Separation and Analysis of Sr-90 and Zr-90 for Nuclear Forensic Applications

Mémoire

Ana Paula Zattoni

Maitrise en chimie Maître ès sciences (M.Sc.)

Québec, Canada

Résumé

Le présent travail porte sur le développement technologique pour déterminer l'âge des sources de radiostrontium à travers du rapport [Zr-90]/[Sr-90], en utilisant les techniques de spectrométrie de masse et scintillation liquide pour quantifier les deux isotopes. Parce que Sr-90 et Zr-90 sont des interférences isobariques en spectrométrie de masse, une séparation radiochimique est nécessaire pour isoler du Zr-90 avant son analyse. Parmi quatre résines commerciales, la résine DGA a fourni la meilleure performance pour isoler le Zr-90 du Sr-90. Des récupérations supérieures à 99% pour le Zr-90 ont été obtenues. La résine DGA était aussi l'approche la plus rapide et la plus efficace pour éliminer les interférences isobariques du Sr-90 et aussi de l’Y-90 potentiellement présents dans des échantillons contenant des niveaux élevés de radioactivité. Des expériences impliquant l’utilisation d’une cellule de collision pour éliminer des interférences isobariques ont fourni des facteurs de décontamination insuffisants pour des applications en criminalistique nucléaire.

iii

Abstract

In this work, a technological development to determine the age of radioactive strontium sources through the [Zr-90]/[Sr-90] ratio using mass spectrometry and liquid scintillation to quantify both isotopes is presented. Because Sr-90 and Zr-90 are isobaric interferences in mass spectrometry, a radiochemical separation to isolate Zr-90 has been shown to be mandatory prior to analysis. Four commercial resins (AG50W-X9, Dowex1-X8, Sr and DGA resins) were tested to isolate Zr-90 from Sr-90. Best performance was observed for the DGA resin, including recoveries higher than 99% for Zr-90. DGA has also demonstrated to be the faster approach and the most efficient not only to eliminate isobaric interferences from Sr- 90, but also from Y-90, potentially present in samples containing high levels of radioactivity. Experiments using a collision cell to eliminate isobaric interferences in a triple quadrupole mass spectrometer (ICP-QQQMS) have also been carried out, but results have demonstrated insufficient decontamination factors for nuclear forensic applications.

Table of Contents

RÉSUMÉ ...... III ABSTRACT ...... V TABLES LIST ...... IX PICTURES LIST ...... XI ABBREVIATIONS LIST ...... XIII ACKNOWLEDGMENTS ...... XIX INTRODUCTION ...... 1

1. RADIOSTRONTIUM ...... 5 1.1. Occurrence and radiological properties of strontium-90 ...... 5 1.2. Applications of strontium-90 ...... 8 1.3. Instability of strontium-90 and the origin of its radioactivity ...... 9 1.4. Hazardous effects of strontium-90 ...... 10

2. NUCLEAR THREATS OF SR-90 AND RADIOCHRONOMETRY FOR AGE-DATING APPLICATIONS ...... 13 2.1. Nuclear threats and risks involving orphaned sources ...... 13 2.2. Radiochronometry for nuclear forensic applications ...... 15

3. ANALYTICAL TECHNIQUES TO QUANTIFY SR-90 AND ZR-90 ...... 21 3.1. Principles of mass spectrometry ...... 21 3.1.1. Advantages and disadvantages of MS for the analysis of Zr-90 ...... 23 3.1.2. Triple quadrupole mass spectrometers to minimize isobaric interferences ...... 24 3.1.3. Separation of Sr-90 from Zr-90 using reaction cells ...... 26 3.2. Analysis of Sr-90 by liquid scintillation ...... 27

4. CHROMATOGRAPHIC TECHNIQUES TO SEPARATE SR-90 AND ZR-90 ...... 31 4.1. Principles of chromatography...... 31 4.2. Distribution ratio (D) ...... 33 4.3. Column performance and efficiency of separation ...... 34 4.4. Measurement of peak asymmetry ...... 36 4.5. Ion exchange chromatography (IEC) ...... 37 4.5.1. Ion exchange resins ...... 39 4.6. Extraction chromatography (EXC) ...... 40 4.6.1. Extraction process in EXC ...... 41 4.7. IEC and EXC for radiochemical separations and potential applications for Sr-90 and Zr-90 43

5. EXPERIMENTAL ...... 47 5.1. Chemicals ...... 47 5.2. Digestion of SrTiO3 ...... 47 5.3. Separation tests ...... 48 5.4. Omnifit® glass column preparation ...... 49 5.5. Methodology ...... 49

vii

5.6. Mass spectrometry analysis ...... 50 5.6.1. Performance of reaction cells to separate strontium from zirconium ...... 52 5.7. Analysis of Sr-90 by liquid scintillation ...... 54

6. RESULTS AND DISCUSSION ...... 55 6.1. Digestion of SrTiO3 ...... 55 6.2. Separation of Sr and Zr using a cation-exchange resin ...... 57 6.3. Resin shrinkage and issues for Zr recovery ...... 61 6.3.1. Effect of method downscaling on separation efficiency ...... 62 6.4. Separation of Sr and Zr using an anion-exchange resin ...... 64 6.5. IEC versus EXC for the separation of Sr and Zr ...... 66 6.6. Addition of HF in samples ...... 69 6.7. Summary of the efficiency of all resins tested ...... 72 6.8. Performance of DGA method for the recovery of trace levels of Zr ...... 73 6.9. Determining the age of a radiostrontium source ...... 74 6.10. Potential of reaction cell to separate strontium from zirconium ...... 76 CONCLUSIONS ...... 81 REFERENCES ...... 83 ANNEXE 1 ...... 87

viii

Tables List

Table 1.1 – Radiological properties of threatening radionuclides ...... P.7 Table 2.1 – Accidents involving RTGs reported by the IAEA ...... P.15 Table 2.2 – Radiological information from nuclear or radioactive materials .. P.16 Table 3.1 – Minimum resolution required to discriminate isobaric interferences at m/z 90 for the analysis of Zr-90 in MS ...... P.24 Table 3.2 – Typical chemical reactions in reaction cells ...... P.25 Table 3.3 – Theoretical binding properties of Zr and Sr with oxygen atoms .. P.27 Table 4.1 – Common commercial IEC resins ...... P.39 Table 4.2 – Common commercial EXC resins...... P.41 Table 4.3 – Distribution ratios (D) for strontium and zirconium in the AG50W-X8 resin ...... P.44 Table 5.1 – Instrumental setting for SrTiO3 digestion (Mars 5, Easy PrepTM vials) ...... P.48 Table 5.2 – Acquisition parameters for analysis of Sr and Zr by ICPQQQ- MS ...... P.51 Table 5.3 – Comparison of ionization energies between measured elements and internal standard ...... P.52 Table 5.4 – Acquisition parameters for the analysis of Sr and Zr using reaction cell and O2 as reaction gas ...... P.53 Table 5.5 – Acquisition parameters for the analysis of Sr-90 by liquid scintillation...... P.54 Table 6.1 – Acid mixtures used for SrTiO3 digestion tests ...... P.55 Table 6.2 – Digestion efficiency of SrTiO3 under different acidic conditions.. P.56 Table 6.3 – Performance of alternative eluents for Zr ...... P.62 Table 6.4 – Sample loading volumes according to the mass of dry resin used for separations ...... P.64 Table 6.5 – Recovery of Zr in DGA Resin according to HNO3/HF ratio in samples ...... P.72 Table 6.6 – Summary of resins performance to isolate Zr prior MS analyses ...... P.72

Pictures List

Figure 1.1 – Brief description of the origin of radioactivity in the environment ...... P.6 Figure 1.2 – Decay chain of strontium-90 ...... P.8 Figure 1.3 – Means of uptake and bioaccumulation for strontium-90 ...... P.11 Figure 2.1 – Number of nuclear and radioactive incidents reported by the IAEA for the last years ...... P.13 Figure 2.2 – Forensic Science...... P.16 Figure 2.3 – Decay process of Sr-90 as function of elapsed time ...... P.19 Figure 3.1 – Basic components of ordinary mass spectrometers ...... P.21 Figure 3.2 – Quadrupole mass spectrometer ...... P.23 Figure 3.3 – Triple quadrupole mass spectrometer mechanism...... P.25 Figure 3.4 – Mechanism of energy transfer and detection of beta particles by liquid scintillation ...... P.28 Figure 3.5 – Growth rate of Y-90 and secular equilibrium with Sr-90 ...... P.29 Figure 4.1 – Equilibrium in chromatographic separations ...... P.32 Figure 4.2 – In column chromatography technique ...... P.33 Figure 4.3 – Experimental variables to determine resolution in chromatography ...... P.35 Figure 4.4 – Parameters for the determination of peak asymmetry ...... P.37 Figure 4.5 – Separation of cations and anions by IEC ...... P.38 Figure 4.6 – Schema of extraction chromatography ...... P.40 Figure 4.7 – D values for strontium and zirconium in the Dowex 1-X10 resin ...... P.44 Figure 4.8 – Capacity factor for strontium and zirconium in the DGA resin ... P.46 Figure 5.1 – AF Omnifit® Column Design ...... P.49 Figure 5.2 – Method applied for separation tests ...... P.50 Figure 6.1 – Reproducibility of SrTiO3 digestion using HNO3/HF mixture ..... P.57 Figure 6.2 – Elution profile of Sr and Zr in 4M HCl (10 g AG50W-X8, 100- 200 mesh) ...... P.58 Figure 6.3 – Elution profile of Sr and Zr in 3M HCl (10 g AG50W-X8, 100- 200 mesh) ...... P.59 Figure 6.4 – Separation of Sr and Zr using a 2M to 6M HCl gradient (10 g AG50W-X8, 100-200 mesh) ...... P.59 Figure 6.5 – Elution curves of Sr as function of HCl molarity (10 g AG50W-X8, 100-200 mesh) ...... P.60 Figure 6.6 – Elution curves of Sr at 2M HNO3 and 2M HCl (10 g AG50W- X8, 100-200 mesh) ...... P.61 Figure 6.7 – Separation of Sr and Zr using a 2M HNO3 to 6M HCl gradient in 2 g AG50W-X8 (100-200 mesh) ...... P.63 Figure 6.8 – Volume of eluent for Sr elution as function of mass of AG50W-X8 ...... P.63 Figure 6.9 – Separation efficiency for Sr and Zr using Dowex1-X8 resin ...... P.65

Figure 6.10 – Zirconium retention in Dowex1-X8 as function of HCl concentration ...... P.65 Figure 6.11 – Maximum recovery of Zr according to HCl concentration in Dowex1-X8 ...... P.66 Figure 6.12 – Comparative of separation of Sr and Zr using ion exchange and extraction resins (a. AG50W-X8, b. DOWEX1-X8, c. Sr-Resin, d. DGA-Resin) ...... P.67 Figure 6.13 – Proposed extraction mechanism for Sr for its separation from Zr by EXC ...... P.68 Figure 6.14 – Tailing effect as a function of Sr concentration (AG50W-X8) . P.69 Figure 6.15 – Separation of Sr and Zr using Dowex1-X8 for samples containing HF ...... P.70 Figure 6.16 – Separation of Sr and Zr using DGA for samples containing HF (a. 0.01%, b. 0.2%) ...... P.71 Figure 6.17 – Complete methodology to separate Sr and Zr using DGA resin ...... P.73 Figure 6.18 – Comparative between experimental and expected results for the recovery of trace levels of Zr using DGA resin ...... P.74 Figure 6.19 – Procedure for determining the age of a radiostrontium source ...... P.75 Figure 6.20 – Comparative between theoretical and experimental concentrations for the analysis of Sr-90 by liquid scintillation ...... P.76 Figure 6.21 – Zr and Sr oxides formation in mass spectrometry as function of O2 concentration in the reaction cell ...... P.77 Figure 6.22 – Predominant species of Zr (a) and Sr (b) at 6% O2 in the reaction cell ...... P.78 Figure 6.23 – Correlation between results for the analysis of Zr at m/z 90 and m/z 106 ...... P.79 Figure 6.24 – Predominant species of Y at 6% O2 in the reaction cell ...... P.79

xii

Abbreviations List

D – Alpha particles D – Separation factor E – Beta particles ߣ – Decay constant a – asymmetry portion of a peak ܣ – Final activity ܣ଴ – Initial activity ܣௌ – Peak asymmetry A+ – Charged analyte AG50W-X8 – Cationic resin AMS – Accelerator Mass Spectrometry b – Back portion of a peak ܥ – Final concentration ܥ଴ – Initial concentration ܥ௘ – Concentration of a solute in the extractant phase ܥ௜,ெ – Concentration in the mobile phase ܥ௜,ௌ – Concentration in the stationary phase cpm – Count per minute cps – Count per second D – Distribution ratio DGA – Diglycolamide resin Dowex1-X8 – Anionic resin ܧ – Extractant EPA – Environmental Energy Agency EXC – Extraction chromatography F – Force F- – Fluoride G – Gas G+ – Charged gas ܪ – Height of the theoretical plate H+ – Proton H2C2O4 – Oxalic acid H2O – Water H2O2 – Hydrogen peroxide H2SO4 – Sulphuric acid HCl – Hydrochloric acid HEU – High-enriched uranium HF – Hydrofluoric acid HNO3 – Nitric acid ǻHR – Enthalpy ݅ – Given compound I – Interference i.d – Internal diameter

xiii

I+ – Charged interference IAEA – Internation Atomic Energy Agency ICP-MS – Inductively Coupled Plasma Mass Spectrometry ID – Identification IEC – Ion exchange chromatography IUPAC – International Union of Pure and Applied Chemistry K – Distribution coefficient ݇Ԣ – Capacity factor ݇௘ – Coulomb’s constant ܮ – Length ܮ – Ligand LL – Lower limit LOD – Detection limit LOQ – Quantification limit m – Mass ο݉ – Mass difference M – Molar m/z – Mass-to-charged ratio M+ – Charged metal ܯܯ – Molar mass MS – Mass spectrometry M: – Megaohm ܰ – Number of theoretical plates N/Z – Neutron-to-proton ratio N2O4 - Nitrogen tetroxide - NO3 – Nitrate ܰ஺ – Avogadro’s number O – Atomic oxygen O2 – Molecular oxygen Pb – Lead ppt – part per trillion Psi – lbf/square inch Pu – Plutonium ݍ – Charge Q – Quadrupole ܴௌ – Resolution ݎ – Distance between two charges R2 – Correlation factor RDDs – Radiological dispersion devices RF – Radio-frequency RTGs – Radiothermal generators s – Standard deviation S – Stationary phase SI – International system Sr – Strontium SrO+ – Charged strontium oxide

xiv

SrTiO3 – Strontium Titanate ݐ – Age of a radioactive source t½ – Half-life Ti – Titanium ݐ௠ – Dead time ݐԢ௜ – Adjusted retention time TIMS – Thermal Ionization Mass Spectrometry ݐ௜ – Retention time u – Mass unit U – Uranium U.S – United States UP – Upper limit ܸெ – Volume of the mobile phases ܸௌ – Volume of the stationary phases v/v – Volume-to-volume ratio y – Years Y – Yttrium ݓ௜, – Width at the base of a peak Zr – Zirconium Zr+ – Charged zirconium ZrO+ – Charged zirconium oxide

xvi

“Science can only be created by those who are thoroughly imbued with the aspiration toward truth and understanding ”

x Albert Einstein

xvii

xviii

Acknowledgments

My completion of this project would not have been possible without the kind support of my director Dominic Larvière. So, I would like to thank him to be always open to discuss and share ideas while guiding me to successfully achieve the goals of this project.

I also would like to thank Serge Groleau for all the support in the laboratory, my office partners Annie Michaud, Pablo Lebed, and Marie-Ève Lecavalier for their pleasant company during all the time we spent together. Charles Labrecque, Kenny Nadeau, Jean-Michel Benoit, Solange Schneider, Laurence Whitty-Léveillé, Sabrina Potvin, Julien Légaré Lavergne, Justyna Florek, and Maela Choimet who I had the opportunity to work with and, in some cases, the opportunity to struck up a close friendship.

Likewise, I would like to thank Health Canada, the Research and Technology Initiative, and Agilent to make this project possible. And finally, I would like to thank Sherrod Maxwell for the interest in this project as well as for the suggestions and the encouragement, which have been all important to the accomplishment of this work.

xix

Introduction

Incidents involving illicit trafficking and smuggling of nuclear and radioactive material have been object of concern since the early 90s, when the first cases involving unauthorized activities started being reported in Switzerland and Italy, then years later in Germany, Czech Republic, and Hungary [1]. Today, more than 2,400 cases have been already confirmed since 1995, and 155 cases have been reported for the period between July 2012 and June 2013 [2].

Before the 90s, the main concern for nuclear security was to protect only high- enriched uranium (HEU) and plutonium in nuclear facilities. However, the increasing number of cases implicating illegal possession, theft, or loss involving other radioactive sources since 1995, forced authorities to establish a new concept of nuclear security, while triggering efforts towards eliminating nuclear and radioactive threats.

Such new concept became synonym of both protection and control over not only nuclear but also any kind of radioactive material that could give rise to malicious actions, including unpredictable terrorist activities and utilization of radiological weapons known as dirty bombs.

Contrary to nuclear bombs, dirty bombs are relatively easier to fabricate and are mainly characterized by their dispersive effect. The purpose of dirty bombs is not to destroy but contaminate, while spreading a radioactive material through the utilization of conventional explosives.

Usually, radionuclides that have long half-lives or a high specific activity are potentially more interesting for the production of radiological dispersion devices (RDDs). Ranked in a short list of these radionuclides, strontium-90 has a half-life of about 29 years and a specific activity of about 518 X 1010 Bq/g. This corresponds, for example, to a specific activity of about 1.5 times higher than for cesium-137, which has an equivalent half-life (i.e. 30 years) [3].

Sources of strontium-90 can be found in laboratories of research or hospitals for the production of yttrium-90 and cancer treatment as well as in wastes of nuclear facilities. The main concern, however, is associated to sources of strontium-90 found in orphaned radiothermal generators (RTGs) widely used in the 50s to provide energy in areas of difficult accessibility. It is estimated that hundreds of orphaned RTGs containing high levels of activity are still lost around the world. Actually, the lower degree of security surrounding these sources is assumed to be appealing for nuclear terrorists.

Following an alleged terrorist attack, where the presence of a nuclear or a radioactive source is detected, a nuclear forensic investigation takes place. Working with other forensic sciences, nuclear forensics aims not only to answers questions about the radiological hazard but also provide complementary radioisotopic information to determine the origin of a seized source. Isotopic composition, for example, could provide information about the fabrication date or last purification (i.e. age of the radioactive source) and, in conjunction with other chemical and physical data, provide clues about the facility responsible for its production.

In practice, the age of a radioactive source could be determined using principles of

the decay law. Actually, Sr-90 is an unstable radioisotope that undergoes beta decay to form Y-90 which in turn decays into Zr-90, a stable nucleus. Thus, the age of a source containing strontium-90, for example, would be a function of the [Zr- 90]/[Sr-90] ratio, where concentrations of both isotopes could be determined, respectively, by liquid scintillation and mass spectrometry, as demonstrated over the present work. To be successfully used to date nuclear materials, however, this approach requires an efficient method for radionuclide separation to isolate Sr-90 from Zr-90 from the radioactive source as well as a sensitive method of analysis to provide accurate results while reducing age uncertainty.

Mass spectrometry has been widely used for analytical purposes because of its sensitivity, accuracy, and possibility to discriminate isotopic species. The major inconvenient of this technique is the isobaric interferences caused by ionic atoms or molecules having the same m/z, as for strontium-90 and zirconium-90. Such interferences cause peak overlap and an overestimation of compounds of interest. Sometimes, even high-resolution devices are not sufficient to overcome this problem and pre-treatments (e.g. separation) are often mandatory prior to analysis.

Besides liquid-liquid extraction and precipitation techniques, ion exchange (IEC) and extraction chromatography (EXC) have gained extensive attention in the past years, especially because of their potential to be used in radiochemical separations. Previous works have demonstrated, for example, their efficiency for removing and determining trace amounts of Sr-90 in environmental, food, and seawater samples [4-7].

In terms of age-dating applications, Charbonneau et al. have recently reported results for the separation of Co-60 from Ni-60 using both anionic and extraction

resins [3]. Likewise, Steeb et al. have presented a method to separate Sr-90 from Zr-90 using the Sr Resin [8]. For this last procedure, however, no information has been found regarding the possibility of using cationic, anionic, or DGA resins. Actually, different distribution coefficients available in the literature for both elements suggest that high levels of selectivity could also be achieved using those resins [9-14].

In this context, this work aims to compare the performance of the AG50W-X8, Dowex1-X8, DGA, and Sr Resin and, eventually, propose one or more alternatives to separate Sr-90 and Zr-90 for nuclear forensic applications. To make a good comparison, experimental conditions like mass of resin and volume of eluents have been kept constant to assess recovery and resolution of peaks after chromatographic separations. Assuming that real samples could contain high levels of radioactivity, significant amounts of yttrium-90 could cause isobaric interferences that should not be neglected. For that reason, the possibility to completely isolate Y-90 has also been considered to evaluate the efficiency of resins.

Chapter 1 1. Radiostrontium

As previously mentioned, potential interest in radiostrontium for nuclear threats is a consequence of peculiar radiological characteristics of Sr-90. Thus, this chapter aims to detail these characteristics, while explaining the radiological risks associated to Sr-90 and its hazardous effects for humans and for the environment.

1.1. Occurrence and radiological properties of strontium-90

Radioactive sources can exist in the environment naturally (i.e. primordial and cosmogenic radionuclides) or via accidental or deliberate anthropogenic activities (Figure 1.1). According to astrophysics theories, primordial radionuclides have been produced in the course of nucleosynthesis and have been presented on Earth from the beginning.

Cosmogenic radionuclides, on the other hand, are continuously produced by the interaction of cosmic irradiation with gases in the atmosphere (e.g. N2, O2, Ar, etc.), and brought to the earth by rainwater. In general, both primordial and cosmogenic radionuclides contribute to the harmless levels of radioactivity in the environment.

The occurrence of worrisome levels of radioactivity, however, is a consequence of the release of significant amounts of radioisotopes through nuclear tests or nuclear accidents. It has been reported, for example, that about 8,000 TBq of Sr-90 have been released around the Chernobyl area in 1986 causing damage that, even almost 30 years later, still holds the attention of numerous scientists [15,16].

Figure 1.1 – Brief description of the origin of radioactivity in the environment

As presented in Figure 1.1, the origin of radionuclides in the environment is multifaceted. In the case of Sr-90, it has an anthropogenic origin. Actually, Sr-90 is a by-product of the fission of uranium and plutonium, continuously produced in nuclear power plants. According to the U.S Environmental Energy Agency (EPA), strontium-90 is considered one of the more hazardous constituents of nuclear wastes [17].

As for any other isotopes of strontium, Sr-90 can form other chemical compounds

(e.g. halides, oxides, sulphides) and its dispersion through the environment would be strongly influenced by the chemical form and solubility.

In terms of radiological properties (Table 1.1), Sr-90 shows a specific activity of about 140 Ci/g. Comparatively to other threatening radioisotopes, it accumulates more reactivity per unit of mass than Ra-226, Am-241, Pu-238, and Cs-137. Also, Sr-90 has a half-life (i.e. time that takes for the radioactivity to decay to one-half of its original value) of about 29 years, which is longer than the half-life of Cf-252, Co- 60, Po-210, and Ir-192.

Table 1.1 – Radiological properties of threatening radionuclides [3] Specific Half-life Radionuclide Activity Decay mode (Ci/g) (y) Ra-226 1 1600 D Am-241 3.5 430 D Pu-238 17 88 D Cs-137 88 30 E Sr-90 140 29 E Cf-252 540 2,6 D Co-60 1100 5.271 E Po-210 4500 0.4 (140d) D Ir-192 9200 0.2 (74d) E

As presented in Table 1.1, the specific activity is inversely proportional to its half- life, which means that the higher is the specific activity, the shorter is the half-life. In practice, short-lived isotopes are less harmful to the environment than long-lived isotopes as they decay away faster and completely. However, short-lived isotopes can be fatal, once humans have been directly exposed to the high-energy emitted. For Sr-90, which is considered a long-lived isotope, long-term damage is expected due to its slower decay rate that will take years. For those radionuclides, human

exposure to ionizing radiation occurs over an extended period of time due to the fact that lower but more persistent quantities of radioactivity will remain in the environment [18,19].

1.2. Applications of strontium-90

Currently, controlled amounts of strontium-90 have been extensively used in medicine as radioactive tracers. As illustrated in the Figure 1.2, Sr-90 is a neutron- rich nucleus that, through a decay process, forms yttrium-90, an intermediate decay product that is often used for cancer treatment.

Figure 1.2 – Decay chain of strontium-90

Due to its capacity to produce heat, Sr-90 in the form of strontium titanate (SrTiO3), has also been widely used in the past for the production of portable power supplies. Known as radioisotope thermoelectric generators (RTGs), these devices have been manufactured to provide energy in remote sites where electricity was quite limited (i.e. navigational beacons, weather stations, and space vehicles).

1.3. Instability of strontium-90 and the origin of its radioactivity

In general, two factors including nucleus mass and neutron-to-proton ratio (N/Z) contribute to nucleus instability and, in practice, to influence the mode of radiation emitted.

It is normally observed, for example, that heavier nuclei (i.e. usually heavier than Pb) are more likely to emit alpha particles (D), while lighter nuclei tend to achieve stability through the emission of positive beta particles or positrons (E+) to compensate repulsive forces caused by an excess of protons. Also, it is noticed that when the number of neutrons becomes more important than the number of protons (i.e. increase in the N/Z ratio), it is the emission of negative beta particles (E-) or electrons that are rather detected.

As already mentioned, strontium-90 is an unstable neutron-rich nucleus and for that reason it undergoes E- decay, which is generally represented as follows:

࡭ ࡭ ି ࢆࢄ ืࢆା૚ ࢄ + ࢼ ଽ଴ ଽ଴ ି ଽ଴ ି ଷ଼ܵݎ ืଷଽ ܻ + ߚ ืସ଴ ܼݎ + ߚ

Thus, to achieve stability, Sr-90 liberates the excess of neutrons in the form of protons and very energetic E- particles. Each rearrangement per second corresponds to the activity of the radioactive source in Becquerel (Bq) according to SI.

As indicated above, the proton gives rise to a decay product, in this case, Y-90, an intermediate decay product. Yttrium-90 is also an unstable nucleus and, as for Sr- 90, it also undergoes E- decay to form Zr-90, which this time is a stable non- radioactive isotope.

1.4. Hazardous effects of strontium-90

Major radiological risks and hazardous effects of strontium-90 sources are associated to the energetic contributions of Sr-90 and Y-90 beta particles. As beta- emitters, Sr-90 and Y-90 penetrates the skin, while interacting with cells and discharging their energy that are, respectively, 546 keV and 2280 keV [20].

In practice, strontium-90 absorption in humans can result from direct exposure to radiation, inhalation of fine particles in air or, as in most situations, from the consumption of both contaminated food and water (Figure 1.3).

Figure 1.3 – Means of uptake and bioaccumulation for strontium-90

Chemically, strontium-90 demonstrates analogue properties with calcium and, once in the organism, it tends to be incorporated in bones and teeth increasing risks of cancer. Actually, a major portion of absorbed strontium-90 is excreted during the first year after exposure with a biological half-life (i.e. the time an organism takes to eliminate one half the amount of a compound or chemical on a strictly biological basis) of 40 days. However, there is about 10% of Sr-90 that is tightly bound to the bones and with a biological half-life of 50 years it is slowly excreted from human’s body [21].

Chapter 2

2. Nuclear Threats of Sr-90 and Radiochronometry for Age-Dating Applications

The significant number of incidents involving nuclear and radioactive material has forced authorities not only to increase the control over those materials but also motivated nuclear forensic experts to develop techniques able to provide important radiochemical information for criminal investigations. In this context, this chapter aims to present the terrorist potential involving orphaned sources, including those of Sr-90, as well as to explain the role of nuclear forensics and how radiochronometry could help to determine the origin of a seized source eventually used in nuclear attacks.

2.1. Nuclear threats and risks involving orphaned sources

Despite international’s effort to monitor and regulate the utilization of nuclear and radioactive materials, the number of incidents and illicit trafficking involving them is still significant (Figure 2.1).

300 243 250 215 222

200 171 172 163 155 150

100

0 2007 2008 2009 2010 2011 2012 2013

Figure 2.1 – Number of nuclear and radioactive incidents reported by the IAEA for the last 7 years

In total, the International Atomic Energy Agency (IAEA) has already reported 2407 incidents from 1995 to 2013, including cases of illegal possession or attempts to sell nuclear or radioactive material, theft or loss, and unauthorized activities apparently without criminal nature [2]. Actually, millions of radioactive sources are available worldwide and inadequate control over usage, storage, and production in different countries seems to contribute to the number of incidents.

One of the biggest issues is probably associated to orphaned sources, that means sources that were abandoned, lost, or misplaced in the past without authorization and, today, are outside of regulatory control. Thousands of radiothermal generators (RTGs) like those using Sr-90 (Chapter 1), for example, have been discovered in the Russia coast containing extremely high levels of radioactivity. Unfortunately, there are about nearly a hundred pieces that have not been yet recovered and remain unprotected against unauthorized interference [22].

In practice, only a few numbers of accidents involving RTGs have been reported (Table 2.1) [23], but authorities do not rule out the risks of nuclear threats resulting from the lower degree of security surrounding these sources. Main concerns started arising after the United States discovered documents in Afghanistan with real intentions of Al Qaeda in developing radiological dispersion devices (RDDs), vulgarly known as dirty bombs [24].

As previously described, dirty bombs consist of conventional explosives combined with a radioactive material. Once detonated, the radioactive material is dispersed, while contaminating the environment, killing, injuring, and exposing people directly to radiation. The degree of damages would depend on many factors like physical and chemical form of the radioactive material, size of explosives, and proximity of

people to the explosion.

Table 2.1 – Accidents involving RTGs reported by the IAEA Year Case 1999 A stolen radioactive heat source was found emitting radioactivity at a bus stop in Kingisepp, in Russia. The source was then recovered.

2001 Three radioisotope heat sources were stolen from lighthouses located in the Kandalaksha Bay area, in Russia. After being found, the sources were sent to Moscow.

2001 Three woodsmen have been diagnosed with radiation sickness after finding two unshielded radioactive heat sources near the Inguri River valley, in Georgia. Two victims have experienced nausea, vomiting, and dizziness after hours of exposure to sources of Sr-90 containing about 30,000 Ci. They were treated for many months before recovering from severe radiation burns. The sources were recovered in 2002.

2002 Three shepherds were exposed to high radiation doses after they stumbled upon a number of RTGs in the Tsalenjikha region. Eight generators were recovered.

2003 An RTG was found 200 meters in the shoals of the Baltic Sea, which was recovered later by a team of experts.

2003 The theft of metals from an RTG has been discovered in the White Sea region, in Russia. The six radioactive sources have not been taken.

2.2. Radiochronometry for nuclear forensic applications

Nuclear forensics is the science responsible for providing radiological properties of radioactive sources that could be complementary to other biological, digital, and chemical properties used in criminal investigations.

Assuming, for example, that a terrorist attack takes place and the presence of a nuclear or a radioactive material is confirmed, nuclear forensic experts are put in

charge to work in conjunction with other forensic sciences to identify the alleged responsible (Figure 2.2).

Figure 2.2 – Forensic Science

In general, radiological information of nuclear or radioactive material includes the appearance, structure, and isotopic composition. As indicated in Table 2.2, an important parameter is the age, which can provide valuable information about the date of fabrication or the last purification [25].

Table 2.2 – Radiological information from nuclear or radioactive materials Parameter Information Appearance Material type (powder, pellet) Dimensions Reactor type U, Pu content Chemical composition Isotopic composition Enrichment (reactor type) Impurities Production process, geolocation 18O/16O ratio Geolocation Surface roughness Production plant Microstructure Production process Age Date of production or last purification

To determine the age, radiochronometry is a technique often used in fields such archaeology, anthropology, and geology to date samples like human bones, corals, and other artefacts preserved even over a billions of years. This technique has also been widely used in environmental research for tracing climate changes [26] and recently started receiving increasingly attention in nuclear forensics.

The principle of the radiochronometry technique is based on the fact that activity of a radionuclide decays exponentially with time. According to the decay law, the activity of a radioactive source (ܣ) is a function of three variables: the initial activity

of the radioisotope (ܣ଴), its decay constant (ߣ) also represented by ln2/t1/2 ratio, and the elapsed time or also called the age (ݐ) in nuclear forensic applications (Equation 2.1).

ିࣅ࢚ ࡭ = ࡭૙ࢋ (2.1)

Here, the age (ݐ) can be isolated in the equation and be expressed in terms of both final and initial activities (Equation 2.2):

૚ ࡭ ࢚ = െ ܔܖ ቀ ቁ (2.2) ࣅ ࡭૙

When the radioactive source is unknown, however, it is impossible to know its original activity (Charbonneau, 2012) and a change of variables in the equation becomes necessary. In this case, respectively activities can be converted in units of concentration (ܥ) as follows (equation 2.3):

࡯×ࡺ ×࢒࢔૛ ࡭ = ࡭ (2.3) ࡹࡹ×࢚૚/૛

.is the molar mass, and ݐଵ/ଶ is half-life ܯܯ ,Where, ܰ஺ is the Avogadro’s number

So, Equation 2.2 can be expressed in terms of final ( ܥ ) and initial ( ܥ଴ ) concentrations of the radioactive species (Equation 2.4).

૚ ࡯ ࢚ = െ ܔܖ ቀ ቁ (2.4) ࣅ ࡯૙

As in Figure 2.3, as the time passes, the radioactive species (in this case Sr-90) tends to decay at the same time a more stable decay product (Zr-90) is build up.

Figure 2.3 – Decay process of Sr-90 as function of elapsed time

For that reason, equation 2.4 could also be expressed as:

૚ [࢙࢚ࢇ࢈࢒ࢋ ࢏࢙࢕࢚࢕࢖ࢋ] ࢚ = െ ܔܖ ቀ ቁ (2.5) ࣅ [࢘ࢇࢊ࢏࢕ࢇࢉ࢚࢏࢜ࢋ ࢏࢙࢕࢚࢕࢖ࢋ]

Or, in the case of Sr-90, as:

૚ [ࢆ࢘ିૢ૙] ࢚ = െ ܔܖ ቀ ቁ (2.6) ࣅ [ࡿ࢘ିૢ૙]

Briefly, the decay process as that illustrated for Sr-90 could serve as a chronometer, where the age of an unknown source could be estimated by the determination of respective concentrations of both the radioactive and the stable .isotopes at a given time ݐ

Chapter 3 3. Analytical Techniques to Quantify Sr-90 and Zr-90

In order to achieve maximum accuracy and precision for age-dating purposes, the analysis of high levels of radioactivity from Sr-90 and trace levels of Zr-90 could be performed using, respectively, liquid scintillation and mass spectrometry techniques. In this chapter, principles, advantages, and/or limitations of these two techniques have been discussed.

3.1. Principles of mass spectrometry

Mass spectrometry is a multi-element technique widely used for obtaining quantitative or qualitative information about a sample containing inorganic or organic material. This technique covers nearly all the elements that are discriminated by their difference in the mass-to-charge ratio (m/z).

Basically, all mass spectrometers are composed of an inlet and an ionization system, a mass analyzer, and a detector (Figure 3.1).

Figure 3.1 – Basic components of ordinary mass spectrometers

Depending on its nature, inlet systems can accommodate samples under solid, liquid, or gas state. In typical setups, liquid samples (usually more homogeneous)

pass through a nebulizer to transform the sample into an aerosol that is driven towards the ion source.

Established mass spectrometry techniques such as Accelerator Mass Spectrometry (AMS), Thermal Ionization Mass Spectrometry (TIMS), and Inductively Coupled Plasma Mass Spectrometry (ICP-MS) have been proved to reduce the amount of sample necessary for the analysis of inorganic compounds. It has been reported that ICP-MS has become a dominating technique especially for

the determination of long-lived radionuclides (t1/2 > 10 years) present at trace levels in different samples (e.g. water, soils, biological, and medical samples) [27,28].

In ICP-MS, for example, inorganic materials are positively ionized the ion source. Usually, the ionization takes place in an inert atmosphere using argon, under a set radio frequency, and a plasma temperature of up to 8,000 K.

Argon is commonly used because its first ionization potential (15.8 eV) is higher than the first ionization potential of almost all other elements (except fluorine, neon, and helium), which ensure the maximum ionization of the elements of interest during analyses. Since the energy required for the second ionization is usually too high for most part of elements, second ionization is less likely to happen.

After elements have been converted into ions, they are sent to the mass analyser that acts as a mass filter to separate different masses. Common mass analysers are called quadrupoles, which are consisted of four rods operating in an oscillating electrical field capable of guiding ions towards the detector (Figure 3.2).

Figure 3.2 – Quadrupole mass spectrometer [29]

Once the ions reach the detector, they are measured as a current and, then, converted into a series of peaks that form the mass spectrum.

3.1.1. Advantages and disadvantages of MS for the analysis of Zr-90

Mass spectrometry offers some advantages such as short analysis time, low sample consumption, high sensitivity, reduced background interference due to the possibility of using efficient mass analyzers as filters [30], and the ability of discriminating different isotopes.

The major disadvantage, however, consists of isobaric interferences caused by ions or molecules having the same mass-to-charge (m/z) ratio. Actually, maximum resolution provided by typical quadrupoles (less than 5,000) can be not sufficient to avoid peak overlapping and overestimation of the compound of interest.

For example, isobaric interferences for the analysis of Zr-90 could be caused by the presence of its parent isotopes like Sr-90 and Y-90 [31,32]. In this case, even high-resolution instruments that provide resolutions of about 15,000 would not be sufficient to discriminate between the peaks of those three isotopes (Table 3.1), suggesting the use of a pre-treatment strategy, usually a chromatographic separation prior to analysis to chemically separate them.

Table 3.1 – Minimum resolution required to discriminate isobaric interferences at m/z 90 for the analysis of Zr-90 in MS Atomic mass Resolution required Ions ࢓ (u) ቀࡾ = ቁ ࡿ ο࢓ Sr-90 89.908 29,668 Y-90 89.907 36,772 Zr-90 89.905 - ݉ = Mass to be analyzed (In this case, 90 for Zr-90) ο݉ = Difference between two atomic masses

3.1.2. Triple quadrupole mass spectrometers to minimize isobaric interferences

Recently designed, a triple quadruple is a tandem mass spectrometer consisted of two quadrupoles (Q1 and Q3) placed at the two extremities of a reaction cell (Q2). The first quadrupole is normally used as a filter to reduce the number of species

entering in Q2 containing a reaction gas such as He, H2, O2 etc. At the end, products from the reaction cell are driven to Q3 that serves to eliminate remaining interferences and guide only the isotope or compound(s) of interest towards the detector (Figure 3.3).

Figure 3.3 – Triple quadrupole mass spectrometer mechanism

As for any quadrupole, a reaction cell operates in a radio-frequency (RF) mode. However, RF is usually adjusted to focus ions and favour either a collision or a chemical reaction with the reaction gas.

Using nonreactive gases, collisions are favoured and the species are discriminated by the difference in their kinetic energy [33]. Using highly reactive gases on the other hand, different reactions can take place to produce different polyatomic species that are discriminated according to their different masses (Table 3.2) [34].

Table 3.2 – Typical chemical reactions in reaction cells Mechanism General Form(s) Advantage Charge ܫା + ܩ ՜ ܩା + ܫ x Formation of uncharged interferences exchange that are not detected.

Proton (a) ܫܪା + ܩ ՜ ܩܪା + ܫ x Formation of uncharged interferences transfer that are not detected (a).

x Formation of charged interferences (b) ܫା + ܩܪ ՜ ܫܪା + ܩ that are heavier than the analyte (b).

Adduct ܣା + ܩ ՜ ܣܩା x Formation of analytes heavier than Formation the interferences. ܫ = Interference, ܩ = Reaction gas, ܣ = Analyte

In general, interferences can be converted into non-detectable species or in species of different masses that are shifted from the region of interest. When the interaction with the reaction gas is stronger with the analyte, this last can be converted in a heavier compound to be shifted to a less overladen region of the spectrum.

Briefly, reaction cells technology had been developed not only to improve the performance of mass spectrometers, but also to reduce background or eliminate isobaric interferences impossible to be removed using ordinary instruments or high-resolution devices.

3.1.3. Separation of Sr-90 from Zr-90 using reaction cells

Successful applications using O2 as a reaction gas for solving problems of isobaric interferences between Zr-90 and Sr-90 have already been reported [31,35-40]. Actually, the formation of zirconium oxide is more likely to happen than the formation of strontium oxide, which makes possible to perform the analysis of Zr at m/z = 106 (ZrO+) rather than m/z = 90 (Zr+).

According to Eiden et al., Zr seems to react at least 200 times faster with oxygen

than Sr through addition of O2 into the reaction cell. Theoretical positive enthalpy

YDOXHV ǻHR) for strontium oxides formation suggest that this reaction is less favourable than that for zirconium oxides6. Also, the covalent bond for ZrO+ seems to be stronger than that for SrO+ due to the differences in both electronic densities and metal–oxygen bond lengths (Table 3.3.).

Table 3.3 – Theoretical binding properties of Zr and Sr with oxygen atoms [32] ǻHR (M+ + O o MO+) Bond length Electronic Ion 2 (Å) Density* Experimental Calculated

ZrO+ -249 -186 1.74

SrO+ +199 +215 2.35

*Zr and Sr atoms on top; O atom at the bottom

3.2. Analysis of Sr-90 by liquid scintillation

Even if it is possible to analyse radioactive isotopes by mass spectrometry, conventional counting techniques, such as liquid scintillation, are still recommended for radioisotopes like Sr-90 that have high specific activities.

In typical liquid scintillation methods, the radioactive sample is mixed with a cocktail containing a solvent and a soluble scintillator. In the cocktail, the solvent occupies from 60-99% of the total solution while the scintillator, only 0.3-1% [41]. For this reason, beta particles usually transfer their energy first to solvent molecules. The energy passes between solvent molecules until it reaches the scintillator. Once excited, the scintillator releases the absorbed energy in the form of photons that can be easily detected (Figure 3.4).

Figure 3.4 – Mechanism of energy transfer and detection of beta particles by liquid scintillation [41]

In some cases, however, beta particles are enough energetic (i.e. energy higher than 0.6 MeV) to cause disturbance of adjacent molecules in matter followed by a photon emission that can be detected without the need to introduce a scintillator in the sample. This phenomenon known as Cerenkov effect occurs when a charged particle travels at constant velocity in a medium characterized by its index of refraction markedly larger than 1 at a speed exceeding that of the light in that medium. Actually, in gaseous, liquid, or solid media, the velocity of light will be less than its velocity in a vacuum, and the beta particle will be able to travel in such media at speeds exceeding that of light [42].

In practice, Sr-90 is not enough energetic to produce Cerenkov radiation. In this case, Sr-90 is detected indirectly through the signal emitted by Y-90 eventually present at secular equilibrium in the sample, which means, with the same activity of Sr-90.

As indicated in Figure 3.5 [42], secular equilibrium between Sr-90 and Y-90 is achieved after approximately 20 days. Once activities have been determined, they can be converted in units of mass or concentration using equation 2.3 presented in chapter 2.

Figure 3.5 – Growth rate of Y-90 and secular equilibrium with Sr-90

Chapter 4 4. Chromatographic Techniques to Separate Sr-90 and Zr- 90

Due to the importance to separate Sr-90 from Zr-90 to avoid isobaric interferences in mass spectrometry, this chapter presents the potential of both ion exchange and extraction chromatography techniques to be used in the separation of those isotopes. Principles of chromatography, theoretical definitions of distribution coefficients, retention factors, distribution ratios, resolution, and the number of theoretical plates to assess the performance of separation have also been addressed.

4.1. Principles of chromatography

Chromatography is the term used to designate a set of techniques implicating a mobile phase (i.e. gas or liquid) and a stationary phase (i.e. solid and/or liquid) for the separation of mixtures. In practice, the mobile phase carries the sample through the stationary phase that interacts with species in the sample.

Thermodynamically, chromatographic separations consist in an equilibrium process (Figure 4.1) where the number of ions of a given species (݅) in either the mobile phase (ܯ) and in the stationary phase (ܵ) is given by the distribution coefficient (ܭ) (Equation 4.1).

஼ ܭ = ೔,ೄ (4.1) ஼೔,ಾ

Figure 4.1 – Equilibrium in chromatographic separations

In general, species having less affinity with the stationary phase (i.e. lower K values) experience shorter retention times and tend to move faster than those having stronger affinities (i.e. higher K values).

Sometimes, the time by which a component is retarded by the stationary phase is expressed in terms of capacity or retention factor (݇Ԣ) as follows:

஼ ௏ ݇ᇱ = ೔,ೄ ೄ (4.2) ஼೔,ಾ௏ಾ

Where, ܥ௜,ௌ and ܥ௜,ெ are the component concentration, ܸௌ and ܸெ are the respective volumes of the stationary and mobile phases.

In typical column chromatographic techniques, the sample is usually introduced at the top of a column packed with the stationary phase. Once the mobile phase is poured into the column, compounds in the sample move at different speeds as a consequence of the magnitude of interaction with the stationary phase. Finally, each compound can be recovered in a different fraction. A chromatogram is usually the visual output of a chromatographic separation, where each different peak generated corresponds ideally to a specific compound in the mixture (Figure 4.2).

Figure 4.2 – In column chromatography technique

4.2. Distribution ratio (D)

The distribution ratio (ܦ) is commonly used to express the distribution of a solute between two phases for a specific mobile phase. According to IUPAC, ܦ corresponds to the ratio of the total concentration of a solute in the extractant

phase (ܥ௘) to the total initial concentration (ܥ଴) (Equation 4.3) [43].

஼ ܦ = ೐ (4.3) ஼బ

In practice, distribution ratios are very useful to compare the degree of selectivity, for example, of a resin for two different species dissolved in the same solvent. The selectivity is usually expressed in terms of separation factors (ߙ), which correspond

to the ratio of D values of two different compounds that should be separated (Equation 4.4).

஽ಲ ߙ஺,஻ = (4.4) ஽ಳ

By convention, ܦ஺ > ܦ஻.

4.3. Column performance and efficiency of separation

One way to assess the performance of a chromatographic column is determining, for example, the resolving power or the ability of a column to separate two or more peaks (Equation 4.5).

ଶ(௧ᇱమି௧ᇱభ) ܴௌ = (4.5) ௪మା௪భ

(In the above equation, ݐԢ௜ corresponds to the adjusted retention time (ݐԢ௜ = ݐ௜ - ݐ௠

of each compound and ݓ௜, the respective widths at the base of each peak. ݐ௠ is usually subtracted and corresponds to the time required for the mobile phase to travel the length of the column without any interaction with the stationary phase.

As demonstrated in Figure 4.3, both ݐԢ௜ and ݓ௜ could be determined experimentally through the chromatogram obtained for a given chromatographic separation.

Figure 4.3 – Experimental variables to determine resolution in chromatography

Theoretically, a separation is considered complete when ܴௌ > 1.5.

The efficiency of separation, on the other hand, could be assessed through the determination of the number of theoretical plates (ܰ) using equation 4.6:

௧ ଶ ܰ = 16 ቀ ೔ ቁ (4.6) ௪೔

Again, ݐ௜ and ݓ௜ are respectively, the retention time and the width at the base of the peak for a given compound ݅.

The notion of theoretical plates was introduced in 1941 by Martin and Synge through the Plate Theory that supposes that a chromatographic column contains a large number of imaginary and thin sections called plates within each analyte is found to be at equilibrium between the stationary and mobile phase. As the notion

of the theoretical plate is now well established and it is applicable to all types of

chromatographic columns, it is convenient to express the performance of

chromatographic columns in terms of number of theoretical plates [44-46].

More efficient methods are obtained for greater ܰ values. In general, ܰ is

associated to the length (ܮ) of the column (Equation 4.7), but can be affected by

experimental factors such as: technique of column and sample preparation, solute

property, temperature, and flow rate.

௅ ܰ = (4.7) ு

Here, ܪ is the height equivalent to one theoretical plate.

4.4. Measurement of peak asymmetry

Normally, perfect Gaussian peaks are rarely obtained. In general, peak asymmetry is frequently observed and the main causes include at least one of the following conditions: nature of the packing material, nature of the analytes to be separated, and chromatographic system [47].

As defined by Bayne, peak asymmetry can be expressed as:

௕ ܣ = (Equation 4.8) ௌ ௔

where ܾ and ܽ represent, respectively, the back and the front portion at 10% of the peak height (Figure 4.4).

Figure 4.4 – Parameters for the determination of peak asymmetry

In practice, symmetrical peaks have asymmetry factors between 0.9 and 1.2.

4.5. Ion exchange chromatography (IEC)

Among numerous chromatographic techniques, ion exchange chromatography is used to separate ions based on electrostatic interactions between a charged surface and the ionic species in the sample. Repulsive electrostatic forces are expected for charges of the same sign and attractive forces, for opposite signs.

In practice, this technique allows the separation, for example, of anions from cations in a mixture. Counter-ions tend to be attracted to the surface while co-ions tend to be repelled (Figure 4.5).

Figure 4.5 – Separation of cations and anions by IEC [48]

In a system formed only by counter-ions, the magnitude of interaction with the

stationary phase will be proportional to the magnitude of the free charges (ݍଶ)

competing for the charge on the surface (ݍଵ) (Equation 4.8).

|௤ ௤ | |ܨ| = ݇ భ మ (4.8) ௘ ௥మ

In equation 4.6 ݇௘ is the Coulomb’s constant and ݎ the distance between ݍଵ and ݍଶ.

Briefly, for a negligible distance (ݎ) in a chromatographic column and a constant

value for the surface charge (ݍଵ), ions carrying larger charges tend to be stronger retained by the stationary phase than smaller charges. In this scenario, separations become possible as the stoichiometric process implicated allows counter-ions to be replaced by equivalent amounts of other counter-ions to preserve electrical

neutrality of the system [49]. Efficient separations could be achieved through the reversible exchanges of counter-ions at the surface of the stationary phase.

4.5.1. Ion exchange resins

Ion exchange resins are solid materials containing active and charged sites covalently bounded to the stationary phase. Depending on the group attached, those resins can be classified as cationic or anionic resins. Anionic resins carry positive charges and are designed to uptake negative counter-ions, while cationic resins, carrying negative charges, are designed to up take positive counter-ions [30] (Table 4.1).

Table 4.1 – Common commercial IEC resins AG50W-X8 DOWEX1-X8 Resin (Cationic) (Anionic) - + Active site - SO3 - N(CH3)3

Structure

Normally, active sites are arranged to form cross-linked chains. Resins with high crosslink percentages show a more rigid structure and provide a greater number of active groups. Common resins are usually available from 2% up to 12% or even 16% of crosslink percentage. In practice, performance of resins is mainly affected by crosslink percentage since it has an impact on the degree of selectivity.

4.6. Extraction chromatography (EXC)

Another chromatographic technique that has been receiving increasingly attention in recent years is the extraction chromatography. This technique combines the selectivity of liquid-liquid extractions with the speed, resolving power, and simplicity of chromatographic procedures.

Figure 4.6 – Schema of extraction chromatography [50]

As presented in figure 4.6, the liquid stationary phase or organic extractant is usually adsorbed on the surface of an inert solid support, usually porous silica or an organic polymer. The nature of the extractant usually determines the selectivity of the resin, but diluents are often employed to change the selective properties of the resin. Two examples of commercial extractants are presented in Table 4.2.

Table 4.2 – Common commercial EXC resins Resin Sr DGA

Extractant

18-crown-6 ether N,N,N’,N’-tetra-n-octyldiglycolamide

Extraction chromatography differs from partitioning chromatography because equilibrium takes place between an aqueous solution that corresponds to the mobile phase and an organic solution, in this case, the stationary phase.

Extraction chromatography is also different from ordinary liquid-liquid extractions due to the presence of the solid support that influences both the distribution coefficient (K) and the efficiency of extraction.

4.6.1. Extraction process in EXC

The basis of successful separations in extraction chromatography depends to a great extent on the ability of some species to undergo chemical transformations while other species do not. For example, metals are usually found in aqueous solutions under their ionic form. However, some metals in the presence of ligands can form neutral complexes that can be further solvated in the organic phase.

Different models have already been proposed to describe the overall mechanism and equilibrium processes implicated in the extraction chromatography technique

[51]. A first model, for example, assumes that the neutral complex is first formed in the aqueous phase (Equation 4.9) and then transferred to the organic phase (Equation 4.10), where the extraction process takes place.

Model 1: Complex formation in aqueous phase

ା௓ ି (௓,௔௤ (4.9ܮܯ ௔௤ ֎ܮ௔௤ + ݖܯ

ܯܮ௓,௔௤ ֎ ܯܮ௓,௢௥௚ (4.10)

A second model suggests that the species are first transferred to the organic phase (Equations 4.11) under their ionic form and then they form the neutral complex in the organic phase (Equation 4.12) to be extracted.

Model 2: Complex formation in organic phase

ା௓ ି ା௓ ି (௢௥௚ (4.11ܮ௢௥௚ + ݖܯ ௔௤ ֎ܮ௔௤ + ݖܯ

ା௓ ି ( (4.12 ௓,௢௥௚ܮܯ ௢௥௚ ֎ܮ௢௥௚ + ݖܯ

In both cases, however, the general equation for the extraction process can be expressed as follows:

(௬,௢௥௚ (4.13ܧ௓ܮܯ ௢௥௚ ֎ ܧ௓,௢௥௚ + ݕܮܯ

ܯܮ௓ܧ௬,௢௥௚ ֎ ܯܮ௓ܧ௬,௔௤ (4.14)

where, ܯା represents a metal, ܮି a ligand, ܯܮ the neutral complex formed, and ܧ the extractant that usually has an electron donor property.

4.7. IEC and EXC for radiochemical separations and potential applications for Sr-90 and Zr-90

Besides precipitation and solvent extraction, ion exchange chromatography is one of the most traditional methods used for radiochemical separations especially for the separation of actinides. In general, ion exchange chromatography has a multi- element character and usually shows better performance and higher recovery rates than other separation techniques [52, 53].

Earlier studies have demonstrated the possibility, for example, of using IEC resins to isolate fission products to evaluate their toxicity even when they are presented at trace levels in samples [54]. Some studies have also showed the efficiency of using ion exchange resins to separate radiostrontium from a variety of matrix [31, 36, 37].

Specifically for a given Sr-Zr system, Strelow had showed that regardless the acidic conditions, zirconium usually experiences stronger affinity with a cationic resin than numerous other elements, including Sr (Table 4.3) [9, 10].

Table 4.3 – Distribution ratios (D) for strontium and zirconium in the AG50W-X8 resin 0.5 M 1.0 M 2.0 M 3.0 M 4.0 M Eluent Zr Sr Zr Sr Zr Sr Zr Sr Zr Sr HCl 105 217 7250 60.2 489 17.8 61 10 14.5 7.5 4 HNO3 10 146 6500 39.2 652 8.8 112 6.1 30.7 4.7

Likewise, zirconium seems to have a stronger affinity with anionic resins while strontium, does not have any affinity in hydrochloric or nitric acid conditions (Figure 4.7) [55].

Figure 4.7 – D values for strontium and zirconium in the Dowex 1-X10 resin

Recent studies, however, have demonstrated that extraction chromatography is now starting to compete with ion exchange in many separation problems, including radiochemical applications where trace levels of analytes are eventually implicated.

Most part of those applications has been focused on the separation and analysis of radionuclides in environmental samples [56,62]. However, Maxwell and Culligan have reported the separation performance of extraction resins for urine samples containing actinides and Sr-90 [63]. Likewise, Kim et al. have presented a separation method to isolate Sr-89 and Sr-90 from calcium, barium, and yttrium in milk samples [64].

Currently, there are extraction resins designed to extract specific radionuclides. This is the case of the Sr resin that has been developed to extract strontium while other elements could be easily eluted from the chromatographic column. It has been demonstrated that in a solution of 3M HNO3 - 0.01M oxalic acid, Sr is completely retained, while Zr, for example, is rapidly eluted from the column [65].

As for ion exchange resins, distribution ratios or also capacity factors for most part of elements in extraction resins have already been reported [14]. As an example, figure 4.8 presents the difference between the capacity factors of strontium and zirconium in the DGA resin. As demonstrated, the potential to separate those two elements using nitric acid solutions becomes possible, as the capacity factor for Sr at 1M HNO3 is at least three times higher than that for Zr. In other words, Sr is more likely to experience a longer retention time than Zr, while this last can be faster eluted from the column.

Figure 4.8 – Capacity factor for strontium and zirconium in the DGA resin [66]

Chapter 5 5. Experimental

This chapter describes the methodology and list the materials used throughout the present work in order to determine the best condition to separate and quantify strontium and zirconium for age-dating applications.

5.1. Chemicals

Certified standard solutions (PlasmaCal ICP/ICPMS, 4% HNO3) of strontium, zirconium, and yttrium have been purchased from SPC Science. Stock solutions for separation tests and calibration curves have been prepared using high-purity deionized water (18.2 M:*cm) from Millipore Bedford, MA, USA, and environmental grade acids (Anachemia Science). All solutions have been conserved at 4°C in centrifuge polypropylene tubes until their utilization.

5.2. Digestion of SrTiO3

SrTiO3 (Aldrich, 99%) has been used for digestion tests to simulate true solid samples containing radiostrontium in the titanate form. The digestion protocol has been adapted from Parker et al [67]. The tests have been performed under high- pressure (Easy PrepTM vials) using a microwave oven (Mars 5) from CEM Corporation with temperature and pressure controls. The digestion program is presented in Table 5.1.

Table 5.1 – Instrumental setting for SrTiO3 digestion (Mars 5, Easy PrepTM vials) Power Ramp Pressure Temperature Hold Stage Max % Min Psi °C Min 1 1600 100 30 250 160 5 2 1600 100 30 250 200 40

Precise masses of SrTiO3 were first weighed in 5 mL Teflon vials and then diluted in concentrated nitric acid before being transferred to the microwave vessels. Poor

recoveries have been observed when SrTiO3 was weighed directly in the

microwave vessels. Actually, the strong electrostatic interactions between SrTiO3 and the surface of microwave vessels resulted in significant losses of the product during sample preparation.

5.3. Separation tests

Four different commercial resins (AG50W-X8, DOWEX1-X8, Sr Resin, and DGA) have been tested to separate strontium and zirconium. The performance of these resins has been assessed using a glass column from Omnifit®. In some cases, pre-packed columns (Eichrom) have also been applied to compare or validate

results. Different HCl, HNO3, H2SO4, and H2C2O4 solutions at different molarities have been tested as eluents. Before elution, mixtures containing strontium, zirconium, and eventually yttrium that would be expected to be present in real samples were evaporated to dryness and then diluted in the appropriate solvent that could contain traces of HF depending on the nature of the test.

5.4. Omnifit® glass column preparation

The Omnifit® glass column was dismantled following manufacturer’s instructions and washed using a laboratory detergent. The column was then well rinsed and packed with the resin suspended in water. A vacuum was applied to slowly drain the excess of water. Dryness was avoided to prevent air bubbles in the column. A plunger and an adjusting nut (Figure 5.1) at the upper side of the column helped to compact the resin. Strong compaction has shown to affect flow rate. Finally, the column was connected to a peristaltic pump and flow rate was set to 2 mL/min.

Figure 5.1 – AF Omnifit® Column Design

5.5. Methodology

After packing, resins have been washed with water and then conditioned with an acid solution, usually the eluent used to elute the first component from the column. Before injecting the mixture to be separated, a blank was recovered. It consisted of

the same eluent used for the conditioning step. After sample injection, strontium and zirconium were recovered in polypropylene centrifuge tubes using the appropriate eluent (Figure 5.2).

Figure 5.2 – Method applied for separation tests

5.6. Mass spectrometry analysis

Analyses of fractions recovered during separation tests were performed in a triple quadrupole mass spectrometer (Agilent 8800). After the separation, fractions were evaporated to dryness to eliminate any trace of HF possibly present that could

damage glass pieces in the instrument. Fractions were then diluted with 4% HNO3 as much as necessary to fit the concentration within the quantification range. Table

5.2 summarizes optimal acquisition settings for the quantification purpose.

Table 5.2 – Acquisition parameters for analysis of Sr and Zr by ICPQQQ-MS Scan Type Single Quad Plasma Mode Hot

Lenses Extract 1 5.3 V Extract 2 -225.0 V Omega Bias -200 V Omega Lens 28.0 V Q1 Entrance 3.0 V Q1 Exit -2.0 V Cell Focus 1.0 V Cell Entrance -50 V Cell Exit -50 V Deflect 15.2 V Plate Bias -50 V Q1 Q1 Bias -6.0 V Q1 Prefilter Bias -20.0 V Q1 Postfilter Bias -30.0 V Q1 Ion Guide SLS Factor 0.40 SLG 0.90 V Cell Gas mode No gas OctP Bias -8.0 V OctP RF 180 V Energy Discrimination 5.0 V Wait Time Offset Wait Time Offset 0 msec Spectrum mode options Replicates 3 Sweeps 100

Strontium and zirconium recoveries have been monitored respectively at m/z 88 and m/z 90. Blanks and standard solutions were used to ensure quality control of results. Signal fluctuations from the instrument were corrected through addition of

Indium as internal standard. According to Kozuka et al., ideal internal standards should have ionization energies close to that for measured elements [68]. Table 5.3 presents the ionization energy for strontium, zirconium, and indium.

Table 5.3 – Comparison of ionization energies between measured elements and internal standard Energy Element (eV)

Strontium 5.69 Zirconium 6.63 Indium (internal standard) 5.79

Both the detection limit (ܮܱܦ) and the quantification limit (ܮܱܳ) were estimated through the determination of the standard deviation (ݏ) obtained for the analysis of 10 blanks in 4% nitric acid (Equation 5.1 and 5.2) .

ܮܱܦ = 3ݏ (5.1)

ܮܱܳ = 10ݏ (5.2)

5.6.1. Performance of reaction cells to separate strontium from zirconium

The potential to use oxygen in reaction cells to separate strontium and zirconium has been assessed through the analysis of standard solutions at different gas compositions. In this case, strontium has been monitored at m/z 88 (Sr+) and m/z 104 (SrO+) as well as zirconium at m/z 90 (Zr+) and m/z 106 (ZrO+). The interaction between yttrium and oxygen has also been evaluated and both m/z 89 and m/z

105 have also been monitored. The acquisition parameters used during reaction- cell tests are presented in Table 5.4.

Table 5.4 – Acquisition parameters for the analysis of Sr and Zr using reaction cell and O2 as reaction gas Scan Type MS/MS Plasma Mode Hot

Lenses Extract 1 5.3 V Extract 2 -225.0 V Omega Bias -200 V Omega Lens 28.0 V Q1 Entrance 3.0 V Q1 Exit -2.0 V Cell Focus 1.0 V Cell Entrance -50 V Cell Exit -50 V Deflect 15.2 V Plate Bias -50 V Q1 Q1 Bias -4.0 V Q1 Prefilter Bias -22.0 V Q1 Postfilter Bias -20.0 V Cell Gas mode Use gas th 4 Gas Flow (O2) 0 – 100 % OctP Bias -8.0 V OctP RF 180 V Energy Discrimination 5.0 V Wait Time Offset Wait Time Offset 60 msec Spectrum mode options Replicates 3 Sweeps 100

5.7. Analysis of Sr-90 by liquid scintillation

For separation tests implicating a solution of radiostrontium (NIST Standard, 30 Bq/mL, 14/04/2000), recovered fractions have been diluted in water and then analysis of Sr-90 has been conducted by liquid scintillation using a Perkin Elmer Tri-Carb 2900TR instrument. Acquisition parameters are listed in Table. 5.5.

Table 5.5 – Acquisition parameters for the analysis of Sr-90 by liquid scintillation Quench Indicator tSIE/AEC External Std Terminator (sec) 0.5 2s% Pre-Count Delay (min) 0.00 Quench Set n/a Count Time (min) 240.00 Count Mode Normal Assay Count Cycles 1 Repeat Sample Count 1 #Vials/Sample 1 Calculate % Reference Off Background Subtract Off Low CPM Threshold Off 2 Sigma % Terminator Off LLa 0.0 Region A (keV) ULb 50.0 LL 0.0 Region B(keV) UL 100.0 LL 0.0 Region C(keV) UL 2000.0

Counting corrections Static Controller On Luminescence Correction Off Colored Samples n/a Heterogeneity Monitor n/a Coincidence Time (nsec) 18 Delay Before Burst (nsec) 75 Half Life Correction Off a. lower limit b. Upper limit

Chapter 6 6. Results and Discussion

This chapter highlights the performance of the ion exchange and extraction resins for the separation of strontium and zirconium. It also presents the potential of using a reaction cell in mass spectrometry to eliminate the need for a prior separation by chromatography. A method for SrTiO3 digestion has also been suggested as a sample preparation step for solid sources of strontium-90.

6.1. Digestion of SrTiO3

Due to its refractory character, strontium titanate is a compound usually very difficult to decompose. For this reason, four different mixtures with different H2O2,

HNO3, and HF acid ratios were tested to assess the best condition to achieve successful SrTiO3 digestion (Table 6.1). As reported by Packer et al., addition of

HF would help the solubilisation of Ti while H2O2 would contribute to reduce the formation of N2O4 as well as to provide a cloudless solution [67].

Table 6.1 – Acid mixtures used for SrTiO3 digestion tests HNO H O HF Mixture ID 3 2 2 (% v/v) (% v/v) (% v/v) 1 100 Absent Absent 2 89 11 Absent 3 84 11 5 4 95 Absent 5

As presented in Table 6.2, efficient digestion has been obtained using mixtures containing HF. Even if concentrated nitric acid is considered as a strong oxidizing agent, addition of HF is usually necessary for complete dissolution of oxide compounds [69].

Table 6.2 – Digestion efficiency of SrTiO3 under different acidic conditions

HNO3 HNO3/H2O2 HNO3/HF/H2O2 HNO3/HF Element Mixture 1 Mixture 2 Mixture 3 Mixture 4 (%) (%) (%) (%) 92 101 100 101 Sr 98 96 98 94 97 92 97 109 Average 95 ± 3 97 ± 4 98 ± 1 101 ± 8 Ti 0 31 ± 8 96 ± 1 96 ± 6

The presence of H2O2 does not improve the performance of SrTiO3 digestion.

Likewise, the appearance of the SrTiO3 solution after digestion in HNO3/HF

demonstrates that samples containing approximately 10,000 mg/L of SrTiO3 were

very transparent, which has made the usage of H2O2 unnecessary to obtain satisfactory results.

In general, digestion using only HNO3/HF has not only been proven to be the most convenient and efficient method and also a very reproducible approach. For nine replicates, obtained recovery for Sr was 102 r 2 %, while for Ti, 97 r 1 %, which means that up to 3% of titanium could not be dissolved under those conditions (Figure 6.1).

110 100 90 80 70 60 50 Sr 40 Ti

% in solution solution in % 30 20 10 0 123456789 Replicate ID

Figure 6.1 – Reproducibility of SrTiO3 digestion using HNO3/HF mixture

Due to the difficulty to generate SrTiO3 pellets in the laboratory, digestion tests have been performed using only SrTiO3 under the powder form.

6.2. Separation of Sr and Zr using a cation-exchange resin

The method to separate zirconium from strontium using the cationic AG50W-X8 resin has been adapted from the method proposed by Strelow [70]. A similar method has been already proven to provide good reproducibility making possible its application in geochronology work for age determination [71].

Preliminary tests have been performed using about 10 g of AG50W-X8 resin, 100- 200 mesh, H+ form (Eichrom) poured into a 15 i.d. X 100 mm glass column. Experimental elution curves have been obtained using weighed solutions

containing equivalent amounts of Sr and Zr (50 µg/mL) eluted in 2, 3 and 4M HCl. Molarities lower than 2M have been disregarded due to the high distribution ratio values reported in the literature [9]. In practice, extremely long retention times would be expected for these conditions.

As presented in Figure 6.2, incomplete separation has been achieved using 4M HCl and separation factor (D) of about 2 seemed not to be sufficient to obtain complete resolution of peaks.

50 45 40 35 30 25 Sr 20 15 Zr Recovery Recovery (%) 10 5 0 0 10 20 30 40 50 60 70 80 90 100 4M HCl (mL)

Figure 6.2 – Elution profile of Sr and Zr in 4M HCl (10 g AG50W-X8, 100-200 mesh)

Although significant improvement has been obtained for a separation factor of 6 using 3M HCl, incomplete separation has also been observed (Figure 6.3).

50 45 40

35 30 25 Sr 20 15 Zr Recovery Recovery (%) 10 5 0 0 50 60 70 80 90 100 110 120 130 140 150 3M HCl (mL)

Figure 6.3 – Elution profile of Sr and Zr in 3M HCl (10 g AG50W-X8, 100-200 mesh)

Briefly, 2M HCl has shown to be the most efficient condition to separate Sr and Zr among the three scenarios tested (Figures 6.4). A separation factor of 27 has been calculated in this case.

50 45 2M HCl 6M HCl

40 35 30 25 Sr 20 Zr Recovery Recovery (%) 15 10 5 0 0 50 100 110 120 130 140 150 160 170 180 190 200 250 300 350 400 450 500 550

HCl (mL) Figure 6.4 – Separation of Sr and Zr using a 2M to 6M HCl gradient (10 g AG50W-X8, 100-200 mesh)

Apparently, in 2M HCl Zr is completed retained in the column during Sr elution and Zr removal was only possible after the addition of 6M HCl. Actually, the increase in solvent concentration increased the number of H+ ions in the system and to displace zirconium that was then easily stripped off the column.

As presented in Figure 6.5, the usage of 2M HCl, however, resulted in longer retention times and broader peaks for Sr. Likewise, the volume of eluent required to completely elution of Sr increased from about 80 mL to about 200 mL.

45 40 35 4M HCl 30 3M HCl 25 20 2M HCl Sr (%) Sr 15 10 5 0 0 20 40 60 80 100 120 140 160 180 200 HCl (mL) Figure 6.5 – Elution curves of Sr as function of HCl molarity (10 g AG50W-X8, 100-200 mesh)

Since the distribution ratio of Sr in nitric acid eluent was theoretically two times smaller than for HCl (Table 4.3), better performance has been obtained by

replacing HCl by HNO3. As demonstrated in Figure 6.6, 2M HNO3 eluent produced narrower peaks and it shifted the maximum peak from about 150 mL to about 80 mL.

15 2M HCl

Sr (%) Sr 10 2M HNO3

0 0 50 100 150 200 250 Eluent (mL) Figure 6.6 – Elution curves of Sr at 2M HNO3 and 2M HCl (10 g AG50W-X8, 100-200 mesh)

6.3. Resin shrinkage and issues for Zr recovery

In general, poor recoveries for Zr have been obtained using the AG50W-X8 resin probably resulting from problems of resin shrinkage after increasing the eluent concentration from 2M to 6M. To avoid sudden changes that could disturb equilibrium in the column, the usage of stronger eluents at low molarities was also assessed. Higher cross-linking resins, however, could also help reducing shrinkage issues.

Tests have demonstrated that sulphuric and oxalic acids could be used as alternative eluents of Zr. Based on results presented in Table 6.3, 2M H2SO4, 3M

H2SO4, and 0.2M H2C2O4 have provided both peak asymmetry and reasonable recovery.

Table 6.3 – Performance of alternative eluents for Zr Zr Recovery Eluent Peak Asymmetry (%)

2M H2SO4 1.2 104 ± 3

3M H2SO4 1.0 101 ± 2

0.2M H2C2O4 1.2 123 ± 3

Utilization of oxalic and sulphuric acids, however, has been limited to elution tests involving high concentrations of Zr where the evaporation steps were not required

to concentrate the analyte. Due to the high boiling point of H2SO4, the utilization of Teflon vials became impracticable and glass beakers have been avoided since they could contain significant amounts of Zr able to contaminate the samples.

6.3.1. Effect of method downscaling on separation efficiency

As presented in Figure 6.7, efficient separation of strontium and zirconium was successfully achieved using a smaller column bed containing approximately 2 g of AG50W-X8 resin.

70 2M HNO3 6M HCl 60 50 40 Sr 30 Zr

Recovery (%) (%) Recovery 20 10 0 0 102030405060708090100110120130140150 Eluent (mL)

Figure 6.7 – Separation of Sr and Zr using a 2M HNO3 to 6M HCl gradient in 2 g AG50W-X8 (100-200 mesh)

Among the advantages, downscaling allowed to reduce the volume of eluent required and consequently elution time for complete separation. As presented in

Figure 6.8, the volume of 2M HNO3 required for the elution of Sr was decreased from about 150 mL to about 20 mL.

150 130 110 (mL)

3 90 70 50 2M HNO 30 10 10g 2g 1g 0.5g AG50W-X8 (g)

Figure 6.8 – Volume of eluent for Sr elution as function of mass of AG50W-X8

Reducing column beds, however, has demanded the adjustment of the volume of loaded sample in order to prevent extra-column effects [72]. In general, compatible loading volumes have been proved to ensure peak shape and avoid peak overlap. Suggested loading volumes according to the mass of resin used are presented in Table 6.4.

Table 6.4 – Sample loading volumes according to the mass of dry resin used for separations

Mass of resin Sample (g) (µL) 2 1000 1 500 0.5 250

6.4. Separation of Sr and Zr using an anion-exchange resin

As demonstrated in Figure 6.9 satisfactory separation has also been achieved using the Dowex1-X8 resin (Acros Organics, 100-200 mesh).

120 110 11M HCl 4M HCl 100 90 80 70 60 50 Sr 40 Zr Recovery (%) (%) Recovery 30 20 10 0 0 5 10 15 20 25 30 35 40 45 50 HCl (mL)

Figure 6.9 – Separation efficiency for Sr and Zr using Dowex1-X8 resin

In agreement with the literature, separation was possible as Zr uptake increased with HCl concentration while strontium that does not have any affinity with the resin was immediately eluted from the column [55]. As presented in Figure 6.10, maximum Zr uptake could be obtained at 11M HCl.

100

70 Zr uptake (%) (%) Zr uptake 60

50 8 9 10 11 HCl (M) Figure 6.10 – Zirconium retention in Dowex1-X8 as function of HCl concentration

At the end, 4M HCl has been proved to be the best condition for Zr elution. Under this condition, however, recoveries higher than 100% have been observed, which was interpreted as possible column contaminations or interferences in mass spectrometry caused by traces of HCl in samples (Figure 6.11).

130 120 110

100 90 80 70 60 50 Zr recovery Zr recovery (%) 40 30 20 0 1 2 3 4 HCl (M) Figure 6.11 – Maximum recovery of Zr according to HCl concentration in Dowex1-X8

6.5. IEC versus EXC for the separation of Sr and Zr

The performance of ion exchange and extraction resins has been carried out using approximately 0.5 g of resin, 250 µL of sample loading, and 40 mL of eluent (i.e. 20 mL for Sr eluent and 20 mL for Zr eluent). In general, eluents have been chosen according to their capability to provide the best resolution and/or maximum recovery. Optimal results for each one of the four resins tested (AG50W-X8 – 100- 200 mesh, Dowex1-X8 – 100-200 mesh, Sr – 50-100 µm, and DGA – 50-100 µm) are presented in Figure 6.12.

100 2M HNO3 3M H2SO4 80 60 Sr 40 Y Zr

Recovery (%) (%) Recovery 20 a 0 0 5 10 15 20 25 30 35 40 45 50 100 11M HCl 4M HCl 80 60 40

Recovery (%) (%) Recovery 20 b 0 0 5 10 15 20 25 30 35 40 45 50 100 0.05M HNO3 3M HNO3 80 60 40

Recovery (%) (%) Recovery 20 c 0 0 5 10 15 20 25 30 35 40 45 50 100 0.05M HNO3 1M HNO3 H2O 80 60 40

Recovery (%) (%) Recovery 20 d 0 0 5 10 15 20 25 30 35 40 45 50 Eluent (mL) Figure 6.12 – Comparative of separation of Sr and Zr using ion exchange and extraction resins (a. AG50W-X8, b. DOWEX1-X8, c. Sr-Resin, d. DGA-Resin)

Contrary to results obtained for IEC resins, Zr and Sr have exhibited opposite behaviours in both Sr and DGA resins. Actually, the ability of zirconium to form charged complexes has certainly increased its affinity with the aqueous mobile phase, which probably contributed to reduce its retention time in EXC. Such an effect could not be possible for Sr since it cannot form charged complexes and, for that reason, it has shown stronger affinity with the organic stationary phase and a longer retention time in Sr and DGA resins (Figure 6.13).

Figure 6.13 – Proposed extraction mechanism for Sr for its separation from Zr by EXC

In practice, the possibility to recover zirconium before strontium reduced the risks of peak overlap caused by tailing problems at high concentrations of Sr potentially found in real samples. Results presented in Figure 6.14 show that tailing problems started to become important once the Sr concentration increased from 10 µg/mL to 1000 µg/mL.

100 10 µg/mL 90 50 µg/mL 80 100 µg/mL 70

60 1000 µg/mL 50

Sr (%) Sr 40 30 20 10 0 40 60 80 100 120 140 160 180 200 2M HNO (mL) 3 Figure 6.14 – Tailing effect as function of Sr concentration (AG50W-X8)

In terms of efficiency to eliminate possible isobaric interferences from yttrium, limitations have been encountered specially for the AG50W-X8 and Sr resins, where peak overlapping has been observed. The main issue for Dowex1-X8 and DGA resins, on the other hand, has been associated to insufficient Zr recovery. For DGA, Zr concentration has been found to be below the quantification limit.

6.6. Addition of HF in samples

Since the separation of Sr and Zr in real samples could be performed in presence of trace levels of HF from digestion step, hydrofluoric acid has been added to samples in order to assess its impact on separation efficiency.

For the anionic resin, separation was compromised after addition of 0.01% HF to the samples. As demonstrated in Figure 6.15, peak overlapping has been

observed. Repulsion forces between neutral or positive complexes of zirconium and the active site on resin surface probably contributed to accelerate Zr elution.

100 Sr 90 80 Zr 70 60 50 40

Recovery (%) (%) Recovery 30 20 10 0 0 5 10 15 20 Eluent (mL) Figure 6.15 – Separation of Sr and Zr using Dowex1-X8 for samples containing HF

For DGA resin, on the other hand, the utilization of HF in samples has not only helped to increase Zr recovery but also ensured separation efficiency. As demonstrated in Figure 6.16, recovery of zirconium was increased from 11 1% to 102 10% and no overlapping has been detected after HF concentration was increased from 0.01% to 0.2%.

100 1M HNO3 0.05M HNO3 80

40 Recovery (%) (%) Recovery 20 a 0 0 5 10 15 20 25 30 35 40 45 50 100

40 Recovery (%) (%) Recovery 20 Sr b Zr 0 0 5 10 15 20 25 30 35 40 45 50 Eluent (mL)

Figure 6.16 – Separation of Sr and Zr using DGA for samples containing HF (a. 0.01%, b. 0.2%)

In general, the HNO3/HF ratio of 5 has proved to be efficient to obtain a satisfactory recovery of zirconium (Table 6.5). According to results, about 50% of zirconium recovery was lost after increasing HNO3/HF ratio from 5 to 10. It was assumed that - the increase in NO3 concentration in the system has provoked competition against F- ions to form zirconium complexes. Actually, even if fluoride complexes formation was theoretically more favoured, in this case nitrate complexes were more likely to form as a consequence of an excess of nitrate ions in the sample.

Table 6.5 – Recovery of Zr in DGA Resin according to HNO3/HF ratio in samples

HNO3/HF [HNO3] [HF] Recovery Ratio (M) (M) (%) 5 1 0.2 100 5 0.5 0.1 112 10 1 0.1 52 10 2 0.2 59

Due to the toxicity and risks of damaging glass components in mass spectrometers, minimal HF concentrations have been considered and the possibility of using HF as pure eluent has been totally discarded.

6.7. Summary of the efficiency of all resins tested

Table 6.6 summarizes the performance of the fours resins tested to separate strontium from zirconium. The respective recoveries for each element under all the experimental conditions tested over this work can be find in the Annexe 1.

Table 6.6 – Summary of resins performance to isolate Zr prior MS analyses Separation efficiency in Resin Sr elimination Y elimination presence of HF in samples AG50W-X8 Dowex1-X8 Sr-Resin DGA-Resin

In general, DGA resin has proven to be the most efficient alternative to obtain satisfactory separation and acceptable zirconium recovery. As presented in Table 6.6, DGA was the only resin able to eliminate both potential interferences from Sr and Y at m/z 90 even under conditions where HF was present. Due to its superior performance over other resins, a complete methodology using DGA resin for isolating zirconium prior to MS analysis has been proposed (Figure 6.17).

Figure 6.17 – Complete methodology to separate Sr and Zr using DGA resin

Before starting the separation, the resin was usually cleaned with 50 mL of water and conditioned with 5 mL 1M nitric acid. A blank was then recovered before loading the sample. The sample load solution, was prepared in a mixture of 1M

HNO3 / 0.2M HF and finally, zirconium was recovered using a 10 mL 1M HNO3 as the eluent.

6.8. Performance of DGA method for the recovery of trace levels of Zr

The method presented in Figure 6.17 has been tested in standard solutions

containing zirconium concentrations at ppt levels and concentrations for strontium 100 times higher than those for Zr (Figure 6.18).

20 ExpectedE values Experimental Results

10 Zr (ng/mL)

0 123456 Sample ID Figure 6.18 – Comparative between experimental and expected results for the recovery of trace levels of Zr using DGA resin

Except for samples 1 and 2, other samples exhibited a good correlation with expected values. Average recovery obtained for zirconium was 94 ± 6%. The detection limit for zirconium has been determined as being 92 pg/mL. No strontium has been detected in the zirconium fractions. The detection limit for strontium has been determined as being 48 pg/mL.

6.9. Determining the age of a radiostrontium source

In order to determine the age of radioactive source, the proposed separation method using DGA-Resin was also applied to isolate zirconium-90 from its parent Sr-90. The complete procedure used for the separation and analyses of Sr-90 and Zr-90 is presented in Figure 6.19.

Figure 6.19 – Procedure for determining the age of a radiostrontium source

As indicated, a solution containing 30 Bq/mL of radiostrontium (Sr-90 NIST Standard, 14/04/2000) has been first separated in two different portions. The first fraction was used to isolate Zr-90 with the DGA resin for further analyses by ICP- MS, while the second was simply diluted in 15 mL of water for the analysis of Sr-90 by liquid scintillation.

As the exact concentration of Zr-90 was not available, blanks and standard solutions have been used to ensure the quality of results for Zr analysis. For Sr-90, it was observed that the experimental concentration was about 30 % lower than the theoretical concentration (Figure 6.20).

40 y = 220.5756x R² = 0.9993

20 cpm

10 Theoretical [Sr-90] Experimental [Sr-90] 0 0.00 0.05 0.10 0.15 0.20 Bq/mL

Figure 6.20 – Comparative between theoretical and experimental concentrations for the analysis of Sr-90 by liquid scintillation

Although the true age of the radiostrontium source was unknown, an age of 68 ± 11 years was proposed. Calculations have been based on the principles of the decay law treated in section 2. The error of 11 years, however, has been considered too high for nuclear forensic applications. Problems associated to the quality of Sr-90 solution have been considered as the main cause to the lack of accuracy encountered.

6.10. Potential of reaction cell to separate strontium from zirconium

As zirconium could be analyzed at m/z 106 to avoid isobaric interferences at m/z 90, some tests to assess the ability of Zr and Sr to form oxides in mass

spectrometry under different O2 concentrations have been carried out. Results for

standard solutions prepared in 4% HNO3 containing 5 ng/mL of Sr and 5 ng/mL of Zr is presented in Figure 6.21.

6000

5000

4000

3000 SrO+

Signal (cps) 2000 ZrO+ 1000

0 0 102030405060708090100

O2 (%)

Figure 6.21 – Zr and Sr oxides formation in mass spectrometry as function of O2 concentration in the reaction cell

The formation of ZrO+ (m/z 106) showed to be favoured at 10% oxygen, where a maximum peak was detected. At this point, a decontamination factor of about 80% against strontium was obtained.

Maximum decontamination factor has been achieved at 6 % O2. Under this condition, there was about 17% of Zr that was transformed in zirconium oxide (90Zr16O+) and about 2% of strontium oxide (88Sr16O+) that could be formed (Figure 6.22a and 6.22b). Such condition has allowed increasing decontamination factor from 80% to 98 % to meet the minimum of 97% expected for date-aging applications.

100 90 90Zr+ 80 70 90Zr16O+ 60 90Zr16O1H+ 50 90 16 + Zr O2 40 90 16 1 + 30 Zr O H2 Zr species (%) (%) Zr species 20 10 a 0 85 90 95 100 105 110 115 120 125

100 88 + 90 Sr 80 70 60 88 16 + 88 16 + 50 Sr O Sr O2

40 88 16 1 + 30 Sr O H

Sr species (%) (%) Sr species 20 10 b 0 85 90 95 100 105 110 115 120 125

Figure 6.22 – Predominant species of Zr (a) and Sr (b) at 6% O2 in the reaction cell

Although only 17% of Zr was expected to be formed at 6% O2, Figure 6.23 shows that comparable results to m/z 90 could always be obtained. After analyzing a series of solutions containing different concentrations of Zr, a good correlation (R2 = 0.997) between results obtained at m/z 90 and m/z 106 has been achieved.

100 R² = 0.997 90 80 70 60 50 (ng/mL) 40 Zr at m/z = 106 m/z = Zr at 30 20 10 0 0 102030405060708090100 Zr at m/z = 90 (ng/mL)

Figure 6.23 – Correlation between results for the analysis of Zr at m/z 90 and m/z 106

The only limitation found at 6% O2 was the maximum decontamination factor of 90% achieved to eliminate potential isobaric interferences from yttrium. In this case, approximately 10 % of yttrium has demonstrated to react with oxygen to form yttrium oxide (89Y16O+) that could interfere in the analysis of Zr at m/z 106 (Figure 6.24).

100 89 + 90 Y 80 70 60 89Y16O+ 89 16 + 50 Y O2 89Y16O1H+ 40 89Y1H+ 30 89 16 1 +

Y species (%) (%) species Y Y O H 20 2 10 0 85 90 95 100 105 110 115 120 125 m/z Figure 6.24 – Predominant species of Y at 6% O2 in the reaction cell

Conclusions

The present work, demonstrates that the utilization of collision cells in MS without prior chromatographic separation does not provide sufficient resolution to completely isolate Zr from all its isobaric interferences. However, the results suggest that triple quadrupole instruments (ICPQQQ-MS) have a potential to significantly minimize the level of isobaric interferences while reducing both the duration and the complexity of sample preparation procedures.

The technique was demonstrated to eliminate 98% of interferences from Strontium and 90% of interferences from Yttrium that are eventually present at high levels in seized sources of radiostrontium. Although the present work has concentrated all the efforts on the development of a radiochronometric method to isolate Sr-90 from Zr-90, it is believed that such a technique could also been applied to other longer- lived radionuclides also of interest for nuclear security experts (e.g. Cs-137 (t1/2 =

30 y), Pu-238 (t1/2 = 88 y)).

Among four commercial resins tested, DGA has been proved to provide the best performance for the radiological separation. Recoveries higher than 99% for Zr have been obtained. The DGA approach has also been demonstrated to be the faster approach and the more efficient to eliminate both the isobaric interferences from Strontium and Yttrium. None of these two elements have been detected by mass spectrometry after the chromatographic separation.

Finally, the proposed method using DGA combined to MS and liquid scintillation for the respective analysis of Zr-90 and Sr-90 has been applied to determine the age of an aqueous solution of radiostrontium. Although the true age of the radioactive

source was unknown, an age of 68 ± 11 years was calculated. The uncertainty of 11 years observed, however, has been considered too high for nuclear forensic applications. Problems associated to the quality of Sr-90 solution have been considered as the main cause to the lack of accuracy encountered.

More tests using sources of radiostrontium having well known ages would help to identify the major sources of error. They would also validate the proposed method while helping to determine the precision of the age determined experimentally. Likewise, the utilization of Teflon material and ultra-pure acidic solutions would help to minimize systematic errors by preventing any trace of unwanted contamination.

References

[1] Mayer, K.; Wallenius, M.; Fanghaenel, T. Journal of Alloys and Compounds 2007, 444, 50-56. [2] International Atomic Energy Agency. Nuclear Security Report 2013, 1-16. [3] Charbonneau, L.; Benoit, J.-M.; Jovanovic, S.; St-Amant, N.; Kiser, S.; Cooke, M. W.; Mercier, J.-F.; Nielsen, K.; Kelly, D.; Samuleev, P.; Galea, R.; Moore, K.; Saull, P. R. B.; Chamberlain, D. B.; Steeb, J. L.; Graczyk, D. G.; Tsai, Y.; Sullivan, V. S.; Dimayuga, I. C.; Shi, Y.; Rao, R.; Larivière, D. Analytical Methods 2014, 6, 983-992. [4] Amano, H.; Yanase, N. Talanta 1990, 37, 585-590. [5] Froidevaux, P.; Geering, J.-J.; Valley, J. F. Science of the Total Environment 2006, 367, 596-605. [6] Grahek, Z.; Eskinja, I.; Kosutic, K.; Lulic, S.; Kvastek, K. Journal of Radioanalytical and Nuclear Chemistry 1999, 241, 617-626. [7] Maxwell, S. L.; Culligan, B. K.; Utsey, R. C. Journal of Radioanalytical and Nuclear Chemistry 2013, 298, 867-875. [8] Steeb, J. L.; Graczyk, D. G.; Tsai, Y.; Mertz, C. J.; Essling, A. M.; Sullivan, V. S.; Carney, K. P.; Finck, M. R.; Giglio, J. J.; Chamberlain, D. B. Journal of Analytical Atomic Spectrometry 2013, 28, 1493-1507. [9] Strelow, F. W. E. Analytical Chemistry 1960, 32, 1185-1188. [10] Strelow, F. W. E.; Rethemey. R; Bothma, C. J. C. Analytical Chemistry 1965, 37, 106-111. [11] Huffman, E. H.; Iddings, G. M.; Lilly, R. C. Journal of the American Chemical Society 1951, 73, 4474-4475. [12] Bunney, L. R.; Ballou, N. E.; Pascual, J.; Foti, S. Analytical Chemistry 1959, 31, 324-326. [13] Faris, J. P. Analytical Chemistry 1960, 32, 520-522. [14] Pourmand, A.; Dauphas, N. Talanta 2010, 81, 741-753. [15] Lallanilla, M. Live Science Retrieved March 24, 2014 from < http://www.livescience.com/44318-chernobyl-trees-barely- decomposed.html>. [16] Institut de Radioprotection et de Santé Nucléaire Retrieved August 22, 2014 from < http://environnement.irsn.fr/document/site_1/fckfiles/File/dossiers/tchernobyl/ tchernobyl_16.pdf>. [17] Environmental Protection Agency Retrieved August 22, 2014 from < http://www.epa.gov/radiation/radionuclides/strontium.html> [18] Environmental Protection Agency Retrieved August 31, 2014 from < http://epa.gov/radiation/radionuclides/iodine.html>. [19] United Nations Scientific Committee on the Effects of Atomic Radiation Retrieved August 31, 2014 from < http://www.unscear.org/unscear/en/chernobyl.html>. [20] Nielsen, E.; Greve, K.; Ladefoged, O. The Danish Environmental Protection Agency 2013, 40 p.

[21] Glasstone, S.; Dolan, P. J.; United States Department of Defense and the Energy Research and Development Administration 1977, 3rd edition, 605 p. [22] International Atomic Energy Agency. Retrieved August 21, 2014 from < http://www.iaea.org/OurWork/ST/NE/NEFW/News/2012/repository/RTGs- Removed-Successfully-from-the-Baltic-Sea.html>. [23] International Atomic Energy Agency. Retrieved August 22, 2014 from < http://www.iaea.org/Publications/Magazines/Bulletin/Bull481/htmls/rtg_accid ents.html>. [24] BBC Retrieved August 22, 2014 from < https://www.youtube.com/watch?v=opTHVe8fzlg>. [25] Wallenius, M.; Mayer, K.; Ray, I. Forensic Science International 2006, 156, 55-62. [26] Boulyga, S. F. International Journal of Mass Spectrometry 2011, 307, 200- 210. [27] Becker, J. S.; Dietze, H.-J. Spectrochimica Acta Part B: Atomic Spectroscopy 1998, 53, 1475-1506. [28] Hou, X.; Roos, P. Analytica Chimica Acta 2008, 608, 105-139. [29] Shimadzu. Retrieved August 31, 2014 from < http://www.shimadzu.com/an/lcms/support/intro/lib/lctalk/61/61intro.html>. [30] Patnaik, P. McGraw-Hill Handbooks 2004, 2nd edition, 1280 p. [31] Vonderheide, A. P.; Zoriy, M. V.; Izmer, A. V.; Pickhardt, C.; Caruso, J. A.; Ostapczuk, P.; Hille, R.; Becker, J. S. Journal of Analytical Atomic Spectrometry 2004, 19, 675-680. [32] Favre, G.; Brennetot, R.; Chartier, F.; Vitorge, P. International Journal of Mass Spectrometry 2007, 265, 15-22. [33] Thomas, R. CRC Press 2008, 2nd edition, 376 p. [34] Dean, J. R. Wiley 2005, 1st edition, 208 p. [35] Epov, V. N.; Taylor, V.; Lariviere, D.; Evans, R. D.; Cornett, R. J. Journal of Radioanalytical and Nuclear Chemistry 2003, 258, 473-482. [36] Isnard, H.; Aubert, M.; Blanchet, P.; Brennetot, R.; Chartier, F.; Geertsen, V.; Manuguerra, F. Spectrochimica Acta Part B 2006, 61, 150-156. [37] Feuerstein, J.; Boulyga, S. F.; Galler, P.; Stingeder, G.; Prohaska, T. Journal of Environmental Radioactivity 2008, 99, 1764-1769. [38] Amr, M. A.; Abdel-Lateef, A. M. International Journal of Mass Spectrometry 2011, 299, 184-190. [39] Sakama, M.; Nagano, Y.; Saze, T.; Higaki, S.; Kitade, T.; Izawa, N.; Shikino, O.; Nakayama, S. Applied Radiation and Isotopes 2013, 81, 201-207. [40] Eiden, G. C.; Barinaga, C. J.; Koppenaal, D. Rapid Communications in Mass Spectrometry 1997, 11, 37-42. [41] Pennsylvania State University Environmental Health & Safety. Retrieved August 22, 2014 from < http://www.ehs.psu.edu/radprot/LSC_Theory2.pdf>. [42] L'Annunziata, M. F. Academic Press 2012, 3rd edition, 1418 p. [43] IUPAC. Retrieved August 22, 2014 from < http://goldbook.iupac.org/D01817.html>. [44] Littlewood, A.B. Elsevier 2013, 560 p. [45] Naushad, M. Khan, M. R. CRC Press 2014, 480 p. [46] Weston, A. Brown, P. R. Academic Press 1997, 280 p.

[47] Bayne, S.; Carlin, M. CRC Press LLC 2010, 272 p. [48] GE Healthcare. Retrieved September 1, 2014 from < https://www.uni- ulm.de/uploads/media/Ion_exchange_chromatography_and_Chromatofocusi ng_02.pdf>. [49] Helfferich, F. G. McGraw-Hill 1962, 624 p. [50] Triskem. Retrieved August 22, 2014 from < http://www.triskem- international.com/full_extraction_chromatographie.asp>. [51] Braun, T.; Ghersini, G. Elsevier Science 1975, 2, 584 p. [52] Lamb, J. D.; Nordmeyer, F. R.; Drake, P. A.; Elder, M. P.; Miles, R. W.; Lash, R. P. Journal of Radioanalytical and Nuclear Chemistry 1989, 134, 317-331. [53] Kraus, K. A.; Nelson, F. Annual Review of Nuclear Science 1957, 7, 31-46. [54] Tompkins, E. R.; Khym, J. X.; Chon, W. E. Journal of the American Chemical Society 1947, 69, 2769-2777. [55] Saito, N. Pure and Applied Chemistry 1984, 56, 523-539. [56] Lee, M. H.; Ahn, H. J.; Park, J. H.; Park, Y. J.; Song, K. Applied Radiation and Isotopes 2011, 69, 295-298. [57] Harrison, J. J.; Zawadzki, A.; Chisari, R.; Wong, H. K. Journal of Environmental Radioactivity 2011, 102, 896-900. [58] Maxwell, S. L. Journal of Radioanalytical and Nuclear Chemistry 2006, 267, 537-543. [59] Wang, J.-J.; Chen, I.-J.; Chiu, J.-H. Applied Radiation and Isotopes 2004, 61, 299-304. [60] Roane, J. E.; DeVol, T. A.; Leyba, J. D.; Fjeld, R. A. Journal of Environmental Radioactivity 2003, 66, 227-245. [61] Mellado, J.; Llaurado, M.; Rauret, G. Analytica Chimica Acta 2001, 443, 81- 90. [62] La Rosa, J. J.; Burnett, W.; Lee, S. H.; Levy, S. H.; Gastaud, J.; Povinec, P. P. Journal of Radioanalytical and Nuclear Chemistry 2001, 248, 765-770. [63] Maxwell, S. L.; Culligan, B. K. Journal of Radioanalytical and Nuclear Chemistry 2009, 279, 901-907. [64] Kim, C. K.; Al-Hamwi, A.; Torvenyi, A.; Kis-Benedek, G.; Sansone, U. Applied Radiation and Isotopes 2009, 67, 786-793. [65] Horwitz, E. P.; Dietz, M. L.; Chiarizia, R. Journal of Radioanalytical and Nuclear Chemistry 1992, 161, 575-583. [66] Eichrom. Retrieved August 22, 2014 from < http://www.eichrom.com/products/info/dga_resin.aspx>. [67] Packer, A.-P.; Lariviere, D.; Li, C.; Chen, M.; Fawcett, A.; Nielsen, K.; Mattson, K.; Chatt, A.; Scriver, C.; Erhardt, L. S. Analytica Chimica Acta 2007, 588, 166-172. [68] Kozuka, S.; Yokote, Y.; Abe, K.; Hayashi, M.; Matsunaga, H. Fresenius Journal of Analytical Chemistry 1995, 351, 801-802. [69] Lahiri, S. New Delhi: Norosa Publishing House 2008, 186 p. [70] Strelow, F. W. E. Analytical Chemistry 1959, 31, 1974-1977. [71] Strelow, F. W. E. Analytical Chemistry 1959, 31, 1201-1203. [72] Neue, U. D. Wiley-VCH 1997, 416 p.

Annexe 1