correction ................................ 117 C.2.2 out-runs .................................... 118 C.2.3 out-blank .................................... 118 C.2.4 out-stds ..................................... 118 C.2.5 out-final-ratios ................................. 119 Bibliography 121 Curriculum Vitae 137 VIII List of Figures 1.1 Basic scheme of the stripping process. ....................... 5 1.2 Basic scheme of an AMS system. ........................... 5 1.3 Layout of several AMS facilities. ........................... 8 2.1 Picture and schematic view of the 1 MV AMS system at CNA Seville. ....... 14 2.2 Picture and schematic view of the 600 kV AMS system at ETH. .......... 15 2.3 Scheme of a high-current sputtering ion source. .................. 16 2.4 Simulation of ions passing through an Einzel lens on SIMION®. ......... 17 2.5 Piston taking a target out of the SARA sample carousel............... 18 2.6 SIMION® simulation of the SO-110B ion source. ................. 18 2.7 Target wheel used by the MC-SNICS ion source. .................. 19 2.8 SIMION® simulation of the ion source from TANDY. ............... 19 2.9 Basic operation principle of the systems used by the 2 different types of particle accelerator to get the high terminal voltage. .................... 22 2.10 Mean charge state of uranium ions as a function of the ion energy for different stripper materials. ................................... 24 2.12 SIMION® view of an electrostatic quadrupole. ................... 26 2.13 Pellet chain used to transport the charge in the pelletron from TANDY. ..... 27 2.14 Focusing in the X axis of magnetic dipoles depending on the inclination of the borders. ......................................... 29 IX LIST OF FIGURES 2.15 Focusing in the Y axis of magnetic dipoles depending on the inclination of the borders. ......................................... 30 2.16 Scheme of an ESA with spherical geometry. .................... 31 2.17 Behaviour of different ions with the same mass but different energy in an achromatic system. .................................. 32 2.18 GICOSY simulation of 41Ca2+ ions in the HE side of SARA. ............. 33 2.19 Faraday cups used in AMS measurements with SARA. ............... 34 2.20 GICOSY simulation of 41Ca2+ ions in the HE side of TANDY. ............ 34 2.21 Scheme of the operation principle of a GIC. .................... 35 2.22 Stopping power of 41Ca and 41K ions with an energy of 26.5 MeV in butane according to SRIM. ................................... 36 2.23 Detector signals in a 41Ca measurement with the 5 MV AMS system at SUERC, Glasgow. ........................................ 37 2.24 Scheme of the compact GIC at SARA. ........................ 38 2.25 Comparison between the old HVE GIC, and the new compact GIC. ....... 38 2.26 Picture and scheme of the retractable GIC used in the 41Ca measurements with TANDY. ......................................... 39 3.1 SRIM simulation of the trajectories of several 41Ca ions at two different energies in a GIC detector. ................................. 42 3.2 Spectra simulation with SRIM of an equal number of counts of 41Ca and 41K at different energies. ................................... 42 3.3 Evolution of 39K/40Ca ratios from blank and standard samples in target holders from different materials. ............................... 47 3.4 Time evolution of the 40Ca2+ current, the 41M2+ and 39K2+ rates, and the 39K/40Ca, 41M/40Ca and 41M/39K ratios for a blank sample (blank 1) during a measurement at SARA. ..................................... 49 3.5 Time evolution of the 40Ca2+ current, the 41M2+ and 39K2+ rates, and the 39K/40Ca, 41M/40Ca and 41M/39K ratios for another blank sample (blank 2) during a measurement at SARA. ................................... 50 X LIST OF FIGURES 3.6 Typical scheme of a 41Ca measurement with a compact AMS system implementing the K-correction. ............................... 53 4.1 (40CaF3)-current output from the SO-110B ion source at SARA for different blank targets mixed with silver or niobium at different weight ratios. ...... 63 4.2 (40CaF3)-current from 2 different blank samples over a long time range. . . . . 64 4.3 40Ca2+ transmission and 41M/40Ca background as functions of the He stripper pressure at SARA. ................................... 65 4.4 40Ca3+ transmission as a function of the He stripper pressure at SARA. ..... 67 4.5 Charge state distribution for the (40CaF3)-→40Caqstripping process in He at 225 keV and 400 keV. ................................. 68 4.6 Relative position of the FC3 depending on the isotope whose current is measured and the radionuclide which is tuned to cross to the ESA. ......... 69 4.7 Experimental a simulated spectra for 41Ca2+ ions reaching the compact GIC at SARA. .......................................... 70 4.8 Experimental spectra for 41Ca3+ and 39K3+ ions reaching the compact GIC at SARA. .......................................... 71 4.9 Relationship between the 41M rate reaching the compact GIC and the detector efficiency. ........................................ 72 4.10 Linear correlation between the 41M/40Ca and 39K/40Ca ratios in blank samples at SARA. ......................................... 73 4.11 Scheme of the calcium isolation method for urine samples. ........... 75 4.12 Comparison of the 41Ca/40Ca ratios from the biomedical samples at TANDY and SARA. .......................................... 81 4.13 Temporal evolution of the urinary 41Ca/40Ca ratio in subject 4, including measurements at TANDY and SARA. ........................... 82 4.14 Temporal evolution of the urinary 41Ca/40Ca ratio in subject 5, including measurements at TANDY and SARA. ........................... 83 4.15 Temporal evolution of the urinary 41Ca/40Ca ratio in subject 6, including measurements at TANDY and SARA. ........................... 84 XI LIST OF FIGURES 4.16 Temporal evolution of the urinary 41Ca/40Ca ratio in subject 7, including measurements at TANDY and SARA. ........................... 85 4.17 Temporal evolution of the urinary 41Ca/40Ca ratio in subject 8, including measurements at TANDY and SARA. ........................... 86 4.18 Temporal evolution of the urinary 41Ca/40Ca ratio in subject 9, including measurements at TANDY and SARA. ........................... 87 4.19 Temporal evolution of the urinary 41Ca/40Ca ratio in subject 10, including measurements at TANDY and SARA. ........................... 88 5.1 40Ca neutron capture cross section and thermal neutron energy distribution. . 90 5.2 Scheme of the calcium isolation process for concrete samples. ......... 93 5.3 Effect of the acid digestion of concrete sample depending on the presence of aggregates. ....................................... 94 5.4 Scheme of the segmentation of the bioshield of Jos´ e Cabrera and 3D picture of the shape from each block. .............................. 96 5.5 Schematic view of the position of the drill cores. .................. 97 5.6 Position of the samples taken by scrapping the lateral surface of the sub-blocks. 98 5.7 41Ca/40Ca results from the drill core samples and correspondence with thermal neutron fluence. .................................... 99 5.8 Cross section of the Jos´ e Cabrera reactor vessel and bioshield when the plant was operative at the height of maximum neutron fluence. ............ 100 5.9 41Ca/40Ca results from the lateral samples. ..................... 102 XII List of Tables 1.1 Common radionuclides measured by AMS. ..................... 6 1.2 41Ca performance parameters from different AMS facilities. ........... 7 3.1 Energy loss straggling and average energy difference with 41K ions of 3.4 MeV 41Ca ions after a silicon nitride window as a function of the foil thickness. . . . 43 3.2 Rates from different ions in an aluminum dummy and a blank sample during the experiment with the 600 kV AMS system at ETH. ............... 46 4.1 Voltages at the accelerator terminal and the ESA in order to select different ions with the HE spectrometer at SARA. ....................... 73 4.2 Blank corrected 41Ca/40Ca ratios of the standard samples used in one of the 41Ca measurements at SARA using the 2+ state. .................. 77 4.3 Comparison of the 41Ca performance parameters at SARA and TANDY. ..... 78 4.4 Measurement of the ETH in-house 41Ca standards B10 and B8 at SARA. . . . . 79 XIII LIST OF TABLES XIV Acronyms AMS Accelerator Mass Spectrometry CNA Centro Nacional de Aceleradores ESA Electrostatic Analyzer ETH LIP ETH Laboratory of Ion Beam Physics FC Faraday Cup GIC Gas Ionization Chamber HE High Energy HEM High Energy Side Magnet HVE High Voltage Engineering Europa B.V. LE Low Energy LEM Low Energy Side Magnet LSC Liquid Scintillation Counting NEC National Electrostatics Corp. NPP Nuclear Power Plant NUS National University of Singapore PWR Pressurized Water Reactor XV Abstract The accelerator mass spectrometry (AMS) technique, developed 40 years ago for 14C dating, was soon used to measure other long-lived radionuclides. One of the radionuclides measured by AMS since these early years is 41Ca and for a variety of applications. Some of these applications are: the calculation of the terrestrial age of meteorites, the study of nuclear reactions of astrophysical interest, the understanding of the calcium metabolism, and the characterization of low-level nuclear waste. The main challenge in 41Ca AMS is dealing with the interference caused by its stable isobar, 41K. This interference is reduced by using calcium fluoride (CaF2) samples and the extraction of the (CaF3)-ion. At large AMS facilities, besides, 41K ions can be discriminated from 41Ca ions using different detection techniques based on the energy loss dependence on the atomic number. The 41Ca measurement with low energy AMS systems, like the 1 MV system at the Centro Nacional de Aceleradores (CNA), is quite challenging, since this discrimination is not possible. Nevertheless, the 41K contribution can be estimated and, therefore, corrected, thanks to the sequential detection of the other stable isotope of potassium, 39K (K-correction). Although the sensitivity achieved in 41Ca AMS at low energies is 3-4 orders of magnitude lower than those achieved at larger facilities, it allows the competitive measurements for biomedical applications, and the characterization of concrete samples from nuclear reactor bioshields. Since 41Ca AMS at low energies is limited to the estimation of 41K interference, it is advisable to study the different ways toward the production of this interference. Some factors related to it have been studied in different experiments performed with the 1 MV AMS system at CNA (SARA) and the 600 kV AMS system at the ETH Laboratory of Ion Beam Physics in Zurich (TANDY). For instance, we could demonstrate that 41K can be injected also as the (41K57Fe)- molecular ion. As a consequence, 41K interference is dependent on the materials used during the sample pressing. We also proved that, even when both 41K/40Ca and 39K/40Ca ratios change over time, the relation between both, 41K/39K remains constant. Therefore, the K- correction is a robust method to estimate the 41K interference. XVII CHAPTER 1. INTRODUCTION Nowadays, though, the most common application of 41Ca does not come from its natural cosmogenic or anthropogenic production, but from its biomedical capability as a tracer of calcium metabolism [Elmore et al.,1990;Freeman et al.,1997]. The use of 41Ca has none of the disadvantages from other calcium radioisotopes or its stable isotopes: •In the first case, the radiological hazard of the administration of a dose of 45Ca or 47Ca, whith half-lives of 165 days and 4.5 days respectively, would be higher, and they allow much lower tracing times because of their decay, while 41Ca is only limited by its biological excretion. By contrast, 41Ca oral doses up to 3700 Bq are actually been used [Denk et al.,2006]; taking into account that the dose coefficient for ingested 41Ca is 2.9 ×10-10 Sv·Bq-1 [ICRP,2012], the dose in 50 years would be around 1 µSv. This dose is totally neglible, since it is much lower than, for instance, the annual natural background exposure (typically higher than 1 mSv). •In the second case, stable isotopes of calcium such as 42Ca or 44Ca require high chemical doses so that their isotopic abundances change significantly. Since 41Ca is not naturally present, if the dose has a relatively high 41Ca/40Ca ratio very low calcium content has to be administrated in order to be able to measure 41Ca concentrations in tissues were calcium is incorporated. A very recent compendium of calcium metabolism studies performed with 41Ca tracing and AMS measurements is found in Weaver et al. [2017]. 1.2 Accelerator Mass Spectrometry (AMS): finding a needle in a haystack Contents from any radionuclide with a relatively long half-life, such as 41Ca, are intrinsically hard to measure by radiometric techniques because of the very low related activities. Its direct detection with conventional mass spectrometry systems meet with the interference of molecular ions with the same mass that the radionuclide of interest. All the mass spectrometry systems follow the same basic scheme: ionization of the sample, the use of electromagntic separators to select its mass/charge ratio, and detection of these ions. Since the contents of long-lived radionuclides are typically orders of magnitude below the part per million (ppm) level, even a trace production of isobar molecular ions would make it impossible to measure of the radionuclide of interest. Accelerator Mass Spectrometry (AMS) solves this problem thanks to the electron stripping 4 1.2. ACCELERATOR MASS SPECTROMETRY (AMS): FINDING A NEEDLE IN A HAYSTACK process with a solid or gaseous target that takes place at energies available with particle accelerators. In this process, the molecular background is destroyed (see Figure 1.1). Figure 1.1: Basic scheme of the stripping process. Other supression processes, though, are necessary to solve the possible interference from stable isobars. Figure 1.2 shows the basic scheme of a common AMS system, together with the other isobar suppression techniques used in AMS: •Ionization: negative ionization provides, in some cases, a first supression or reduction of the interference from the stable isobar, because of the instability of negative ions from elements or molecules with a negative electron affinity (EA). This is the case, for instance, of 14C, since 14N does not form negative ions [Bennett et al.,1977]. Even if that EA is positive, but very low, its interference is reduced several orders of magnitude, since the negative ion production efficiency increases exponentially with the EA [Alton, 1993]. Figure 1.2: Basic scheme of an AMS system. While the possibility of supression of stable isobars depends on the radionuclide and experimental system, the use of the stripping process to destroy the molecular isobars is common to all AMS measurements. 5 CHAPTER 1. INTRODUCTION •Detection: techniques based on the different energy loss can be used because of the high energies of the ions. Since the energy loss in a material depends on the atomic number Zof the ion, the difference in the energy loss between the radionuclide of interest and its stable isobar can be used to reduce the interference entering the detector and separating both signals. The AMS technique, initially developed for 14C measurements, was soon applied to many other long-lived radionuclides, including 41Ca [Gove et al.,1979;Elmore and Phillips,1987]. Table 1.1 shows some of the most important ”AMS radionuclides”. Table 1.1: Common radionuclides measured by AMS. Half-lives from National Nuclear Data Center [n.d.], except for 10Be [Korschinek et al.,2010]. The main technique used to suppress/reduce the interference caused by the stable isobar is also shown: ”Ionization”refers to those cases where the negative ion of the stable isobar is unstable or metastable; ”Detection”, to those cases where the different energy loss is used to distinguish the radionuclide from its stable isobar. Radionuclide Half-life (My) Negative ion Stable isobar Suppression / reduction 10Be 1.387 (BeO)- 10B Detection 14C 0.0057 C- 14N Ionization 26Al 0.717 Al- (AlO)- 26Mg Ionization Detection 36Cl 0.301 Cl-36Ar 36S Ionization Detection 41Ca 0.0994 (CaH3)- (CaF3)- 41K Ionization + detection 129I 15.7 I- 129Xe Ionization 236U 23.42 (UO)-@ 1.2.1 41Ca AMS In the particular case of 41Ca, two different chemical forms and their respective negative ions can be used in order to deal with its stable isobar, 41K: calcium hydride (CaH2), and extraction of the (CaH3)-ion; or calcium fluoride (CaF2), and extraction of the (CaF3)-ion. The choice of one material or the other depends on several factors, since each one has both advantages and disadvantages: •CaH2:the main advantage of selecting the (41CaH3)-ion in the LE side is the inestability of the (KH3)-ion. A 41K/40Ca interference in the level of 10-13 is typically found from the 6 1.2. ACCELERATOR MASS SPECTROMETRY (AMS): FINDING A NEEDLE IN A HAYSTACK contribution of (41KH2)-ions where one of the 2 hydrogen nuclides is a deuterium (2H) [Fink et al.,1990]. The main disadvantage is the high hygroscopy of CaH2, which makes its handling problematic, making it necessary to minimize its contact with air [Sharma and Middleton,1987]. This choice is suitable for applications where few samples are measured and very high sensitivities are required, like the measurement of accretion rate of extraterrestrial 41Ca in Antarctica [G´ omez-Guzm´ an et al.,2015]. •CaF2:in contrast, calcium fluoride chemical preparation is more simple and it is a chemically stable salt [Kubik and Elmore,1989]. However, the (KF3)-ion is stable, even when its stability is much lower than that from (CaF3)-, leading to a higher 41K/40Ca interference in the level of 10-12-10-11 [Zhao et al.,2010]. Ion source output currents are also slightly lower [Middleton,1990], but in this case the difference is not so important. This choice is suitable for applications where large numbers of samples are involved and relatively lower sensitivities are required, like the biomedical applications detailed in section 1.1 [Freeman et al.,1995]. In any case, in conventional AMS facilities, which typically use terminal voltages higher than 3 MV, the different energy loss on a material, as it was already stated, can be used to separate the signals of 41Ca and 41K in the detector [Fink et al.,1990;Vockenhuber et al.,2005;Wallner et al.,2010;Hosoya et al.,2017]. Better separation between both signals can be achieved at higher energies, resulting in a lower fraction of 41K counts that interfere with the 41Ca counts. The performance parameters for 41Ca from several AMS facilities are presented in Table 1.2. The layouts from these systems are shown in Figure 1.3. Table 1.2: 41Ca performance parameters from different AMS facilities. Only data from measurements using calcium fluoride are used, even when Wallner et al. [2010] presents also performance parameters with hydride. Facility VERA (Vienna, Austria) DREAMS (Dresden, Germany) 14UD ANU (Canberra, Australia) References [Wallner et al.,2010][Akhmadaliev et al.,2013] [Rugel et al.,2016][Fifield et al.,2010] Terminal voltage (MV) 3 6 14 Stripper material O2gas Ar gas O2gas + C foil HE side charge state 4+ 4+ 7+ Background (×10-15)50 2 1 7 CHAPTER 1. INTRODUCTION (a) VERA (b) DREAMS (c) 14UD ANU Figure 1.3: Layout of several AMS facilities: the 3 MV system at the Universit¨ at Wien (a), the 6 MV system at the Helmholtz-Zentrum Dresden-Rossendorf (b) and the 14 MV system at the Australian National University (c). 1.3 AMS at low energies During the first decades of AMS it was believed that, in order to destroy the molecular background, charge states higher or equal to 3+ had to be selected (”golden rule”of AMS). Molecular ions with those charge states become unstable because of the electrostatic repulsion [Suter,1990]. With compact particle accelerators with terminal voltages ≤1 MV, though, the most populated charge states for carbon ions are 1+ and 2+ [Jacob et al.,2000] and molecular interferents would still survive. Despite this fact, Lee et al. [1984] and Suter et al. [1997] showed that, if the stripper mass thickness of the stripper was high enough, this molecular background was reduced to neglible levels. Based on these results, the first low-energy AMS system for 14C measurements was developed: the 600 kV AMS system at ETH Zurich, TANDY, developed by the ETH Laboratory of Ion Beams Physics (LIP) in collaboration with National Electrostatics Corp. (NEC, USA) [Synal et al.,2000]. Soon, the capabilities of this system for other AMS radionuclides were studied 8 1.4. MOTIVATION AND OUTLINE and, since then, the system has also performed routine measurements of 10Be, 26Al, 41Ca, 129I and actinides (236U, Pu isotopes...) [Stocker et al.,2004,2005]. The first 1 MV AMS system designed by High Voltage Engineering Europa B.V. (HVE, Netherlands) was initially designed to measure not only 14C, but also other typical AMS radionuclides, such as 10Be and 26Al [Klein et al.,2006,2007], and installed at the Centro Nacional de Aceleradores (CNA) in Seville. The peculiarities from these systems and the developments implemented on them are presented in chapter 2. These small accelerators, nevertheless, still have some properties in common with larger machines, like the need of using some dieletric gas, as sulfur hexafluoride, to electrically insulate the terminal from the accelerator tank. This necessity does not apply any more when even lower terminal voltages are used, and vacuum insulation is enough to avoid discharges. This is the case of the MICADAS system, a system developed by the ETH LIP whose terminal voltage is provided by a 200 kV HV power supply [Synal et al.,2007]. The MICADAS system is designed only for 14C measurements. A new multi-element AMS system whose terminal voltage is provided by a 300 kV HV power supply is now being developed also by the ETH LIP [Maxeiner,2016]. The use of low-energy AMS systems reduce the operating and maintainance costs of these facilities. The original cost of the system is also lower. Furthermore, their sizes allow them to be placed in conventional laboratory rooms and no huge and dedicated halls are required to be built. 1.4 Motivation and outline The main goal of this thesis has been the setting up and optimization of the performance parameters of the 1 MV AMS system at CNA for 41Ca. Since its operation started in 2006 [Klein et al.,2006], the AMS research group at CNA has looked for exploiting the multielement capacity of the system [Klein et al.,2007;Chamizo et al.,2008b]. The same setting up and optimization of the performance has been performed, for instance, for 10Be and 26Al [Ruiz-G´ omez et al.,2010;Padilla,2015;Scognamiglio,2017], for 129I [G´ omez-Guzm´ an,2010; G´ omez-Guzm´ an et al.,2012], for 236U [Chamizo et al.,2015a], and for Pu isotopes [Chamizo et al.,2008a;Chamizo,2009]. Several technical changes and developments have been performed in collaboration with both, HVE and the ETH LIP [Chamizo et al.,2015c;Scognamiglio et al.,2016]. Nowadays, there exist several AMS facilities around the world similar to the 1 MV system at CNA. Thus, the study of the performance of its different elements are also useful to other 9 CHAPTER 1. INTRODUCTION research groups, so they can exploit the capacities of their HVE 1MVAMS systems. Some of these groups actually included 41Ca in their acceptance tests [Klein et al.,2013;Heinemeier et al.,2015], so it can be said that they are interested on performing routine 41Ca measurements in the future. The results provided by this thesis, specially those also published in Vivo-Vilches et al. [2017], are useful to these groups as much as it is to ours. This thesis included a stay of four months at ETH LIP in order to perform some experiments related to the 41K interference in 41Ca AMS at low energies. The stay also included the collaboration with the ETH LIP in the 41Ca measurements for biomedical applications performed by them for the NutriTrace group at the National University of Singapore. The information and experience acquired during this stay was really helpful in order to optimize the 41Ca measurement performance of the 1 MV AMS system at CNA Seville. Even when the study of the capabilities of the system for new radionuclides and technical developments themselves are two of its main research lines, our group has also focused on some important application research lines. These research lines include the development of chemical procedures for the sample preparation. Nowadays, the most important applications performed by our group are: •Environmental applications, specially related to oceanography [Chamizo et al.,2015b, 2016;L´ opez-Lora et al.,2018;Vivo-Vilches et al.,2018]. •Characterization of Low-Level Waste (LLW) from operating and decommissioning nuclear power plants, in collaboration with the Spanish radioactive waste management agency, ENRESA [L´ opez-Guti´ errez et al.,2013]. Radiocarbon applications are no longer performed with the 1 MV AMS system since the group acquired a MICADAS system in 2012. The 41Ca AMS measurements in the group have been implemented initially in the LLW characterization project. 41Ca/40Ca ratios have been measured in concrete samples from several locations different in both radial and vertical coordinates from the primary shield of a decommissioning nuclear power plant (NPP): the Jos´ e Cabrera NPP. The results from these measurements will be presented in chapter 5. Even if we did not perform measurements for biomedical applications at CNA Seville, we did chemical tests with urine samples. We also showed that the performance of the 1 MV AMS system is well suited to perform those measurements for any interested group. This first chapter was provided in order to introduce the reader to the applications of 41Ca measurements, the AMS technique and its evolution to low energy systems. 10 1.4. MOTIVATION AND OUTLINE In chapter 2, a description of the fundamental principles of the most important elements in an AMS system is presented. The two systems used in the experiments carried out during this PhD thesis are also concisely described, from which a more detailed description can be found in Chamizo [2009]: the 1 MV AMS system at CNA Seville, SARA (Spanish Accelerator for Radionuclide Analysis); and the 600 kV AMS sytem at ETH Zurich, TANDY. The specific challenges related to 41Ca AMS at low energies and how the 41K interference can be estimated thanks to the detection of the other stable isotope of potassium, 39K (K- correction) [Vockenhuber et al.,2015] are presented in chapter 3. Some experiments performed on both compact AMS systems (SARA and TANDY) providing information about the production pathways of the 41K interference are also presented, as well as the specific complexity introduced to the data analysis by the K-correction. Because of this complexity, 2 similar programs, one for each AMS system, were written in FORTRAN code to perform the data analysis. The experiments in order to set-up and optimize the 41Ca measurements with the SARA system are presented in chapter 4. These experiments include the characterization and optimization of the ion source current output, the transmission using He as a stripper gas, the detection efficiency, and the performance with blank samples (with neglible content of 41Ca) and standard samples (with known 41Ca/40Ca ratios). A comparison of the performance paramenters with those from the TANDY system is shown, together with the measurements of some of the ETH in-house standard samples used in the measurements with that system and an intercomparison experiment carried out on some of the biomedical samples which were measured during the stay at ETH LIP. Results from the 41Ca measurements with the SARA sytem of the concrete samples from the primary shield of the Jos´ e Cabrera Nuclear Power Plant are presented in chapter 5. These results are preceded by the exposition of the way the 41Ca is produced in the concrete of the primary shield, and how the 41Ca/40Ca can be give information about the thermal neutron fluence the shield was exposed to during the operation of the power plant. The last chapter summarize the most important conclusions of this work. 11 CHAPTER 1. INTRODUCTION 12 Chapter 2 Comparative description of the experimental systems 2.1 General overview of the systems In this chapter, the basic physical principles from the typical elements of AMS systems are presented, together with the characteristics of each one of these elements for the systems used during this thesis (the SARA system at CNA Seville and the TANDY system at ETH Zurich). Both systems are described in detail in Chamizo [2009]. Additional information about the TANDY system can be found in Jacob [2001]. This chapter, therefore, focuses on the changes implemented since then and the features which are specially important for 41Ca measurements. This first section presents a general overview of both systems. The next sections present: •The production of negative ions by Cs sputtering (section 2.2). •The operation and optical properties of tandem accelerators, together with the basic characteristics of the stripping process (section 2.3). •The kinematic filters used in AMS, and the two different sectors: the LE injector and the HE spectrometer (section 2.4). •The charged particle detection with gas ionization chambers (section 2.5). In some of this sections, simulations performed with GICOSY [Weick,n.d.] are presented. The input files from these simulations are shown in Appendix A. 13 CHAPTER 2. COMPARATIVE DESCRIPTION OF THE EXPERIMENTAL SYSTEMS The lack of electrostatic lens before the accelerator suppose a restriction in the V0 VTratio so that the LE acceleration tube focus the ion beam in the stripper center. The reason for this is explained in the next section.Jacob [2001] calculated that the optimal ratio was V0 VT= 0.102 ≃ 0.1. This means that, for those radionuclides where a terminal voltage of ∼500 kV is used (10Be, 26Al and 41Ca), the energy of the negative ions in the LE side should be ∼50 keV, while for those radionuclides where a terminal voltage of ∼300 kV is used (129I and actinides), this energy should be 30 keV. 2.3 Tandem accelerators In a tandem accelerator, the negative ions are attracted to its center (terminal) by a large positive potential VTV0; the same potential accelerates again the positive ions after the stripping process which takes place in the terminal, by electrostatic repulsion. At each side of the accelerator terminal there is a set of electrodes connected with resistors which avoid a sudden change in the electrostatic potential (acceleration tubes). The potential in these tubes can be supposed to increase linearly if all the resistors of each tube have the same resistance. In this case, the electrostatic force which the negative ions, with a charge −e, experience in the first acceleration tube (LE tube) is FLE−tube =eVT LLE ,(2.2) where LLE is the length of the LE tube. The positive ions produced after the stripping process, with a charge qe :q∈Z+, experience a force in the second acceleration tube (HE tube) of FHE−tube =qeVT LHE ,(2.3) where LHE is the length of the HE tube. In the rest of the zones (before the LE tube, in the terminal and after the HE tube) the potential is constant, so the electrostatic force is 0. If the negative ions selected in the LE side have a mass M0, and the positive atomic ions have a mass of M≤M01, the kinetic energy Eof the ions is: •eV0at the entrance of the accelerator. 1M=M0if the negative ion was the same atomic species, e.g. 14Cqions from the stripping of 14C-.M < M0 if this positive atomic ion is one of the fragments from a molecular negative ion, e.g. 41Caqions from the stripping of (41CaF3)-. 20 2.3. TANDEM ACCELERATORS •e(V0+VT)at the exit of the LE tube, this is, in the terminal before the stripping process. •M M0e(V0+VT)at the entrace of the HE tube, this is, in the terminal after the stripping process. This M M0factor comes from assuming that the molecular ion breaks elastically, so the velocity of its fragments is the same, distributing the kinetic energy among those fragments proportionally to their masses. •M M0e(V0+VT) + qVTat the exit of the HE tube. In each one of these 4 locations, there is a change of the electrostatic force ∆Fwhich has a focusing effect whose optical power 1 fis directly proportional to ∆F E[Larson,1974]2. If we define the effective charge state of the positive ions as q∗≡M0 Mq[Maxeiner,2016], and taking into account that VTV0, it can be deduced that the optical power is directly proportional to: •VT V0 1 LLE at the entrance of the LE tube. •−1 LLE at the exit of the LE tube. •q∗1 LHE at the entrace of the HE tube. •−q∗ 1+q∗ 1 LHE at the exit of the HE tube. The VT V0relationship becomes the key factor in the optical focusing of the negative ions in the terminal, while the effective charge state q∗defines the optical behaviour of the positive ions after the terminal, in the the HE side of the AMS system. Depending on how the terminal voltage is achieved, there are two main types of particle accelerators used in AMS (not taking into account the HV power supplies used by the MICADAS system): •The tandetrons: these accelerators use a Cockcroft-Walton circuit based on a RF power supply connected to a series of diodes and capacitors which, after some time, increase the output voltage to a maximum value. In the case a of Figure 2.9 the most basic version of this kind of circuit is shown, but the number of diode-capacitor stages can be increased, being the output voltage directly proportional to the number of stages. The main disadvantage of this kind of accelerator is that only terminal voltages up to 6 MV have been achieved with them. On the other hand, their terminal voltages are very stable and they require little maintenance. 2Even when this approximation to a thin lens is not suitable to get the actual behaviour of a certain accelerator tube, the proportionality can be used to compare the behaviour of different accelerator tubes or different ions. 21 CHAPTER 2. COMPARATIVE DESCRIPTION OF THE EXPERIMENTAL SYSTEMS •The pelletrons: these accelerators use a Van de Graaf generator consisting in a chain of metallic pellets which are charged positively outside the terminal, and deposit this charge on it, increasing the terminal voltage (see the case b of Figure 2.9). The pellets can also be charged negatively before leaving the terminal, so the positive charge deposited by each pellet after passing through doubles. The voltages that some of these accelerators can achieve are even higher than 20 MV. The main disadvantage from the pelletrons is that the use of the pellet chains requires a regular maintenance. In both cases, the high electric field between the terminal and the accelerator tank (at ground potential) requires the use of a gaseous dieletric material in order to avoid the loss of the voltage by electric discharges. The most common material for this purpose is sulfur hexafluoride (SF6). (a) Tandetron (b) Pelletron Figure 2.9: Basic operation principle of the systems used by the 2 different types of particle accelerator to get the high terminal voltage. The stripping process Several large AMS systems use both solid and gaseous materials as stripper materials. The angular straggling caused by solid strippers is not so important in these systems since this effect is inversely proportional to the kinetic energy. In contrast, compact AMS systems, like the 1 MV AMS system at CNA Seville and the 600 kV AMS system at ETH Zurich, only use gaseous strippers, since the posibility of easily change the mass thickness of the material is critical to destroy the molecular background without severely increasing unwanted angular straggling or scattering effects. With gaseous strippers, the desing of the pumping system of the gas in and out of the stripper tube is extremely important in order to minimize the leakage of the stripper gas to the acceleration tubes (residual gas). This reduces the background 22 2.3. TANDEM ACCELERATORS produced by ions with different mass/charge ratio which are scattered by collisions with the residual gas, leading to a trajectory which makes them reach the detector. The stripping process is caused by the charge exchange reactions between the stripper material and the ions. If the stripper mass thickness is high enough, these charge exchange reactions reach an equilibrium where there are no negative ions, and the positive ions and the atomic species are distributed in several charge states q≥0. The fraction Φqbetween the number of ions with charge state qproduced and the total number of negative ions stripped is called ”charge state fraction”. It is clear, then, that Z X q=0 Φq= 1, where Zis the atomic number of the species. The ”mean charge state”is, so, defined as ¯q≡ Z X q=0 qΦq, and depends on the stripping material, but also on: •The kinetic energy which the ions have in the terminal Estrip =M M0e(V0+VT), also called stripping energy. The higher this energy is, the higher the populated charge states are. The most populated states at low energy AMS systems like SARA and TANDY for most of the radionuclides are typically 1+, 2+ and/or 3+, being Φq≃0 : ∀q > 3. This is the minimum charge state reached at large AMS facilities using terminal voltages higher than 2.5 MV. This dependence actually is on the velocitiy; therefore, for the same stripping energy, two isotopes from the same element have slightly different ¯q. •The atomic species of the ions. The higher the atomic number is, the higher the populated charge states are. At SARA, for instance, the most populated charge state in the He gas stripper for 10Be is 1+, but 2+ for 41Ca and 3+ for 236U. The stripper transmission for a charge state qis defined as the ratio between the rate of positive ions at this charge state that exit the accelerator and the total rate of negative ions which enter it. Therefore, this transmission is slightly lower than the charge state fraction, because of the optical loses due to the angular straggling introduced by the stripper. It is estimated by measuring the current −I0of the negative ion from the stable isotope entering the accelerator, and the current Iqfrom the positive ion of charge state qfrom stable isotope ion. The transmission, then, is Tq=Iq qI0 .(2.4) Both systems described in this chapter initially used argon as stripper gas, but changed to helium in the recent years [Vockenhuber et al.,2013;Scognamiglio et al.,2016]. 23 CHAPTER 2. COMPARATIVE DESCRIPTION OF THE EXPERIMENTAL SYSTEMS The use of He gas as stripper material means an improvement at low-energy AMS because of the dependence of the mean charge state with the velocity. An example of this can be seen in Figure 2.10, which shows data from Wittkower and Betz [1973] using uranium ions (because of the high mass, velocities are much lower than for other typicial ions as carbon for similar energies): while for high energies the mean charge states for other materials are higher than for helium, for low energies the case is the opposite. This makes He gas the best option for AMS at low energies and/or for heavy ion AMS [Vockenhuber et al.,2013;Winkler et al.,2015]. In addition, the angular straggling is reduced because of the low mass of helium. Figure 2.10: Mean charge state of uranium ions as a function of the ion energy for different stripper materials. Data from Wittkower and Betz [1973], presented also in Vockenhuber et al. [2013]. 2.3.1 The 1 MV tandetron at SARA Figure 2.11 shows the scheme of the elements within the 1 MV tandetron used by the SARA system. All the resistors from each of the acceleration tubes are equal, so it can be supposed that inside them there is a constant electric field =VT L, where L=LLE =LHE = 72.5cm. In both cases, the electric field is in direction from the terminal to the accelerator extremes. At the entrance of the accelerator there is an electrode (Q-Snout), whose potential VQ−Snout >0 can be changed. The Q-Snout lens is used to increase the energy with which the ions enter the LE accelerator tube and, so, reducing the focusing power from the aperture of this tube. 24 2.3. TANDEM ACCELERATORS The convergent power of the tube entrance is, in this case, proportional to VT V0+VQ−Snout . Chamizo [2009] calculated that, in order to focus the beam in the center of the stripper tube, this voltage has to be VQ−Snout = 0.082 ·VT−V0.(2.5) In the stripper tube, the gas pressure is measured only at its center (p0). In order to calculate the stripper mass thickness, it is assumed that the gas pressure follows a triangular distribution, so the average pressure would be p0 2, and the mass thickness, ρA=MgasL RT p0 2,(2.6) where Mgas is the molar mass of the stripper gas, Lthe length of the stripper tube (30 cm for SARA), R= 83145 mbar ·cm3·K−1·mol−1, and Tthe gas temperature, which is at a room temperature of 294.15 K (21 ºC). Since the focusing power of the entrance of the HE accelerator tube is proportional to q∗≡ M0 Mq, a quadrupole triplet (Q-Pole) is placed at the exit of the accelerator to refocus the beam. An electrostatic quadrupole is a set of four electrodes, two of them facing in the X direction and connected to a potential VQ, and the other two facing in the Y direction and connected to a potential −VQ(see Figure 2.12). The electrodes have a geometry so that the potential φ in places close to the center is approximately parabolic, this is, φ∝(x2−y2)(so it follows the Laplace equation ∆φ= 0). Figure 2.11: Scheme of the system of the 1 MV accelerator from SARA. Picture from Chamizo [2009], modified by Scognamiglio [2017]. 25 CHAPTER 2. COMPARATIVE DESCRIPTION OF THE EXPERIMENTAL SYSTEMS Since each quadrupole focuses the beam in one axis, but unfocuses it in the other, sets of multiple quadrupoles (multiplet) are used, where each quadrupole uses the opposite potentials as the previous one. In the case of the quadrupole triplet at SARA, the potentials are, for aVQ>0: •VQin the X axis for the central quadrupole, and in the Y axis for the first and third quadrupoles. •−VQin the Y axis for the central quadrupole, and in the X axis for the first and third quadrupoles. Figure 2.12: SIMION® view of an electrostatic quadrupole. The arrows represent the electrostatic force which a positively charged particle would experienced for a VQ>0. 2.3.2 The 600 kV pelletron at TANDY TANDY uses a 600 kV pelletron from NEC, which has only one charging pellet chain, shown in Figure 2.13. All the resistors of the HE tube are equal, so it can be assumed that there is a constant electric field HE =VT L, where L=LLE =LHE = 28 cm is the length of each one of the acceleration tube. The resistors of the LE, instead, have a resistance in the first third part of the acceleration tube, but 3 2times that resistance in the rest [Jacob,2001], leading to two different electric fields: a field 1=3 4 VT Lin the first third part, and a field 2=3 21=9 8 VT Lin the other two third parts, so 1L 3+22L 3=VT. Thanks to this lower electric field at the entrance of the LE tube, the optimal V0 VTratio to focus the beam in the center of the stripper can be achieved with the energies available with the ion source even for terminal voltages of 500 kV3. The electric field difference after this first 3If the electric field were =VT L=4 3ε1, the optimal V0 VTwould be also higher in order to have the same optical power. 26 2.4. ELECTROMAGNETIC KINEMATIC FILTERS third part of the tube also has a focusing effect, but this is much lower than the one at the entrance because of the higher energy of the ions in that place. In this case, the stripper gas pressure is measured not only in the center, but also in the stripper apertures. Together with simulations, this allows a more accurate estimation of the stripper mass thickness. Nevertheless, the simulation on Jacob et al. [2000], for instance, estimates an Ar mass thickness of 2 µg·cm-2 for a central pressure of 0.058 mbar, whereas using Equation 2.6, a mass thickness of 2.04 µg·cm-2 is estimated, showing the agreement of the triangular approximation with more accurate estimations. Figure 2.13: Pellet chain used to transport the charge in the pelletron from TANDY. 2.4 Electromagnetic kinematic filters The kinematic filters most commonly used by AMS systems to select only the ions with a certain mass/charge ratio are: •Magnetic dipoles, which use a magnetic force to select the product of both the mass/charge ratio and the energy/charge ratio (EM q2). •Electrostatic analyzers, which use an electrostatic force to select the energy/charge ratio (E q). •Wien filters, which confront two forces (an electrostatic one and a magnetic one) to select the velocity of the ions (v∝qE M). 27 CHAPTER 2. COMPARATIVE DESCRIPTION OF THE EXPERIMENTAL SYSTEMS The AMS systems described in this thesis use only magnetic dipoles and electrostatic analyzers. Magnetic dipoles are always used in both, the LE side and the HE side. A common characteristic of these elements is that ions with both the same mass/charge and energy/mass ratio have the same behavior. In order to distiguish the radionuclide of interest from other ions reaching the detector, it has to be able to measure the kinetic energy of these. Magnetic dipoles A magnetic dipole is an electromagnet which generate a magnetic field ~ Bin a specific area. For ions with a charge qe which are moving with a linear momentum ~p perpendicular to ~ B, the magnetic force will make these ions follow a curved trajectory, also perpendicular to ~ B, with a radius ρ=p qeB .(2.7) If the trajectory of these non-relativistic ions of mass m(p=√2Em) must have a radius ρ=R in order to exit the magnet and continue its trajectory through the system, the product of the mass/charge ratio and the energy/charge ratio must be Em q2e2=B2R2 2.(2.8) In an AMS system, mass selection is needed before (LE side) and after (HE side) the ion acceleration in the tandem accelerator. Taking into account the ion kinetic energy and charge state in each of the sides, the mass selected by the LE side magnet (LEM) will be M0=eB2 LEMR2 LEM 2V0 ,(2.9) while the mass/charge ratio selected by the HE side magnet (HEM) must satisfy that M qV0+VT M0 M q+VT=eB2 HEMR2 HEM 2.(2.10) It can be demonstrated that this relationship between BHEM and M qis bijective so, theoretically, for a constant BHEM, the use of the HEM would be enough to select the required mass/charge ratio. However, some energy dispersion or charge exchange processes with the 28 2.4. ELECTROMAGNETIC KINEMATIC FILTERS residual gas in the HE accelerator tube could make some ions with different mass/charge ratio have the same EM q2relationship. That is the reason that, for the sensitivities required in AMS measurements, an additional filter, like the electrostatic analyzer, must be used in the HE side. Magnetic dipoles also play an optical role. This ion focusing effect depends on the angle between the beam and the borders of the magnet [Wollnik,1999] (see Figure 2.14). If the entrance and the exit of the magnet are perpendicular to the beam direction (Z axis), ions diverging in the X axis (perpendicular to the beam but parallel to its curvature radius) from a point (object) before the magnet will converge in a point (image) that is lined up with both the object and the center of curvature of the ion beam. Ions diverging in the Y axis (perpendicular to both the beam and its curvature radius), nevertheless, would not be focused (see Figure 2.15) since Lorentz force, perpendicular to both the magnetic field (in the Y direction) and the ion direction, would be in the X direction. Even in the borders, where the magnetic field has a little component perpendicular to them, the magnetic force would still remain in the X direction. If the borders are not perpendicular to the Z axis, this component perpendicular to the borders do has a focusing effect in the Y direction, since the magnetic field in them would now have components in the 3 axis [Wollnik,1999]. This convergent power from each border is, approximately, 1 fy =tan ε R,(2.11) where εis the angle which the border is inclined respect to the plane normal to the beam, (a) Borders perpendicular to the beam. (b) Borders oblique to the beam. Figure 2.14: Focusing in the X axis of magnetic dipoles depending on the inclination of the borders. 29 CHAPTER 2. COMPARATIVE DESCRIPTION OF THE EXPERIMENTAL SYSTEMS peak energy (27.5 MeV for 41Ca in butane, according to SRIM), this stopping power is ∝Z2 p·Zt E (Bethe-Bloch formula), where Zpand Ztare the atomic numbers of the ion (projectile) and the counting gas (target), respectively. For lower energies, as those reached at SARA and TANDY for 41Ca (one order of magnitude lower) the stopping power becomes ∝Zp·Zt (Z2/3 p+Z2/3 t)3/2√E (LSS formula). In both cases, the higher the atomic number of the projectile, the higher this stopping power is4. This can be useful to distinguish the radionuclide of interest from its stable isobar. For instance, in Figure 2.22 it can be seen the different dependence of the stopping power with the distance covered by the ion for 41Ca and 41K. Sectioning the GIC anode in different ones, therefore, the signals from both isobars can be distinguished. If the gas ionization chamber has two or more anodes, the energy deposited on each one of them by the stable isobar ion will be different that the one deposited by the radionuclide of interest. The total energy deposited in both anodes will still be the same for both ions. An example of this separation with 41Ca and 41K ions with an energy of 27 MeV can be seen in Figure 2.23. Figure 2.22: Stopping power of 41Ca and 41K ions with an energy of 26.5 MeV in butane according to SRIM. The butane density was set to stop the 41Ca ions before covering a distance of 500 mm. 4Even when this is not so trivial for the LSS formula, it can be easily proven that the function f(Z) = Z (Z2/3+x)3/2fulfills that df dZ >0. 36 2.5. GAS IONIZATION CHAMBERS At low energies, though, the high angular and energy loss stragglings within the counting gas leads to a higher energy straggling of the signals from the different anodes, and the distinction between 41Ca and 41K is not possible (see chapter 3). For heavier ions, the transmission of the kinetic energy to the whole gas atom/molecule (nuclear stopping power) becomes important at low energies. Figure 2.23: Detector signals in a 41Ca measurement with the 5 MV AMS system at SUERC, Glasgow. Picture from Schulze-K¨ onig et al. [2010]. 37 CHAPTER 2. COMPARATIVE DESCRIPTION OF THE EXPERIMENTAL SYSTEMS 2.5.1 The compact gas ionization chamber at SARA The compact GIC at SARA, which is presented in Figure 2.24, was designed at the ETH LIP, and it is located after the ESA. It has two anodes, each one of them 5 cm long. Comparing with the old HVE detector (see Figure 2.25), the lower volume of the compact GIC reduce the electronic noise and the design minimizes the non-active area. Figure 2.24: Scheme of the compact GIC at SARA. Picture from Scognamiglio et al. [2016]. Figure 2.25: Comparison between the old HVE GIC (left), and the new compact GIC (right). 38 2.5. GAS IONIZATION CHAMBERS 2.5.2 The retractable gas ionization chamber at TANDY The retractable GIC at TANDY (see Figure 2.26) can be inserted after the ESA, in the image of the 3+ charge state. It has two anodes, 3 cm long each one. For other radionuclides, a compact GIC similar to the one at SARA placed after the HEM 2 is used [M¨ uller et al.,2012]. Figure 2.26: Picture (left) and scheme (right) of the retractable GIC used in the 41Ca measurements with TANDY. Picture from Stocker et al. [2005]. Scheme adapted from a figure presented on Stocker [2006]. 39 CHAPTER 2. COMPARATIVE DESCRIPTION OF THE EXPERIMENTAL SYSTEMS 40 Chapter 3 41Ca AMS at low energies In this chapter, the main challenges related to 41Ca AMS at low energies are detailed, and some experiments related to this performed during this thesis with both systems, SARA and TANDY, are also presented. The most important results were also published in Vivo-Vilches et al. [in press]. 3.1 The 41K interference Kinematic filters commonly used at AMS only select ions with a certain mass/charge ratio. Nevertheless, the measurement of the energy of the ions in the detector allows the discrimination of ions with different mass but the same M/q. Therefore, the main challenge of 41Ca AMS measurements is dealing with its stable isobar, 41K. As it is stated in chapter 1, large AMS facilities deal with this potential interference using the different stopping power in the gas of the GIC [Hosoya et al.,2017]. At low energies, on the other hand, scattering processes with the gas itself produce an important angular straggling. This straggling, together with the one on the energy loss itself, also entails a lower energy resolution, since different trajectories lead to different energy losses in each one of the anodes. Figure 3.1, for instance, shows a simulation performed with SRIM [Ziegler,n.d.] of the behaviour of 41Ca ions in a GIC with 2 anodes: scattered ions will cover a longer path in the first half of the detector, leaving more energy in the first anode. This way, at lower enegies, 41K ions scattered an angle θKcould leave more energy in the first anode than some 41Ca ions scattered and angle θCa < θK, even when the stopping power of the potassium ions is lower. The simulated spectra corresponding to 41Ca and 41K ions at different energies in a 2-anode 41 CHAPTER 3. 41CA AMS AT LOW ENERGIES GIC detector like the one presented in Figure 3.1 are presented in Figure 3.2. At relatively high energies, the spectrum shows 2 count areas: 41K ions deposit less energy than 41Ca ions in the first anode (ΔE) and more energy in the second one (Eres). At lower energies, like those reached with the 1 MV AMS system at CNA, the 41K and 41Ca counts are mixed in the spectrum. The higher relative energy and angular stragglings at low energies also avoids the use of tech- (a) 26.5 MeV (b) 2.3 MeV Figure 3.1: SRIM simulation of the trajectories of several 41Ca ions at two different energies in a GIC detector with a 50 nm silicon nitride window, filled with isobutane and with 2 anodes 2.5 cm long each. Isobutane pressures were adjusted on each case to stop the ions just before the GIC end. (a) 26.5 MeV (b) 2.3 MeV Figure 3.2: Spectra simulation with SRIM of an equal number of counts of 41Ca and 41K at different energies. The ΔE axis represents the energy deposited in the first anode; the Eres axis, the energy deposited in the second one. 42 3.1. THE 41K INTERFERENCE niques based on the different energy loss on silicon nitride foils, such as the ΔTOF technique [Vockenhuber et al.,2005]. As it can be seen in Table 3.1, the energy straggling introduced by these foils in 41K and 41Ca ions with an energy of 3.4 MeV (41Ca3+ for a terminal voltage of 1 MV) would be, in most of the cases, higher, than the difference on the energy loss. Even at the very best, this difference is only slightly higher than the energy loss straggling. For thicker foils, besides, there would be an additional contribution to the energy straggling because of the angular one. The use of CaH2, selecting the (CaH3)-ion in the LE side, would reduce this potassium interference because of the instability of the (KH3)-ion. This only leaves the minor contribution to the 41K interference of the (41KH2)-ions where one of the hydrogen atoms is a deuterium one. However, because of the hygroscopic character of CaH2, its chemical preparation and handling is problematic. This makes this chemical form not suitable for applications which involve large number of samples, like those typically performed with low energy AMS. The possibility of extracting the (CaH3)-ion from other matrices, like the mix of calcium salts with titanium hydride, have been tested in the past [Fink et al.,1990;Baba et al.,1997], and recently here at CNA, in the last year of this thesis. Unfortunately, only unstable currents of few nA of (40CaH3)-were measured in all these cases. More information about these experiments is provided in Appendix B. Therefore, at SARA and TANDY, CaF2samples, extracting the (CaF3)-ion in the LE side, are used, as it is also done in most of the large AMS facilities. The stability of this compound makes it very suitable for handling of a large number of samples. The negative point, is that the 41K interference is increased by 2 orders of magnitude [Fink et al.,1990], this is, typically in the 10-11 level. Nevertheless, most of the samples measured with these systems present 41Ca/40Ca ratios higher than 10-10. First 41Ca measurements at TANDY were performed treating this interference as part of the background [Schulze-K¨ onig et al.,2010]. However, since the relative content of potassium could differ from one sample to another, an estimation of Table 3.1: Energy loss straggling and average energy difference with 41K ions of 3.4 MeV 41Ca ions after a silicon nitride window as a function of the foil thickness. The energy straggling (full width at half maximum of the Gaussian distribution) has been calculated with the semi-empirical Sun’s formula [Sun et al.,2007;M¨ uller,2009]. Average energies after the foil were obtained from SRIM simulations. Foil thickness (nm) 41K-41Ca energy difference (keV) Energy loss straggling (keV) 50 11 23 100 30 32 250 70 51 43 CHAPTER 3. 41CA AMS AT LOW ENERGIES the 41K interference would be more suitable. The best way to estimate this interference is the measurement of the other stable isotope of potassium, the 39K [Vockenhuber et al.,2015]. Potassium has 2 stable isotopes: 39K, with a natural isotopic abundance of 93.2581%; and 41K, with a natural isotopic abundance of 6.7302% [National Nuclear Data Center,n.d.]. Apart from slight variations, these isotopic abundance are always the same, independently from the sample, so the 41K/39K ratio should be 0.072167 in any material. Therefore, since the 41K/40Ca interference is typically in the 10-11-10-10 range, 39K/40Ca ratios are in the level of 10-10-10-9, ratios very suitable to be measured by AMS. Consequently, measuring not only 41M/40Ca1, but also 39K/40Ca ratios, the actual 41Ca/40Ca ratio could be calculated as R41,K−corr =R41 −αR39,(3.1) where R41 denotes the uncorrected 41M/40Ca ratio; R39, the 39K/40Ca ratio; and α, the measured 41K/39K ratio. Slight differences on the stripper and optical transmission between both potassium isotopes would make αnot being exactly the natural 41K/39K ratio. This experimental 41K/39K ratio (K-correction factor) can be calculated performing a linear regression where the values of the dependent variable xiare the 39K/40Ca ratios from different measurements of the blank samples, and those of the independent variable yi, the 41M/40Ca ratios. The experimental 41K/39K ratio, α, is the slope of this linear regression. More information about the related data analysis is presented in section 3.4. In order to sequentially inject 41M and 39K ions into the detector, avoiding changes in the magnetic field of the HEM, the accelerator terminal and ESA voltages have to be changed sequentially, as it is done in actinide measurements [Chamizo et al.,2008a;Fifield,2008]. The terminal voltage to inject the 39Kqions should make its energy 41 39 times that one of the 41Mqions when these are injected. However, this switching has to be much slower than the one typically used to inject the negative ions into the accelerator, because of both: the need of counting times of several seconds for both ions, and the relatively slow change of the terminal voltage. The most populated states at the stripping energies achieved with these compact AMS systems are the 2+ and 3+ states. Using the latter, there is no need of using relatively high stripper pressures in order to destroy the molecular background, since molecular ions with charge state 3+ are unstable. Up until 2018, it was thought that the K-correction would not allow the use of this charge state: the detection of 39K3+ would not be possible because of the rates of 13C+were expected to be so high that they would saturate the detector. Our recent studies at SARA, nevertheless, show that 13C+rates are not so high, and the K-correction can 1Where 41M denotes all the ions with mass 41, this is, 41M≡41K+41Ca 44 3.2. STUDIES OF THE SOURCES OF THE 41K INTERFERENCE be also used when the 3+ state is selected (see chapter 4). Using the K-correction, the 39K/40Ca ratio becomes a relevant factor involved in the final uncertainty of the results. Besides, since 39K and 41M are not measured simultaneously, but sequentially, huge changes in the 39K/40Ca ratio should be avoided. It is clear, then, that understanding the factors related to the production of the 41K interference is extremely important for 41Ca AMS at low energies. In the next sections, several experiments related to this interference, performed during this thesis with both systems, SARA and TANDY, are presented. 3.2 Studies of the sources of the 41K interference During the first tests of 41Ca AMS using calcium fluoride, it was thought that the (KF3)-ion is a totally unstable ion, and the 41K interference is not caused by the production of (41KF3)- ions, but from other molecular ions where 41K is present and with the same mass, 98 u [Fink et al.,1990]. Zhao et al. [2010] proved, though, that F+ions were produced from the stripping of ions with a mass of 96 u, consequently identifying those ions as (39KF3)-, confirming the existence of the ion. The slight stability from the (KF3)-ion, the fluorine present in the CaF2 sample, and the relatively high trace amount of potassium typically found in Cs, make possible the production of (41KF3)-ions, even in the theoretical case where there is no potassium trace amount within the cathode. The existence of the (KF3)-ion, nevertheless, does not mean that this ion is the only negative ion leading to 41K interference. The relatively high mass of the (41KF3)-ion would allow, also, the formation of other several different ions with the same mass,98 u, and where potassium is present. For instance, a common observation during experiments during this thesis is the detection of relatively high rates of both 41K and 39K ions when material with no fluorine was sputtered. This has been systematically observed with both systems, SARA and TANDY. Zhao et al. [2010] proposed the theory that another important negative ion related to 41K interference would be the (41K57Fe)-ion, since it has also, a mass of 98 u. They also checked that adding iron to the samples increased the 41K rates. This would explain the high 41K rates when metal is sputtered. In order to prove that this theory is correct, and that their increase on the 41K rates was not due to a higher trace amount of potassium in their iron material, an experiment was performed with the TANDY system during thw stay at the ETH LIP. This experiment consisted in the measurement of the 41M2+,39K2+,57Fe2+ and 19F+rates in a blank sample and an aluminum dummy. 41M2+ and 39K2+ rates were measured sequentially using the slow switching 45 CHAPTER 3. 41CA AMS AT LOW ENERGIES 41Ca AMS at low energies, as it was stated in section 3.1, the only way to estimate the whole contribution of this interference is during the data analysis. This K-correction becomes one of the key steps of this analysis. The specific features of 41Ca AMS data analysis at low energies depend on the system used, since available information slightly depends on the elements present on the system; for instance, the possibility of measuring the (40CaF3)-current on the LE side at TANDY makes possible using discard criteria based on the stripper transmission. However, the basic steps are the same. In order to reduce the time consumed by this data analysis, 2 programs were written in FORTRAN code during this thesis: one for the measurements at TANDY, written during the 4 months stay at ETH LIP; and another one, based on it, for the measurements at SARA. The code changes are only related to these slight differences in the information provided by each AMS system. More information about these programs is provided in Appendix C. In this section, an overview of the steps followed by these programs is presented. 3.4.1 AMS measurement In order to understand the data analysis of the measurements, it is very important to know how the measurements are performed. Figure 3.6 shows the typical measurement procedure in 41Ca AMS at low energies. Each sample i3is introduced in the machine a number nruns of runs. During each one of these (i, j)runs, both ions, 41M and 39K, are sequentially measured using different sequence times t41 and t39 for each one. Both ions are measured the same number of sequences nseqs. Each 41M sequence (i, j, k), except the first one from each run, (i, j, 1), is preceded by the 39K sequence (i, j, k −1) and followed by the 39K sequence (i, j, k). Typically, 5 sequences of t41 = 30 sfor 41M, and t39 = 10 sfor 39K, are used in the measurements at SARA and TANDY. This time can be lower if a certain number of ion counts in the detector is reached. The measurement of the 40Ca current is performed in a different way at SARA than at TANDY. In the first case, a 40Ca sequence is added just after the 41M one, using the same terminal and ESA voltage than those used in 39K sequences; the reason for this will be explained in chapter 4. In the second case, the 40Ca current is measured using the fast switching of the LEM bouncer during the 41M sequence. Theoretically, a system with 2 FCs in the HE side and with the proper software could mea- 3This includes not only the samples whose 41Ca/40Ca we want to know, but also blank and standard samples. 52 3.4. DATA ANALYSIS FOR 41CA AMS AT TANDY AND SARA sure the 40Ca current during both, the 41M and the 39K sequences, and it would be possible to directly get the 41M/40Ca ratio Rseq 41 (i, j, k)and the 39K/40Ca ratio Rseq 39 (i, j, k)from each sequence. At SARA and TANDY, the 40Ca current is assumed to change linearly from one sequence to the next in order to calculate these ratios. If the uncertainties of the 40Ca current and the measuring time are disregarded, the statistical uncertainty from each of these Rseq A(i, j, k)ratios, where Ais either 41 or 39, is Useq A(i, j, k) = Rseq A(i, j, k) pNA(i, j, k),(3.2) where NA(i, j, k)is the total number of ions with mass Adetected in that sequence4. Figure 3.6: Typical scheme of a 41Ca measurement with a compact AMS system implementing the K-correction. 4The uncertainty from NA(i, j, k)is pNA(i, j, k) 53 CHAPTER 3. 41CA AMS AT LOW ENERGIES 3.4.2 Calculation of the experimental 41K/39K ratio (K-correction factor) In order to get the experimental 41K/39K ratio α(or K-correction factor), the 41M/40Ca and 39K/40Ca ratios from each run of the blank samples are used in order to perform the linear regression. These ratios are calculated as Rrun A(i, j) = nseqs P k=1 wseq A(i, j, k)·Rseq A(i, j, k) nseqs P k=1 wseq A(i, j, k) ,(3.3) where wseq A(i, j, k)is a weight function dependent on the 40Ca current in that sequence IA(i, j, k) and the sequence time tA(i, j, k), defined as wA(i, j, k)≡ IA(i, j, k)·tA(i, j, k)·Cseq A(i, j, k).(3.4) This weight function makes more important those sequences with higher statistics. The Cseq A(i, j, k)is a logical factor which is 0 if any of the selection criteria (minimum 40Ca current, for instance) is not passed, and 1 otherwise. The common selection criteria used at SARA and TANDY are presented in subsection 3.4.5. Once all the Rrun 41 (i, j)and Rrun 39 (i, j)ratios from the blank samples have been calculated, a linear regression is performed by the method of least squares, using the Rrun 41 (i, j)ratios as independent variable x, and Rrun 39 (i, j)as dependent variable y5. The slope from this linear regression is the K-correction factor α±U(α). 3.4.3 Application of the K-correction and sample average Since each 41M measurement sequence is performed between two 39K sequences, the 39K/40Ca ratio which should be applied to correct the Rseq 41 (i, j, k)ratio is not exactly Rseq 39 (i, j, k), neither Rseq 39 (i, j, k −1). Nevertheless, since the expected changes in this 39K/40Ca ratio are slight in the time scale between two sequences, during the 41M sequence it can be supposed to be the average from the previous and the following one. The K-corrected 41Ca/40Ca ratio from each sequence, therefore, is 5[x;y](i, j) = [Rrun 41 (i, j); Rrun 39 (i, j)] 54 3.4. DATA ANALYSIS FOR 41CA AMS AT TANDY AND SARA Rseq K−corrected(i, j, k) = =Rseq 41 (i, j, k)−α·Rseq 39 (i, j, k −1) + Rseq 39 (i, j, k) 2 :k= [2, nseqs].(3.5) The uncertainty from this K-corrected ratio is calculated as Useq K−corrected(i, j, k) = v u u u u u t [Useq 41 (i, j, k)]2+hα·Rseq 39 (i, j, k −1) + Rseq 39 (i, j, k) 2pN39(i, j, k −1) + N39(i, j, k)i2+ +hU(α)·Rseq 39 (i, j, k −1) + Rseq 39 (i, j, k) 2i2.(3.6) Using the same weight functions w41(i, j, k)used to calculate the uncorrected 41M/40Ca ratios from each blank sample run, the K-corrected 41Ca/40Ca ratios from each run can be calculated for every sample as Rrun K−corrected(i, j) = nseqs P k=2 wseq 41 (i, j, k)·Rseq K−corrected(i, j, k) nseqs P k=2 wseq 41 (i, j, k) .(3.7) and the standard deviation is Srun K−corrected(i, j) = =v u u u u u u t Neffseqs(i, j) Neffseqs(i, j)−1· nseqs P k=2 wseq 41 (i, j, k)·[Rseq K−corrected(i, j, k)−Rrun K−corrected(i, j)]2 nseqs P k=2 wseq 41 (i, j, k) ,(3.8) where Neffseqs(i, j)is defined as the number of sequences in the run (i, j)where Cseq 41 (i, j, k) = 1. The standard deviation is only calculated if Neffseqs(i, j)>1. Assuming a virtual number of 41Ca counts in each sequence equal to hRseq K−corrected(i,j,k) Useq K−corrected(i,j,k)i2, the uncertainty of Rrun K−corrected(i, j)is Urun K−corrected(i, j) = Rrun K−corrected(i, j) snseqs P k=2 hCseq 41 (i, j, k)·Rseq K−corrected(i,j,k) Useq K−corrected(i,j,k)i2 ,(3.9) The weight function wrun(i, j)from each run is defined as 55 CHAPTER 3. 41CA AMS AT LOW ENERGIES wrun(i, j)≡( 0if Crun(i, j)=0 hRrun K−corrected(i, j) Urun K−corrected(i, j)i2if Neffseqs(i, j)=1 minnhRrun K−corrected(i, j) Srun K−corrected(i, j)i2;hRrun K−corrected(i, j) Urun K−corrected(i, j)i2oif Neffseqs(i, j)>1 ,(3.10) where, evidently, at least one of the criteria making Crun(i, j) = 0 is that Neffseqs(i, j) = 0. This way, those runs where the behavior was more stable and, therefore, with lower relative uncertainties, take more importance. The sample average of the K-corrected 41Ca/40Ca ratio RK−corrected(i)and its standard deviation SK−corrected(i)are calculated from the results of the runs in an analogous way as these results of the runs are calculated from the results of the sequences. The uncertainty is calculated as UK−corrected(i) = RK−corrected(i) snruns P j=1 wrun(i, j) .(3.11) 3.4.4 Blank and standard correction The next corrections are common to almost all the AMS measurements, no matter the system or the radionuclide. The background Bis calculated as the non-weighted average of the K- corrected ratios from all the blank targets. This way, the blank corrected ratio from each sample is RBLK−corrected(i) = RK−corrected(i)−B.(3.12) In this case, no weight function is added since no reason makes some blank samples more scientifically relevant than others. The uncertainty from this ratio is UBLK−corrected(i) = =(p[UK−corrected(i)]2+ [S(B)]2if Zs 0 χ2[x;Neffruns(i)−1]dx < 0.95 p[SK−corrected(i)]2+ [S(B)]2if Zs 0 χ2[x;Neffruns(i)−1]dx ≥0.95 ,(3.13) where S(B)is the standard deviation of the background, Neffruns(i)is the total number of runs from the sample iwith wrun(i, j)>0, and s≡[SK−corrected(i) UK−corrected(i)]2·[Neffruns(i)−1]. 56 3.4. DATA ANALYSIS FOR 41CA AMS AT TANDY AND SARA The next step is the correction of the stripping and optical fractionation, or standard correction. The Rexp(`)from each standard material `, with a nominal 41Ca/40Ca ratio of Rnom(`), is calculated as the non-weighted average of the BLK-corrected ratios from all the targets of that standard material. If that average has a standard deviation Sratio(`), the nominal/experimental factor f(`) = Rnom(`) Rexp(`)will have a standard deviation σf(`) = Sratio(`) Rnom(`)·f(`). Since, typically, several different standard materials are used during measurements, the final nominal/experimental factor Fapplied for the standard correction will be the non-weighted average of the different f(`)factors. No weight function is applied because the same reasons as those for blank correction: no reason makes some standard samples more scientifically relevant than others. If this produces a standard deviation σ(F), the final uncertainty of the standard correction factor Fis U(F) = max nσ(F); F P ` [f(`) σf(`)]2o.(3.14) The final 41Ca/40Ca from each sample is, therefore, RSTD−corrected(i) = F ·RBLK−corrected(i),(3.15) with an uncertainty USTD−corrected(i) = p[U(F)·RBLK−corrected(i)]2+ [F ·UBLK−corrected(i)]2.(3.16) 3.4.5 Common selection criteria used at SARA and TANDY The only selection criteria used only in 41Ca measurements at TANDY, and not at SARA, are those related to the stripper transmission, since that information is available in the measurements with that system. Typically, sequences (i, j, k)where the stripper pressure has a difference of 5% or higher with either the average tranmission through the whole measurement, or the average transmission during that run (i, j), are discarded. Nevertheless, these possible changes in the transmission are not expected to be relevant at SARA, where the optical losses are lower for 41Ca. Therewith, the rest of selection criteria, used also in the measurement at SARA, are presented, and the reasons to apply them are explained. 57 CHAPTER 3. 41CA AMS AT LOW ENERGIES Ignoring first run As it can be seen in section 3.3, during the first sputtering minutes the ion source current output typically presents a quite unstable behaviour. It is advisable not to use these results in the data analysis, so a common selection criterion is that Crun(i, 1) = 0 ∀i. (3.17) Minimun 40Ca current output A very low ion source current ouput within a sequence (i, j, k)can be symptomatic of a temporal failure, which makes the results of the ratios from that sequence not representative of the reality. Besides, a low current output is a common feature of time intervals where that same output is more unstable; for instance, when the target is starting to run out of sample material. In order to work with sequences where the behaviour of this output is stable, a minimum 40Ca current Imin can be established. For any sequence (i, j, k) : IA(i, j, k)<Imin, the weight function will be set as 0. This minimum current should be established taking into account the average currents observed during that measurement, as well as the typical currents observed historically with that AMS system. At SARA for instance, we typically use a minimum 40Ca2+ current of 20 nA. Maximum ion rate GIC detectors typically do not work properly when the total ion rates are too high. In these cases, mainly 2 effects reduce the detection efficiency: •The pile-up, or combination of the signals from 2 or more ions, producing a signal proper to an ion with a higher energy. •The saturation of the electronic chain that transform the signal produced by the ion into a digital one. It is suitable, therefore, to establish a maximum ion rate, so the sequences where that 41M or 39K rates are higher are discarded. To establish this limit, a study of the relationship between the ion rate and the GIC detector efficiency should be performed. More information about the efficiency of the SARA GIC and the typical maximum rates in our measurements is presented in chapter 4. 58 3.4. DATA ANALYSIS FOR 41CA AMS AT TANDY AND SARA Maximum 39K/40Ca ratio In order to take the measurements of each sample where the 41K interference was the least, and to avoid the unstable behaviour of the 39K/40Ca ratio when it si higher, a maximal 39K/40Ca ratio RMAX 39 can be set, so all the runs (i, j)where the average 39K/40Ca ratio Rrun 39 (i, j) is higher than RMAX 39 are discarded. This also avoids the extrapolation of the K-correction factor for Rrun 39 (i, j)higher than those found in blanks samples, provided that this maximum 39K/40Ca ratio is lower or equal than the highest 39K/40Ca ratio found in blank samples. This maximum 39K/40Ca ratio is typically set in 1-2 ×10-9 range, so the 41K/40Ca interference is expected to stay in the 10-11-10-10 range. 59 CHAPTER 3. 41CA AMS AT LOW ENERGIES 60 Chapter 4 Setup of the 1 MV AMS system at CNA for 41Ca measurements In this chapter, several experiments performed throughout this thesis related to the performance parameters for 41Ca measurements at SARA are detailed. These parameters are related to the final efficiency of the 41Ca ion detection in order to provide a strong statistics, and the final 41Ca/40Ca background and sensitivity. The most important results presented here were also presented in Vivo-Vilches et al. [2017]. The specific parameters studied and/or optimized have been: •The (40CaF3)-current output with the SO-110B ion source (section 4.1). •The transmission in the He gas stripper of the charge states 2+ and 3+ (section 4.2). •The detection efficiency and signal identification ΔE-Eres spectra of the compact GIC (section 4.3). •The linear correlation between 41K and 39K ions and the final 41Ca background after the K-correction (section 4.4). •The optical transmission of 41Ca through the HE spectrometer by the comparison of nominal and experimental ratios in standard samples (section 4.5). A comparison of these performance parameters with those at TANDY is also presented in section 4.6. In order to complete this comparison, measurements of the ETH LIP in-house 41Ca standards [Christl et al.,2013] and several samples of biomedical interest previously measured at TANDY are also shown. 61 CHAPTER 4. SETUP OF THE 1 MV AMS SYSTEM AT CNA FOR 41CA MEASUREMENTS higher mean charge state with He as stripping gas for low energies agrees with which has been observed in uranium [Wittkower and Betz,1973]. On the other hand, semi-empirical formulas applied to higher energies, like the one from Schiwietz and Grande [2001]3, predict the opposite: that, at large facilities, these mean charge states would be higher using Ar as stripping gas than using He. Therefore, it is clear that Schiwietz-Grande formula cannot be applied at low energies: for 40Ca with an energy of 400 keV, the mean charge state according to this formula would be 0.9, less than the half of the one we calculate. Therefore, the measurements of this kind of parameters are truly important in order to provide data about the stripping process at low energies. Figure 4.5: Charge state distribution for the (40CaF3)-→40Caqstripping process in He at 225 keV and 400 keV. The charge state fraction of the 4+ state was supposed to be 0 at TANDY. 4.3 40Ca current measurement with the FC3 and 41Ca detection with the compact GIC A special feature of the 41Ca measurements at SARA is that the measurement of the 40Ca current with the FC3 cannot be performed when the accelerator terminal voltage is the same 3This formula can only be applied to 40Ca with energies ≥800 keV. 68 4.3. 40CA CURRENT MEASUREMENT WITH THE FC3 AND 41CA DETECTION WITH THE COMPACT GIC which is used when the 41Ca is being detected with the GIC. The reason is the geometry of this FC. As it can be seen in Figure 4.6, the hole for 129I measurements involves that, when the terminal voltage is tuned to detect 41Ca in the GIC, part of the FC measuring 40Ca current would block the way of 41Ca ions to the ESA. Fortunately, this feature is no problem, since the FC can be placed where 40Ca is deviated when 39K is being detected. This way, 40Ca current is measured during 39K sequences, instead of during 41Ca sequences. In order to get the 41Ca/40Ca ratios for each sequence, the 40Ca current is supposed to be the average from those measured just before and just after that sequence. As expected, the 41Ca ΔE-Eres spectrum at the GIC when the 2+ state is selected in the HE side is very similar to the one at TANDY for the same charge state [Vockenhuber et al.,2015] being 82Se4+ the only other ion reaching the detector. Figure 4.7 shows how, when the isobutane pressure is low enough to let the 41Ca ions reach the second anode, the 82Se4+ ions lose only a part of its kinetic energy in the gas, matching up with the spectrum predicted by SRIM simulations. Both signals can still be distinguished in the spectrum, but there exists a remote possibility of some 82Se ions interfering with the 41Ca signal. Nevertheless, since at these energies the signals from 41K and 41Ca cannot be distinguished, there is no advantage in using this ΔE-Eres spectrum. In order to have a 82Se4+ peak with well-defined energy, the gas pressure inside the GIC can be increased so that both ions are stopped before reaching the second anode. This way, the signal from the first anode would be directly related to the total energy with which these ions reach the detector. As it is shown in Figure 4.7, with this configuration, the 82Se4+ peak presents a well-defined deposited energy of 2 times the energy of 41Ca2+, as expected. This configuration is, therefore, the one used (a) 129I (b) 41Ca (c) 39K Figure 4.6: Relative position of the FC3 depending on the isotope whose current is measured and the radionuclide which is tuned to cross to the ESA. 69 CHAPTER 4. SETUP OF THE 1 MV AMS SYSTEM AT CNA FOR 41CA MEASUREMENTS during routine 41Ca measurements at SARA. The 41Ca ΔE-Eres spectrum when the 3+ state is used (see Figure 4.8) does not present many of the peaks which can be seen at TANDY for that charge state, where ions with different M q ratio reach the detector because of the worse optical transport through the HE sprectometer. At SARA, only 82Se6+ ions reach the detector with 41Ca, as it happens when the state 2+ is used. The relative production of 82Se6+ ions is much lower in this case, presumably because of a much lower stripper transmission for such a high charge state. The integer M qratio of the 39K3+ ion implies the detection of several other ion species, being the most important of them 13C+,26Mg2+ and, when copper target holders are used, 65Cu5+. 13C+rates when sputtering a blank sample in copper targets are in the 1000-2000 s-1 range; (a) Experimental ΔE-Eres (b) Simulated ΔE-Eres (c) Experimental total E Figure 4.7: Experimental a simulated spectra for 41Ca2+ and 82Se4+ ions reaching the compact GIC at SARA. In the experimental ΔE-Eres spectrum, the dashed lines represents the average energy deposited by the 41Ca ions, and twice that energy. In the In the SRIM simulation, each dot represents a single ion. 70 4.3. 40CA CURRENT MEASUREMENT WITH THE FC3 AND 41CA DETECTION WITH THE COMPACT GIC this is several orders of magnitude below the 13C+found when trying to detect 26Al2+ [M¨ uller et al.,2015]. This is not totally unexpected, since the formation of negative ions where 13C is present and with a mass of 96 u, when injecting (39KF3)-, should not be as important as the formation of (13C2)-and (12C13CH)-clusters, when injecting 26Al-.13C+and 26Mg2+ when Al target holders were used were much higher, leading to important pile-up interfering with 39K counting region. The relationship between the total 41M rate and the compact GIC efficiency has also been studied for charge state 2+, in order to provide a reliable maximum rate for the data analysis (see chapter 3). This was done by connecting a pulse generator to the detector, and comparing the measured rate from this pulse with its nominal value, previously measured with an osciloscope. Figure 4.9 shows that, for 41M rates lower than 1000 s-1, the efficiency keeps higher than 90%. For higher rates, efficiency becomes inversely proportional to this 41M rate. It is convenient setting the maximum 41M rate in 3000-5000 s-1, for instance, and using the pulse generator during the measurements to control the detector efficiency. This way, the efficiency will stay stable and higher than 80%, but it will be possible to measure 41Ca/40Ca ratios up to 10-8 even when the source output current is relatively high. (a) 41Ca (b) 39K Figure 4.8: Experimental spectra for 41Ca3+ and 39K3+ ions reaching the compact GIC at SARA. Different amplifier gains were used during the adquisition of these spectra. 71 CHAPTER 4. SETUP OF THE 1 MV AMS SYSTEM AT CNA FOR 41CA MEASUREMENTS Figure 4.9: Relationship between the 41M rate reaching the compact GIC and the detector efficiency. The 41M rates presented have been already corrected dividing the measured rate by the efficiency. 4.4 K-correction and 41Ca background As it was stated in chapter 3, the sequential detection of ions of two different masses, 41 and 39 u, requires the use of different terminal and ESA potentials. At SARA, the best way to do so is tuning, first, the voltages which would make 40Ca ions reach the GIC, thanks to an insertable Faraday cup which can be placed just after the ESA (FC5). This also allows a better tuning of the other parameters, like the Q-Pole triplet voltage. From these voltages, it is relatively easy to calculate the terminal voltage to make another ion have the same EM product so that it is selected by the HEM. Since the energy is changed, so has to be the ESA voltage. Table 4.1 shows typical voltages used at SARA for 41M and 39K, based on the previous tuning of 40Ca in the FC5. One of the first studies at SARA of the relationship between the 41M/40Ca and 39K/40Ca ratios in blank samples is presented in Figure 4.10. As it happened in this case, most of the 41Ca measurements at SARA have shown a R2value higher than 0.9 in this linear regression, with a K-correction factor typically between 0.065 and 0.09, this is, around the natural 41K/39K ratio (0.0721). 72 4.4. K-CORRECTION AND 41CA BACKGROUND The 41Ca/40Ca background for the 2+ state, after applying the K-correction, is usually on the (2-8) ×10-12 range, and only has been higher than 10-11 when samples with 41Ca/40Ca ratios higher than 10-8 have been measured. This background levels are similar to those found at TANDY. Since 41K interfence is corrected, the main cause of this background, still much higher than those found in larger AMS facilities, would be the one produced by scattering processes Table 4.1: Voltages at the accelerator terminal and the ESA in order to select different ions with the HE spectrometer at SARA. Ion 40Ca 41M39K LE side ion mass (u) 97 98 96 Bouncer magnet voltage (V) 0 -357 364 Terminal voltage (kV) 940 915 965 2+ energy (keV) 2280 2225 2335 2+ ESA voltage (kV) 44.7 43.6 45.9 3+ energy (keV) 3220 3140 3300 3+ ESA voltage (kV) 42.1 41.1 43.2 Figure 4.10: Linear correlation between the 41M/40Ca and 39K/40Ca ratios in blank samples at SARA. 7 different blank samples were measured during 3 runs. 73 CHAPTER 4. SETUP OF THE 1 MV AMS SYSTEM AT CNA FOR 41CA MEASUREMENTS on residual gas, apertures, etc. This effect, which is increased because of the lower energies and the need of using higher stripper pressures to destroy the 2+ molecular background [Suter,1998], takes also additional importance in 41Ca measurements because of the high mass difference between the (41CaF3)-ion and 41Ca. This fact makes easier the production of ions with mass 98 u where stable isotopes of calcium with masses close to 41 u, mainly 40Ca and 42Ca, are present. Theoretically, the cross section of this effect should be lower at SARA than at TANDY, because of the higher energy. However, at TANDY, the presence at the exit of the stripping tube of one of the pumps which recirculates the stripping gas could reduce the presence of residual gas in the HE acceleration tube, compensating the higher scattering cross section. The 41Ca/40Ca background for the 3+ state has not been studied yet. 4.4.1 Tests with blank urine samples: radiochemical method One of the applications where 41Ca backgrounds found in low-energy AMS systems allow to perform measurements completely competitive with large facilites is the biomedical one. For instance, administrating oral doses of 41Ca of 100 nCi (28.23 nmol), minimum urinary 41Ca/40Ca ratios4are in the 5-10 ×10-11 range [Denk et al.,2006,2007]. As it was stated in chapter 1, such a dose can be considered radiologically exempt. Even studies where 41Ca doses are one order of magnitude higher (and so would be 41Ca/40Ca ratios) are still considered to have no radiological hazard [Weaver et al.,2009], since the dose over 50 years would be 10 µSv5[ICRP,2012]. The capabilities of the SARA system to measure this kind of samples was tested in an intercomparison measurement with the TANDY system (see section 4.6). These samples were sent to the ETH LIP in CaF2form by the NutriTrace group at the National University of Singapore. Nevertheless, in order to be able to perform 41Ca measurements for biomedical groups at CNA, including the chemical preparation of the samples in our service, chemical preparation tests with blank urine samples have been performed. The chemical procedure was based on those presented on the bibliography [Freeman et al., 1995;Miller et al.,2013;Jackson et al.,2013], but with two important changes: the initial precipitation of calcium carbonate instead of calcium oxalate; and the omission of the use of ion-exchange resins. The use of calcium carbonate avoids the presence of this salt in the final 4Urine samples are the most common biomedical samples use to control the changes of the 41Ca concentration in the extracellular fluids on this kind of studies. 5The yearly background dose because of natural sources is typically higher than 1 mSv. 74 4.4. K-CORRECTION AND 41CA BACKGROUND CaF2precipitate, since the carbonate ion is destroyed if the salt is dissolved in an acid6. The ion-exchange step was discarded because our tests with carbonate samples did not show a clear reduction of the final 39K/40Ca ratios. The time consumption and sample loss because of this process are, therefore, not worthy. The steps of the first part of our chemical procedure, the calcium carbonate isolation (see Figure 4.11), designed for 24 h samples and tested with 1 l collections from volunteers7, are: 1. Addition, per 100 ml urine, of 5 ml of concentrated HCl and settling of the solution for 1 h. 2. Decanting of the supernatant in a 2 l (or larger) beaker and addition of concentrated NH3until pH=11. 3. Addition, per 100 ml urine, of 10 ml Na2CO3solution (0.4 g/ml), heated until it almost boils. 4. Settling of the beaker overnight with some inclination. It is also possible to let the beaker upright, but heating without boiling to take advantage of the indirect relationship between temperature and the solubility of calcium carbonate8. 5. Discarding of most of the supernatant and transfer of the rest, together with the precipitate, to 50 ml centrifuge tubes. 6. Centrifugation and discarding of the supernatant. Transfer precipitates to an unique 50 ml centrifuge tube per sample. Figure 4.11: Scheme of the calcium isolation method for urine samples. 6(CO3)2- +2H+−→ CO2+ H2O 7Since these volunteers were not dosed with 41Ca these samples are considered blanks. 8This also avoids the presence of other calcium salts since, unlike calcium carbonate, most of the compounds present a direct relationship between temperature and solubility. 75 CHAPTER 4. SETUP OF THE 1 MV AMS SYSTEM AT CNA FOR 41CA MEASUREMENTS 7. Washing with 20 ml saturated Na2CO3solution (0.2 g/ml). Centrifugation and discarding of the supernatant. This step should be performed twice. 8. Washing with 20 ml Na2CO3solution (0.02 g/ml). Centrifugation and discarding of the supernatant. This step should be performed twice. 9. Washing with 20 ml of MilliQ water. Centrifugation and discarding of the supernatant. 10. Dissolving of the precipitate with 5 ml of concentrated HCl. 11. Addition of concentrated NH3until pH=9. 12. Centrifugation and discarding of the precipitate (mostly magnesium hydroxide). 13. Addition of some drops of concentrated Na2CO3solution (0.2 g/ml). 14. Centrifugation and discarding of the supernatant. 15. Washing with acetone. 16. Centrifugation and discarding of the supernatant. 17. Drying at 70 ºC for 1 h. The CaF2precipitation part is a much more simple procedure. The carbonate is dissolved with 1 M HCl, and just some drops of concentrated HF are necessary to precipitate the calcium fluoride. After centrifuging and discarding the supernatant, the precipitate is washed with MilliQ water. Then, the centrifugation and discarding of the supernatant is repeated. The final precipitate is dried at 105 ºC overnight. K-corrected 41Ca/40Ca ratios from these blank urine samples was equivalent to those found in our typical blank samples, around 5 ×10-12. This agrees with the fact that, in large facilities where background levels are in the 10-15-10-14, the 41Ca/40Ca ratio from processed urine blanks is typically lower than 5 ×10-13 [Jackson et al.,2013], one order of magnitude lower than our background values. During this thesis, we have been in contact with several biomedical groups from different universities in Spain in order to perform these kind of 41Ca measurements. Unfortunately, these contacts were not successful. Nevertheless, our system is totally capable of performing the measurements in the case that any group is interested in the future. 76 4.5. MEASUREMENT OF THE HE SIDE OPTICAL TRANSMISSION WITH STANDARD SAMPLES 4.5 Measurement of the HE side optical transmission with standard samples The presence of a quadrupole triplet at the exit of the 1 MV tandetron allows an optimal optical focusing no matter the effective charge state q∗≡M0 Mqof the ions (see chapter 2). This is one of the most important differences of SARA with the TANDY system at ETH LIP for 41Ca AMS. The non-optimal optical behaviour of 41Ca through the HE side spectrometer at TANDY implies optical losses of, at least, 15% for the charge state 2+, and 33% for the charge state 3+ [Christl et al.,2013;Vockenhuber et al.,2015]. At SARA, therefore, there should exist an improvement on this transmission relative to the one at TANDY, and no such a dependency on the charge state. An approximation to this optical transmission can be obtained by checking the measured 41Ca/40Ca ratios from the standard samples, provided that the detector efficiency is corrected. At SARA, this study has only been performed, yet, with the 2+ state. The AE701 standards from the Institute for Reference Materials and Measurements (IRMM) of the European Commission’s Joint Research Centre [Hennessy et al.,2005] have been used for standard correction in 41Ca measurements at SARA during this thesis. Minimum measured/nominal 41Ca/40Ca ratios higher than 80% were typically consequence of slight changes in the terminal voltages, while the optimal conditions lead to measured/nominal relationship even higher than 99%. Taking, for instance, all the 41Ca measurements performed during the year 2017, the average of this measured/nominal factor is (93.8 ±5.5)%, using the standard deviation as uncertainty. Table 4.2 shows an example of the blank corrected 41Ca/40Ca ratios from each of the 3 standards used during one of our 41Ca measurements at SARA; this is, the 41Ca/40Ca ratios calculated in order to perform the standard correction (see chapter 3) During this measurement, 2 cathodes from each standard material were measured. The optical transmission for the 3+ state has not been studied yet. Table 4.2: Blank corrected 41Ca/40Ca ratios of the standard samples used in one of the 41Ca measurements at SARA using the 2+ state. Standard Nominal 41Ca/40Ca Measured 41Ca/40Ca Measured/nominal (%) AE701/3 (1.0181 ±0.0069) ×10-8 (0.988 ±0.019) ×10-8 97.0 ±2.0 AE701/4 (1.0479 ±0.0071) ×10-9 (1.029 ±0.049) ×10-9 98.2 ±4.8 AE701/5 (1.0520 ±0.0071) ×10-10 (1.000 ±0.013) ×10-10 95.1 ±1.5 77 CHAPTER 4. SETUP OF THE 1 MV AMS SYSTEM AT CNA FOR 41CA MEASUREMENTS Figure 4.15: Temporal evolution of the urinary 41Ca/40Ca ratio in subject 6, including measurements at TANDY and SARA. 84 4.6. COMPARISON MEASUREMENTS WITH TANDY Figure 4.16: Temporal evolution of the urinary 41Ca/40Ca ratio in subject 7, including measurements at TANDY and SARA. 85 CHAPTER 4. SETUP OF THE 1 MV AMS SYSTEM AT CNA FOR 41CA MEASUREMENTS Figure 4.17: Temporal evolution of the urinary 41Ca/40Ca ratio in subject 8, including measurements at TANDY and SARA. 86 4.6. COMPARISON MEASUREMENTS WITH TANDY Figure 4.18: Temporal evolution of the urinary 41Ca/40Ca ratio in subject 9, including measurements at TANDY and SARA. 87 CHAPTER 4. SETUP OF THE 1 MV AMS SYSTEM AT CNA FOR 41CA MEASUREMENTS Figure 4.19: Temporal evolution of the urinary 41Ca/40Ca ratio in subject 10, including measurements at TANDY and SARA. 88 Chapter 5 41Ca measurement in concrete samples from the bioshield of a nuclear power plant Previous chapters are dedicated to the technical aspects of the 41Ca AMS measurements. This last chapter is dedicated to the first application performed at CNA Seville of the 41Ca measurements: the characterization of concrete samples from the bioshield of a NPP for the Spanish radioactive waste management agency, ENRESA. The direct relationship between the 41Ca/40Ca ratio in the concrete bioshield and the thermal neutron fluence is explained in section 5.1. Most of the information from this section is based on Chapter 3 of Lewis [2008]. The samples had to be chemically treated in order to have the CaF2to be measured by AMS. The chemical method developed for that, and applied to all the samples measured, is detailed in section 5.2. The final description of the samples and the results obtained are shown in section 5.3. 5.1 Production of 41Ca in the concrete bioshield of nuclear reactors 41Ca is produced in the concrete bioshield of nuclear reactors by the neutron activation of its stable isotope, 40Ca. The energy spectrum from these neutrons reaching the shield presents 89 CHAPTER 5. 41CA MEASUREMENT IN CONCRETE SAMPLES FROM THE BIOSHIELD OF A NUCLEAR POWER PLANT two main groups: •Fast neutrons (energies in the 1 MeV range) produced after the fission processes which escape the reactor vessel before being thermalized by the reactor moderator. •Thermal neutrons (energies much lower than 1 eV) from the moderation of the neutrons produced in the fission which escape the reactor vessel before either producing a new fission reaction, or being absorbed by another material within the vessel (control rods, walls, etc.). Nevertheless, the production of 41Ca by fast neutron capture can be disregarded since, as it can be seen in Figure 5.1, the cross section of the 40Ca(n,γ) reaction is around 2 orders of magnitude lower for these neutrons than for thermal energies. In fact, the main contribution of fast neutrons to the production of 41Ca would come from its moderation in the concrete material itself, due to the elastic scattering processes with the hydrogen present on it. This way, it can be supposed that 41Ca is entirely produced by the thermal neutron capture of 40Ca, being these either original thermal neutrons reaching the shield, or fast neutrons which have been thermalized by the material. The elastic scattering not only thermalize the neutrons to room temperature, but also causes a diffusion of the thermal neutrons. Figure 5.1: 40Ca neutron capture cross section and thermal neutron energy distribution. Cross section data from IAEA [n.d.]. 90 5.1. PRODUCTION OF 41CA IN THE CONCRETE BIOSHIELD OF NUCLEAR REACTORS The 41Ca/40Ca ratio in any location of the shield, therefore, would be 41Ca/40Ca =Z∞ 0 σ(E)dΦ(E) dE dE, (5.1) where σ(E)is the 40Ca(n,γ) cross section as a function of the energy, and dΦ(E) dE is the energy distribution of the neutron fluence. The energy distribution of thermal neutrons can be considered to be Maxwellian, this is, dΦ(E) dE = Φth E (kT)2e−E kT ,(5.2) being Φth the total thermal neutron fluence, kthe Boltzmann constant, and Tthe temperature of the medium where the neutrons have been thermalized. In pressurized water reactors (PWR), for instance, the water moderator is typically at a temperature around 600 K. However, these neutrons are easily thermalized again to room temperature (∼300 K) when they reach the concrete shield, mainly because of the hydrogen content on it1. Therefore, in this area it can be considered that kT = 0.0253 eV. At thermal energies, neutron absoption cross sections are inversely proportional to the velocity of the neutrons. Typically, the cross section for neutrons with an energy E0= 0.0253 eV is reported in the bibliography, which is called ”thermal neutron capture cross section”; for instance, Cranston and White [1971] report a thermal neutron capture cross section of 400 ±40 mb for 40Ca. Consequently, the neutron capture cross section for other energy in the thermal range is σ(E) = σ(E0)rE0 E.(5.3) Therefore, the 41Ca/40Ca ratio produced by a thermal neutron fluence Φth would be 41Ca/40Ca = Φth ·σ(E0)·1 (kT)2Z∞ 0rE0 EEe−E kT dE = = Φth ·σ(E0)·rE0 kT Z∞ 0 √te−tdt =√π 2rE0 kT σ(E0)Φth. (5.4) Defining the thermal average cross section as ¯σ(T)≡√π 2rE0 kT σ(E0),(5.5) 1Neutron energy loss is more important when the scattering process takes place with light nuclei. 91 CHAPTER 5. 41CA MEASUREMENT IN CONCRETE SAMPLES FROM THE BIOSHIELD OF A NUCLEAR POWER PLANT the total thermal neutron fluence in a location is then directly related to the measured 41Ca/40Ca there, being Φth = 41Ca/40Ca ¯σ(T).(5.6) In concrete, where neutrons are thermalized to room temperature kT =E0, the thermal average cross section would be ¯σ=√π 2σ(E0). For the 40Ca(n,γ) reaction, this cross section would be, therefore, 354 ±35 mb = (3.54 ±0.35) ×10-25 cm2. Note that, since production of 41Ca is directly proportional to 40Ca content, the 41Ca/40Ca ratio is independent on the calcium concentration in the material. The only other important factor related to this ratio, besides the thermal neutron fluence, is the temperature to which these neutrons are thermalized. The regions of maximum neutron flux in bioshields from commercial nuclear reactors can suffer thermal neutron fluences close to 1019 cm-2. This would produce 41Ca/40Ca ratios higher than 3.5 ×10-6. In order to calculate the 41Ca specific activity2from 41Ca/40Ca measurements, it is necessary to know the calcium concentration in the concrete, which can strongly vary from one type of concrete to another. The 41Ca/40Ca ratio can be used, also, to calculate atom ratios from other radionuclides produced by neutron capture. For instance, the 36Cl/35Cl ratio in the same location would be 36Cl/35Cl =σ[E0,35Cl(n, γ)] σ[E0,40Ca(n, γ)] 41Ca/40Ca,(5.7) independently of the temperature, in this case, since the same factor would correct both the thermal neutron capture cross sections of 35Cl and 40Ca. 5.2 Radiochemical treatment of concrete samples for 41Ca AMS measurements Our chemical method to treat the concrete is, in some aspects, based on those methods used for 41Ca LSC or intercomparisons between this technique and AMS [Hampe et al.,2013;Hou, 2005;Warwick et al.,2009]. Nevertheless, many steps which are not necessary for 41Ca AMS 2The 41Ca specific activity in any material is defined as the 41Ca activity per material mass unit. 92 5.2. RADIOCHEMICAL TREATMENT OF CONCRETE SAMPLES FOR 41CA AMS MEASUREMENTS have been eliminated, and some steps which improve the potassium suppression have been added. The method, as the one described for urine samples in chapter 4, is divided in two parts: the calcium isolation as calcium carbonate, and the calcium fluoride precipitation. A simple scheme of the calcium isolation process is presented in Figure 5.2. Each sample is first digested in 20 ml of aqua regia during 2 hours. Typically, between 100 and 500 mg of each sample is used. The amount depends on the aspect of the sample: •When most of it has a pale color and is easily ground, most of the sample comes from the binder, which is strongly dissolved by acid digestion, so less amount of concrete sample is needed. •When the sample includes hard and dark aggregates, which are typically silicates which are not dissolved by aqua regia, a relatively large amount of sample, between 300 and 500 mg, is treated. The 41Ca/40Ca ratio does not depend on the chemical form in which the calcium is found. Consequently, the fact that the acid digestion does not dissolve part of the calcium on the sample, like that calcium present in silicates, has no huge relevance. In Figure 5.3 the difference between how the acid digestion attacks the sample depending on the aggregate/binder relationship can be observed. In one of the 2 samples, it can be observed how an aggregate is not affected by the acid digestion. This step eliminates most of the Si present in the sample. The acid solution is centrifuged, discarding the precipitate. 20 ml of concentrated NH3is added to the supernatant in order to precipitate hydroxides from other elements present in concrete, like iron or magnesium. If necessary, more concentrated NH3is added until pH is between 9 and 11. The solution is centrifuged again, discarding the precipitate. Figure 5.2: Scheme of the calcium isolation process for concrete samples. 93 CHAPTER 5. 41CA MEASUREMENT IN CONCRETE SAMPLES FROM THE BIOSHIELD OF A NUCLEAR POWER PLANT concrete structure or even a higher moderation of the fast neutrons at some points. The case of those drill cores at 30 cm from the central ones in the horizontal direction (CDL) is special, since they are extremely close to the apertures toward the instrumentation wells (see Figure 5.8). This apertures were filled with air, so they could be considered ”transparent”to the neutrons, which reached the side walls of these aperture. The diffusion of the neutrons in the concrete would cause another ”neutron front”, perpendicular to this wall. For deeper locations of the CDL drill cores, the neutron fluence added by this front, relative to the one coming from the cavity surface, would be more important than for those locations closer to the reactor vessel. In any case, this apertures add complexity to the behavior of the neutrons fluence in these locations, so the non ideal behavior of the 41Ca/40Ca depth profiles in these drill cores is not unexpected. The lack of symmetry between the results of both blocks, on the other hand, is quite unexpected, but could be also a consequence of the huge heterogeneity of the material. Due to the 30% Ca content of the concrete, the samples closest to the vessel, with 41Ca/40Ca ratios around 3 ×10-6, would present 41Ca specific activities of 3000 Bq/g. On the other hand, the sample with the minimum 41Ca/40Ca ratio, 3 ×10-10, would present 41Ca specific activities of 0.3 Bq/g. According to the results, maximum thermal neutron fluences on the bioshield surface are very close to 1019 cm-2. This result has a great relevance, since it is known that concrete Figure 5.8: Cross section of the Jos´ e Cabrera reactor vessel and bioshield when the plant was operative at the height of maximum neutron fluence. The positions where drill cores CDC and CDL were taken, for blocks 1, 4 and 6. For the rest of the blocks, CDL core was taken from the opposite side. 100 5.3. MEASUREMENTS OF 41CA IN THE BIOSHIELD OF THE JOS´ E CABRERA NUCLEAR POWER PLANT materials start losing some of the mechanical properties when they undergo total neutron fluences above 1019 cm-2 [Field et al.,2015]. Taking into account that our results only assess the thermal neutron fluence, it is almost sure that the total neutron fluence reached levels much higher that 1019 cm-2 at these points. These effects on the mechanical proporties, nevertheless, may strongly depend on the type of concrete. Consequently, the study of the mechanical properties of this concrete, which ENRESA will perform, together with our results of the the thermal neutron fluences and additional studies of the fast neutron fluence, could help to a better assessment of the effects of neutron radiation in these properties. Another capability of these results is the estimation of the specific activities from other radionuclides which are produced in concrete mainly by thermal neutron activation. This is the case, for instance, of 36Cl. Chlorine is only found in trace amounts in concrete, mainly caused by the Cl-presence in water. Typical chloride concentrations in concrete are in the range of 100 ppm. The much lower presence of chlorine in concrete, nevertheless, is compensated by the relatively high thermal neutron capture cross section, of 41.8 ±1.2 b [Sims and Juhnke, 1969], this is, 2 orders of magnitude higher than the one for 40Ca. In any case, even assuming chloride concentrations near limits, 36Cl specific activities would not exceed 35 Bq/g, which is well below the limit for exempt material, 104Bq/g [CSN,2003;IAEA,2012]. 5.3.3 Results from the lateral samples Samples from the left side of the block 4 were not sent to be measured by AMS at CNA since these would be equivalent to those from the right side of the block 3. A scheme of the location of these samples was presented in Figure 5.6. The heights of these locations, relative to the center of the reactor core, were +360 cm, +280 cm, +200 cm, +120 cm, -260 cm and -300 cm. The depths were 22.5 cm and 67.5 cm for sub-blocks A and B, and 35 cm and 105 cm for sub-block E. The 41Ca/40Ca results in these locations is presented in Figure 5.9. The main uncertainty for the interpretation of these results is the possibility of cross-contamination during the bioshield segmentation process. Nevertheless, in most of the cases, an expected indirect relationship between the vertical distance to the core and the 41Ca/40Ca ratio is found; so is between depth and 41Ca/40Ca. Taking into account that these samples are relatively far from the areas of maximum neutron fluence, the exemptions to these indirect relationships are not unexpected. These exemptions may be caused by the thermal neutron diffusion processes or the presence of higher fast neutron fluences near these areas which are moderated by the concrete, as it happened in the case of the drill cores. It is, nevertheless, quite unexpected that the results from the 3 sides are not equivalent since, because of the geometry 101 CHAPTER 5. 41CA MEASUREMENT IN CONCRETE SAMPLES FROM THE BIOSHIELD OF A NUCLEAR POWER PLANT of the reactor core, the neutron spatial distribution should have a 45º rotational symmetry. On the other hand, this rotational symmetry should only be expected on the neutron fluence for an homogeneous material. As it was previously stated, this is not the case of concrete walls. One of the elements where this heterogeneity is expected to be high is hydrogen, since additional water is typically incorporated during the hardening process, and in the different fuel changes during the operation of the NPP. Hydrogen presence is key for thermal neutron diffusion and fast neutron moderation. The minimum 41Ca specific activity in this samples is found to be close to 0.15 Bq/g. For most of these lateral samples the maximum mass available was of several hundred of milligrams in order to keep the sample package below 137Cs exemption limits. On the other hand, the most sensitive studies of 41Ca measurements by LSC report detection limits of 0.1 Bq. According to this, 41Ca characterization by LSC would have not been possible, and AMS becomes necessary. Besides, even in the case where samples could have been measured by LSC, as those from the drill cores, the sample mass which should be destroyed and statistical uncertainties would have been much higher. Figure 5.9: 41Ca/40Ca results, in ×10-9 units, from the lateral samples. The -I and -D makes reference to which of both sides of the block the samples come from, as seen from the reactor cavity: I, for left side; D, for right side. The dashed line represents the height where the center of the reactor core was. 102 Chapter 6 Summary and conclusions This thesis has fulfilled its main goal: the setting-up of the 41Ca measurements with the 1 MV AMS system at CNA, SARA. Most of the results obtained with this system can also be applied to other 1 MV AMS systems from HVE. Besides, the data collected from the different measurements performed during this thesis on this system and the 600 kV AMS system at ETH Zurich, TANDY, have provided information useful to any 41Ca AMS measurement, specially at low energy systems. Routine AMS measurements of 41Ca with SARA have been implemented at CNA, applied to the characterization of low level nuclear residues. In chapter 1, an initial idea about the different applications of 41Ca measurements was presented, together with the basic ideas related to the AMS technique. The presentation of the SARA and TANDY systems in chapter 2 served also to describe the physical processes involved in the different elements of AMS systems. These descriptions focus on the recent changes on both systems and in those differences which are important to the 41Ca AMS performance. Different studies related to the 41K interference, which is the main challenge in 41Ca AMS at low energies, are presented in chapter 3. The experiments in order to setting-up and optimize the performance of 41Ca measurements at SARA are detailed in chapter 4. The first application of 41Ca AMS at CNA, the study of the 41Ca spatial distribution in the bioshield of a NPP, can be found in chapter 5. 41Ca AMS at low energies Even when the 41K interference can be estimated by sequentially measuring the other stable isotope of potassium, 39K (K-correction), it is important to assess the possible ways by which this interference can vary. 103 CHAPTER 6. SUMMARY AND CONCLUSIONS The existence of the (41K57Fe)-ion and its relationship with the 41K interference has been proven thanks to several experiments performed at TANDY during the stay at ETH LIP. This is the cause of the relatively high 41K ion rates when metallic elements are sputtered. The main consequences, which have been observed in both systems, SARA and TANDY, are: •The dependence of the 41K interference on the materials used in the cathode, like the target holder or the binder. For instance, higher 41K/40Ca has been observed when Al target holders are used, instead of Ti or Cu. Al materials typically present higher Fe trace content. •The 41K/40Ca is not constant in time. Because of this, the K-correction must be applied sequentially during the data analysis. A detailed study of this evolution at SARA is presented, showing how the 41K interference behavior differs even for different targets from the same sample. Applying this K-correction manually can be time consuming. Therefore, two programs have been written in FORTRAN code for the 41Ca measurements at SARA and TANDY. The use of these programs means a huge reduction of the time dedicated to this analysis, from around one hour to several minutes. 41Ca measurements at SARA A detailed study of the performance parameters of 41Ca measurements at SARA has been performed throughout this thesis. The main parameters studied have been: •Ionization efficiency: mixing the calcium fluoride samples with silver in a CaF2:Ag weight ratio of 1:4, the (40CaF3)-output current of the HVE SO-110B ion source is stable and between 50 and 150 nA. The ionization efficiency in this case is around 0.1%. •Stripper transmission: the charge state with the highest transmission in the He stripper is the 2+ state. Using a He gas pressure of 0.03 mbar, this is, a He mass thickness of ∼0.075 µg·cm-2, transmission for this charge state is 40%, and a full supression of the molecular background is achieved. The possibilities of the charge state 3+, with a lower transmission of 20%, are being studied because of the possibility of reducing the final background. •GIC detection efficiency: the compact GIC presents detections efficiencies higher than 90% for 41Ca rates lower than 1000 s-1. For higher rates, the efficiency decreases linearly. 104 •41Ca/40Ca background: while the typical 41K/40Ca ratios are in the 10-11 range, the 41Ca/40Ca in blank samples, after applying the K-correction, is one order of magnitude lower. •HE side optical transmission: thanks to the independence of the object of the HEM on the charge state, due to the quadrupole triplet at the exit of the accelerator, the 41Ca beam is optimally transported through the HE spectrometer. The measurements of 41Ca standards show that the optical transmission in this sector is in the 90-100% range. The comparison of the performance parameters at SARA with those at TANDY show that the capabilities of both systems for 41Ca AMS are equivalent. SARA is perfectly fit to perform, for instance, 41Ca measurements for biomedical applications and for characterization of low level nuclear residues. 41Ca in the bioshield from a nuclear power plant The first application of the 41Ca measurements at SARA has been the study of the 41Ca spatial distribution in the bioshield of the Jos´ e Cabrera NPP. The 41Ca/40Ca ratio in this structure is directly related to the thermal neutron fluence suffered by it. A radiochemical method to treat large number of concrete samples has been developed and applied to samples from a huge variety of depths and heights of Jos´ e Cabrera’s bioshield. Only 100-500 mg per sample are needed to be destroyed in this treatment, one order of magnitude lower than the sample mass needed for LSC measurements. Besides, AMS not only improves the limit of detection, but also improves the precision of this measurements. The results from the 41Ca measurements at SARA show that, because of the complexity of the geometry and the heterogeneity of the material, the behavior of the thermal neutron fluence becomes less ideal the farer the samples are from the position of maximum neutron fluence. Consequently, neutron transport simulations may be not enough to estimate the inventories of the radionuclides produced by the neutron radiation. 105 CHAPTER 6. SUMMARY AND CONCLUSIONS Relevant publications within this PhD thesis Two articles have been published related to this thesis: The studies of the relationship with the 41K interference of the iron presence and the time evolution, and the specific features of the measurement analysis relates to the K-correction, detailed in chapter 3, are presented in C. VIVO-VILCHES, J.M. L´ OPEZ-GUTI´ ERREZ, M. GARC´ IA-LE´ ON, C. VOCKENHUBER Factors related to 41K interference on 41Ca AMS measurements Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, in press doi: 10.1016/j.nimb.2018.04.031 The main results and experiments presented on chapter 4 about the optimization of the performance parameters of the 41Ca measurements with SARA can be found in C. VIVO-VILCHES, J.M. L´ OPEZ-GUTI´ ERREZ, M. GARC´ IA-LE´ ON, C. VOCKENHUBER, T. WALZCYK 41Ca measurements on the 1 MV AMS facility at the Centro Nacional de Aceleradores (CNA, Spain) Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 413 (2017) 13-18 doi: 10.1016/j.nimb.2017.10.003 106 Appendix A GICOSY files Figures 2.18 and 2.20 show simulations of the ion optics in the HE spectrometers of SARA and TANDY, performed with GICOSY [Weick,n.d.]. The geometry data from both systems were obtained from Jacob [2001] and Chamizo [2009]. The meaning of the different codes can be found in the manual, which is in Dr. Helmut Weick’s webpage. Simulation from Figure 2.15 has also been performed with GICOSY. Below, the files written for the HE spectrometers of SARA and TANDY are presented. A.1 SARA HE spectrometer S N SARA-HEsector ; R P 0.395 41 2 ; P X 0.00 0.014 ; P Y 0.00 0.014 ; D P 0.005 0.0025 ; ; VQ2=7.8; VQ1=-VQ2; ; SS; ; D L 0.24 ; ; FF0; 107 APPENDIX A. GICOSY FILES A C 0.8 1.1375 0.0475; FF0; ; D L 0.095 ; ; F F 1; ; E Q 0.059 VQ1 0.025; ; F F 1; ; D L 0.0405 ; ; F F 1; ; E Q 0.132 VQ2 0.025; ; F F 1; ; D L 0.0405 ; ; F F 1; ; E Q 0.059 VQ1 0.025; ; F F 1; ; D L 0.52 ; ; F F 3 25.6 0.0 ; ; M S 0.85 90 0.0215 0 0 0 0 0; ; FF3310.0; ; D L 1.094 ; ; 108 A.2. TANDY HE SPECTROMETER D H E (X,Y) 100000 0.003 0.003 1 10 ; ; D L 0.375 ; ; FF100; ; E S 0.65 120 0.0125 1 -1 1; ; FF100; ; D L 0.375 ; ; D H E (X,Y) 100000 0.003 0.003 1 10 ; ; D L 0.375 ; ; D S 0.025 10 100 1 3 1 3;; DB0.0350.03511313131; END ; ; A.2 TANDY HE spectrometer S N TANDY-HEsector ; R P 0.23 41 2 ; P X 0.00 0.014 ; P Y 0.00 0.014 ; D P 0.025 0.0025 ; ; SS; ; D L 0.38 ; ; FF0; A C 0.24 2.083 0.0475; FF0; 109 APPENDIX C. FORTRAN PROGRAMS FOR 41CA MEASUREMENT ANALYSIS Number of cycles (=seq)/run: 5 min. 40Ca HE (nA): 20 max. 39K/40Ca in run: 1.25e-9 max rate (cts/s): 3000.0 Ignore first run? (yes/no=1/0) 0 Sample scatter (%): 1 Charge state: 2 Number of standards of interest: 3 Standard 1 nom. val, number of samples, positions: 1.0181e-8 2 3 15 Standard 2 nom. val, number of samples, positions: 1.0479e-9 2 4 16 116 C.2. OUTPUT FILES Standard 3 nom. val, number of samples, positions: 1.0520e-10 2 5 17 Number of blanks: 4 Position: 2 7 14 19 Pulser (1/0=yes/no) 1 Pulser frequency (Hz) 100.131 C.2 Output files C.2.1 out-K-correction This file is written while performing the K-correction. The information provided is: •39K/40Ca and 41M/40Ca ratios from the different runs from all the blank samples, which are the (x,y) values for the linear regression. •The K-corretion factor α, this is, the slope from the linear regression. •The relative uncertainty of α. •The R2value of the linear regression. 117 APPENDIX C. FORTRAN PROGRAMS FOR 41CA MEASUREMENT ANALYSIS C.2.2 out-runs The file providing information about the different runs from all the samples is written between applying the K-correction and the other important corrections: blank and standard correction. The information provided for each run is: •The number of sequences which passed the discard criteria during that run. •The total number of 41M counts detected in that run. •The average 40Ca current. •The average 39K/40Ca ratio. •The K-corrected 41Ca/40Ca ratio. •The relative uncertainty of this ratio. •The relative standard deviation of this ratio. In the case of measurements at TANDY, this file also provides information about the stripper transmission. C.2.3 out-blank This file is written while the background is being calculated. It just shows: •The average 41Ca/40Ca from each blank sample after K-correction. •The 41Ca/40Ca background, this is, the average from all the blank samples. •The relative standard deviation from this background. C.2.4 out-stds This file is written while calculating the nominal/experimental factor from the standard samples. It shows, for each different standard material: •The nominal 41Ca/40Ca ratio from that material. 118 C.2. OUTPUT FILES •The average of the measured 41Ca/40Ca in all the cathodes from that material. •The factor between this nominal and measured ratios. •The relative standard deviation from this factor. It also shows that average of the nominal/measured factor from different standard materials, and its final relative uncertainty. C.2.5 out-final-ratios In this file, the final results from each sample are presented. The information provided for each sample is: •The number of runs from that sample which passed the discard criteria. •The average K-corrected 41Ca/40Ca ratio. •The relative uncertainty from that ratio. •The relative standard deviation from that ratio. •The integral of the χ2distribution which is used to select the uncertainty of the blankcorrected ratio. •The blank-corrected 41Ca/40Ca ratio. •The relative uncertainty from that ratio. •The final standard-corrected 41Ca/40Ca ratio. •The relative uncertainty from that ratio. •The total number of 41M counts detected in all the runs form that sample. •The average 39K/40Ca ratio. •The average 40Ca current. In the case of measurements at TANDY, this file also provides information about the stripper transmission. 119 APPENDIX C. FORTRAN PROGRAMS FOR 41CA MEASUREMENT ANALYSIS 120 Bibliography Akhmadaliev, S., Heller, R., Hanf, D., Rugel, G., Merchel, S., (2013). The new 6 MV AMS- facility DREAMS at Dresden. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 294, 5–10. http://linkinghub.elsevier.com/retrieve/pii/S0168583X12000961 Alton, G., (1993). Ion sources for accelerators in materials research. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 73, 221–288. http://linkinghub.elsevier.com/retrieve/pii/0168583X9395738Q Arnold, M., Merchel, S., Bourl` es, D. L., Braucher, R., Benedetti, L., Finkel, R. C., Aumaˆıtre, G., Gottdang, A., Klein, M., (2010). The French accelerator mass spectrometry facility ASTER: Improved performance and developments. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 268, 1954–1959. http://linkinghub.elsevier.com/retrieve/pii/S0168583X10002028 Baba, T., Funaya, N., Miyazaki, N., Takahashi, T., Kaikura, T., Seki, R., Sueno, S., Jiang, S., Nagashima, Y., (1997). Detection of 41Ca with a 7Li37Cl molecular pilot beam. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 123, 183–185. http://linkinghub.elsevier.com/retrieve/pii/S0168583X96007847 Bennett, C. L., Beukens, R. P., Clover, M. R., Gove, H. E., Liebert, R. B., Litherland, A. E., Purser, K. H., Sondheim, W. E., (1977). Radiocarbon Dating Using Electrostatic Accelerators: Negative Ions Provide the Key. Science 198, 508–510. http://www.sciencemag.org/cgi/doi/10.1126/science.198.4316.508 Bergqvist, I., Drake, D., McDaniels, D., (1974). Radiative capture of fast neutrons by 40Ca. Nuclear Physics A 231, 29–44. http://linkinghub.elsevier.com/retrieve/pii/0375947474902917 121 BIBLIOGRAPHY BMJV, (n.d.). Verordnung ¨ uber den Schutz vor Sch¨ aden durch ionisierende Strahlen. https://www.gesetze-im-internet.de/strlschv_2001/BJNR171410001.html Chamizo, E., Enamorado, S. M., Garc´ıa-Le´ on, M., Suter, M., Wacker, L., (2008)a. Plutonium measurements on the 1 MV AMS system at the Centro Nacional de Aceleradores (CNA). Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 266, 4948–4954. http://linkinghub.elsevier.com/retrieve/pii/S0168583X08009506 Chamizo, E., L´ opez-Guti´ errez, J. M., Ruiz-G´ omez, A., Santos, F., Garc´ıa-Le´ on, M., Maden, C., Alfimov, V., (2008)b. Status of the compact 1 MV AMS facility at the Centro Nacional de Aceleradores (Spain). Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 266, 2217–2220. http://linkinghub.elsevier.com/retrieve/pii/S0168583X08002218 Chamizo, E., (2009).Medida de is´ otopos de plutonio, 239Pu y 240Pu, mediante espectrometr ´ ıa de masas con aceleradores de baja energ ´ ıa. Ph.D. thesis, Universidad de Sevilla. Chamizo, E., Christl, M., Fifield, L., (2015)a. Measurement of 236U on the 1 MV AMS system at the Centro Nacional de Aceleradores (CNA). Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 358, 45–51. http://linkinghub.elsevier.com/retrieve/pii/S0168583X15004796 Chamizo, E., L´ opez-Lora, M., Villa, M., Casacuberta, N., L´ opez-Guti´ errez, J. M., Pham, M. K., (2015)b. Analysis of 236U and plutonium isotopes, 239,240Pu, on the 1 MV AMS system at the Centro Nacional de Aceleradores, as a potential tool in oceanography. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 361, 535–540. http://linkinghub.elsevier.com/retrieve/pii/S0168583X15001913 Chamizo, E., Santos, F. J., L´ opez-Guti´ errez, J. M., Padilla, S., Garc´ıa-Le´ on, M., Heinemeier, J., Schnabel, C., Scognamiglio, G., (2015)c. Status report of the 1 MV AMS facility at the Centro Nacional de Aceleradores. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 361, 13–19. http://linkinghub.elsevier.com/retrieve/pii/S0168583X15001391 Chamizo, E., L´ opez-Lora, M., Bressac, M., Levy, I., Pham, M., (2016). Excess of 236U in the northwest Mediterranean Sea. Science of The Total Environment 565, 767–776. http://linkinghub.elsevier.com/retrieve/pii/S0048969716308452 Chamizo, E., L´ opez-Lora, M., (in press). Accelerator mass spectrometry of 236U with He stripping at the Centro Nacional de Aceleradores. Nuclear Instruments and Methods in 122 BIBLIOGRAPHY Physics Research Section B: Beam Interactions with Materials and Atoms. http://linkinghub.elsevier.com/retrieve/pii/S0168583X18302659 Christl, M., Vockenhuber, C., Kubik, P. W., Wacker, L., Lachner, J., Alfimov, V., Synal, H.- A., (2013). The ETH Zurich AMS facilities: Performance parameters and reference materials. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 294, 29–38. http://linkinghub.elsevier.com/retrieve/pii/S0168583X12001577 Cranston, F., White, D., (1971). Thermal neutron capture cross sections in calcium. Nuclear Physics A 169, 95–100. http://linkinghub.elsevier.com/retrieve/pii/0375947471905628 CSN, (2003). INSTRUCCI´ ON IS/05, de 26 de febrero de 2003, del Consejo de Seguridad Nuclear, por la que se definen los valores de exenci´ on para nucleidos seg´ un se establece en las tablas A y B del anexo I del Real Decreto 1836/1999. Bolet ´ ın Oficial del Estado 86, 14221–14229. https://www.boe.es/buscar/doc.php?id=BOE-A-2003-7520 CSN, (n.d.). General information - Jos´ e Cabrera. https://www.csn.es/en/central-nuclear/jose-cabrera Denk, E., Hillegonds, D. J., Vogel, J., Synal, A., Geppert, C., Wendt, K., Fattinger, K., Hennessy, C., Berglund, M., Hurrell, R. F., Walczyk, T., (2006). Labeling the human skeleton with 41Ca to assess changes in bone calcium metabolism. Analytical and Bioanalytical Chemistry 386, 1587–1602. http://link.springer.com/10.1007/s00216-006-0795-5 Denk, E., Hillegonds, D. J., Hurrell, R. F., Vogel, J., Fattinger, K., H¨ auselmann, H. J., Kraenzlin, M., Walczyk, T., (2007). Evaluation of 41Calcium as a New Approach to Assess Changes in Bone Metabolism: Effect of a Bisphosphonate Intervention in Postmenopausal Women With Low Bone Mass. Journal of Bone and Mineral Research 22, 1518–1525. http://doi.wiley.com/10.1359/jbmr.070617 Dillmann, I., Domingo-Pardo, C., Heil, M., K¨ appeler, F., Wallner, A., Forstner, O., Golser, R., Kutschera, W., Priller, A., Steier, P., Mengoni, A., Gallino, R., Paul, M., Vockenhuber, C., (2009). Determination of the stellar (n,γ) cross section of 40Ca with accelerator mass spectrometry. Physical Review C 79, 065805. https://link.aps.org/doi/10.1103/PhysRevC.79.065805 123 BIBLIOGRAPHY D¨ obeli, M., Kottler, C., Stocker, M., Weinmann, S., Synal, H.-A., Grajcar, M., Suter, M., (2004). Gas ionization chambers with silicon nitride windows for the detection and identification of low energy ions. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 219-220, 415–419. http://linkinghub.elsevier.com/retrieve/pii/S0168583X04001211 Dunai, T. J., (2010). Cosmogenic nuclides : principles, concepts and applications in the earth surface sciences. Cambridge University Press. Elmore, D., Phillips, F. M., (1987). Accelerator Mass Spectrometry for Measurement of Long- Lived Radioisotopes. Science 236, 543–550. http://www.sciencemag.org/cgi/doi/10.1126/science.236.4801.543 Elmore, D., Bhattacharyya, M. H., Sacco-Gibson, N., Peterson, D. P., (1990). Calcium-41 as a long-term biological tracer for bone resorption. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 52, 531–535. http://linkinghub.elsevier.com/retrieve/pii/0168583X90904716 ENRESA, (n.d.). Dismantling of the Jose Cabrera Nuclear Power Plant. http://www.enresa.es/eng/index/activities-and-projects/dismantlin g-and-environmental-restoration/dismantling-of-the-jose-cabrera-n uclear-power-plant Field, K., Remec, I., Pape, Y. L., (2015). Radiation effects in concrete for nuclear power plants –Part I: Quantification of radiation exposure and radiation effects. Nuclear Engineering and Design 282, 126–143. http://linkinghub.elsevier.com/retrieve/pii/S0029549314005603 Fifield, L. K., (2008). Accelerator mass spectrometry of the actinides. Quaternary Geochronology 3, 276–290. http://linkinghub.elsevier.com/retrieve/pii/S1871101407000581 Fifield, L., Tims, S., Fujioka, T., Hoo, W., Everett, S., (2010). Accelerator mass spectrometry with the 14UD accelerator at the Australian National University. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 268, 858–862. http://linkinghub.elsevier.com/retrieve/pii/S0168583X09010982 Fink, D., Paul, M., Hollos, G., Theis, S., Vogt, S., Stueck, R., Englert, P., Michel, R., (1987). Measurements of 41Ca spallation cross sections and 41Ca concentrations in the grant meteorite by accelerator mass spectrometry. Nuclear Instruments and Methods in Physics 124 BIBLIOGRAPHY Research Section B: Beam Interactions with Materials and Atoms 29, 275–280. http://linkinghub.elsevier.com/retrieve/pii/0168583X87902497 Fink, D., Middleton, R., Klein, J., Sharma, P., (1990).41Ca: Measurement by accelerator mass spectrometry and applications. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 47, 79–96. http://linkinghub.elsevier.com/retrieve/pii/0168583X9090049Z Fink, D., Klein, J., Middleton, R., Vogt, S., Herzog, G. F., (1991).41Ca in iron falls, Grant and Estherville: production rates and related exposure age calculations. Earth and Planetary Science Letters 107, 115–128. http://linkinghub.elsevier.com/retrieve/pii/0012821X9190048M Freeman, S., Serfass, R., King, J., Southon, J., Fang, Y., Woodhouse, L., Bench, G., McAninch, J., (1995). Biological sample preparation and 41Ca AMS measurement at LLNL. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 99, 557–561. http://linkinghub.elsevier.com/retrieve/pii/0168583X95003428 Freeman, S. P., King, J. C., Vieira, N. E., Woodhouse, L. R., Yergey, A. L., (1997). Human calcium metabolism including bone resorption measured with 41Ca tracer. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 123, 266–270. http://linkinghub.elsevier.com/retrieve/pii/S0168583X9600701X G´ omez-Guzm´ an, J. M., (2010).Medida de 129I en el sistema de Espectrometr ´ ıa de Masas con Aceleradores de 1 MV del Centro Nacional de Aceleradores. Aplicaci´ on a problemas medioambientales. Ph.D. thesis, Universidad de Sevilla. G´ omez-Guzm´ an, J. M., L´ opez-Guti´ errez, J. M., Pinto-G´ omez, A., Holm, E., (2012).129I measurements on the 1 MV AMS facility at the Centro Nacional de Aceleradores (CNA, Spain). Applied Radiation and Isotopes 70, 263–268. http://linkinghub.elsevier.com/retrieve/pii/S0969804311004040 G´ omez-Guzm´ an, J. M., Bishop, S., Faestermann, T., Famulok, N., Fimiani, L., Hain, K., Jahn, S., Korschinek, G., Ludwig, P., Rodrigues, D., (2015). Accretion rate of extraterrestrial 41Ca in Antarctic snow samples. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 361, 620–626. http://linkinghub.elsevier.com/retrieve/pii/S0168583X15004875 Gove, H. E., Fulton, B. R., Elmore, D., Litherland, A. E., Beukens, R. P., Purser, K. H., Naylor, H., (1979). Radioisotope Detection with Tandem Electrostatic Accelerators. IEEE 125