Rapid Radiochemical Separations of Americium from Actinides, Mixed

Fission Products and Matrix Elements

A thesis submitted to The University of Manchester for the degree of Doctor of Philosophy in the Faculty of Engineering and Physical Sciences

Matthew Alan Higginson

2015

Table of Contents

1.0 Table of Contents ...... 2 1.1 Table of Figures ...... 6 1.2 Table of Tables ...... 8 1.3 Abbreviations ...... 9 2.0 Abstract ...... 12 3.0 Declaration ...... 13 4.0 Copyright Statement ...... 14 5.0 Acknowledgments...... 15 6.0 OUTLINE AND AIMS OF PROJECT...... 16 6.1 Research Outline ...... 16 6.2 Aims of Project...... 19 6.3 Thesis Structure ...... 20 7.0 LITERATURE SURVEY ...... 20 7.1 Scope of Review...... 20 7.2 General Actinide/Lanthanide Chemistry ...... 21 7.2.1 Lanthanide Chemistry ...... 21 7.2.2 Actinide Chemistry ...... 23 7.3 Alpha, Beta and Gamma Decay ...... 24 7.4 Nuclear Forensics ...... 25 7.5 Current Actinide Separation Technology ...... 27 7.5.1 General Principles of Actinide Separation ...... 27 7.5.2 Current Actinide Separation Processes ...... 28 7.5.2.1 Rapid Radiochemical Separations ...... 31 7.5.3 Methods of Actinide/ Lanthanide Separation ...... 32 7.5.4 Separation of Americium...... 35 7.5.5 Soft Anionic Donor Ligands ...... 38 8.0 References ...... 43 9.0 Chapter 1 – Rapid Am/Cm separation from nuclear forensic matricies using four triazine ligands ...... 48 9.1 Abstract ...... 50 9.2 Key Words: ...... 50 9.3 Introduction ...... 51

9.4 Experimental ...... 53 9.41 Ligand Synthesis...... 54 9.42 Am/Cm Separation from Normal Nuclear Forensics Matrix ...... 54 9.5 Results and Discussion ...... 56 9.51 Ligand Selection ...... 56 9.52 Application of Ligands to Am/Cm Extraction from Simulated Nuclear Forensic Matrices ...... 57 9.53 Am/Cm Extraction from Simulated Nuclear Forensic Matrices in the Presence of Fe, Ca and Al ...... 60

9.54 Am/Cm Extraction from Simulated Nuclear Forensic Matrices by CyMe4BTPhen without Prior Chromatography ...... 63 9.6 Conclusions ...... 65 9.7 Acknowledgements ...... 65 9.8 Notes and References ...... 66 10.0 Appendix to Paper One ...... 68 10.1 Method implementation with UTEVA, TEVA and TEVA-TRU and comparison with DGA resin method for Am (III)/Ln (III) ...... 68 10.2 Resin investigation results ...... 71 10.3 Investigating the effect on the addition of relevant fission and activation products on the BTPhen separation from a typical nuclear forensic matrix ...... 73 10.4 BTPhen separations with Co, Pb and Tc ...... 75 10.5 Appendix 1 References ...... 77 11.0 Chapter 2 - Separation of Am from complex mixtures using a BTPhen extraction chromatography resin...... 79 11.1 Abstract ...... 81 11.2 Introduction ...... 82 11.3 Experimental ...... 84 11.31 General Experimental Detail ...... 84 11.32 Synthesis of PVB-Me4BTPhen ...... 85

11.33 Conjugation of Me4BTPhen derivative 9 to PVBC ...... 86 11.34 Am/Eu separation method ...... 87 11.35 Generation of Complex Nuclear Forensic Matrix ...... 87 11.4 Results and Discussion ...... 88

11.41 Synthesis of PVB- Me4BTPhen ...... 88

11.42 Initial development of americium extraction using PVB-Me4BTPhen ...... 91

11.43 Americium extraction from nuclear forensic matrix using PVB-Me4BTPhen...... 93 11.44 Variation of acidity and contact time in americium extraction from nuclear forensic matrix ...... 95 11.5 Conclusions ...... 97 11.6 Acknowledgements ...... 98 11.7 Notes and References ...... 98 12.0 Appendix to Paper 2...... 100 12.1 SEM-EDX of polymer 2...... 100 12.2 SSNMR ...... 102 12.3 Polymer Disc Work ...... 103 12.4 Polymer disk method development ...... 103 12.5 Adsorption versus covalently bound EC resins………………………………… 103

13.0 Chapter 3 – Am separation using extraction chromatography resins based on adsorped triazine ligands ...... 108 13.1 Abstract ...... 110 13.2 Key Words...... 110 13.3 Introduction ...... 111 13.4 Experimental ...... 113 13.41 Ligand Synthesis...... 114 13.42 Polymer adsorption method ...... 114 13.43 Initial Am/Eu separation method for triazine EC Resins ...... 114 13.44 General Am/Eu separation method for triazine EC Resins ...... 115 13.45 Generation of Nuclear Forensic/Complex Nuclear Matrix ...... 115 13.46 General column Am/Eu separation method for triazine EC Resins ...... 116 13.5 Results and Discussion ...... 117 13.51 Resin Preparation ...... 118 13.52 Resin loading investigation for AMB-BTPhen/BTBP ...... 120 13.53 Matrix elements separations using AMB-BTPhen/BTBP ...... 121

13.54 KD Data for Poly-AMB-BTPhen ...... 122

13.55 KD Data for Poly-AMB-BTBP ...... 124 13.6 Conclusions ...... 126 13.7 Acknowledgements ...... 127 13.8 Notes and References ...... 127

14.0 Chapter 4 – Synthesis of BTPhen derivatives; effects on solubiliity and Am extraction ...... 130 14.1 Abstract ...... 132 14.2 Introduction ...... 132 14.3 Experimental ...... 132 14.31 Ligand Synthesis...... 136 14.32 Solubility screening of BTPhen Ligands ...... 136 14.33 Am/Eu Separations with BTPhen Ligands ...... 136 14.34 Complex matrix separations with benzil-BTPhen ligands ...... 137 14.35 Generation of Complex Nuclear Matrix ...... 137 14.4 Results and Discussion ...... 137 14.41 Ligand Synthesis...... 137 14.42 Solubility Testing in Common Reprocessing Solvents ...... 138 14.43 Acid Stability of Functionalised BTPhens ...... 141 14.44 Am/Eu Separations by Functionalised BTPhens ...... 142 14.45 Complex Nuclear Matrix Separations with Benzil-BTPhens ...... 145 14.5 Modelling of [AmL]3+ complexes of benzil-BTPhen ligands ...... 147 14.6 Conclusions ...... 148 14.7 Acknowledgements ...... 148 14.8 Notes and references ...... 149 15.0 Appendix to section 4 ...... 151

15.1 Complex Matrix SFAm/Eu calculated results (by gamma spectroscopy) ...... 152 15.2 Triazine ligand experimental data ...... 152 16.0 Conclusions and recommendations for future work ...... 158 17.0 Methodology Section ...... 162 17.1 Analysis Techniques Applicable to Americium Separation ...... 162 17.17 References ...... 166

1.1 Table of Figures and Schemes from all chapters

Chapter 1 Figure 1. AWE americium separation process (Red = Point at which BTPhen separation of Am may be applied)...... 18 Figure 2. An example AWE americium separation process using BTPhen...... 19 Figure 3. The f orbitals...... 21 Figure 4. Ligands used for actinide separations...... 29 Figure 5. Types of extraction...... 34 Figure 6. Calculation of separation factor...... 34 Figure 7. Synthesis of americium...... 35 Figure 8. Ligands used for americium extraction...... 42

Chapter 2 Figure 1. Example of Am separation for a nuclear debris sample...... 52 Figure 2. Structures of the ligands used in this comparison study...... 53 Figure 3. Method for extraction of Am/Cm from complex nuclear forensic sample using

CyMe4BTPhen (this work)...... 59

Chapter 3 Figure 1 TEVA-TRU-BTPhen separation scheme ...... 69 Figure 2 TEVA-TRU-DGA Separation Scheme ...... 70 Figure 3 UTEVA-BTPhen separation scheme ...... 70 Figure 4 TEVA-BTPhen separation scheme...... 71 Figure 5. Recovery of Ln matrix elements using EC resin separations schemes involving

CyMe4BTPhen...... 73 Figure 6. Potential β/γ and activation products that may require analysis in a nuclear forensic investigation...... 74 Figure 7. Recovery of relevant fission/neutron activation products from BTPhen matrix separation from a typical nuclear forensic matrix ...... 75 Figure 8. Matrix element recovery from our devised BTPhen separation method represented (with errors) in a Periodic Table (without noble gases). All isotopes apart from U, Pu and Np were 1 mg mL-1 and analysed by ICP-MS. U, Pu and Np were determined after AG1-X8 separation and alpha spectrometry analysis...... 78 Figure 8. Matrix element recovery from our devised BTPhen separation method represented (with errors) in a Periodic Table (without noble gases). All isotopes apart from U, Pu and Np

6 were 1 mg mL-1 and analysed by ICP-MS. U, Pu and Np were determined after AG1-X8 separation and alpha spectrometry analysis...... 78

Figure 1. Structures of CyMe4BTPhen 1 and PVB-Me4BTPhen 2...... 84

Figure 2. Recovery of matrix elements after PVB-Me4BTPhen extraction (by ICP-MS)...... 93

Figure 3. Variation of Am (III) and Eu (III) extraction by PVB-Me4BTPhen with aqueous phase (HCl) acidity ...... 95

Figure 4. Variation of Am (III) and Eu (III) extraction by PVB-Me4BTPhen with contact time (N.B. experiments run in triplicate, error calculated from instrument and method error plus one sigma.) ...... 96

Figure 1. Representative image of PVB-Me4BTPhen 2 ...... 100 Figure 2. EDX image of polymer 2 with the BSE detector...... 101

Figure 1. Alpha spectrum from 10 Bq Am/Eu separation by Poly-Aniline-Me4-BTPhen bonded to an alpha planchette...... 105 Figure 1. Structures of the ligand 1, 2 and 3 used to synthesise the EC resins used in this study ...... 113

Figure 1: Acid dependencies (KD) of the uptake of Am (III), Ln (III) and FP/Matrix elements on AMB-BTPhen at 0.1-5 M HNO3, values determined by ICP-MS and gamma spectroscopy (no errors >2.0% from raw data)...... 122

Figure 2: Acid dependencies (KD) of the uptake of Ln (III) and FP/Matrix elements on AMB- BTPhen at 0.1-5 M HNO3, values determined by ICP-MS and gamma spectroscopy (no errors >2.0% from raw data)...... 123

Figure 3: Acid dependencies (KD) of the uptake of Am (III), Ln (III) and FP/Matrix elements on AMB-BTBP at 0.1-5 M HNO3, values determined by ICP-MS and gamma spectroscopy (no errors >2.0% from raw data)...... 124

Figure 4: Acid dependencies (KD) of the uptake of Ln (III) and FP/Matrix elements on AMB- BTBP at 0.1-5 M HNO3, values determined by ICP-MS and gamma spectroscopy (no errors >2.0% from raw data)...... 125

Scheme 1. Synthesis of istain-BTPhens. Reagents: (i) diketone, Et3N, THF, Δ, <72 h...... 137 Scheme 2. Synthesis of benzil-BTPhens. Reagents: (i) diketone, Et3N, THF, Δ, <72 h...... 138

Scheme 3: Synthesis of aliphatic (and indan) BTPhens. Reagents: (i) diketone, Et3N, THF, Δ, <72 h...... 138

1.2 Table of Tables Chapter 1

Table 1: Recoveries of eluted matrix elements from AG1-X8 column chromatography from analysis by alpha spectrometry and ICP-MS...... 58 Table 2: Recoveries of eluted matrix elements for AG1-X8 column chromatography/triazine ligand separation from analysis by alpha spectrometry, gamma spectroscopy and ICP-MS.. 60 Table 3a and 3b: Recoveries of eluted matrix elements for high Al and high Ca from AG1-X8 column chromatography from analysis by alpha spectrometry, gamma spectroscopy and ICP- MS...... 62

Table 4: Recoveries of eluted matrix elements for high Al/Ca/Fe initial CyMe4BTPhen (10 mM, 90 min) separation of nuclear forensic matrix, analysis by alpha spectrometry, gamma spectroscopy and ICP-MS...... 64 Table 1. Am/Eu recoveries (by gamma spectroscopy) for the Am and Ln fractions from each separation method, N/D is given as a value if the recovery was determined to be less than 1% and the error being larger than or equal to the net area of the peak...... 72 Table 2. Pu, Np and U recoveries calculated by alpha spectrometry for the four separation methods with associated errors ...... 72 Table 3: Pb and Co recoveries from a BTPhen Am separation determined by ICP-MS), Tc recovery determined by liquid scintillation counting of aqueous phase before and after separation with 10 mM CyMe4BTPhen...... 76 Chapter 2

Table 1. Recoveries of Am (III)/Eu (III) from PVB-Me4BTPhen 2 extraction of Am (III)/Eu (III) matrix (N.B. experiments run in triplicate, error calculated from instrument and method error plus one sigma.) ...... 92

Table 2. Recoveries of Am (III)/Eu (III) from PVB-Me4BTPhen 2 extraction of nuclear forensic matrix (N.B. experiments run in triplicate, error calculated from instrument and method error plus one sigma.) ...... 94

Table 3. Recoveries of U, Np, Pu after PVB-Me4BTPhen 2 extraction and AG1-X8 chromatography of nuclear forensic matrix , determined by alpha spectrometry (N.B. error calculated from instrument and method error plus one sigma.)...... 94 Chapter 3

Table 1: Preliminary amberlite polymer Am/Eu separation data, determined by gamma spectroscopy against a known activity standard in the same geometry ...... 119 Table 2: Am/Eu recoveries from w/w loading AMB-BTPhen/BTBP 15 min polymer separations (after back stripping) determined by gamma spectroscopy ...... 120

Chapter 4

Table 1. Solubilities of isatin- (and indan) BTPhens in industrially relevant solvents ...... 139 Table 2: Solubilities of benzil-BTPhens in industrially relevant solvents...... 140 Table 3: Solubilities of aliphatic BTPhens in industrially relevant solvents...... 141

Table 4: SFAm/Eu for isatin-BTPhens, determined by gamma spectroscopy...... 142

Table 5: SFAm/Eu for aliphatic BTPhens, determined by gamma spectroscopy...... 143

Table 6: SFAm/Eu for benzil-BTPhens, determined by gamma spectroscopy...... 144 Table 7: Percentage recovery for matrix elements from separation of complex matrix using benzil-BTPhen, determined by ICP-MS (samples run in triplicate, error calculated from instrument error plus one sigma)...... 146 Table 8: Calculated additional electron population on triazine nitrogens calculated by DFT (Gaussian 09, B3LYP) ...... 148

Table 1. SFAm/Eu from separation of complex matrix using benzil-BTPhen ligands ...... 152 1.3 LIST OF ABBREVIATIONS

α-HIBA α-Hydroxyisobutyric acid AWE Atomic Weapons Establishment BTBP Bis-triazinylbipyridines BTP Bis-triazinylpyridines BTPhen 2,9-Bis(5,5,8,8-tetramethyl-5,6,7,8-tetrahydro-1,2,4-benzotriazin-3-yl)- 1,10-phenanthroline (CyMe4BTPhen) CA-BTP Bis-2,6-(5,6,7,8-tetrahydro-5,9,9-trimethyl-5,8-methano- 1,2,4-benzotriazin-3-yl)pyridine CA-BTPhen 2,9-Bis(5,9,9-trimethyl-5,6,7,8-tetrahydro-5,8- methanobenzo[e][1,2,4]triazin-3-yl)-1,10-phenanthroline CHON Carbon, hydrogen, oxygen, nitrogen CMPO Carbamoyl methyl phosphine oxide DEHSO Di-(2-ethylhexyl)sulfoxide DIAMEX Diamine extraction DOSY Diffusion ordered NMR spectroscopy DOWEX Dow® Extraction DTPA Diethylenetriamine-N,N,N',N'',N''-pentaacetic acid EDTA Ethylenediaminetetraacetic acid EXAFS Extended X-ray absorption fine structure GANEX Group actinide extraction HDEHP Di-(2-ethylhexyl)phosphoric acid HLLW High level liquid (radioactive) waste HPLC High pressure/high performance liquid chromatography ICP-MS/OES Inductively coupled plasma mass spectrometry/optical emission spectroscopy MALDI Matrix assisted laser desorption ionisation NMR Nuclear magnetic resonance n-PrBTP 2,6-Di(5,6-dipropyl-1,2,4-triazin-3-yl)pyridine PUREX Plutonium uranium extraction

PVBC Polyvinylbenzyl chloride PVC Polyvinyl chloride RTIL Room temperature ionic liquids SANEX Selective actinide extraction SIMS Secondary ion mass spectrometry TALSPEAK Trivalent actinide-lanthanide separation by phosphorus reagent extraction from aqueous complexes TBDMS Tert-butyldimethylsilyl TBP Tri-n-butyl phosphate TERPY Terpyridine TODGA Tetraoctyl diglycolamide TPTZ 2,4,6-Tripyridyl-s-triazine TRLFS Time-resolved laser fluorescence spectroscopy TRPO Tertiary R group phosphine oxide TRUEX Trans-uranium elements extraction UTEVA Uranium tetravalent extraction UV-VIS-NIR Ultra-violet-visible-near infra red spectroscopy XRD X-ray diffraction

2.0 Abstract

In analysis of complex nuclear forensic matrices containing lanthanides, actinides and matrix elements, rapid selective extraction of Am/Cm for quantification is challenging due to the difficult separation of Am/Cm from lanthanides. This project attempts to develop novel separation processes for Am/Cm separation utilising liquid-liquid extraction and extraction chromatography based on soft N-donor triazine extractants. Selective extractants were identified and synthesized and a liquid-liquid separation procedure was developed. Of these ligands, CyMe4-BTPhen, CyMe4-BTBP, CA-BTP and CA-BTPhen were compared for application to complex matrices. The developed process allows for purification and quantification of Am and Cm (recoveries 80–100%) and other major actinides in <2 days without the use of multiple columns or thiocyanate, yielding a full data set. The process developed was shown to be unaffected by Ca/Fe/Al (10 mg mL-1) and thus requires little pre- treatment of samples.

Due to limited availability of media for separation of americium using EC, we synthesized and tested a novel covalently-linked EC resin, utilising a triazine soft N-donor

(Me4BTPhen) extractant for americium extraction. The resin was generated by conjugation of a Me4BTPhen derivative with poly(vinylbenzyl) chloride to generate PVB-Me4BTPhen.

PVB-Me4BTPhen was shown to extract americium from a complex matrix simulating nuclear forensic samples, and containing lanthanides, actinides and matrix elements with high Am

(III) recovery (>90%) and low extraction of other elements, and provides an alternative separation process for Am (III) extraction. Adsorption was also investigated as an alternative, more flexible approach to resin preparation. BTBP/BTPhen Am selective triazine ligands were adsorbed onto Amberlite XAD-7, then characterised and tested for Am/Eu selectivity, complexation kinetics and polymer loading. These polymers were tested with complex

12 matrices in conjunction with AG1-X8 anion exchange chromatography to achieve a complete isotope separation and quantification method. From these results, the resin capacity factor

(KD) as a function of HNO3 concentration was calculated, allowing potential separation methods to be designed.

Selective americium extractants from the BTPhen ligand family have been identified through this work, and we also report a study of functionalisation of BTPhen ligands to help design new selective Am extractants by determining the effects on solubilities and americium extraction capabilities of variations in substituents. The data obtained show trends that could assist in future ligand design.

3.0 Declaration

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

4.0 Copyright Statement

i. The author of this thesis (including any appendices and/or schedules to this thesis) owns

certain copyright or related rights in it (the “Copyright”) and s/he has given The

University of Manchester certain rights to use such Copyright, including for administrative

purposes.

ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy,

may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as

amended) and regulations issued under it or, where appropriate, in accordance with

licensing agreements which the University has from time to time. This page must form

part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trade marks and other intellectual

property (the “Intellectual Property”) and any reproductions of copyright works in the

thesis, for example graphs and tables (“Reproductions”), which may be described in this

thesis, may not be owned by the author and may be owned by third parties. Such

Intellectual Property and Reproductions cannot and must not be made available for use

without the prior written permission of the owner(s) of the relevant Intellectual Property

and/or Reproductions. iv. Further information on the conditions under which disclosure, publication and

commercialisation of this thesis, the Copyright and any Intellectual Property and/or

Reproductions described in it may take place is available in the University IP Policy (see

http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=487), in any relevant Thesis

restriction declarations deposited in the University Library, The University Library’s

regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The

University’s policy on Presentation of Theses.

5.0 Acknowledgments

Thanks go to AWE for the funding of this project

Many thanks go to Sarah and Francis for being great supervisors and for all the help and

guidance throughout the PhD. I also wish to thank them for choosing me to undertake this work and for making the last few years a lot of fun and letting me travel to some lovely parts

of the world.

Lots of appreciation and thanks are owed to all the people from AWE who have helped

support and guide this work. In particular I want to thank Olivia, Paul, Andy and Jun for

making my times in Reading feel very welcoming and for all the help with the project.

Thanks go to my all my family and personal friends, especially Faith, Alan, Steve, Charlene,

Simon, John and Margaret for their continued love and support over the years.

I want to thank everyone past and present in the CRR and the ‘Heath Group’ and in the small office for all the wonderful times we spent together, especially at football on Monday nights

and in all the impromptu coffee breaks. There are too many of you for me to individually

mention everyone but thanks to you all for making the CRR an enjoyable place to work.

6.0 OUTLINE AND AIMS OF PROJECT

6.1 Research Outline

Radiochemical separations are required to in order to purify radioisotopes, and hence are essential for nuclear science. Without radiochemical separation technology, all applications of nuclear isotopes in areas such as energy, medicine and military applications would not be viable.

This research project attempts to develop new methods for the separation and quantification of americium from matrices containing numerous radioisotopes. Most current separation techniques for americium are complex, as americium is usually found with lanthanide fission products which have similar chemical properties (and with stable elements, uranium, curium and/or plutonium). Thus, most separations involve laborious extraction processes in order to remove all other elements and isotopes, and isolate a pure americium fraction. The Atomic Weapons Establishment (AWE) supports the UK’s wider nuclear security programmes. AWE’s role in maintaining the UK’s capability in nuclear forensics, which supports national and international investigations and treaties, most notably the

Nuclear Non-Proliferation Treaty and Comprehensive Nuclear Test-Ban Treaty is also relevant to this research project.

The radiochemical analysis of complex matrices is a challenging task, and the rapid measurement of americium in such samples is essential for nuclear forensics. Gamma spectroscopy cannot be used to analyse these samples fully due to the complex gamma spectrum and relative insensitivity of the technique, so separation is essential. A process for rapid selective extraction of Am which does not interfere with the recovery of other elements is desired and the main challenge in extraction of pure americium from nuclear forensic samples is its separation from the lanthanides. Quicker separation processes would also allow information to be obtained before any short lived isotopes present in the material decay below 16

detectable limits, and allow a full data set to be obtained quickly. Such methods would improve nuclear forensic capability, but development of these methods will require research into novel separation systems applicable to complex mixtures of nuclear materials. Currently, the lengthy separations techniques for americium and curium are the limiting factor in obtaining (in a timely manner) a full isotopic data set for an unknown sample. The aim of this research project is to develop a procedure that fulfils AWE’s requirements for the separation and quantification of americium from complex matrices to obtain americium data within 2 days. This separation will then be integrated with the overall separation procedure provided by AWE (Figure 1). This project will use ligands which are able to separate americium from lanthanides in other contexts such as nuclear fuel recycle.1 Of particular interest for this work is the N-donor heterocyclic ligand CyMe4BTPhen (hereafter referred to as BTPhen), as this has been shown to have a separation factor >250 for americium over europium2 and can be incinerated/recycled after use. Four points in the AWE separation procedure at which the matrix is aqueous and may be contacted with an organic solution, and where BTPhen separation may offer a time advantage, have been identified as potentially suitable for americium extraction with BTPhen.

Sample in 4M HCl containing Be, Sr, Y, Zr, Mo, Ag, Cd, Cs, Ba, Ce, Lanthanides, U, Np, Pu, Am & Cm

Convert to nitrates, precipitate Sr & Ba Supernate nitrates with fuming HNO3

Insoluble material Be, Y, Zr, Mo, Ag, Cd, Cs, Ce Precipitate Sr & Ba Lanthanides, U, Np, (Pu), (Am) & (Cm)

Convert to chlorides & dissolve in conc HCl Dissolve in water 1. Elute Be Y, Ag, Cs, Ce, Lanthanides, Am & Cm

Solution Sr & Ba with conc HCl (Pu/Am separated) 2. Elute Pu with conc HCl/ 1% HI (Np/Pu separated) Anion

Exchange Combine both fractions 3. Elute Zr & Np with 4 M HCl 4. Elute U & Mo with both 1 M & 0.05 M HCl Precipitate hydroxides using ammonia 5. Elute Ag & Cd with water & 1 M Ammonia

Precipitate contains Be, Y, Ce, Dissolution 2M HNO3 Precipitate Dissolution HNO3, H3BO3 Supernate contains Ag, Lanthanides, Am & Cm contains Y, Ce, Cd & Cs Reduced Fluoride Lanthanides, Am Fluoride-hydroxide cycle precipitation & Cm Precipitate Iodate (Ce/Pr Precipitate using Supernate contains Be separated) H2S

Lanthanides, Y Supernate Precipitate α-HIBA contains Cs contains Ag Cm Y, Lanthanides, Precipitate Separation & Cd Am & Cm Ce Am 18 Figure 1. AWE americium separation process (Red = Point at which BTPhen separation of Am may be applied).

6.2 Aims of Project

 The primary project aim is to improve capability by developing a process that

separates americium (or co-extracts americium and curium) selectively from nuclear

forensic samples with a specified decontamination factor faster than the current 7 day

method (Figure 1).

 There is also a subsidiary aim to improve AWE’s matrix separation process (Figure

1), as this does not currently utilise cartridge columns or new selective resins

applicable to radiochemical separations.

Specifically, the process will initially utilise the ligand BTPhen for americium extraction,

while other ligands which have shown high selectivity for americium will also be tested.

A possible americium separation process using BTPhen extraction and EC resins is

shown in Figure 2.

Sample in 4M HCl containing Be, Sr, Y, Zr, Mo, Ag, Cd, Cs, Ba, Ce, Lanthanides, U, Np, Pu, Am & Cm

Multiple contact with BTPhen in 1-octanol to give decontamination factor for Am >1000 by gamma spectroscopy

Contact organic phase with 0.1 M HCl Matrix with Am removed

1. 6 M HNO3, 0.3% H2O2 (An (III), Ln(III), Strontium fission products separated) extraction with 18- Chromatographic 2. 2 M HNO , 2 mM ascorbic acid, 2 mM crown-6 3 Resin (UTEVA) NH2OH (Pu separated) conditioned with

6 M HNO3, 0.3% 3. 2 M HNO3, 0.1 mM oxalic acid (Np H2O2 Sr An(III), separated) Ln(III),

fission 4. 7 mM (NH4)C2O4 (U separated) products

Figure 2. An example AWE americium separation process using BTPhen. 19

6.3 Thesis Structure

This thesis has been prepared in the alternative format. It contains four chapters which are either published in the scientific literature or are accepted at a relevant scientific journal. The first chapter outlines chemistry relevant to, and separation techniques for analytical Am/Ln separations. Chapter two describes synthesis of Am selective triazine ligands and application to complex nuclear forensic mixtures to develop rapid Am/Cm separation methods in liquid- liquid separations. Chapter 3 outlines the synthesis, characterisation and application of a covalently bound, triazine based EC resin. The resin was screened against a range of radionuclides and stable elements to develop an Am(III)/Ln(III) separation method from complex matrices. The fourth chapter discusses an alternative, adsorption based approach to generating triazine based, Am/Cm selective EC resins for rapid separations of

Am(III)/Ln(III). The fifth chapter describes the functionalization of a set of BTPhen ligands to probe the effects of solubility, acidic stability and Am extraction and hence aid future design of selective Am(III)/Cm(III) extractants for both analytical and industrial applications.

The final chapter draws overall conclusions from this work and suggests future work.

7.0 LITERATURE SURVEY

7.1 Scope of Review

This review will include a summary of actinide and lanthanide chemistry applicable to this project as well as a brief overview of nuclear forensics. The review will then provide a critical analysis of the currently available Am separation capabilities.

7.2 General Actinide/Lanthanide Chemistry

7.2.1 Lanthanide Chemistry

The lanthanides and actinides are the two groups of elements that have electrons in f-orbitals.

Lanthanide chemistry is defined by the f-orbital electrons, with the f-orbitals having high orbital confinement and are therefore highly directional (Figure 3). Furthermore the energy and radial extent of the f-orbitals relative to the 6s/7s and 5p/6p orbitals means they are less likely to interact as valence electrons; rather, they shield the nuclear charge less and are better considered core orbitals. This means that the f-elements are more likely to lose both s and p electrons demonstrating the importance of the +3 oxidation state.12

Figure 3. The ‘general set’ of f orbitals.

Lanthanides are often found together in mixtures and occur in many minerals (for example cerite and bastnäsite). They are very difficult to separate from each other as they have similar ionic radii for a given charge (usually 3+) and very limited access to other oxidation states.

Many lanthanide isotopes are formed as fission or neutron activation products in nuclear reactions. The alternate name of ‘rare earths’ reflects the difficulty of separating the lanthanides from each other although, ironically, they actually have a relatively high 21

abundance in the Earth’s crust. As expected, lanthanide isotopes which have an even proton number and even neutron number are more abundant in Nature than those with even-odd or odd-odd numbers.3

Variation in the chemistry of the lanthanides is attributed to a slight decrease in ionic radius across the series. Exceptions are generally due to half or fully filled orbitals which give rise to +2 or +4 oxidation states, for example europium exhibits a stable +2 oxidation state as this has a half filled valence orbital. The principal industrial scale method of separating the lanthanides from each other is to use extremely long cation exchange separation columns, exploiting the linear decrease of complexation with chelating ligands such as EDTA across the series to elute the lanthanides individually.4 This trend is also utilised in anion exchange for separation of lanthanide/actinide mixtures.

The Ln3+ ions form strong complexes with hard donor ligands which are bonded electrostatically. Lanthanides also have high co-ordination numbers in solution due to their size and relatively high charge (+3) and form fewer organometallic complexes than seen for d-block metals. The lanthanides form weakly coloured solutions due to Laporte forbidden f-f optical transitions.

Lanthanides can have up to seven unpaired f-electrons and hence exhibit some of the largest magnetic moments amongst the elements. Lanthanides are commonly used in misch metal for igniters, as catalysts in the petrochemicals industry and as dopants in specialised glass for use in televisions and lenses. Current research aims to develop single molecule magnets from the lanthanides in complexes with optimal geometries to give slow relaxation of magnetisation at non-cryogenic temperatures.5

7.2.2 Actinide Chemistry

The actinides, particularly uranium, neptunium and plutonium, show more varied chemistry than the lanthanides. Early actinides (thorium to neptunium) show similar chemistry to the d- block metals and the later actinides (americium to lawrencium) can be considered as analogous to their lanthanide homologues and are redox stable.

The actinides have heavy, unstable nuclei and therefore exhibit numerous modes of radioactive decay. This arises from unstable nuclear configurations; specifically alpha decay and the associated gamma rays can be used to fingerprint the isotope if it is of measurable activity. Post uranium, the elements are essentially synthetic and are produced through the nuclear fuel cycle or atom by atom (resulting in mixtures of actinides). Their electronic configuration is not easy to confirm and is usually dynamic;6 further, the energies and radial extent of the 5f, 6d, 7s and 7d orbitals are similar in the early actinides, meaning electrons can move easily between orbitals to give many possible modes of bonding. The early actinides exhibit multiple possible stable oxidation states (from +2 to +7). Their electropositive nature means they are reactive, and study of their chemistry requires consideration of pH, hydrolysis, side reaction products, dry solvents and use of inert atmospheres. Early actinides show interesting co-ordination chemistry due to their size and varying oxidation states.

Complexation with numerous ligands has yielded many novel crystal structures showing new types of bonding interactions and unexpected oxidation states.7,8

Pure metallic actinides are generally pyrophoric and react with oxygen to form stable

+/2+ actinyl species (AnO2 ) in aqueous solution. The early actinides are easier to separate from each other as different oxidation states can be exploited. However, actinides also have complex electrochemistry and will disproportionate and hydrolyse readily in solution which can complicate separation. Actinides can form multiple allotropes depending on the

temperature and chemical environment, resulting in complexes with variable symmetry, co- ordination and useful properties in the solid state.

The 5f orbital energy decreases as Z increases across the actinide series meaning the f orbitals contract and become core, causing the +3 oxidation state to dominate for americium and beyond. This means that the later actinides are very difficult to separate from the 3+ oxidation state lanthanides and hence more difficult to obtain in a pure form than early actinides.

7.3 Radioactive Decay

Radioactive decay can be split into two general modes (alpha and beta) with the excess energy released in a gamma ray photon. Alpha decay is the release of a two proton, two neutron helium nucleus with a specific energy equal to the energy gap between parent and daughter nuclei. Alpha particles are the most energetic type of radiation, but due to their size they dissipate their energy easily to their surroundings. Alpha decay is characteristic of heavy nuclei. The particle energy is quantised and excess energy is released as one or more quantised gamma photons. These specific energies both allow identification of specific radioactive isotopes.

Beta decay can be split into three modes: negatron decay, positron decay and electron capture. Negatron decay is the release of an antineutrino and an electron, in which the electron energy can range up to an Emax value (maximum energy of the electron from the specific transition), the remainder of the energy belonging to the antineutrino. Positron decay is the opposite process that takes place with the emission of a positron and a neutrino.

Electron capture is the third process, where an orbital electron and nuclear proton combine to form a neutron in the atom. These phenomena caused by the instability of the nuclei have allowed many analytical methods to be devised to help in the identification of the actinides.

The energy balance is maintained by the emission of gamma rays. There are pure gamma emitters, these are usually metastable states such as technetium (Tc-99 metastable) caused by the excited state generated being separated in time from the gamma emission.

7.4 Nuclear Forensics

Nuclear forensics is a branch of nuclear science that aims to carry out the legal and judicial process to determine the origin and intended use of radioactive materials out of regulatory control. It uses a combination of radiochemistry, analytical chemistry and traditional forensics to accomplish its aims.

Currently there is a need to be able to respond to any event involving complex radioactive materials as quickly as possible. Nuclear forensics aims to optimise the process of obtaining the evidence needed to make an informed decision about an event. Several case studies have been published of investigations.9,10,11,12

Mayer et al. provide a review of the techniques and the methodology that has been developed for nuclear forensics.13 Nuclear forensic methodology has to work within the timeframe of the legal process, e.g. 24 hours to hold a suspect, 1 week to charge a suspect and

2 months to arrange a trial (the times quoted are example timescales that have been used in

International Exercises). Nuclear forensics aims to build up a body of evidence iteratively to draw conclusions about the material based on databases of real and modelled data.

Furthermore, the physical properties of the material determine the relevant analytical techniques that can be applied. Every point of processing of a material can leave indicators of its origin and the techniques applied can be drawn from many scientific fields.

The initial 24 hours are used to categorise whether the material is weapon, reactor or medical grade, and this will determine the investigation path.14 The main focus in the initial steps is the use of non-destructive techniques for hazard evaluation and maximum sample

retention; identification of the nuclear isotopes and their activity will usually be prioritised.

Hence, gamma spectroscopy is first used on an unprocessed sample still within containment

(to minimise any unknown alpha activity) to identify all significant gamma emitting radionuclides.

Usually alpha spectrometry will then be used for determination of the alpha emitters and, when combined with gamma spectroscopy, will give information on almost all radioisotopes present. Beyond this point in the investigation, many complementary techniques including mass spectroscopy and physical characterisation are used.15 These aim to give the important isotope ratios for determining information of the sample origin as well as forensic information on the sample. Determining these isotope ratios requires radiochemical separations of important nuclides (U, Pu, Am, Np) which is another driver in the renewed interest into actinide separations. Hot particles may be analysed using Scanning

Electron Microscopy (SEM), Surface Ionisation Mass Spectrometry (SIMS) and X-Ray

Diffraction (XRD).

The need to determine the origin of a seized material is the main focus of nuclear forensics and the associated development of rapid radiochemical separations. Radiochemical separations are currently the lengthiest part of the investigation, and modelling of the data can only be undertaken after analysis of pure isotopic solutions obtained from separations (or from physical analysis of the material). Ideally, the ultimate aim in nuclear forensics would be a process in which no radiochemical purification of the sample is necessary and all required data to model the material could be obtained using rapid analytical techniques requiring little to no sample preparation.

7.5 Current Actinide Separation Technology

7.5.1 General Principles of Actinide Separation

Separations technology for the actinides can be divided into those applicable to the industrial or the analytical scale. The requirements for separation in a given system can vary widely, but separations ultimately require ligands which have high selectivity for specific ions.

Processes developed in the 1950s still dominate separation technology used for industrial nuclear applications worldwide. The need to develop familiarity with, and gain regulatory approval of a process, the infrastructure requirements and high level of control desired when remote handling large amounts of radioactive materials can mean separation technology requires decades in development before industrial scale use. For example, the TBP-based

PUREX system is widely used as it is well established and meets most of the requirements of a large scale nuclear separation. A review by Nash covers many of the important requirements needed in separation technology.16

Separation techniques exploit a wide range of chemical principles. Actinides will generally be handled in an aqueous acidic medium on dissolution to avoid hydrolysis and other unwanted processes. Therefore, in a separation process actinides will exist as aqueous ions and any selective ligands must be stable in strongly acidic solutions and intense radiation fields.

An industrial scale separation requires highly specialised equipment and shielding as gram to kilogram quantities of irradiated fuel are lethally radiotoxic. A ligand for this type of separation need not have high extracting strength per unit amount because it may be recycled in an industrial process, but it should be selective for the target actinide(s). An entire separation process involves numerous reagents/ligands working together. Separation processes also need to be suitable for remote handling and continuous separation (i.e. predictable dissolution profile and chemistry). The process must allow addition of other 27

separating reagents and if possible utilise the same solvents throughout to decrease waste.

Importantly, any complexes formed in the separation must be labile to allow dissociation of the actinide(s) prior to the next separation stage, a reagent/ligand with slow reversible phase transfer is not desired.

Separations on the analytical scale have different requirements; Ekberg et al. discuss the criteria with examples for An/Ln separations.1 The ideal separation process on an analytical scale is one which completes a highly efficient separation quickly and with the lowest number of stages, as impurities at low concentration can greatly perturb results. Analytical separations aim to work on the smallest reasonable scale for safety reasons and to limit the amount of sample required. However, some desired isotopes are minor components and are therefore most easily obtained by removal of major components followed by further purification of the remaining mixture, which requires a larger amount of sample.

7.5.2 Current Industrial Actinide Separation Processes

The chemistry of the processes that govern actinide separations can be divided according to the chemistry of the specific functional groups of the ligands used in a separation process.

Ultimately the worldwide effort to develop new actinide separations to recycle fuel is an important research area. Many nations are working towards a closed fuel cycle where individual recovery of each element is possible, leading to many separation processes. An example of this is the French diamine extraction process (DIAMEX).17 This aims to separate the minor actinides specifically whilst avoiding phosphorus/sulphur based reagents as set out in their CHON strategy.18 Another important area of research is work toward a trans-uranium elements extraction (TRUEX) process from high level radioactive liquid waste (HLLW), which has been investigated by all nuclear nations, most notably the USA, in their search for a new system that could expand the PUREX process.19 This process aims to use carbamonyl

methyl phosphine oxide (CMPO, Figure 4) together with tri-n-butylphosphate (TBP) to extract all trans-uranium elements to one phase. Interest in the partitioning and transmutation of the later actinides has led to the development of the selective actinide (SANEX) process, which has been studied in China with a focus on sulfur based selective trans plutonium extractants20 This uses the soft S-donor ligand Cyanex 301 (Figure 4) to perform separations of americium/curium from HLLW. The possible limitation of this work is that sulfur/phosphorus based reagents are not as selective as the current soft N-donor ligands for americium/curium in analytical separations.21 Cyanex 301 has also been trialled in a simulated HLLW separation.22 Concern over toxic degradation products and formation of toxic waste, as well as the weaker extraction capabilities, encouraged German and French scientists to investigate an alternative n-SANEX process. This process separates actinides from lanthanides in HLLW using neutral N-donor extractants, notably the bis- triazinylpyridines (BTPs, Figure 4) and more recently the bis-triazinylbipyridines (BTBPs,

Figure 4) which are now the European benchmark ligand for the SANEX process. The USA is developing the TALSPEAK process, which was initiated in the 1970s using di-(2- ethylhexyl)phosphoric acid (HDEHP, Figure 4) as extractant and diethylenetriamine-

N,N,N',N'',N''-pentaacetic acid (DTPA, Figure 4) as the selective actinide complexing agent.

Other processes in development include the group actinide separation process (GANEX) which aims to separate out isotopes in HLLW using agents which are selective for multiple elements to allow the nuclear fuel cycle to be proliferation-safe. These research efforts, most notably the SANEX process, offer many potential candidates applicable to modernising the

AWE separation process.

Figure 4. Ligands used for actinide separations.

The most successful processes have used phosphorus reagents, for example TBP for extraction of uranium and plutonium. Other relevant functional groups include amides23 such as tetraoctyl diglycolamide (TODGA, Figure 4) which is used in the Japanese DIAMEX process. TBP has also been investigated as a possible extraction reagent for americium, notably by Weaver et al.24 They showed that extraction was possible in concentrated nitric acid, but the separation factor was low, and in a mixture of actinides this would be inhibited by plutonium and uranium extraction. TBP can however be used as a specific extractant with the use of synergists such as di-(2-ethylhexyl)sulfoxide (DEHSO, Figure 4). Notable approaches to using phosphorus-based ligands for americium extraction include the TRPO phosphine oxide ligand (Figure 4). Japanese research efforts have led to trials of the ligand

for americium separation from HLLW.25 After ten stages of separation, TRPO (in kerosene) was able to separate 99% of all relevant actinides (U, Pu, Np, Am). However, the need for synergists and the slow extraction kinetics mean this reagent is not applicable for americium separation within a 2 day timescale.

The most useful and adaptable separation technique applicable for the actinides is liquid-liquid separation. This process aims to control solvation of the actinide ion, complexion kinetics (to a selective ligand/s) and exploit the redox potentials of the ions to enable selective separation. It is fundamentally a mass transfer process, where multiple phases are contacted and, by altering variables, an equilibrium change is achieved to give a separation.19 Other methods that are receiving renewed or new attention are applicable to the next two generations of nuclear reactors. These methods still utilise fundamental actinide chemistry but the separation is undertaken in extreme media. Notable examples include pyrochemical processes (e.g. molten salt separations), supercritical fluids (most notably CO2), room temperature ionic liquids (RTIL) or methods based around volatilising mixtures of actinides.26 These currently do not lend themselves to a rapid separation procedure.

7.5.2.1 Rapid Radiochemical Separations

Rapid radiochemical separation techniques have been developed using specific extractants for a given radionuclide. Resins loaded with selective ligands in pre-manufactured columns that can be attached to each other has increased the potential throughput of radiochemical separations to the point that they are sometimes termed as ‘rapid’ radiochemical separations.

This was exemplified by Pilvvo et al. separating Th, U, Pu, Am, and Cm from one solution with the TRU resin.27 The actinides are eluted separately and relatively pure solutions of specific isotopes can be obtained. Theoretically this process could be scaled up for industrial use. The majority of the work in rapid radiochemical separations has come from the group of 31

Horwitz. Development of CMPO ligand derivatives for transuranium extraction allowed development of the TRU resin, which is applicable for rapid separations of Fe, Th, Pa, U, Np,

Pu, Am, and Cm.28 Horwitz and co-workers also developed the UTEVA resin which is applicable to Th, U, Np, Pu separations.29 The ligand in this column is diamyl aminephosphonate (DAAP) which forms nitrato complexes at varying nitric acid concentrations allowing retention and selective elution of actinides on the column. However, an americium and curium selective resin is an interesting gap in the range of currently available resins (used for separating U, Th, Pu and Np). This could be a useful resin in the nuclear forensic investigations where a rapid result for americium is desired. However with a complex radioactive sample, considerable pre-treatment and purification/multiple columns would be needed to control a separation using these resins. Furthermore no current extraction chromatography (EC) methods are selective for Am or can achieve individual Ln purification from complex lanthanide mixtures.

7.5.3 Methods of Actinide/ Lanthanide Separation

7.5.3.1 Precipitation

Historically the lanthanides were extracted from their ores from each other using precipitation methods. These methods have been used as final purification methods in large scale production of actinides, as shown in the work of Heckmann et al.30,31 However, these methods are no longer utilised in actinide separations on industrial scale. Precipitation techniques can be useful in aqueous media as it is possible to change the anion source and precipitate the desired actinide as a chloride, nitrate or oxalate complex. If hydroxide is present with no competing ligands, the actinide hydroxides ranging from +3 to +6 oxidation state (assuming the oxidation state is stable to disproportionation) can be readily precipitated as shown by Smith et al.32 32

7.5.3.2 Liquid-Liquid Separation

This type of separation is arguably the most flexible and is currently the only viable separation technology for industrial actinide/lanthanide separations. A comprehensive review of solvent extraction has been performed by Mackenzie.33 Fundamentally it involves separation of a highly concentrated aqueous phase (usually acidic due to leaching from ores) from a selective immiscible phase which has an affinity for one or more ions in the aqueous phase. The disadvantage of this type of separation is the potential to generate large volumes of highly toxic organic and aqueous waste. However, with counter-current separators and recycling of the organic extractant, the levels of waste can be minimised. Liquid-liquid separations are a mass transfer equilibrium process within a usually biphasic

(aqueous/organic) system. Adjustable factors include polarity of the solvent, redox potential, potential for hydrolysis, selectivity of reagents and lability of complex formed, and the presence of synergists. There are multiple types of equilibria in involved in these separations

(Figure 5).

Figure 5. Types of extraction.

The extent of separation is usually quantified by the metal ion distribution ratio, which is the ratio of the amount of complexed ion in the desired phase over the amount in the initial phase

(measured by techniques such as gamma spectroscopy). If this is known for all ions in the solution a separation factor can be determined which gives a measure of the selectivity of the ligand in a separation for one element over another (Figure 6).

Figure 6. Calculation of separation factor.

In liquid-liquid separation processes, a complex mixture of species is often present and requires elucidation in order to understand the process and ensure a robust separation. This usually involves many complementary analysis techniques; single crystal X-ray diffraction,

X-ray absorption spectroscopy, and modelling are usually the most important for probing speciation. Once the species required for the separation and the mechanism of separation have been established, this information can be used to help guide design of other ligand systems and the separation can be optimised by alteration of variables and further ligand design.19

7.5.4 Separation of Americium

Americium has historically been difficult to separate. It is a synthetic element, and is predominantly made as a by-product of nuclear power as a daughter of plutonium upon successive neutron captures (Figure 7).

Figure 7. Synthesis of Americium.

Americium was first isolated as the hydroxide Am(OH)3 from irradiated plutonium using co- precipitation. Seaborg and co-workers initially named the element delirium due to the challenge it represented to isolate. To this date 13 synthetic isotopes of americium are known,34 but only three have either the required abundance or half-life needed to find applications (241Am, 242m Am and 243Am). The other ten isotopes are not produced in quantity in reactors and/or decay with short half-lives to neptunium, plutonium or curium isotopes.35

The historical way of acquiring americium was via continuous purification of plutonium (e.g. 35

PUREX) followed by application of radiochemical separation methods (liquid/liquid, co- precipitation, column separation).

241Am is the most abundant isotope of americium. It is the beta decay daughter of

241Pu and has a half-life of 432.7 years. 241Am decays to 237Np with release of an alpha particle (5.486 MeV) together with 59.5 keV gamma ray (36% abundance).

The other two principal isotopes are 242Am (242Am and 242mAm) and 243Am, which are formed from 241Am upon successive neutron captures, or via deliberate neutron activation of

241Am or similar heavy targets (e.g. 242Pu). The majority of americium which is in use today is separated from spent nuclear fuel where Am makes up 0.03% and Cm 0.01% by mass of light water reactor spent fuel. 36 Despite this, the two represent over 90% of the long term radiotoxicity, justifying the development of rapid extractions of these ions.

Americium chemistry is generally well defined and ionic in character. Few comprehensive reviews are available; the review of Schultz (1976) is still widely used.37-38

The +3 oxidation state defines americium chemistry. Americium and the later actinides have limited redox chemistry compared to the early actinides, meaning Th, U and Pu can readily be separated from Am based on redox chemistry and sequestration by selective ligand systems, usually driven by favourable complexation energetics.

The solution chemistry of americium is generally well understood, and it is regarded as a hard Lewis acid in simple systems. The hydrolysis of americium has been widely studied and it is generally accepted that americium is complexed by hydroxide above pH 5.39 When in contact with organic complexants, high oxidation state americium tends to reduce to the +3 oxidation state, which can reduce the separation capability of organic extractants. The solution spectra of americium (e.g. UV-VIS-NIR) can be useful to identify the oxidation state of americium in solution.40 The americium co-ordination environment can be determined from luminescence spectra as well as techniques such as EXAFS and TRLFS.41 36

Americium is in a theoretically interesting position in the Periodic Table, being at the start of the later actinides yet available for study in reasonable amounts. Americium has found use both by theoreticians and experimentalists trying to understand the actinides, notably in the work of Johansen et al., who showed that americium’s electronic configuration is static (5f 7,7s2) under normal conditions with the 5f electrons being inert but, when subjected to pressure, the configuration and energy levels alter.42

Historically the most important complexes of americium are the oxalate and α- hydroisobutyric acid (α-HIBA) complex. Oxalate was the first organic ligand to show strong complexation with americium and was investigated by Markin,43 allowing selective precipitation of Am oxalate, which was later employed for radiochemical separations. When an improvement was desired in actinide separation, α-HIBA was shown to elute actinides

(especially americium) slowly and in a controlled manner, and thus to separate americium reliably from mixtures containing lanthanides/actinides. This process and all An/Ln (III) separations are extremely time consuming. The current AWE separation of americium

(Figure 1) still uses this reagent for the final purification of americium despite the age of the technique.44

Despite renewed interest in actinide co-ordination chemistry few americium compounds have been synthesised and characterised by XRD due to challenges in handling reasonable quantities of americium.19 Characterisation of americium-organic ligand complexes have not been widely reported, with only ten complexes characterised by XRD.

This may be due to difficulty in crystallisation of americium complexes and the fact that handling reasonable quantities of americium is extremely difficult due to its radiotoxicity. In most cases Eu3+ will be used as a structural homologue.45

Oxygen donor ligands, most notably carboxylate, have been used to separate americium from the lanthanides. Carboxylate is interesting in larger americium co-ordination environments 37

due to its ability to have multiple bonding modes. In addition, americium complexes with carboxylate, hydroxyl and other simple ligands can be used as a model for the interaction of actinides in the environment with humic acids.46

7.5.5 Soft Anionic Donor Ligands

Current research aims to develop selective soft donor ligands that take advantage of the enhanced covalency of the 5f orbitals to form actinide complexes preferentially to complexes of other metals. This research is mainly based in Europe, for example the EUROPART project which aims to develop advanced separation techniques for the next generation of nuclear reactors to close the fuel cycle. A comprehensive review of neutral N-donor ligands has been carried out by Geist and Panak so this section will focus only on the systems most directly relevant to the work discussed in this thesis.47

The processes being developed in EUROPART aim to separate out, for separate management, actinides that lead to long term radiotoxicity. Removal of americium and curium would reduce the time the waste needs to be stored in a repository from thousands to hundreds of years. A PUREX process followed by either a group actinide extraction process

(GANEX) process or a selective actinide extraction (SANEX) would allow control over all the actinide isotopes in the fuel cycle. These could then be used to provide stocks of americium, curium etc., or the materials could be destroyed by transmutation to shorter half- life species.

The enhanced covalency of the actinide 5f orbitals in complexes with soft N-donor ligands has been shown by Adam et al. using NMR experiments with a 15N labelled americium complex of n-PrBTP.48 This is just one of a number of soft donor ligands which have high affinity for the actinides over the lanthanides.49 If this distinction can be exploited

there is the potential to yield new separation systems that allow rapid and selective control over actinide separations.

Research into soft N- and S-donor ligands has yielded some promising ligands for use in actinide/lanthanide separations. This field has been reviewed by Ekberg et al.1 Development of soft N-donor ligands aims to improve on the use of small chelating molecules such as

EDTA or α-HIBA, as such molecules are not selective enough to carry out a rapid actinide/lanthanide separation.

Nitrogen donor ligands were first identified for their ability to separate metal ions; the

TERPY ligand (Figure 8) was one of the first N-donor ligands to be shown to form complexes with actinides and lanthanides. Despite being synthesised in the 1930s by Morgan et al.,50 it was not until the 1970s that the ligand was first applied to actinide/lanthanide separation. The pyridine unit is known to bind to trivalent cations, and was therefore desired in the design of subsequent soft N-donor ligands.

The first attempt to develop a more selective ligand for actinides was to adapt TERPY to allow use at lower pH. This was achieved by lowering the pKa of the ligand by increasing the size of the aromatic system to give the TPTZ ligand (Figure 8), which has a separation factor of 10 for Am/Eu compared to 7 for TERPY.51 TPTZ and TERPY could be combined with synergists to enhance separation. Any variations to the process gave the same increase in separation factor for both TERPY and TPTZ. Both ligands form 1:1 and 1:2 complexes in varying acidity with the actinides and lanthanides. Derivatives of these ligands were also synthesised, but most were not utilised due to poor solubility and similar separation factors.52

The higher order complexes were seen to be more hydrophobic sothe 1:2 species was the target in many of these systems.

Replacement of the outer pyridine rings of TERPY with terpyridine rings gave much more noticeable improvements, with the first ligand of this type being BTP (Figure 4).53 Higher 39

distribution ratios were observed for BTP than known ligands, attributed to the basicity of the ligand which reduces the ability to bind protons, thus favouring the formation of a ligand to metal complex.54 The complexes formed could be precipitated as nitrate salts which allowed isolation and removed the need for synergists. The novel 1:3 actinide-BTP complex was also observed as a heteroleptic complex (via proton NMR).

The weakness of the BTP derivatives was that linear aliphatic groups on the ligand were susceptible to radiolysis, hence preventing use in an industrial SANEX process. When an aromatic or branched group was added (as shown in TPTZ) the kinetics of extraction and solubility of the ligand were reduced. This led to a need for a synergist molecule such as co- ordinating malidomides. Subsequently it was found that CyMe4 groups on the outer triazine rings (i.e. CyMe4BTP, Figure 8) resulted in a 5000-fold selectivity increase although the extraction kinetics slowed to the hour scale. This has not been fully explained, but is attributed to the complex no longer being labile as the actinide ions were not released into the aqueous phase, (hence CyMe4BTP is unsuitable for a radiochemical separation.) The next development with the BTP molecules came from Hudson et al. who showed that it is possible to decrease the probability of radiolytic and acidic degradation with substitution of the α-CH2 positions on the triazinyl moiety.55

Drew et al. published a QSAR (Quantitative Structure Activity Relationship) study of

N-heterocyclic donor molecules in 2004 and showed that a method usually applied to drug discovery can be applied to many parameters related to extraction and ligand design for actinide extraction. The study predicted the separation factor of theoretical extractants that could improve on the BTP derivatives synthesised at the time and was a useful tool to guide

SANEX synthetic research.56

Harwood and co-workers attempted to solve the challenges associated with the BTP ligand by adjusting the structure. They developed the bistriazinyl bipyridine (BTBP) ligand group 40

(Figure 4). Addition of an extra nitrogen atom to the ring allowed expansion of the ion cavity and increased the SFAm/Eu (160±16) which was higher than the best BTP derivative molecules.57 The design aimed to use groups proven to be resistant to hydrolysis and radiolysis, and to produce ligands which were soluble in a wider range of solvents than previously used CHON ligands. The ligand was shown to rotate freely in solution around the bipyridine bond and is most stable in the trans conformation. When in solution with a metal ion, it co-ordinates via the cis conformer and is relatively stabilised due to the sterically demanding phenyl groups.

Harwood et al. showed that the BTBP derivatives form 1:1 or 1:2 complexes with the lanthanides and actinides, depending on the metal, though the metal is always ten co- ordinate.58,59 It was also shown that extraction of americium was independent of thorium, uranium and neptunium in the aqueous phase. This made these ligand classes applicable to selective separation of americium from complex actinide mixtures.

Phenanthroline based ligands were identified in the 1960s for their ability to complex metal ions including copper and have been used as synergists in actinide separations. The first report of the selectivity of phenanthroline ligands for the separation of americium from competing radionuclides was in 1988.60 No further work was carried out to explain the selectivity observed. The cis-locked 1,10-phenanthrolines have found many uses in inorganic complexes and in ion sequestration.61 It was reasoned that the locking of the bi-pyridine ring would improve separation kinetics as non-locked ligands such as BTBP must overcome the barrier to rotation to co-ordinate with an ion. The dipole moment of the phenanthroline ligands would also aid in the phase separation with the polar aqueous phases. The first phenanthroline ligand reported by Harwood et al. was CyMe4BTPhen (BTPhen, Figure 8).

This ligand again formed 1:1 and 1:2 complexes with actinides; the 1:2 complex was observed to form more quickly than the lanthanide 1:1 complex. The crystal structure of the 41

europium-BTPhen 1:2 complex showed a 10 co-ordinate Eu ion with a labile co-ordination sphere which can accommodate other ions. The separation factor of BTPhen is >250 for americium/europium separation, and this separation can be accomplished within an hour. The

BTPhen can also be recycled, and the americium can be transferred back to an aqueous phase due to the labile complex formed. The BTPhen and BTBP ligands were therefore a landmark in progress towards CHON SANEX process ligands. BTPhen was shown to be selective in a

SANEX process HLLW trial. However, industrial application of BTPhen has not yet been further developed despite the promising initial analytical separations. This has been attributed to the americium complex being too selective as seen above with CyMe4BTP. In addition,

BTPhen has not been applied to a separation using representative sample matrices to separate americium from mixed fission products, although many speciation studies are underway on the BTBP and BTPhen ligands.

Figure 8. Ligands used for americium extraction.

Whittaker et al. further progressed this work by investigating speciation in these separations.

For BTBP and BTPhen, crystal structures of complexes of all the lanthanides (excluding Pm) were obtained as the 1:2 complex and their stability was shown to be less than that of the corresponding americium complexes.62 Furthermore, TRLFS has shown the formation of the

1:2 species for Cm with the two ligands.63 This work has also explored the species formed in individual reactions with uranyl, neptunium (+4), plutonium (+4) and thorium (+4) ions, where the species can be proven using a combination of analytical techniques (EXAFS, UV-

Vis titration, NMR titration, XAS).64 All except thorium formed interesting 1:1 complexes with the ligands, with thorium forming the complex [Th(CyMe4-

BTPhen)2(NO3)2][Th(NO3)6].4CH3CN. These results have helped to visualise the species that can form in solution and imply BTPhen is suitable for selective americium and curium co- extraction.

8.0 References

1 C. Ekberg, A. Fermvik, T. Retegan, G. Skarnemark, M. R. S. Foreman, M. J. Hudson, S.

Englund and M. Nilsson, Radiochim. Acta., 2008, 96, 225–233.

2 F. W. Lewis, L. M. Harwood, M. J. Hudson, M. G. B. Drew, J. F. Desreux, G. Vidick, N.

Bouslimani, G. Modolo, A. Wilden, M. Sypula, T.-H. Vu and J.-P. Simonin, J. Am. Chem.

Soc., 2011, 133, 13093–13102.

3. W. D. Loveland, D. J. Morrissey and G. T. Seaborg, Modern Nuclear Chemistry, John

Wiley & Sons, Hoboken (New Jersey), 1st edn., 2006, pp. 35–52.

4. K. L. Nash, D. Brigham, T. C. Shehee, A. Martin, Dalton Trans., 2012, 41, 14547–14556.

5. R. J. Blagg, C. A. Muryn, E. J. L. McInnes, F. Tuna and R. E. P. Winpenny, Angew. Chem.

Int. Ed., 2011, 50, 6530–6533.

6. H. I. Skriver, O. K. Andersen and B. Zohansson, Phys. Rev. Lett., 1978, 41, 42–45.

7. J. L. Sessler, P. J. Melfi and G. Dan Pantos, Coord. Chem. Rev., 2006, 250, 816–843.

8. L. Natrajan, F. Burdet, J. Pécaut and M. Mazzanti, J. Am. Chem. Soc., 2006, 128, 7152–

7153.

9. K. J. Moody and P. M. Grant, J. Radioanal. Nucl. Chem, 1999, 241, 157-167.

10. R.E. Heft, J. Adv. Chem., 1970, 93, 254-259.

11. J.J. Bellucci, A. Simonetti, C. Wallace, E.C. Koeman and P.C. Burns, Anal. Chem., 2013,

85, 4195-4198

nd 12. K. J. Moody, P. M. Grant and I. D. Hutcheon, Nuclear Forensic Analysis (2 Edition),

CRC Press (FL), 2014. (ISBN 978-1-4398-8061-6)

13 . K. Mayer, M. Wallenius and I. Ray, Analyst, 2005, 130, 433–441.

14. E. P.How, M. L. Dletz, R. Chlanzla and H. Diamond, Anal. Chim. Acta, 1992, 266, 25–37.

15. H. Kurosaki, J. R. Cadieux, and S. B. Clark, J. Anal. At. Spectrom., 2014, 29, 2419-2423

16. K. L. Nash, Solvent Extr. Ion Exch., 1993, 11, 729–768.

17. O. Courson, R. Malmbeck, G. Pagliosa, K. Romer, B. Satmark and J.-P. Glatz, Radiochim.

Acta, 2000, 88, 885–864.

18. G. Thiollet. and C. Musikas., Solv. Extr. Ion Exch., 1989, 7, 813–827.

19. J. D. Law, K. N. Brewer, R. S. Herbst, T. A. Todd, Demonstration of the TRUEX Process for Partitioning of Actinides from Actual ICPP Tank Waste Using Centrifugal Contactors in a Shielded Cell Facility, National Engineering and Environmental Laboratory Report INEL-

96/0353; Idaho Falls, 1996.

20. J. Chen, Y. Zhu and R. Jiao, Sep. Sci. Technol. (Philadelphia, PA, U.S.), 1996, 31, 2723–

2721.

21. E. Aneheim, C. Ekberg, A. Fermvik, M. R. S. J. Foreman, T. Retegan and G. Skarnemark,

Solv. Extr. Ion Exch., 2010, 28, 437–458.

22. D. R. Peterman, J. D. Law, T .A. Todd and R. D. Tillotson, Separations for the Nuclear

Fuel Cycle in the 21st Century, American Chemical Society, Idaho Falls (Idaho), 1st edn.,

2009, pp. 251–259.

23. S. A. Ansari, P. N. Pathak, V. K. Manchanda, M. Husain, A. K. Prasad and V. S. Parmar,

Solvent Extr. Ion Exch., 2005, 23, 463–479.

24. J. Weaver, Solvent Extr. Ion Exch., 1974, 6, 189–277.

25. J. Chen, J. Wang and C. Song, Tsinghua Sci. Technol., 2001, 6, 180–184.

26. T. Takao, T. J. Bell and Y. Ikeda, Inorg. Chem., 2013, 52, 3459–3472.

27. R. Pilvio and M. Bickel, J. Alloys Compd., 1998, 271, 49–53.

28. E. P. Horwitz, Anal. Chim. Acta., 1993, 281, 361–372.

29. M. L. Dietz, E. P. Horwitz, L. R. Sajdak and R. Chiarizia, Talanta, 2001, 54, 1173–1184.

30. J. Spurny and K. Heckmann, J. Radioanal. Nucl. Chem., 1992, 163, 305–307.

31. K. Heckmann and J. Strnad, J. Radioanal. Nucl. Chem., 1992, 163, 47–57.

32. R. M. Smith and A. E. Martell, Sci. Tot. Environ., 1987, 64, 125–147.

33. M. Mackenzie, The Solvent Extraction of Some Major Metals, Cognis GmbH, 2008.

34. J. J. Katz, J. Fuger and N. M. Edelstein (Eds.), Chemistry of the Actinides and

Transactinides, Springer, Dordrecht, 4th edn., 2006, pp. 1265–1395.

35. K.Tsukada, Y. Ichikawa, Y. Hatsukawa, I. Nichinaka, K. Hata, Y. Nagame, Y. Oura, T.

Ohyama, K. Sueki, H. Nakahari, M. Asai, Y. Kojima and T. Hirose, Phys. Rev. C: Nucl.

Phys., 1998, 57, 2057–2060.

36. R. Cashmore, J. Billowes, W. Bowen, C. Brown, R. Grimes, R. Howsley, F. Livens, J.

Simpson and P. Styles, Fuel Cycle Stewardship in a Nuclear Renaissance, The Royal Society

Science Policy Centre Report, London, 2011.

37. W. W. Schulz, The Chemistry of Americium, United States Energy Research and

Development Administration TID-26971, Springfield (Virginia), 1976.

38. Wolfgang H. Runde and Wallace W. Schulz, Chemistry of the Actinides and

Transactinides (4th Edition, Chapter 8), Springer Netherlands, 2011

39. W. W. Schulz, Trans. Am. Nucl. Soc., 1975, 21, 262–263.

40. W. H. Runde and B. J. Mincher, Chem. Rev., 2011, 111, 5723–5741.

41. T. Yaita, S. Tachimori, N. M. Edelstein, J. J. Bucher, L. Rao, D. K. Shuh and P. G. Allen,

J. Synchrotron Radiat., 2011, 8, 663–665.

42. B. Johannson, J. Phys. Chem. Solids, 1978, 39, 467–471.

43. T. L. Markin, J. Inorg. Nucl. Chem., 1958, 7, 290–292.

44. G. R. Choppin and R. J. Silva, J. Inorg. Nucl. Chem., 1956, 3, 153–154.

45 P. B. Iveson, C. Riviers, D. Guillaneux, M. Nierlich, P. Thuery, M. Ephritikhine and C.

Madic, Chem. Commun., 2001, 1512–1513.

46. J. H. Burns and W. H. Baldwin, Inorg. Chem., 1976, 16, 289–294.

47. A. Geist and P. Panak, Chem. Rev., 2013, 113, 1199–1236.

48. C. Adam, P. Kaden, B. B. Beele, U. Müllich, S. Trumm, A. Geist, P. J. Panak and M. A.

Denecke, Dalton Trans., 2013, 42, 14068–14074.

49. S. G. Minasian, J. M. Keith, E. R. Batista, K. S. Boland, D. L. Clark, S. D. Conradson, S.

A. Kozimor, R. L. Martin, D. E. Schwarz, D. K. Shuh, G. L. Wagner, M. P. Wilkerson, L. E.

Wolfsberg and P. Yang, J. Am. Chem. Soc., 2012, 134, 5586–5597.

50. K. T. Potts, P. Ralli, G. Theodoridis, and P. Winslow, Org. Synth., 1990, 64, 189–195.

51. C. Mesmin, J. O. Liljenzin, Solvent Extr. Ion Exch., 2003, 21, 783–795.

52. P. Y. Cordier, C. Hill, P. Baron, C. Madic, M. J. Hudson and J. O. Liljenzin, J. Alloys

Compds., 1998, 271-273, 738–741.

53. F. H. Case, Heterocycl. Chem. 1971, 8, 1043–1044.

54. Z. Kolarik, U. Müllich and F. Gassner, Solvent Extr. Ion Exch., 1999, 17, 23–28.

55. M. J. Hudson, C. E. Boucher, D. Brackers, J. F. Desreux, M. G. B. Drew, M. R. S. J.

Foreman, L. M. Harwood, C. Hill, C. Madic, F. Marken and T. G. A. Youngs, New J. Chem.,

2006, 30, 1171–1183.

56. M. G. B. Drew, M. J. Hudson and T. G. A. Youngs, J. Alloys Compds., 2004, 374, 408–

415.

57. M. G. B. Drew, M. R. S. J. Foreman, C. Hill, M. J. Hudson and C. Madic, Inorg. Chem.

Comm., 2005, 8, 239–241.

58. L. M. Harwood, F.W. Lewis, M. J. Hudson, J. John and P. Distler, Solvent Extr. Ion Exch.,

2011, 29, 551–576.

59. M. Nilsson, S. Andersson, F. Drouet, C. Ekberg, M. Foreman, M. Hudson, J. O. Liljenzin,

D. Magnusson and G. Skarnemark, Solvent Extr. Ion Exch., 2006, 24, 299–318.

60. M. D. Ogden, C. L. Hoch, S. I. Sinkov, G. P. Meier, G. J. Lumetta and K. L. Nash, Solvent

Extr. Ion Exch., 1988, 6, 439–445.

61. X. L. Zhang and K. Cheng, Synth. React. Inorg., Met.-Org., Nano-Met. Chem., 2013, 43,

1019–1024.

62. D. M. Whittaker, T. L. Griffiths, M. Helliwell, A. N. Swinburne, L. S. Natrajan, F. W.

Lewis, L. M. Harwood, S. A. Parry and C. A. Sharrad, Inorg. Chem., 2013, 52, 3429–3444

63. A. Bremer, D. M. Whittaker, C. A. Sharrad, A. Geist and P. J. Panak, Dalton Trans., 2014,

43, 2684–2694.

64. D. M. Whittaker, PhD Thesis, The University of Manchester, 2014

9.0 Chapter 1

The material in the following section has been published in the journal Radiochimica Acta, it was also presented at the MARC X conference in 2015. DOI: 10.1515/ract-2015-2403

The author synthesised the ligands, designed and performed the radiochemical method development, analysed the data, interpreted the results and wrote first draft and the final version of the manuscript.

Rapid selective separation of americium/curium

from simulated nuclear forensic matrices using

triazine ligands

Matthew A. Higginson,a Olivia J. Marsden,b Paul Thompson,b Laurence M. Harwood,c

Michael J. Hudson, c Frank W. Lewis, c,d Francis R. Livensa and Sarah L. Heath.a,* a Centre for Radiochemistry Research, School of Chemistry, The University of Manchester,

Manchester, M13 9PL, UK. b AWE, Aldermaston, Reading, RG7 4PR, UK. c Department of Chemistry, University of Reading, Whiteknights, Reading, RG6 6AD, UK. d Department of Chemical and Forensic Sciences, Faculty of Health and Life Sciences,

Northumbria University, Newcastle upon Tyne NE1 8ST, UK.(Current address)

* Corresponding author, [email protected], tel. +44 161 275 4696

9.1 Abstract

In analysis of complex nuclear forensic samples containing lanthanides, actinides and matrix elements, rapid selective extraction of Am/Cm for quantification is challenging, in particular due the difficult separation of Am/Cm from lanthanides. Here we present a separation process for Am/Cm which is achieved using a combination of AG1-X8 chromatography followed by Am/Cm extraction with a triazine ligand. The ligands tested in our process were

CyMe4-BTPhen, CyMe4-BTBP, CA-BTP and CA-BTPhen. Our process allows for purification and quantification of Am and Cm (recoveries 80–100%) and other major actinides in <2 days without the use of multiple columns or thiocyanate. The process is unaffected by high level Ca/Fe/Al (10 mg mL-1) and thus requires little pre-treatment of samples.

9.2 Key Words:

Americium separation, nuclear forensics, triazine ligand, BTPhen, BTBP

9.3 Introduction

Nuclear forensic techniques aim to characterise unknown radioactive materials to provide evidence for use in legal cases, for example in seizures of illicit nuclear material. The collection of data must be performed in a timely manner and analytical techniques must be strategically applied to minimise sample loss/damage (i.e. non-destructive techniques applied first). Although the methodology of investigation will be driven by the specific sample, details of elemental and isotopic composition are required to allow comparison with reference data. This requires multiple analytical methods (e.g. radiometric methods, mass spectrometry) to be applied to the sample. However, samples in nuclear forensics can be complex, containing lanthanides, actinides and matrix elements, including fission/activation products with isotopic variation. Radiochemical separation is therefore required in order to avoid isobaric and polyatomic interferences (e.g. 232U/232Th) in mass spectrometry and/or spectral overlap in radiometric analyses [1-3].

The ability to obtain pure fractions of the lanthanides and Am/Cm is currently a bottleneck in nuclear forensic investigations of such samples. While some methods have been reported which achieve this in 1–2 days [4-6], these processes require multiple column steps and/or rely on the use of thiocyanate for Am elution, resulting in toxic wastes. More commonly, separation of Am/Cm and lanthanides from these complex samples requires a process which may take several days (Figure 1).

Figure 1. Example of Am separation for a nuclear debris sample.

We envisioned a simplified process in which Am/Cm, actinides and lanthanides could be extracted and quantified from a complex nuclear forensic sample in < 2 days, whilst also avoiding the production of toxic waste. Specifically, this would entail separation of the major actinides (U, Pu, Np) by chromatography using current methods, followed by extraction of

Am/Cm from the remaining matrix containing lanthanides and matrix elements.

Multiple soft-N donor ligands based on 1,2,4-triazines have shown selectivity for complexation with Am/Cm over other actinides and lanthanides, with separation factors >100

[7]. This can be attributed to the enhanced covalency in bonds formed by the 5f orbitals of the actinides, which confers a preference for softer ligands [8]. In our process, we desired a decontamination factor of >1000 for Am/Cm, which could theoretically be achieved by ligands with such separation factors. Therefore, we identified four ligands based on the triazine motif for potential application in the extraction of Am/Cm from a complex nuclear forensics matrix (Figure 2). Herein we report application of these ligands to the extraction of

Am/Cm to aqueous matrices containing lanthanides and matrix elements which are representative of complex nuclear forensic samples for the development of a rapid Am/Cm separation process.

Figure 2. Structures of the ligands used in this comparison study.

9.4 Experimental

All radionuclides used were provided from calibrated stocks in the School of Chemistry,

University of Manchester. Micropipettes of 100 µL, 0.1–1 mL and 2–10 µL were calibrated on a 4 d.p. balance with >18 MΩ deionised water in the temperature range 18–22 °C and were found to be within their stated range. All acid solutions were made from analytical grade concentrated solutions and were diluted with >18 MΩ deionised water. All solutions were considered to have expired within one month of creation. Gamma counting was performed using a Canberra 2020 coaxial HPGe gamma spectrometer with an Ortec 919E multi-channel analyser. Gamma spectroscopy was performed against a standard of known activity counted in the same geometry and analysed using the diagnostic photon energies of

241Am (59.5 keV) and 152Eu (121.8 keV). Samples were electrodeposited and alpha spectrometry was performed on a Canberra model 7401VR detector with multi-channel analyser. ICP-MS analysis was performed on an Agilent 7500cx spectrometer. Multiple

standards for each element in the range 1-100 ppb were used for ICP-MS quantification. All reagents and solvents used were of standard analytical grade unless otherwise stated.

9.41 Ligand Synthesis

CyMe4BTPhen, CyMe4BTBP and CA-BTP were synthesised as reported previously [9, 10,

11], and purified using the method reported by Whittaker et al [12]. CA-BTPhen was synthesised using a modification of the previously reported procedure [13], in which the solvent was changed to EtOH and the reaction time changed to 3 h reflux followed by 6 h at

RT. CA-BTP was purified in the same manner as the other three ligands. Data collected for all intermediates and ligands was as previously reported.

9.42 Am/Cm Separation from Normal Nuclear Forensics Matrix

Generation of matrix: Using 4 M HCl (VWR, AnalaR NORMAPUR® grade) a solution containing ~1 mg mL-1 of metal was prepared by addition of the following elements in chloride form (Sigma-Aldrich or Fluka, >99.99% trace metal): Be (II), Cs (I), Mo (VI), Ce,

Pr, Nd, Sm, Tb, Y (III) , Zr (IV), Ag (I), Cd (II), Sr (II) and Ba (II). Spike solutions of 241Am

(10-100 Bq range), 152Eu (10-100 Bq range), 244Cm (10-100 Bq range, all III), 232U (VI),

239Pu (IV) and 237Np (IV, 10 Bq each isotope) were added and the solution readjusted to 4 M

HCl. N.B. For high level Fe/Ca/Al matrices, the matrix was generated as above with the addition of ~10 mg mL-1 Fe(III) or Ca(II) or Al (III), added as chloride form.

Anion exchange/extraction chromatography: The matrix was taken to dryness two times with addition of conc. HCl (2 mL). The residue was dissolved in the minimum volume of conc.

HCl. A 50 mm x 4 mm anion exchange column (Bio-Rad AG1-X8, 200-400 mesh) was conditioned with the following procedure: sodium bromate (5 mg) in conc. HCl (5 mL) was added. The column was washed with conc. HCl (4 mL) then aq. HCl (5 mL, 1 M) and conc.

HCl (5 mL). The sample was loaded using conc. HCl to wash (2 mL). Am, lanthanides, Cs,

Be were eluted with conc. HCl (6 mL). Pu was eluted with conc. HCl containing 1% v/v HI

(5 mL). Co, Zr, Np were eluted with HCl (5 mL, 4 M). Mo, Fe, U were eluted with HCl (3 mL, 1 M) followed by HCl (3 mL, 0.5 M). Cd and Ag were eluted using deionised water (1 mL) followed by NH4OH (4 mL, 1 M). N.B. this step may also be achieved using the

UTEVA® based method described by Morgenstern [4].

Liquid-liquid Am/Cm separation procedure: Each ligand was dissolved in 1 mL of 1-octanol to its chosen concentration, mixed for 5 min and sonicated for 1 min. The organic phase was contacted with the Am-containing fraction from chromatography for 90 minutes with vortex mixing. The sample was centrifuged at 1788 g/4000 rpm for 5 min and the phases separated.

The aqueous phase was analysed by gamma spectroscopy and then evaporated to dryness, dissolved in aqueous HNO3 (2% w/v, VWR, AnalaR NORMAPUR® grade) and diluted by a factor of 10000 for ICP-MS analysis. The organic phase was back-stripped with 0.1 M HCl for 90 minutes and the resulting aqueous solution electrodeposited as below. The U, Pu and

Np-containing fractions from chromatography were also electrodeposited for alpha spectrometry.

Electrodeposition procedure: The solution to be electrodeposited was evaporated to dryness.

Aqueous NaHSO4 (5%, 2.5 mL), water (2 mL) and aqueous Na2SO4 (15%, 5 mL) were added to the residue and heated, the solution was transferred to an electrodeposition cell using water

(3 mL) and (NH4)2C2O4 (160 mM, 1 mL) was added. The current was set to 0.5 A for 5 min, then 0.75 A for 60 min. 1 min before the end of electrodeposition, aqueous KOH (25%, 2 mL) was added. The solution was decanted off and the cell washed with aqueous NH4OH

(5%, 2 mL). The plate was rinsed with 5% NH4OH, ethanol and acetone before drying in the air. 55

9.5 Results and Discussion

9.51 Ligand Selection

The family of four ligands used in this study was based around CyMe4-BTPhen 1, previously shown by us to separate Am/Eu with a separation factor of up to 400 [9]. We therefore identified CyMe4-BTPhen as a potential ligand for the separation of Am from aqueous matrices containing actinides, lanthanides and matrix elements. The ligand can be recycled for repeat use, and adheres to the CHON principle such that it can be safely incinerated after use, avoiding the generation of toxic waste. From this, we identified 3 related ligands with similar properties which could be applied to selective extraction of

Am/Cm: CA-BTPhen 2, CA-BTP 3 and CyMe4-BTBP 4 (Figure 2). These ligands encompass the most promising triazine motifs previously used in Am/Eu separation (BTPs,

BTPhens and BTBPs). We also attempted to select compounds which could easily be obtained in an analytical laboratory setting i.e. easily synthesised. In particular, we were interested in ligands containing the camphor (CA) group as this is resistant to radiolysis [10] and the camphor diketone required for synthesis is commercially available (unlike CyMe4 diketone). CA-BTP 3 was chosen as a potential ligand as it is synthesised in only two steps and has increased solubility in 1-octanol compared to BTBPs and BTPhens [9]. CyMe4BTBP

4 is the current benchmark ligand for the proposed n-SANEX process and has been widely

applied to Am extraction in recent years [8]. It is possible that CyMe4BTBP may become commercially available in the near future due to its widespread use.

9.52 Application of Ligands to Am/Cm Extraction from Simulated Nuclear Forensic

Matrices

The four ligands were synthesised as previously reported [9, 10, 11]. Based on previously reported work, our work with the BTPhen ligands and preliminary Am/Eu separations carried out as part of this investigation, we determined a number of parameters for the Am/Cm extraction. The organic phase selected for the separation was 1-octanol, since alternative solvents such as cyclohexanone were shown to cause phase separation issues in initial tests with the nuclear forensic matrix. The aqueous phase selected was 4 M HCl as this would be compatible with further separation processes e.g. for the lanthanides. Acid concentrations in the range 2–4 M (HCl or HNO3) would be suitable for Am extraction from previous Am/Eu separation work [9]. We avoided the addition of phase modifiers in this work as this could increase co-extraction of particular elements such as lanthanides, and also complicate back stripping. (N.B. Separation factors (Am/Eu) were determined to be similar to those reported above, for distribution ratio data and HCl/HNO3 extraction data please see [9,10,11]).

The simulated nuclear forensic matrix was designed to simulate an aqueous acidic matrix obtained from a nuclear debris sample after pre-treatment (e.g. total dissolution). The matrix was generated by dissolution of Be, Mo, Cs, Ce, Pr, Nd, Sm, Tb, Y, Zr, Ag, Cd, Sr and Ba and all the lanthanides (except Pm and Eu) as chlorides in 4 M HCl. The target concentration for each stable element was ~1 mg mL-1 of metal. Tracers of 241Am, 152Eu,

244Cm, 232U, 239Pu and 237Np were added and the solution adjusted to 4 M HCl. The concentration of each element in the representative matrix before Am/Cm extraction was determined by ICP-MS.

The simulated nuclear forensic matrix was then subjected to AG1-X8 anion exchange chromatography (chloride form) to separate the major actinides (U, Pu, Np) and yield an

Am/Cm containing fraction which would be used in the subsequent triazine ligand extraction. 57

The major actinides were analysed by alpha spectrometry and recoveries were found to be >94% (Table 1). Recoveries of the co-eluting matrix elements (Zr, Mo, Ag, Cd) were determined by ICP-MS of the relevant fractions and found to be >91% (Table 1). This first step of the overall separation process allows for immediate separation of the major actinides, allowing their quantification (electrodeposition and alpha spectrometry) to be performed concurrently with the subsequent Am/Cm separation. This would ensure data is collected in a timely manner, which is of vital importance in nuclear forensics.

Isotope 90Zr 95Mo 107Ag 111Cd 232U 239Pu 237Np Recovery ± 97.21 95.23 ± 99.74 ± 92.01 ± 96.76 ± 97.24 ± 96.08 ± Error / % ± 2.01 1.99 1.05 1.35 3.21 2.12 0.71

Table 1: Recoveries of eluted matrix elements from AG1-X8 column chromatography from analysis by alpha spectrometry and ICP-MS. (N.B. error calculated from instrument and method error plus one sigma, sample data averaged from n=20 analysis of all run in triplicate).

The Am/Cm fraction was then subjected to liquid-liquid separation using each of the four ligands in 1-octanol/ 4 M HCl. The phases were vortex mixed for 90 minutes, this was based on the time required for Am/Eu extraction using CyMe4BTPhen in our previous work.

The resulting aqueous phases were analysed by gamma spectrometry (to determine Eu) and

ICP-MS analysis (to determine matrix elements). The organic phase was back-stripped with

0.1 M HCl for 1 h and the resulting aqueous phase was analysed by alpha spectrometry for

Am/Cm. Extraction of Am and Cm was seen with all four ligands (Table 2). CyMe4BTPhen extracted the highest quantity of Am over 90 minutes (96.0 ± 2.3%) and also had high extraction of Cm (94.1 ± 2.2%). The other three ligands extracted Am/Cm to a lesser extent over 90 minutes, but when the mixing time was increased to 3 hours these recoveries were >90%. Over the 90 minute separations, Eu extraction was minimal compared to Am/Cm 58

extraction. ICP-MS analysis showed high recovery of matrix elements including all lanthanides (>88%, Table 2) showing that the method allows for high decontamination (and

SF> 100 for each ligand with respect to the lanthanides) for Am/Cm with respect to these elements.

Figure 3. Method for extraction of Am/Cm from complex nuclear forensic sample using CyMe4BTPhen (this work).

Overall, the combined AG1-X8/triazine ligand separation procedure developed allows for the quantification and determination of the major actinides (U, Pu, Np) and Am/Cm within 2 days. In addition, a lanthanide containing fraction is obtained which can be subjected to further separation. From the results obtained, CyMe4BTPhen is the most promising ligand for a rapid extraction (<90 min) of Am/Cm. We therefore propose a method, as shown in Figure

3, in which this ligand is utilised for Am/Cm extraction from a Am/Cm/Ln solution post

AG1-X8 chromatography.

Isotope Recovery ± Error / % Ligand 241Am 244Cm 152Eu 9Be 88Sr 89Y 133Cs CyMe 4- 96.4 ± 2.3 94.1 ± 2.2 95.0 ± 4.1 98.2 ± 2.2 99.7 ± 0.7 98.7 ± 2.4 100.1 ± 2.0 BTPhen CA-BTP 70.0 ± 2.2 65.0 ± 2.3 90.8 ± 6.2 98.2 ± 2.6 99.2 ± 2.1 93.2 ± 0.5 91.9 ± 2.5

CyMe4BTBP 85.7 ± 2.6 95.2 ± 2.8 95.3 ± 3.9 99.8 ± 1.0 94.5 ± 1.1 99.8 ± 1.9 95.2 ± 0.6 CA-BTPhen 43.5 ± 2.7 56.1 ± 2.6 93.8 ± 5.6 99.9 ± 0.8 92.9 ± 3.1 92.2 ± 0.5 101.3 ± 2.1

137Ba 139La 140Ce 141Pr 146Nd 147Sm 157Gd

CyMe 4- 99.0 ± 1.0 94.4 ± 0.5 95.2 ± 2.1 99.6 ± 1.7 100.6 ± 2.1 102.1 ± 1.0 99.2 ± 1.5 BTPhen CA-BTP 99.6 ± 0.8 93.7 ± 1.0 96.4 ± 1.2 100.3 ± 1.0 97.7 ± 1.0 94.3 ± 0.7 99.1 ± 0.3

CyMe4BTBP 101.0 ± 1.1 102.1 ± 2.8 100.0 ± 1.3 99.6 ± 0.7 98.9 ± 1.1 107.2 ± 9.7 100.1 ± 0.7 CA-BTPhen 91.9 ± 0.3 94.7 ± 1.0 88.7 ± 0.6 92.5 ± 0.7 94.7 ± 0.6 99.3 ± 0.3 99.7 ± 0.4

159Tb 163Dy 165Ho 166Er 169Tm 172Yb 175Lu

CyMe 4- 94.6 ± 1.2 94.3 ± 1.5 98.9 ± 0.5 91.6 ± 1.0 97.7 ± 0.3 98.9 ± 1.2 95.1 ± 1.2 BTPhen CA-BTP 97.8 ± 0.6 98.2 ± 1.1 98.9 ± 0.7 97.6 ± 0.7 101.7 ± 0.7 97.8 ± 0.5 91.3 ± 2.1

CyMe4BTBP 97.4 ± 0.2 100.5 ± 1.3 98.9 ± 0.6 99.2 ± 1.2 99.3 ± 0.2 100.5 ± 1.6 100.3 ± 0.9 CA-BTPhen 98.2 ± 0.2 99.7 ± 0.5 98.7 ± 0.4 95.8 ± 0.4 97.4 ± 0.4 99.9 ± 0.4 102.4 ± 2.2 Table 2: Recoveries of eluted matrix elements for AG1-X8 column chromatography/triazine ligand separation from analysis by alpha spectrometry, gamma spectroscopy and ICP-MS. (N.B. error calculated from instrument and method error plus one sigma, sample data averaged from n=20 analysis run in triplicate, all calibrated against a range of standards in the concentration ranges above).

9.53 Am/Cm Extraction from Simulated Nuclear Forensic Matrices in the Presence of

Fe, Ca and Al

Complex nuclear forensic samples may be derived from any origin, but Fe, Ca and Al are often present in high amounts as these elements are found in high amounts in soils, cements and metals. Typically samples with high levels of these elements require additional purification to avoid interference with separation processes. We wished to assess the impact of realistic (10 mg mL-1) Fe, Ca and Al on our Am/Cm separation procedure. We therefore generated three matrices by separate addition of Fe (III), Ca (II) and Al (III) in chloride form to the matrix described above, and repeated the investigation with each of the four ligands.

Recovery of U, Np, Pu, Zr, Mo, Ag and Cd from AG1-X8 chromatography was unaffected by high levels of Fe, Ca or Al, as expected. Iron was also removed in this part of 60

the process (co-eluted with U, separated by precipitation), with a recovery >93% by ICP-MS across all experiments (errors from SD from n=14 replicates <3.0% in all experiments). After extraction of Am/Cm by the four ligands, the remaining elements were determined as previously, also Ca and Al were determined by ICP-MS (Table 3). Am/Cm and Eu extraction was unaffected for all ligands in the presence of Ca and Al except for CyMe4BTBP in the presence of Ca when Am/Cm extraction was reduced by ~15–20%. Recovery of matrix elements including lanthanides was not greatly affected by the presence of high level Ca or

Al. From ICP-MS analysis, the recoveries of Ca and Al were also determined from each separation. Apart from CA-BTPhen, all ligands were seen to co-extract both Al and Ca to some extent. In particular, for CyMe4BTBP, Ca was extracted (~15%) and Am/Cm extraction decreased, meaning that high Am/Cm recovery could not be obtained even over a 3 hour extraction. However, in cases where Ca/Al co-extraction does not affect Am/Cm separation it will affect electrodeposition of the sample. In addition, pre-treatment steps could be utilised to remove the high level elements, but this could be at the expense of process time.

From investigation of the normal and high level Fe, Ca and Al matrices, we concluded that CyMe4BTPhen is the most suitable ligand for Am/Cm extraction from complex nuclear forensic samples based on the near-quantitative Am/Cm extraction achieved in 90 minutes. In addition, high levels of Ca and Al did not affect the Am/Cm extraction by this ligand. To fully determine the ability of CyMe4BTPhen to extract Am/Cm from complex nuclear forensic matrices, we then wished to perform the extraction in the presence of all elements i.e. without prior AG1-X8 chromatography.

Table 3a (Ca)

Isotope Recovery ± Error / %

241 244 152 9 88 89 43 133 137 139 140 Ligand Am Cm Eu Be Sr Y Ca Cs Ba La Ce

CyMe4- 95.9 93.9 95.2 99.2 98.8 95.7 96.4 107.6 98.3 98.9 92.5 BTPhen ± 2.4 ± 2.4 ± 6.5 ± 2.1 ± 3.2 ± 1.9 ± 1.2 ± 1.2 ± 1.0 ± 1.0 ± 1.3 68.0 63.1 90.3 100.0 94.6 100.8 91.8 99.3 89.6 92.5 87.2 CA-BTP ± 2.6 ± 2.6 ± 6.6 ± 1.6 ± 1.5 ± 3.0 ± 3.2 ± 1.8 ± 2.2 ± 1.0 ± 1.5 CyMe4- 68.6 73.6 91.4 99.7 95.2 99.7 85.1 95.7 93.8 97.1 96.3 BTBP ± 2.3 ± 2.9 ± 5.7 ± 2.5 ± 0.3 ± 0.7 ± 1.2 ± 0.7 ± 1.1 ± 1.0 ± 0.4 CA- 41.5 43.4 95.3 99.7 93.5 102.8 103.0 102.3 100.5 102.9 100.9 BTPhen ± 2.4 ± 2.2 ± 6.8 ± 1.3 ± 1.2 ± 0.9 ± 2.3 ± 2.6 ± 1.5 ± 1.2 ± 0.6

141 146 147 157 159 163 165 166 169 172 175 Pr Nd Sm Gd Tb Dy Ho Er Tm Yb Lu

CyMe4- 96.9 99.2 102.1 103.7 102.9 100.3 96.8 100.5 100.3 100.7 92.7 BTPhen ± 0.5 ± 0.5 ± 2.4 ± 3.9 ± 3.1 ± 1.0 ± 0.8 ± 0.7 ± 0.7 ± 0.8 ± 0.5 91.3 93.0 97.0 97.4 95.3 96.9 99.8 93.1 100.7 99.9 92.1 CA-BTP ± 1.1 ± 0.9 ± 1.6 ± 0.6 ± 1.3 ± 1.4 ± 1.7 ± 1.0 ± 1.3 ± 1.7 ± 2.0 CyMe4- 96.0 93.3 100.4 94.2 100.1 94.2 96.4 89.5 97.2 87.0 101.1 BTBP ± 1.9 ± 1.4 ± 1.6 ± 2.4 ± 1.4 ± 0.5 ± 1.1 ± 1.4 ± 0.7 ± 3.0 ± 2.1 CA- 99.7 100.1 100.5 99.0 99.9 102.3 102.4 99.0 96.7 99.3 101.9 BTPhen ± 0.8 ± 1.6 ± 0.7 ± 1.6 ± 0.4 ± 2.5 ± 2.7 ± 0.5 ± 0.5 ± 0.5 ± 2.1

Table 3b (Al)

Isotope Recovery ± Error / %

241 244 152 9 88 89 27 133 137 139 140 Ligand Am Cm Eu Be Sr Y Al Cs Ba La Ce

CyMe4- 94.8 97.8 95.0 98.8 98.4 98.4 86.5 92.7 92.1 94.3 97.6 BTPhen ± 2.1 ± 2.1 ± 6.2 ± 1.3 ± 0.7 ± 1.9 ± 2.2 ± 1.2 ± 1.0 ± 1.1 ± 1.3 65.4 63.6 91.3 98.2 91.8 100.6 92.5 91.8 94.5 97.5 96.0 CA-BTP ± 2.7 ± 2.7 ± 5.9 ± 1.6 ± 1.3 ± 1.4 ± 1.6 ± 1.8 ± 1.8 ± 1.0 ± 1.5 CyMe4- 84.5 94.2 94.6 99.8 99.9 99.1 84.3 98.8 87.3 91.5 86.2 BTBP ± 2.3 ± 2.7 ± 6.0 ± 0.7 ± 0.8 ± 1.1 ± 2.2 ± 0.7 ± 1.1 ± 1.0 ± 0.3 CA- 43.5 53.1 101.4 99.8 100.0 93.6 100.5 102.9 100.6 101.7 95.6 BTPhen ± 2.7 ± 2.7 ± 5.9 ± 0.9 ± 0.6 ± 0.9 ± 0.8 ± 3.2 ± 1.5 ± 1.8 ± 0.6

141 146 147 157 159 163 165 166 169 172 175 Pr Nd Sm Gd Tb Dy Ho Er Tm Yb Lu

CyMe4- 101.8 99.0 96.5 91.7 96.0 96.9 92.8 98.9 100.2 97.5 98.0 BTPhen ± 2.1 ± 0.3 ± 0.5 ± 0.6 ± 0.3 ± 1.0 ± 0.8 ± 0.7 ± 0.8 ± 0.5 ± 0.9 96.7 94.0 93.2 95.1 92.7 93.2 97.3 94.9 97.9 102.2 101.3 CA-BTP ± 1.1 ± 0.9 ± 1.0 ± 0.5 ± 1.3 ± 1.4 ± 1.7 ± 1.0 ± 2.0 ± 1.9 ± 1.2 CyMe4- 89.3 91.6 94.8 95.8 94.2 95.3 98.3 91.5 93.5 99.2 99.9 BTBP ± 1.4 ± 1.4 ± 1.6 ± 1.2 ± 1.4 ± 0.5 ± 1.1 ± 1.4 ± 0.6 ± 0.6 ± 0.6 CA- 99.1 98.1 95.2 99.6 98.2 100.4 101.0 99.4 99.5 97.3 91.4 BTPhen ± 0.8 ± 0.9 ± 0.7 ± 0.3 ± 0.4 ± 0.5 ± 1.8 ± 0.5 ± 0.5 ± 0.5 ± 0.5

Table 3a and 3b: Recoveries of eluted matrix elements for high Al and high Ca from AG1-X8 column chromatography from analysis by alpha spectrometry, gamma spectroscopy and ICP- MS. (N.B. error calculated from instrument and method error plus one sigma, sample data 62

averaged from n=20 analysis run in triplicate, all calibrated against a range of standards in the concentration ranges above).

9.54 Am/Cm Extraction from Simulated Nuclear Forensic Matrices by CyMe4BTPhen

without Prior Chromatography

CyMe4BTPhen was used to extract Am/Cm in the manner outlined above from the normal matrix and matrices with high level Fe, Ca and Al, without the AG1-X8 chromatography method. Am/Cm extraction was unaffected with recoveries >90% within 90 minutes, with minimal extraction of Eu (Table 4). Recoveries determined for matrix elements including lanthanides by ICP-MS were >88%, including Ag and Mo which were removed by AG1-X8 in our previous investigation. A notable exception is Cd, which was co-extracted from all matrices by CyMe4BTPhen (recoveries 35–45% depending on matrix. This can be rationalised as the closely related N-donor ligand 2,9-di(2-pyridyl)-1,10-phenanthroline has been shown to be a highly selective sensor of Cd(II) [14]. Although in our proposed AG1-X8/

CyMe4BTPhen extraction procedure this would not pose a problem as Cd would be removed in the chromatography step, Cd can also be selectively removed from the Am/Cm solution using known separation techniques [15]. For the high level Fe, Ca and Al matrices, recoveries of Fe, Ca and Al were >86% by ICP-MS (Table 4).

After separation, the aqueous phase was subjected to chromatography to separate U,

Pu and Np for determination by alpha spectrometry. Recoveries >90% were obtained for the major actinides indicating presence of these elements does not affect Am/Cm extraction.

Overall, the initial method of AG1-X8/ CyMe4BTPhen separation would be preferred in order to simplify the matrix and avoid Cd co-extraction, however this investigation shows that CyMe4BTPhen may be used to extract Am/Cm in the presence of many competing ions with minimal co-extraction.

Ligand: CyMe4BTPhen, Isotope Recovery ± Error / % Matrix 241Am 244Cm 152Eu 56Fe 43Ca 27Al Fe(III) 91.3 ± 2.3 90.2 ± 2.3 95.1 ± 5.1 100.1 ± 3.1 ------Ca(II) 90.9 ± 2.4 91.0 ± 2.5 100.3 ± 4.1 ------91.0 ± 0.4 ------Al(III) 90.1 ± 2.1 90.5 ± 2.2 94.3 ± 4.5 ------88.9 ± 2.5

9Be 88Sr 89Y 95Mo 107Ag 111Cd

Fe(III) 99.1 ± 4.4 100.5 ± 1.6 86.6 ± 1.4 100.0 ± 0.4 100.8 ± 0.8 44.7 ± 0.4 Ca(II) 98.1 ± 1.4 98.8 ± 3.7 95.7 ± 1.1 102.1 ± 1.1 98.6 ± 1.0 36.3 ± 1.0 Al(III) 98.8 ± 2.1 98.4 ± 3.2 98.4 ± 1.9 98.1 ± 0.3 92.8 ± 0.7 34.7 ± 1.9

133Cs 137Ba 139La 140Ce 141Pr 146Nd

Fe(III) 98.4 ± 1.1 88.2 ± 0.3 90.3 ± 1.4 94.3 ± 1.0 96.4 ± 1.2 90.4 ± 1.1 Ca(II) 98.3 ± 1.2 98.3 ± 0.5 98.9 ± 1.5 92.5 ± 0.8 96.9 ± 0.9 99.2 ± 1.2 Al(III) 92.7 ± 1.2 92.1 ± 1.0 94.3 ± 1.1 97.6 ± 1.3 101.8 ± 2.0 99.0 ± 0.3

147Sm 157Gd 159Tb 163Dy 165Ho 166Er

Fe(III) 100.8 ± 0.9 94.2 ± 0.7 92.9 ± 0.7 93.8 ± 0.6 88.6 ± 1.1 94.2 ± 0.8 Ca(II) 102.7 ± 2.9 101.1 ± 1.3 101.9 ± 2.0 100.6 ± 1.1 96.8 ±1.8 100.5 ± 1.2 Al(III) 96.5 ± 0.5 91.7 ± 0.6 96.0 ± 0.7 96.9 ± 1.0 92.8 ± 0.8 98.9 ± 0.9

169Tm 172Yb 175Lu 232U 239Pu 237Np

Fe(III) 91.9 ± 0.7 96.7 ± 0.7 100.6 ± 0.7 96.4 ± 3.2 96.3 ± 2.0 96.0 ± 0.7 Ca(II) 100.3 ± 1.5 100.8 ± 1.0 92.7 ± 0.8 96.1 ± 3.3 98.5 ± 2.2 95.4 ± 0.6 Al(III) 99.5 ± 0.7 97.5 ± 0.8 98.0 ± 0.5 96.7± 3.3 98.5 ± 2.1 92.6 ± 0.7

Table 4: Recoveries of eluted matrix elements for high Al/Ca/Fe initial CyMe4BTPhen (10 mM, 90 min) separation of nuclear forensic matrix, analysis by alpha spectrometry, gamma spectroscopy and ICP-MS. (N.B. error calculated from instrument and method error plus one sigma, sample data averaged from n=5 analysis run in triplicate, all calibrated against a range of standards in the concentration ranges above).

9.6 Conclusions

We have developed a selective radiochemical separation process for Am/Cm for application to complex nuclear forensics samples using AG1-X8 chromatography followed by a triazine ligand extraction. In our procedure, purification and quantification of all major actinides can be achieved in < 2 days, which would be valuable in a nuclear forensics investigation. Our method avoids use of multiple column techniques and generation of thiocyanate waste. All four triazine ligands tested were found to extract Am/Cm from the matrix after AG1-X8 chromatography with little complexation of competing ions.

CyMe4BTPhen was the best ligand for the procedure, as the improved Am/Cm extraction kinetics allow recoveries >90% to be obtained in 90 minutes. However, we believe all four ligands tested would be viable alternatives, and many are more easily accessed synthetically which may be of benefit to application in an analytical laboratory.

Overall, our investigation shows that these triazine ligands are capable of extracting

Am/Cm from complex nuclear forensic matrices, including those with high level Fe, Ca and

Al. Of particular note is the ability of the ligands to perform separation of Am/Cm from all the lanthanides, which to our knowledge has not been tested before with ligands based on the triazine motif. We hope that this work can be utilised in nuclear forensic investigations and in the wider field of analytical radiochemistry to provide a new rapid method for Am/Cm separation.

9.7 Acknowledgements

Funding for this project was provided by AWE via a studentship to MAH. ICP-MS analyses were carried out by Mr Paul Lythgoe (School of Earth, Atmospheric and Environmental

Sciences, The University of Manchester).

9.8 Notes and References

1. Mayer, K., Wallenius, M., Ray, I.: Nuclear forensics- A Methodology Providing

Clues on the Origin of Illicitly Trafficked Nuclear Materials. Analyst 130, 433 (2005).

2. Goldstein, S. L., Hensley, C. A., Armenta, C. E., Peters, R. J.: Environmental and

Human Monitoring of Americium-241 Utilizing Extraction Chromatography and α-

Spectrometry. Anal. Chem. 69, 809 (1997).

3. Wallenius, M., Mayer, K.: Age determination of plutonium material in nuclear

forensics by thermal ionisation mass spectrometry. Fresenius J. Anal. Chem. 366, 234

(2000).

4. Maxwell, S. L.: Rapid method for determination of plutonium, americium and curium

in large soil samples. J. Radioanal. Nucl. Chem. 275, 395 (2008).

5. Maxwell, S. L., Culligan, B. K., Kelsey-Wall, A., Shaw, P. J.: Rapid radiochemical

method for determination of actinides in emergency concrete and brick samples. Anal.

Chim. Acta. 701, 112 (2011).

6. Luisier, F., Alvarado, J. A. C., Steinmann, P., Krachler, M., Froidevaux, P. J.: A new

method for the determination of plutonium and americium using high pressure

microwave digestion and alpha-spectrometry or ICP-SMS. Radioanal. Nucl. Chem.

281, 425 (2009).

7. Lewis, F. W., Hudson, M. J., Harwood, L. M.: Development of Highly Selective

Ligands for Separations of Actinides from Lanthanides in the Nuclear Fuel Cycle.

Synlett 2609 (2011).

8. Adam, C., Kaden, P., Beele, B. B., Müllich, U., Trumm, S., Geist, A., Panak, P. J.,

Denecke, M. A.: Evidence for covalence in a N-donor complex of americium(III).

Dalton Trans. 42, 14068 (2013).

9. Lewis, F. W., Harwood, L. M., Hudson, M. J., Drew, M. G. B., Desreux, J. F., Vidick,

G., Bouslimani, N., Modolo, G., Wilden, A., Sypula, M., Vu, T.-H., Simonin, J.-P.:

Highly Efficient Separation of Actinides from Lanthanides by a Phenanthroline-

Derived Bis-triazine Ligand. J. Am. Chem. Soc. 133, 13093 (2011).

10. Trumm, S., Geist, A., Panak, P. J., Fanghänel, T.: Highly Efficient Separation of

Actinides from Lanthanides by a Phenanthroline-Derived Bis-triazine Ligand. Solvent

Extr. Ion Exch. 29, 213 (2011).

11. Hudson, M. J., Harwood, L. M., Laventine, D. M., Lewis, F. W.: Use of Soft

Heterocyclic N-Donor Ligands To Separate Actinides and Lanthanides. Inorg. Chem.

52, 3414 (2013).

12. Whittaker, D. M., Griffiths, T. L., Helliwell, M., Swinburne, A. N., Natrajan, L. S.,

Lewis, F. W., Harwood, L. M., Parry, S. A., Sharrad, C. A.: Lanthanide Speciation in

Potential SANEX and GANEX Actinide/Lanthanide Separations Using Tetra-N-

Donor Extractants. Inorg. Chem. 52, 3429 (2013).

13. Laventine, D. M., Afsar, A., Hudson, M. J., Harwood, L. M.: Tuning the Solubilities

of Bis-triazinylphenanthroline Ligands (BTPhens) and Their Complexes

Heterocycles. 86, 1419 (2012).

14. Turan, M. D., Safarzadeh, M. S.: Separation of zinc, cadmium and nickel from ZnO–

CdO–NiO mixture through baking with ammonium chloride and leaching.

Hydrometallurgy 119-120, 1 (2012).

15. Cockrell, G. M., Zhang, G., VanDerveer, D. G., Thummel, R. P., Hancock, R. D.:

Enhanced Metal Ion Selectivity of 2,9-Di-(pyrid-2-yl)-1,10-phenanthroline and Its

Use as a Fluorescent Sensor for Cadmium(II). J. Am. Chem. Soc. 130, 1420 (2008).

10.0 Appendix to Paper One

10.1 Method implementation with UTEVA, TEVA and TEVA-TRU and comparison with DGA resin method for Am (III)/Ln (III)

The resins AG1-X8 and UTEVA are both applicable to purify all the actinides (apart from

Am/Cm) from a typical nuclear forensic sample matrix. To widen the applicability of our liquid-liquid or polymer americium purification step we planned to develop separation schemes based on AG1-X8, TRU, TEVA, DGA and UTEVA with the Am/Ln separation already developed. DGA resin would be an interesting comparison resin to our developed methods as it is currently one of the most utilised Ln selective EC resins. It utilises N-donor chemistry and forms 3-4 metal to ligand complexes and reverse micelles. It is the ligand

TODGA immobilised on Amberlite XAD-7, and has many published methods achieving Am separation from complex matrices e.g. high level Fe.65 TRU and TEVA resins are also applicable as they are very selective for trans-uranium and tetravalent actinides respectfully and have very little affinity for fission products.66

The combination processes of TEVA-BTPhen, TEVA-TRU-BTPhen, TEVA-TRU-

DGA and UTEVA-BTPhen were all tested with a typical nuclear forensic sample matrix and a U, Pu, Np and Am fraction separation attempted. The methods were all either based or adapted from Maxwell et al.67 and Morgenstern et al.68 The methods were then compiled into a simplified scheme and further purification methods added to produce flow sheets applicable to BTPhen separation (Figures 1-4).

Figure 1 TEVA-TRU-BTPhen separation scheme

Sample in 4M HNO3 containing Lanthanides, U, Np, Pu, Am & Cm

1. Stack TEVA-TRU column 2. Adjust oxidation state with 1.5 M ascorbic acid and 1 mL 3.5 M NaNO2 3. Add sample with 3 M HNO3 (5 mL), wash with 3 M HNO3 (5 mL) 4. Remove TRU column (Am/Cm) 5. Wash TEVA column with 15 mL 3 M HNO3 6. Wash with 9 M HCl (15 mL, Th) 7. Elute Np with 4 M HCl (15 mL) 8. Elute Pu with 0.1 M HCl/HI

1. TRU Column, elute Am/Cm with 4 M HCl (15 mL) 2. Wash with 3 M HCl (15 mL) 3. Elute U with 0.1 M Ammonium Oxalate (15 mL)

Adjust sample to 2-4 M HCl or HNO3 contact with Evaporate U, Pu, Np to dryness and electrodeposited equal volume of 10 mM BTPhen in 1-octanol for alpha analysis

Strip organic phase with 0.1 M HCl and electoplate for Am/Cm analysis

Elute Ln selectively using α-HIBA column or dissolve entire remaining matrix for ICP-MS analysis

Figure 2 TEVA-TRU-DGA Separation Scheme

Sample in 4M HNO3 containing Lanthanides, U, Np, Pu, Am & Cm

1. Stack TEVA-TRU column 2. Adjust oxidation state with 1.5 M ascorbic acid and 1 mL 3.5 M NaNO 2 3. Add sample with 3 M HNO3 (5 mL), wash with 3 M HNO3 (5 mL) 4. Remove TRU column (Am/Cm)

5. Wash TEVA column with 15 mL 3 M HNO3 6. Wash with 9 M HCl (15 mL, Th) 7. Elute Np with 4 M HCl (15 mL) 8. Elute Pu with 0.1 M HCl/HI

1. TRU Column, attach DGA column, elute Am/Cm with 3 M HCl (15 mL), remove DGA column 2. Elute U from TRU resin with 0.1 M Ammonium Oxalate (15 mL) 3. Elute Am/Cm from Ln using 0.25 M HCl (15 mL) 4. Elute Ln selectively using α-HIBA column or dissolve entire remaining matrix for ICP-MS analysis 5.

Evaporate U, Pu, Np, Am, Cm to dryness and electrodeposition for alpha analysis

Figure 3 UTEVA-BTPhen separation scheme

Sample in 4M HCl containing lanthanides, U, Np, Pu, Am & Cm

Chromatographic Resin (UTEVA) conditioned with 6 M HNO3, 0.3% H2O2

1. 6 M HNO3, 0.3% H2O2 (An (III), Ln(III), fission products separated) 2. 2 M HNO3, 2 mM ascorbic acid, 2 mM NH2OH (Pu separated) 3. 2 M HNO3, 0.1 mM oxalic acid (Np separated) 4. 7 mM (NH4)C2O4 (U separated)

Evaporate Am/Cm/Ln fraction to dryness, dissolve in Adjust sample to 2-4 M HCl or HNO3 contact with equal volume of 10 mM BTPhen in 1-octanol

1. 1. Strip organic phase with 0.1 M HCl and electoplate for Am/Cm analysis 2. Evaporate U,Pu and Np samples for alpha analysis 3. Elute Ln selectively using α-HIBA column or dissolve entire remaining matrix for ICP-MS analysis

Figure 4 TEVA-BTPhen separation scheme

Sample in 4M HNO3 containing Lanthanides, U, Np, Pu, Am & Cm

1. Conditioned 4 M TEVA column

2. Adjust oxidation state of sample with 1.5 M ascorbic acid and 1 mL 3.5 MNaNO2 3. Add sample with 3 M HNO3 (5 mL), wash with 3 M HNO3 (5 mL) 4. Retain previous fraction and wash with 3 M HNO3 (15 mL) to elute Am/Cm/Ln 5. Wash TEVA column with 15 mL 3 M HNO3 6. Wash with 9 M HCl (15 mL, Th) 7. Elute Np with 4 M HCl (15 mL) 8. Elute Pu with 0.1 M HCl/HI 9. Elute uranium with 0.1 M HCl with 0.1 mM ammonium oxalate

1. Evaporate Am/Cm/Ln fraction to dryness, dissolve in Adjust sample to 2-4 M HCl or HNO3 contact with equal volume of 10 mM BTPhen in 1-octanol 2. Strip organic phase with 0.1 M HCl and electoplate for Am/Cm analysis 3. Elute Ln selectively using α-HIBA column or dissolve entire remaining matrix for ICP-MS analysis 4. 5.

Evaporate U, Pu, Np fractions to dryness and electrodeposition for alpha analysis

10.2 Resin investigation results

The separations in Figures 1-4 were undertaken with a typical sample matrix solution spiked with 10 Bq of 241Am (III), 152Eu (III), 232U (VI), 237Np (IV) and 239Pu (IV). The separated

Am/Ln/FP fraction were analysed by gamma spectroscopy to determine the recovery of Am

(III)/Eu (III). The fractions containing U, Pu and Np from each method were evaporated to dryness, further purified and analysed by alpha spectrometry. The waste fractions and Eu/Ln elutions in each case were retained analysed by ICP-MS with a standard created from the starting model nuclear forensic matrix. TRU, TEVA and UTEVA all appeared to perform as reported in the literature giving U, Pu and Np fractions for analysis. The recoveries for all the actinides in the solution are tabulated in tables 1-2 below.

241Am Recovery 152Eu Recovery / Separation Method / % Error / % % Error / % TEVA-TRU-BTPhen (Am elution) 100.6 1.6 5.4 4.4 TEVA-TRU-BTPhen (Ln elution) N/D N/D 102.6 8.2 TEVA-TRU-DGA (Am Elution) 103.7 1.7 5.1 4.7 TEVA-TRU-DGA (Ln Elution) N/D N/D 96.7 9.1 UTEVA-BTPhen (Am Elution) 94.9 1.6 4.9 4.7 UTEVA-BTPhen (Ln Elution) N/D N/D 107.0 9.3 TEVA-BTPhen (Am elution) 100.8 1.6 3.0 4.7 TEVA-BTPhen (Ln elution) N/D N/D 93.5 9.3 Table 1. Am/Eu recoveries (by gamma spectroscopy) for the Am and Ln fractions from each separation method, N/D is given as a value if the recovery was determined to be less than 1% and the error being larger than or equal to the net area of the peak.

Separation 239Pu Recovery Error / 232U Recovery Error / 237 Np Recovery Error / Method / % % / % % / % % TEVA-TRU- BTPhen 80.1 1.9 91.8 2.2 91.5 1.8 TEVA-TRU-DGA 85.1 2.2 88.1 2.2 89.9 1.9 UTEVA-BTPhen 86.1 2.3 90.2 2.3 93.1 2.1 TEVA-BTPhen 90.1 1.7 97.2 1.9 99.9 1.7 Table 2. Pu, Np and U recoveries calculated by alpha spectrometry for the four separation methods with associated errors

Recovery of Ln matrix elements from EC resin separations involving CyMe4BTPhen 120

100

40 Recovery / % Recovery/

0 139 La 140 Ce 141 Pr 146 Nd 147 157 Gd 159 Tb163 Dy 165 Ho 166 Er 169 172 Yb 175 Lu Sm Tm Separation scheme and isotope TEVA-TRU-BTPhen TEVA-TRU-DGA UTEVA-BTPhen TEVA-BTPhen

Figure 5. Recovery of Ln matrix elements using EC resin separations schemes involving CyMe4BTPhen.

As can be seen from the ICP-MS data, almost quantitative recoveries of all Ln was observed from the Ln elution’s from all methods, however an overall low trend can be seen for the early Ln isotopes with the TEVA-TRU-BTPhen separation scheme. This could be due to experimental error or minor co-extraction. Overall comparing all the gamma, alpha and ICP-

MS data it can be said that all four schemes are able to achieve a high quality separation of all the major actinides and achieve the challenging Am/Ln separation.

10.3 Investigating the effect on the addition of relevant fission and activation products on the BTPhen separation from a typical nuclear forensic matrix

As stated in detail above, nuclear forensic investigations after the identification of the princical isotopes of interest (e.g. An, FP) will look for further information to build up a body

of evidence from both the concentrations of matrix elements and activation products. Most potential signatures are described in a report from the environmental protection agency in the

USA and can be shown in the reproduced table below.69

Alpha Beta/Gamma Emitters Emitters Ba-140/La-140 Nd-147/Pm-147 Ru-106/Rh-106 U-234 Ce-141 Eu-155 Sb-125 U-235 Ce-143/Pr-143 H-3 Sr-89 U-238 Ce-144/Pr-144 I-131/Xe-131 Sr-90/Y-90 Pu-238 Cs-134 I-133 Tc-99 Pu-239 Cs-137 Np-239 Te-132/I-132 Pu-240 Eu-154 Pm-151/Sm151 Zr-95/Nb-95 Pu-241 Mo-99/Tc-99m Ru-103/Rh-103 Zr-97/Nb-97 Am-241 Activation Products Co-58 Ag-110m Cr-51 Mn-54 Np-239 Co-60 Fe-59 Na-24

Figure 6. Potential β/γ and activation products that may require analysis in a nuclear forensic investigation.

Some of the isotopes in figure 6 have already been evaluated in our separation method; furthermore, Na, K and H have been present in our matrices although not evaluated. Pm would be expected to not interfere based on our lanthanide data. We decided to test our full lanthanide typical matrix with addition of 1 mg mL-1 Mn, Ru, Rh, Sb, Te (chloride form) and

-1 with the addition of 1 mg mL I (as I2). The BTPhen separation procedure was undertaken and performed as reported above. The matrix was submitted for ICP-MS analysis and the recoveries of these additional matrix elements were evaluate and tabulated in figure 7 below.

Recovery of relevant fission/neutron activation products from BTPhen matrix separation from a typical nuclear forensic matrix 120

100

60 Recovery/ Recovery/ % 40

0 127 I 101 Ru 103 Rh 128 Te 55 Mn 122 Sb Isotope

Figure 7. Recovery of relevant fission/neutron activation products from BTPhen matrix separation from a typical nuclear forensic matrix

Interestingly again the selectivity for americium over 90 minutes saw recoveries of >90% for all isotopes. This implies that regular purification methods for these isotopes would be applicable after americium purification. In regards to the matrix, only Pb, Co and Tc have not been tested with the matrix.

10.4 BTPhen separations with Co, Pb and Tc

After our previous complex matrix separations looking at the lanthanides and fission/activation products we decided to look into other relevant isotopes in the periodic table. The following isotopes (Pb, Co, Tc) were chosen for the following reasons, firstly 99Tc as it is a long lived significant fission product of uranium, therefore the concentrations of 99Tc in a material are relevant to the processing of a uranium containing material, furthermore it

would add to the data for the triazine ligands for reprocessing applications. Co was also chosen for this matrix investigation as 60Co is a common neutron activation product that may be present in complex matrices and therefore its affinity for the ligand is relevant. Finally Pb was chosen to investigate as its isotopes form the end of the uranium decay series and will be present in many environmental samples and may require purification from a matrix for lead dating. Furthermore, the interaction of Pb with the ligand along with Co and Tc would cover the fate of almost all potential matrix elements in our process (e.g. An, Ln, F/AP and Matrix elements).

As shown previously our current procedure at 4 M HCl (containing 1 mg mL-1 trace metals including Co and Pb) was repeated with the addition of 10 Bq of 241Am, 152Eu, 99Tc

(130 Bq). BTPhen (10 mM, 1 mL) was added and our 90 minute procedure carried out. The subsequent organic and aqueous phases were counted by a gamma spectroscopy for Am/Eu and recoveries calculated. The subsequent aqueous phase was then analysed by ICP-MS for

Co and Pb and then dissolved in a scintillation cocktail along with a 130 Bq 99Tc standard to calculate the recovery of 99Tc in the process. The results are shown below (Table 11).

99Tc Error/ BTPhen 207Pb Recovery / Error / 59Co Recovery / Error / Recovery/ % % Matrix % % % %

Co, Pb and Tc 99.0 2.4 95.6 4.1 90.1 3.9 Table 3: Pb and Co recoveries from a BTPhen Am separation determined by ICP-MS), Tc recovery determined by liquid scintillation counting of aqueous phase before and after separation with 10 mM CyMe4BTPhen.

As we did not predict any competition with the Am separation from the size of the ions or the oxidation states in aqueous solution and the results above all showed high recoveries for the three isotopes (>90%). The lack of affinity for Co and Pb are promising at these

concentrations for application to environmentally derived nuclear forensic samples. The Tc data may need to be repeated as it would appear that some Tc is complexed within the error of the method. Therefore we have tested a large majority of relevant isotopes that could appear in a typical nuclear forensic sample with our method, apart from high levels of matrix elements (e.g.Ca/Al) and Cd the method is unaffected by all other matrix elements as concentrations ~1 mg mL-1. Below a chart outlines the overall recoveries of all isotopes from our method (Figure 8).

10.5 Appendix 1 References

65Horwitz, E. P., Bloomquist, C. A. A. (1972). Separation, performance and factors effecting band spreading of high efficiency extraction chromatographic columns for actinide separations. Journal of Inorganic and Nuclear Chemistry 34, 3851-3871. 66 Lee, M. H., Park, T. H., Park, J., H., Song, K., Lee, M. S. (2013). Radiochemical separation fo Pu, U, Am and Sr isotopes in environmental samples using extraction chromatography resins. Journal of Radioanalytical and Nuclear Chemistry, 295, 1419-1422. 67 Maxwell, S. L. (1998). Rapid Actinide-Separation Methods, Radioactivity and Radiochemistry. 8(4), 36-44. 68 Morgenstern, A., Apostolidis, C., Carlos-Marquez, R., Mayer, K., Molinet, R. (2002). Single-column extraction chromatographic separation of U, Pu, Np, and Am. Radiochimica Acta, 90, 81-85. 16 Radiological Laboratory Sample Analysis Guide for Incident Response - Radionuclides in Soil. EPA 402-R-12-006. Maryland: Environmental Management Support, INC., 2012.

Figure 8. Matrix element recovery from our devised BTPhen separation method represented (with errors) in a Periodic Table (without noble gases), all isotopes apart from U, Pu and Np were 1 mg mL-1 and analysed by ICP-MS, U, Pu and Np were determined after AG1-X8 separation and alpha spectrometry analysis. 11.0 Chapter 2

The material in the following section has been published in the journal Reactive and

Functional Polymers and represents the final manuscript

The author synthesised the ligand/polymer, designed and performed the radiochemical method development, analysed the data, interpreted the results and wrote first draft and the final version of the manuscript.

DOI: http://dx.doi.org/10.1016/j.reactfunctpolym.2015.05.002

Separation of americium from complex radioactive mixtures using a BTPhen extraction chromatography

resin

Matthew A. Higginson,a Olivia J. Marsden,b Paul Thompson,b Francis R. Livensa and Sarah L.

Heath.a,* a Centre for Radiochemistry Research, School of Chemistry, The University of Manchester,

Manchester, M13 9PL, UK. b AWE, Aldermaston, Reading, RG7 4PR, UK.

* Corresponding author, [email protected], tel. +44 161 275 4696

80 11.1 Abstract

Extraction chromatography (EC) resins are widely used in analytical radiochemical separations, in particular for actinide separation. However, there is currently limited choice for separation of americium using EC, with DGA (N,N,N’,N’-tetra-n-octyldiglycolamide) resin being the preferred option. Here, we describe preparation and testing of a covalently-linked EC resin utilising a triazine soft N-donor (BTPhen) extractant for americium extraction. The resin was generated by conjugation of a Me4BTPhen derivative with poly(vinylbenzyl) chloride to generate

PVB-Me4BTPhen. PVB-Me4BTPhen was shown to extract americium from a complex matrix simulating nuclear forensic samples, and containing lanthanides, actinides and matrix elements with high Am (III) recovery (>90%) and low extraction of other elements, and provides an alternative to the currently used BTPhen liquid-liquid separation process for Am (III) extraction.

The synthetic strategy used will also be applicable to immobilisation of other triazine soft N- donor ligands for americium separation.

Keywords

americium separation, nuclear forensics, extraction chromatography, BTPhen

81 11.2 Introduction

Soft N-donor ligands based on 1,2,4-triazines have shown selectivity for trivalent actinides (such as Am (III) and Cm) over other actinides and lanthanides [1]. Such ligands have high Am

(III)/Eu (III) selectivity factors and are applicable to both the proposed n-SANEX (N-donor

Selective Actinide Extraction) industrial reprocessing methodology and to complex mixtures typically expected in nuclear forensic investigations [2]. Previously, we reported use of the ligand CyMe4BTPhen 1 (SF Am/Eu >250; Figure 1) to achieve americium separation from simulated nuclear forensic matrices containing actinides, lanthanides and matrix elements (e.g. d- block metals, Fe, Ca) [3]. This process and similar separations using soft N-donor ligands on Am

(III)/Eu (III) mixtures [4] employ liquid-liquid organic/aqueous extraction. However, this technique has limitations arising from poor ligand solubility, [5] variation in extraction capabilities with mixing type [3] and slow phase transfer [1]. Immobilisation of the ligand to generate an extraction chromatography (EC) resin could therefore offer advantages over the current liquid-liquid process. Other actinide-selective ligands such as diamyl amine phosphonate

(DAAP) have been immobilised to generate EC resins (e.g. Amberlite XAD-7 (UTEVA®)). The ability to recondition the resin and vary elution methods allows extensive use in radiochemical analysis [6]. Further benefits over liquid-liquid extraction include no requirement for mixing or phase separation, and possible use of a vacuum system to increase flow rate. There are numerous

EC resin methods for separation of U, Pu, Th and Np (e.g. TRU, TEVA®, UTEVA®). However, the only methods for Am extraction on EC resins utilise DGA or LN resin, and these methods cannot typically achieve Am (III)/ Ln (III) separations without additional purification [7]. Also, most EC resins are generated by adsorption of a ligand into a polymeric support such as

82 Amberlite, rather than covalent linkage between ligand and support, as has been shown for the soft N-donor ligand iso-propyl BTP [8]. Therefore, we surmised that a covalently-linked, soft N- donor ligand based EC resin for Am (III) (and/or Cm) could offer new benefits over the current resins for Am (III) separation in radiochemical analysis, especially in nuclear forensic investigations where a rapid result for americium is desired.

Here, we report synthesis of Me4BTPhen-polyvinyl benzyl (PVB) cross-linked polymer 2

(hereafter referred to as PVB-Me4BTPhen, Figure 1) and modification of our previously reported

Am (III) extraction method utilising PVB-Me4BTPhen. Initial experiments showed selective recovery of Am (III) over Eu (III); application of PVB-Me4BTPhen to simulated nuclear forensic samples showed extraction of Am (III) with low levels of co-extraction of matrix elements and other actinides (>90% recovery of Am (III) and low (<10%) extraction of U (VI), Np (IV), Pu

(IV) and all stable matrix elements with the exception of Cd (II) and Pr (III) (>20%)). We propose that the EC resin PVB-Me4BTPhen 2 offers a new medium for the separation of americium from complex radioactive mixtures, in particular the lanthanides and early actinides to Pu. Our method removes the need for liquid-liquid separation and, unlike conventional Am

(III) EC resins, does not require the use of thiocyanate for elution.

Figure 1. Structures of CyMe4BTPhen 1 and PVB-Me4BTPhen 2. 11.3 Experimental

11.31 General Experimental Detail

All radionuclides used were provided from calibrated stocks in the School of Chemistry,

University of Manchester. Micropipettes of 100 µL, 0.1–1 mL and 2–10 µL were calibrated on a

4 d.p. balance with >18 MΩ deionised water in the temperature range 18–22 °C and were found to be within their stated range. All acid solutions were made from analytical grade concentrated solutions and were diluted with >18 MΩ deionised water. All solutions were considered to have expired within one month of creation. Gamma counting was performed using a Canberra 2020 coaxial HPGe gamma spectrometer with an Ortec 919E multi-channel analyser. Alpha spectrometry was performed on a Canberra model 7401VR detector with multi-channel analyser.

ICP-MS analysis was performed on an Agilent 7500cx spectrometer. Multiple standards for each element in the range 1–100 ppb were used for ICP-MS quantification. Percentage recoveries of metal ions are calculated relative to initial amounts from the nuclear forensic matrix. All reagents and solvents used were of standard analytical grade unless otherwise stated. Melting points were determined using a Stuart Scientific SMP10 apparatus and are uncorrected. IR spectra were recorded for solid samples using a Bruker Alpha-P ATR spectrometer. NMR spectra were recorded for solutions in CDCl3 or d6-DMSO on a Bruker Avance III instrument (400 MHz) and

84 were referenced to the residual solvent signal. Assignments were determined using COSY and

HMQC experiments. Coupling constants (J values) are quoted to the nearest 0.1 Hz. Low resolution mass spectra were measured on a Micromass Platform II instrument with electrospray ionisation. Accurate mass measurements were obtained using a Micromass Q-TOF instrument with electrospray ionisation. Preparative column chromatography was performed using Sigma-

Aldrich silica gel (technical grade, 60 Å, 220–240 mesh, 35–75 μm) and the flash technique [9].

All solvents used were of standard laboratory grade unless otherwise specified. Compositions of solvent mixtures are quoted as ratios of volume. Organic solutions were dried with anhydrous magnesium sulfate.

11.32 Synthesis of PVB-Me4BTPhen 2

5-bromo-2,9-bis(5,5,8,8-tetramethyl-5,6,7,8-tetrahydrobenzo[e][1,2,4]triazin-3-yl)-1,10- phenanthroline 8

Diacetal (800 mg, 4.73 mmol) and triethylamine (2.0 mL, 14.2 mmol) were added to a suspension of amidrazone 7 (890 mg, 2.38 mmol) in 1,4-dioxane (75 mL) and the mixture was stirred at reflux for 3 d. After cooling to room temperature, the solvent was evaporated and the remaining semi-solid residue was triturated with ice-cold Et2O (30 mL). The insoluble solid was filtered and washed with further ice-cold Et2O (30 mL), then purified by column chromatography

(CH2Cl2/MeOH gradient) to afford the title compound 8 as a yellow solid (330 mg, 0.70 mmol,

+ 21%; m.p. 130–133°C); HRMS found 473.1114, C22H18BrN8 (M+H) requires 473.0810; δH (400

MHz, CDCl3) 9.01 (d, J = 8.6 Hz, 1 H), 8.93 (d, J = 8.4 Hz, 1 H), 8.88 (d, J = 8.6 Hz, 1 H), 8.39

(d, J = 8.4 Hz, 1 H), 8.30 (s, 1 H); δC (100 MHz, CDCl3): 162.6, 161.3, 161.1, 160.3, 160.1,

154.2, 154.0, 146.9, 146.0, 137.3, 136.3, 130.6, 129.8, 129.0, 123.8, 123.7, 122.0.

85 4-(2,9-bis(5,6-dimethyl-1,2,4-triazin-3-yl)-1,10-phenanthrolin-5-yl)aniline 9

BTPhen derivative 8 (500 mg, 1.06 mmol), 4-aminophenylboronic acid pinacol ester (257 mg,

1.17 mmol), Pd(PPh3)4 (61.5 mg, 53 μmol), and K2CO3 (0.59 g, 4.27 mmol) were suspended under N2. After addition of N2-purged THF (25 mL), methanol (7.5 mL) and water (7.5 mL), the mixture was stirred at 70° C for 24 h. The solvents were removed in vacuo. The resulting crude solid was then dissolved in CH2Cl2 (100 mL) and washed successively with aq. KOH (100 mL,

25% w/v) and water (100 mL). The organic phase was dried (Na2SO4) and the solvent was removed under reduced pressure. The crude product was purified using column chromatography

(CH2Cl2/MeOH 96:4) to give the title compound 9 (200 mg, 0.41 mmol, 39 %); HRMS found

+ 486.2110, C28H24N9 (M+H) requires 486.2101; δH (d6-DMSO): 9.01 (d, J = 8.6 Hz, 1 H), 8.93

(d, J = 8.4 Hz, 1 H), 8.88 (d, J = 8.6 Hz, 1 H), 8.30 (s, 1 H), 8.10 (d, J = 8.4 Hz, 1 H), 7.01 (d, J

= 8.1 Hz, 2 H), 6.71 (d, J = 8.2 Hz, 2 H) 2.78 (s, 6 H), 2.74 (s, 6 H), δC (d6-DMSO): 162.6.

161.1, 161.3, 160.3, 160.1, 156.0, 154.2, 154.0, 146.9, 146.0, 137.3, 136.3, 131.0, 130.6, 129.8,

129.0, 123.8, 123.7, 122.0, 121.8, 118.1, 34.3, 32.5.

11.33 Conjugation of Me4BTPhen derivative 9 to PVBC

Aniline functionalised BTPhen 9 (1.6 g) was dissolved in DMF (50 mL) and KOH added until the solution was ~ pH 10. PVBC (1.6 g) was added and the resulting mixture heated at reflux for

12 h with monitoring by proton NMR. After 24 h, Et3N (100 µL) was added, with further additions every 12 hours (50 µL aliquots). After 3 days no residual free ligand could be observed by NMR. The polymer was filtered and washed with warm DMF (100 mL), warm MeOH (20 mL), warm acetone (20 mL) and cold sodium methoxide (5 mM, 10 mL, in MeOH), dried at

60 °C for 24 h and then under high vacuum to afford polymer 2 (3.01 g, 95%). Analysis of 2 was

86 carried out as described in the text above (see Supplementary Information) and indicated the desired polymer was formed.

11.34 Am/Eu separation method

A known amount of polymer (10-20 mg) was suspended in a solution of 241Am (III)/152Eu (III)

(10-100 Bq each, 1 mL) in 4 M HNO3 for 1.5 h. The polymer was then back-stripped for 1.5 h with 0.1 M HCl. To ensure complete back-extraction of the polymer for re-use, the polymer was suspended in 0.1 M HCl for 24 h. The resulting three acidic phases were counted by gamma spectroscopy against a standard of known activity in the same geometry. For separations of the complex matrix, the procedure was used as above but with 1 h contact and back-extraction times.

11.35 Generation of Complex Nuclear Forensic Matrix

Using 4 M HCl (VWR, AnalaR NORMAPUR® grade) a solution containing ~1 mg mL-1 of metal ion was prepared by addition of the following elements in chloride form (Sigma-Aldrich or

Fluka, >99.99% trace metal): Be, Cs, Mo, Ce, Pr, Nd, Sm, Tb, Y, Zr, Ag, Cd, Sr and Ba. Spike solutions of 241Am (10-100 Bq range), 152Eu (10-100 Bq range), 232U (VI), 239Pu (IV) and 237Np

(IV) (10 Bq each isotope) were added to generate a known molarity solution.

87 11.4 Results and Discussion

11.41 Synthesis of PVB- Me4BTPhen

For synthesis of a BTPhen-based EC resin, a hydrophobic polymer support which was stable to radiolysis and oxidation by acidic media was required. Many common polymers such as polyvinylchloride (PVC) or polyvinylidinechloride (PVDC) are susceptible to degradation and yellowing at low radiation doses and polymethylpentene-based polymer backbones are susceptible to oxidation by acidic media [10]. Functionalised polystyrenes offer a wide range of moieties for ligand attachment and are stable to oxidation and radiolysis due to lack of aliphatic protons. Polyvinyl benzyl chloride (PVBC) was chosen for BTPhen immobilisation due to the presence of a primary alkyl chloride moiety for ligand attachment.

In order to produce a BTPhen EC resin, additional functionality was required on the ligand to allow attachment to a polymer support. Due to the electron rich phenanthroline system in neocuproine 3 (starting material for BTPhen synthesis), many conventional synthetic routes were not suitable to functionalise the molecule for immobilisation. Functionalisation of the 5/6 position of the phenanthroline ring was desired, as this should have the least effect on the triazine moiety which is functional for Am (III) complexation. There is literature precedent for functionalisation at C-5/C-6 with -Br, -Cl, -NO2 (which can be reduced to -NH2) and epoxide

(can be opened to hydroxyl) moieties [11-13]. Harwood et al. have previously derivatised a

BTPhen ligand by addition of a Br moiety at C-5/6. This requires no protecting group as the moiety is unaffected by the synthesis [14]. Subsequent Suzuki reaction with a di-functionalised linker was used to immobilise the ligand onto magnetic nanoparticles. We therefore decided to utilise this method for immobilisation of BTPhen onto PVBC, as variation of the linker would allow synthetic flexibility.

88 Brominated amidrazone 7 was synthesised as previously reported [14] (Scheme 1). We then attempted cross-coupling with 3,3,6,6-tetramethylcyclohexane-1,2-dione to synthesise Br-

CyMe4BTPhen, but we found the product of this reaction required extensive purification and was only obtained in low yield (N.B. this synthesis has since been achieved by Harwood and co-

workers [15]). Therefore, we elected to replace CyMe4 with a less sterically hindered group in order to make a more accessible triazine cavity for Am (III) which may assist complexation on the polymer. We also sought a diketone reagent which was commercially available (unlike

3,3,6,6-tetramethylcyclohexane-1,2-dione) to remove this bottleneck in the synthesis. Coupling of 7 with diacetal afforded Br-Me4BTPhen 8 in low yield (21%). Suzuki coupling of 8 with 4- aminophenylboronic acid pinacol ester afforded 9, which bears the amine functionality for PVB conjugation.

Scheme 1. Reagents: (i) Br2, fuming H2SO4, Δ, 85%; (ii) SeO2, dioxane/H2O, Δ, 74%; (iii)

Hydroxylamine hydrochloride, Et3N, anhydrous MeCN, Δ, then p-toluenesulfonyl chloride, pyridine, Δ, 48%; (iv) NH2NH2.H2O, EtOH, 49%; (v) Diacetal, Et3N, THF, Δ, 21%. (vi) 4- aminophenylboronic acid pinacol ester, Pd(PPh3)4, potassium carbonate, water, MeOH, THF, 70

°C, 39%; (vii) PVBC, Et3N, DMF, Δ, 95%.

Reaction of 9 with PVBC at reflux in DMF was monitored by proton NMR; no free 9 was observed in solution after 3 d. The resulting polymer could not be analysed by conventional mass spectrometry (or MALDI) due to the cross-linked nature. Due to low solubility of the polymer in organic solvents, many analytical techniques were not available (e.g. solution NMR, UV-vis,

90 HPLC, GC-MS). However, the IR spectrum of 2 showed bands at 1450 cm-1 and 1600 cm-1 (as seen for phenanthrolines) and stretches in the NH region at ~1620 cm-1. Elemental analysis of the

PVBC starting material showed a composition of C 70.08%, H 5.91%, Cl 23.32%. Replacement of all Cl moieties would be expected upon conjugation of Me4BTPhen with a predicted composition of C 74.14%, H 5.90%, N 19.95% [16]. Elemental analysis of the product polymer found C 71.93%, H 6.29%, N 19.17%, Cl 2.05%. This loss of Cl and increase in N content suggests that PVB-Me4BTPhen 2 was formed but some Cl functionality (~8%) remained. In addition, SSNMR (CP-MAS) showed key differences in the spectra of polymer 2 and PVBC (see

Supplementary Information), in particular signals at δ 160–165 ppm which were indicative of

C=N. SEM-EDX showed the polymer to be homogeneous in terms of elemental composition

(see Supplementary Information).

11.42 Initial development of americium extraction using PVB-Me4BTPhen

To assess the americium extraction capabilities of PVB-Me4BTPhen 2, we adapted our liquid- liquid separation procedure by removal of the 1-octanol phase, so that 2 would be suspended in an acidic matrix containing Am (III) for a period of time, followed by filtration and stripping of the polymer with low-molarity acid [3]. Firstly, stability of PVB-Me4BTPhen 2 in 4 M HCl and

4 M HNO3 was assessed over 1 month. A comparable elemental analysis was obtained for 2 and no physical change was observed, indicating stability over a long time period and potential for re-use of 2 in acidic radiochemical separations. In addition, extraction of Am (III) was seen to be unaffected after storage in acid for one month.

In an initial Am (III)/Eu (III) extraction investigation, PVB-Me4BTPhen 2 (20 mg) was

241 152 suspended in a solution of Am/ Eu (10 or 100 Bq each, 1 mL) in 4 M HNO3 for 90 minutes

(seen to be the time required for >80% Am (III) complexation in liquid-liquid separations with

91 Me4BTPhen). The polymer was then back-stripped for 90 minutes with 0.1 M HCl. To ensure complete back-extraction of the polymer for re-use, the polymer was suspended in 0.1 M HCl for

24 h. The resulting three acidic phases were counted by gamma spectroscopy against a standard of known activity in the same geometry (Table 1). The experiment was also repeated with PVBC as a control.

% 241Am in aqueous phase ± % 152Eu in aqueous phase ± Polymer Solution error error PVB- Aqueous phase after 90 min polymer contact (4 M 2.3±1.7 96.6±1.7 Me4BTPhen 2 HNO3) Polymer 90 min back extraction (0.1 M HCl) 96.5±1.5 3.4±1.6

Polymer 24 h back extraction (0.1 M HCl) 0.3±1.6 2.1±1.8 Aqueous phase after 90 min polymer contact (4 M PVBC control 99.1±1.9 92.9±2.1 HNO3) Polymer 90 min back extraction (0.1 M HCl) 1.7±1.8 6.7±1.8

Unfunctionalised PVBC was not seen to separate americium although, surprisingly, some extraction of europium was observed (6.7±1.8%). PVB-Me4BTPhen 2 extracted 96.5±1.5% of

Am (III) with minimal extraction of Eu (III). Therefore, this method was deemed to be suitable for Am (III) extraction. PVB-Me4BTPhen 2 was recycled four times and was shown to extract americium with ~90% recovery each time.

92 11.43 Americium extraction from nuclear forensic matrix using PVB-Me4BTPhen

The initial method developed above was then applied to a matrix designed to simulate a nuclear forensic sample, containing Be, Al, Sr, Y, Mo, Ag, Cd, Cs, Ba, La, Ce, Pr, Nd, Sm, Gd, Tb, Dy,

Ho, Er, Tm, Yb, Lu (~1 mg mL-1, initial concentrations determined by ICP-MS) and 241Am,

152Eu, 232U, 239Pu, 237Np (10 Bq each). Recovery of Am (III) and Eu (III) was determined by gamma spectrometry as previously (Table 2). After extraction with PVB-Me4BTPhen 2, the remaining aqueous phase was subjected to AG1-X8 chromatography to separate the remaining actinides for quantification by alpha spectrometry (Table 3), and to ICP-MS to determine recovery of matrix elements (Figure 2).

Figure 2. Recovery of matrix elements after PVB-Me4BTPhen extraction (by ICP-MS).

93 Solution 241Am Recovery± Error / % 152Eu Recovery± Error / % Aqueous phase after 90 min polymer 4.1±1.7 95.6±5.4 contact (4 M HNO3) Polymer 90 min back extraction (0.1 M 95.2±1.7 6.0±4.1 HCl)

232U 98.9 3.1 239Pu 94.9 2.0 237Np 98.3 0.6

Similar extraction of Am (III) was observed from this matrix as from the mixture of Am (III)/Eu

(III) (recovery 95.2±1.7%) indicating little effect on complexation by other elements present. All matrix elements showed recovery of >90% apart from Cd (80%) and Pr (83%). These results are similar to those seen with CyMe4BTPhen in liquid-liquid separation; in particular Cd has been seen to co-extract under these conditions [3,17]. The high recovery of these matrix elements is important for nuclear forensics as quantification of these elements is often required for data modelling in an investigation. Recoveries of >94% were observed for U, Pu and Np after AG1-

X8 separation by alpha spectrometry, indicating that PVB-Me4BTPhen 2 has low affinity for U,

Pu and Np in the presence of Am (III) and lanthanides.

94 11.44 Variation of acidity and contact time in americium extraction from nuclear forensic matrix

To determine the optimal aqueous phase acidity and contact time for americium recovery from the nuclear forensic matrix using PVB-Me4BTPhen, the acidity was altered in the range 0.1 to 5

M HCl with a contact time of 1.5 h (Figure 3) and the contact time in the range 15 to 180 min at

4 M HCl (Figure 4).

Figure 3. Variation of Am (III) and Eu (III) extraction by PVB-Me4BTPhen with aqueous phase (HCl) acidity (N.B. experiments run in triplicate, error calculated from instrument and method error plus one sigma.)

Figure 4. Variation of Am (III) and Eu (III) extraction by PVB-Me4BTPhen with contact time (N.B. experiments run in triplicate, error calculated from instrument and method error plus one sigma.)

For acidities in the range 2.0–4.0 M, americium recovery is comparable within the same timescale (~92%). With an aqueous phase in the range 0.1-1.0 M HCl no appreciable (<2.0 %) americium was extracted. At 5.0 M HCl the americium recovery is reduced to 35%, due to protonation of the ligand as previously observed [4].

Less than 1% americium extraction was observed after 15 min, suggesting there may be preliminary steps required prior to complexation of americium onto the polymer. Limited extraction was seen after 30 min (10.5±1.5%); after 45 min extraction reached a maximum

(~90%). Therefore, the minimum time needed to extract americium is 30 min and 45 minutes would allow for maximum recovery, comparable to the liquid-liquid separation process [3].

96 11.5 Conclusions

We have described the preparation and application of a soft N-donor triazine (BTPhen) ligand

EC resin which is capable of americium extraction from a complex matrix containing lanthanides, actinides and matrix elements. The polymer PVB-Me4BTPhen was generated by synthesis of the C-5 brominated derivative of the ligand, which allowed attachment of an aniline linker and subsequent conjugation to a PVBC solid support. We propose that this general methodology could be applied to other BTPhens and also related triazine ligands e.g. BTBPs.

PVB-Me4BTPhen 2 extracted americium (>90%) from Am (III)/Eu (III) solutions with an

Am (III):Ln (III) decontamination factor of >1000 (calculated as ratio of separation factors, from percentage recoveries) over 90 min. Back-extraction of the polymer to obtain aqueous americium was possible using low molarity acid, as observed for the analogous liquid-liquid separation process. The polymer was shown to be recyclable with no observed effect on extraction. We then applied PVB-Me4BTPhen 2 to our simulated complex nuclear forensics sample under the same conditions and obtained high extraction (>90%) of americium. The major actinides (U, Np, Pu) and matrix elements including lanthanides were obtained after americium extraction and AG1-

X8 separation with recoveries >90% (with the exceptions of Cd and Pr, 80 and 83% respectively). Investigation of extraction efficiency as a function of contact time and aqueous phase acidity showed >45 min is required for Am (III) extraction from the matrix and that acidities in the range 2.0–4.0 M are usable. Based on these results, we propose that PVB-

Me4BTPhen 2 offer an alternative to our previously reported process with very similar outcomes.

The polymer-bound nature of the ligand may enable wider use of BTPhen americium extraction in analytical radiochemistry and allow the development of new separation methods when combined with existing EC methodology. Crucially, immobilised soft N-donor ligands offer an

97 advantage over current methods for Am (III)/Ln separation (DGA) in that there is no thiocyanate used hence no toxic wastestream is generated.

11.6 Acknowledgements

Funding for this project was provided by AWE via a studentship to MAH. ICP-MS analyses were carried out by Mr Paul Lythgoe (School of Earth, Atmospheric and Environmental

Sciences, The University of Manchester). Mohamed Nuruddin Mohamed Nasir is thanked for assistance with synthetic chemistry.

11.7 Notes and References

1. See reviews: (a) M. J. Hudson, L. M. Harwood, D. M. Laventine, F. W. Lewis, Inorg.

Chem., 2013, 52, 3414–3428; (b) P. J. Panak and A. Geist, Chem. Rev., 2013, 13, 1199–

1236.

2. (a) S. L. Goldstein, C. A. Hensley, C. E. Armenta, R. J. Peters, Anal. Chem., 1997, 69, 809–

812, (b) K. Mayer, M. Wallenius, I. Ray, Analyst, 2005, 130, 433–441, (c) D. Magnusson,

B. Christiansen, R. Malmbeck, J.-P. Glatz, Radiochim. Acta, 2009, 97, 497−502, (d) D. M.

Whittaker, T. L., Griffiths, M. Helliwell, A. N. Swinburne, L. S. Natrajan, F. W. Lewis, L.

M. Harwood, S. A. Parry, C. A. Sharrad, Inorg. Chem. 2013, 52, 3429–3444.

3. M. A. Higginson, O. J. Marsden, P. Thompson, L. M. Harwood. M. J. Hudson, F.W.

Lewis, F. R. Livens, S. L. Heath (Submitted)

4. F. W. Lewis, L. M. Harwood, M. J. Hudson, M. G. B. Drew, J. F. Desreux, G. Vidick, N.

Bouslimani, G. Modolo, A. Wilden, M. Sypula, T.-H. Vu and J.-P. Simonin, J. Am. Chem.

Soc., 2011, 133, 13093–13102.

5. D. M. Laventine, A. Afsar, M. J. Hudson and L. M. Harwood, Heterocycles, 2012, 86,

1419–1429.

98 6. (a) M. H. Lee, J. H. Park, S. Y. Oh, H. J. Ahn, C. H. Lee, K. Song and M. S. Lee, Talanta,

2011, 86, 99–102, (b) S. L. Maxwell, J. Radioanal. Nucl. Chem., 2008, 275, 395–402.

7. (a) E. P. Horwitz., D. R. McAlister, A. H. Bond and R. E. Barrans, Solvent Extr. Ion Exch.,

2005, 23, 219–225; (b) D. R. McAlister and E. P. Horwitz., Solvent Extr. Ion Exch., 2007,

25, 757–769.

8. C. Klug and R. Sudowe, Sep. Sci. and Technol., 2013, 48, 2567–2575.

9. W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43, 2923–2925.

10. IAEA, Controlling of Degradation Effects in Radiation Processing of Polymers, IAEA-

TECDOC-1617, Vienna, 2009.

11. J. Mlochowski, J. Rocz. Chem., 1974, 48, 2145–2155.

12. S. Deroo, E. Defrancq, C. Moucheron, A. Kirsch-De Mesmaeker and P. Dumy,

Tetrahedron Lett., 2003, 44, 8379–8382.

13. G. A. Slough, V. Krchňák, P. Helquist, and S.M. Canham, Org. Lett., 2004, 6, 2909–2912.

14. A. Afsar, D. L. Laventine, L. M. Harwood, M. J. Hudson and A. Geist, Heterocycles, 2014,

88, 613–620.

15. A. Afsar, L. M. Harwood, M. J. Hudson, P. Distler and J. John, Chem. Commun., 2014, 50,

15082–15085.

16. Calculated with Chemdraw® software using the predicted repeat unit of the polymer.

17. G. M. Cockrell, G. Zhang, D. G. VanDerveer, R. P. Thummel, R. D. Hancock, J. Am.

Chem. Soc., 2008, 130, 1420–1430.

99 12.0 Appendix to Paper 2

Supplementary Information for Higginson et al.

12.1 SEM-EDX (scanning electron microscopy with energy dispersive x-ray analysis) of polymer 2.

Figure 1. Representative image of PVB-Me4BTPhen 2

The materials were homogeneous throughout the analysed area with sizes of the individual pieces of polymer ranging from 2-200 µm.

100

Figure 2. EDX image of polymer 2 with the BSE detector.

From analysis of the material with the backscattered detector no change could be seen throughout the material which implies that the polymer is all one consistent matrix, furthermore similar EDAX spectra were obtained from the material at every point analysed. As H cannot be characterised by EDAX, the C/N ratios can be qualitatively estimated and a composition at a given point estimated for the material, (every spectra was approx. (3.5:1 C/N taking into account for the carbon background).

101

12.2 Solid State NMR (Proton 1H, CP-MAS, 14000 Hz spin speed) spectra

PVBC

Stretches for aromatic (120-160 ppm) and aliphatic (20-60 ppm) for starting material

PVB-Me4BTPhen 2

Stretches for ligand attachment (162 ppm, 20-60 ppm) maintaining PVBC aromatic structure

102 12.3 Polymer Disk Work

Polymer disk method development

High precision alpha spectrometry is usually dependant on good electrodeposition of the target alpha emitter onto a stainless steel planchette/disk for analysis. Interference of matrix elements in the electrodeposition or the forming of thick layers of the metal is problematic. Both of these problems lead to self-attenuation of the alpha particles within the spectrometer which reduces resolution and causes tailing of peaks which may then overlap within the small range of energies of common alpha emitters. This is why care is usually taken with the current and voltage and time taken when electrodepositing the sample and why separations of the Ln and Fe for example are carried out in order to not interfere with the determination of alpha-emitters. All these processes take time, especially for low level samples purification of the sample and carriers may be required. This process and the purification of Am from the Ln takes considerable time in a nuclear forensic investigation.

We plan on attaching our polymer to a stainless steel disk in order to extract americium from the solution directly onto the alpha plate, this would remove the need to take the eluted solution to dryness and complete an electrodeposition. However, this idea poses a few problems, firstly attaching the polymer to the disk needs to use as thin a layer of adhesive as possible which must also be resistant to up to 4 M HCl and be strong enough to hold the polymer. Secondly, the polymer must be in as thin a layer as possible so that self-attenuation is limited so a spectrum can be obtained. As the polymer also requires synthesis it may not be able to be melted and re-set onto disk without changing the chemical structure of the material.

We identified a liquid epoxy based superglue (Masterbond©) as a potential candidate for the glue, it can be used to stick polymers to metal and it is resistant to acidic conditions and

103 requires time to set allowing for manipulation of the polymer. The polymer (1.45 g) was ground into a fine light brown powder with a pestle and mortar and the liquid glue was applied to the disk, the polymer was then added to the disk, flattened and allowed to set in the oven at 60°C.

The disk was removed and flattened with a mill used to press metal disks to give a thin layer of polymer on the disk (~1.2 g of polymer). The disk was then tested to see if the polymer could be removed in 1 M HCl for 10 minutes. The polymer appeared to not be affected. The disk was put into a 3 mL solution of 4 M HCl containing 10 Bq of Am and Eu and left for 60 minutes. The solution was then counted by gamma spectroscopy (Figure 1) and the recoveries of Am/Eu determined. The disk was then dried for 2 days in the air and then analysed by alpha spectroscopy using the procedure outlined in chapter 2. The solution again showed almost quantitative separation of americium and limited europium separation (<5%).

104 Poly-Me4BTPhen Disk Am Separation Count (12 hours) 2500

2000

1500

1000 Counts

500

0 0 2000 4000 6000 8000 10000 12000 -500 Energy / keV

Figure 1. Alpha spectrum from 10 Bq Am/Eu separation by Poly-Aniline-Me4-BTPhen bonded to an alpha planchette.

The alpha spectrum however shows that the polymer self-attenuates the americium leading to a broad range of energies leading up to the typical americium maximum. This shows that the polymer is not suitable for direct alpha analysis, however the disk can be stripped and gamma counted for recovery which could find utility in a nuclear forensic investigation or the area integrated in a pure solution. Future work with the immobilisation onto a disk will try to obtain a process by melting the polymer and setting it onto an alpha disk. This was initially discounted as we reasoned that the functionality of the polymer would be damaged, preliminary work with

PVBC as shown that it is possible to melt and potentially coat a disk completely with the polymer which may reduce the self-attenuation problem. We attempted to try and thermoset the polymer into a circular mould but the polymer would not melt even at 550°C and discolouration began to occur. The only other way to possibly solve this problem would be to completely re-

105 synthesise the ligand monomer and polymerise this or to bind it to a silica or PVBC functionalised disk. Alternatively, it may be possible to spin coat or adsorb the monomer onto a surface but currently we have no feasible way to accomplish this.

12.31 Adsorption versus Covalently Bonded Extraction Chromatography Resins

The previous and following study highlight the potential advantages and disadvantages from producing resin systems via conventional synthesis and adsorption of a ligand into a porous polymer system. Covalently bound EC resins have the advantage of being stable to many repeat procedures, oxidation/reduction reagents, highly acidic/basic media and solvents that can dissolve extractants. However, adsorption based systems have the benefits of the extractant being held by Van der Waals forces and in solution meaning they can interact in a mobile phase more readily, therefore a quicker separation can be achieved and this is why this type of resin is predominant in the commercially available resins. In this work, we found this to be the case as the adsorbed BTBP/BTPhen ligands have faster kinetics (approx. 15 min to reach equilibrium compared to at least 30 min with PVB-Me4BTPhen) but were susceptible to the problems listed in the manuscript and above. The covalently linked resin has been shown to be more robust than the adsorbed systems. Therefore developing both types of system allows flexibility for application to different matrices and overall separation methodologies. Overall these EC resin systems offer higher decontamination factors than the corresponding liquid-liquid separation systems, as the separation is taking place as more than a single separation step on the column and this is justification for the two studies away from the liquid-liquid systems.

106

107 13.0 Chapter 3

The material in the following section has been accepted to the journal of Radioanalytical and

Nuclear Chemistry for the MARC X issue in 2016 and was presented at MARC X in 04/15.

The author synthesised the ligands, prepared and analysed the resins, designed and performed the radiochemical method development, analysed the data, interpreted the results and wrote first draft and the final version of the manuscript.

108

Americium separation by extraction chromatography

resins using adsorbed triazine ligands

Matthew A. Higginson,a Olivia J. Marsden,b Paul Thompson,b Francis R. Livensa and Sarah L.

Heath.a,* a Centre for Radiochemistry Research, School of Chemistry, The University of Manchester,

Manchester, M13 9PL, UK. b AWE, Aldermaston, Reading, RG7 4PR, UK.

* Corresponding author, [email protected], tel. +44 161 275 4696

109

13.1 Abstract

Extraction chromatography (EC) resins find widespread use in analytical radiochemical separations, in particular for actinide separation. However, there is currently limited choice for separation of americium from lanthanides using EC, with DGA (N,N,N’,N’-tetra-n- octyldiglycolamide) and the LN resin being used for Ln separations in overall methods. Here we present three novel EC resins based on the soft-N donor Am selective BTBP/BTPhen triazine ligands. These ligands were adsorbed into Amberlite XAD-7, characterised and tested for Am/Eu selectivity, complexation kinetics and polymer loading. The two polymers which showed promise required between 15-30 min to achieve complexation of Am from solution. These polymers were then tested with complex matrices in conjunction with AG1-X8 chromatography to achieve a complete isotope separation and quantification method. From these results the resin capacity factor (K) as a function of HNO3 concentration was calculated allowing for potential separation methods to be designed. In particular these resins show excellent potential for the specific extraction of Am from complex matrices. We hope further Am/Ln selective EC resins can be developed and tested based on the easy adsorption method and screening method reported.

13.2 Key Words:

Americium separation, EC Resin, radiochemical separation, nuclear forensics, triazine ligand, BTPhen, BTBP

110

13.3 Introduction

Nuclear forensic methodology aims to characterise unknown radioactive materials to provide evidence to support law enforcement and legal cases. This could be from seizures of illicit nuclear material or from unknown origin radioactive samples. The isotopic composition of a sample is required to allow comparison with reference data and for use in modelling. The collection of data must be performed in a timely manner and analytical techniques must be strategically applied to minimise sample loss/damage. [1-2]

This requires separation and multiple analytical methods (e.g. radiometric methods, mass spectrometry, and microscopy) to be applied to the sample. Samples in nuclear forensics can be inherently complex, containing lanthanides, actinides and matrix elements, including fission/activation products with isotopic variation and high concentrations of matrix elements.

Radiochemical separation is therefore required in order to avoid isobaric and polyatomic interferences (e.g. 237Np/197Au+40Ar) in mass spectrometry and/or spectral overlap in radiometric analyses [3-4].

The ability to obtain pure fractions of the lanthanides and Am/Cm is currently a bottleneck in radioanalytical separations and nuclear forensic investigations. Methods have been reported which achieve this in 1–2 days [5-6]; these processes require extraction chromatography methods and/or rely on the use of thiocyanate for Am elution, resulting in toxic wastes. More commonly, separation and quantification of Am and lanthanides from these complex samples requires a process which may take more than 2 days. These separations typically rely on either

DGA/LN resin or an α-HIBA separation in conjunction with other EC resins, all of which use selective extractants to achieve separation. EC and cation/anion exchange resins also offer

111 benefits over liquid-liquid extraction techniques in that they can be faster, do not need mixing or phase separation, gradients can be used and redox changes can be induced on the resin.

Therefore the development of selective extraction methods and new extraction chromatography resins is of interest in wider analytical radiochemistry analysis.

We aimed to develop extraction chromatography resins from the triazine ligand liquid- liquid separation methods developed previously for complex matrices where Am/Cm results are required in a timely manner. [7] This would build on the synthetic triazine polymer methodology we have also developed but hopefully make the preparation of these resins feasible on a commercial and laboratory scale. [8] These resins would also give flexibility in the application of triazine soft-N donor ligands in analytical radiochemistry and allow them to be used in tandem with other EC resin separation methods in the literature for Th/U/Pu/Np purification allowing challenging separations to be accomplished. Specifically in our work, this would entail separation of the major actinides (U, Pu, Np) by chromatography using reported methods (AG1-

X8, TEVA-TRU, UTEVA) followed by extraction of Am/Cm from the remaining matrix containing lanthanides and matrix elements using an Am/Cm selective polymer with further purification of all elements of interest. Herein we report the synthesis and adsorption of three triazine soft-N donor ligands (CyMe4BTPhen, CyMe4BTBP and CA-BTPhen, Figure 1) to generate novel EC resins. These resins were screened under various conditions and Am/Eu separations were completed. Two polymers were then tested for the separation of Am from complex matrices simulating a nuclear forensic sample in an overall separation scheme to obtain a full data set. These polymers were found to offer an alternate method within the same timescale of the current literature methods for Am/Ln separation and do not require the use of toxic

112 chemicals. This allowed for the purification and quantification of all elements of interest from a complex matrix within a 2 day timescale.

Furthermore, the methodology of adsorbed selective extractants allows an easy route to generating useful EC resins for wider separation science.

Figure 1. Structures of the ligand 1, 2 and 3 used to synthesise the EC resins used in this study

13.4 Experimental

All radionuclides used were provided from calibrated stocks in the School of Chemistry,

University of Manchester. Micropipettes of 100 µL, 0.1–1 mL and 2–10 µL were calibrated on a

4 d.p. balance with >18 MΩ Millipore deionised water in the temperature range 18–22 °C and were found to be within their stated range. All acid solutions were made from Aristar grade concentrated solutions and were diluted with >18 MΩ Millipore deionised water. All solutions were considered to have expired within one month of creation. Gamma counting was performed using a Canberra 2020 coaxial HPGe gamma spectrometer with an Ortec 919E multi-channel analyser. Gamma spectroscopy was performed against a standard of known activity counted in the same geometry and analysed using the diagnostic photon energies of 241Am (59.5 keV) and

152Eu (121.8 keV). Samples were electrodeposited and alpha spectrometry was performed on a

Canberra model 7401VR detector with multi-channel analyser. ICP-MS analysis was performed on an Agilent 7500cx spectrometer. Multiple standards for each element in the range 1-100 ppb

113 were used for ICP-MS quantification. Recovery values were calculated from multiple initial standard concentrations run in triplicate. All reagents were of >99.99% purity unless otherwise stated. IR spectra were recorded for solid samples using a Bruker Alpha-P ATR spectrometer.

13.41 Ligand Synthesis

CyMe4BTPhen 1 and CyMe4BTBP 2 were synthesised as reported previously [9, 10], and purified using the method reported by Whittaker et al [11]. CA-BTPhen 3 was synthesised using a modification of the previously reported procedure [12], in which the solvent was changed to

EtOH and the reaction time changed to a 3 h reflux followed by 6 h at RT. Data collected for all intermediates and ligands was as previously reported.

13.42 Polymer adsorption method

A known quantity of CyMe4BTPhen/CA-BTPhen/CyMe4BTBP ligand was dissolved in either

MeOH or acetonitrile (1-15 mL) and Amberlite XAD-7 (Aldrich, high purity, pre-treated as set out in the manufacturer’s instructions) was added (5-50% w/w). The solution was stirred for 2 hours at room temperature. The mixture was then stirred at 30°C under reduced pressure to remove the solvent and dry the resin, the resin was then dried under high vacuum (1 hour). The

EC resins were contacted with 4 M HNO3 for 10 min then separated and dried under high vacuum before use. Analysis was completed on each synthesis/batch using IR spectroscopy and

CHNS elemental analysis for nitrogen content and C=N stretches against the starting material polymer.

13.43 Initial Am/Eu separation method for triazine EC Resins

A known amount of polymer (20-100 mg) was suspended in a solution of 241Am/152Eu (10-100

Bq each, 1 mL) in 4 M HNO3 for 1.5 h. The polymer was then back-stripped 1.5 h with 0.1 M

HCl. To ensure complete back-stripping of the polymer for re-use, the polymer was suspended in

114 0.1 M HCl for 24 h. The resulting three acidic phases were counted by gamma spectroscopy against a standard of known activity in the same geometry.

13.44 General Am/Eu separation method for triazine EC Resins

A known amount of polymer (20-300 mg, 5-50% w/w) was suspended in a solution of

241 152 Am/ Eu (10-100 Bq each, 1 mL) in 0.1-5 M HCl/HNO3 with 15 min intervals to 1.5 h. The polymer was then separated/filtered and back-stripped for 1 h with 0.1 M HCl. The resulting phases were counted by gamma spectroscopy against a standard of known activity in the same geometry.

13.45 Generation of Nuclear Forensic/Complex Nuclear Matrix

Ag, Cd, Sr and Ba. Spike solutions of 241Am (III) (10-100 Bq range), 152Eu (III) (10-100 Bq range), 232U (VI), 239Pu (IV) and 237Np (IV) (10 Bq each isotope) were added to generate a known molarity solution.

Anion exchange/extraction chromatography: The matrix was taken to dryness two times with addition of conc. HCl (2 mL). The residue was dissolved in the minimum volume of conc. HCl.

A 50 mm x 4 mm anion exchange column (Bio-Rad AG1-X8, 200-400 mesh) was conditioned with the following procedure: sodium bromate (5 mg) in conc. HCl (5 mL) was added. The column was washed with conc. HCl (4 mL) then aq. HCl (5 mL, 1 M) and conc. HCl (3 x 5 mL).

The sample was loaded using conc. HCl to wash (2 mL). Am, lanthanides, Cs, Be were eluted

115 with conc. HCl (6 mL). Pu was eluted with conc. HCl containing 1% v/v HI (5 mL). Co, Zr, Np were eluted with HCl (5 mL, 4 M). Mo, Fe, U were eluted with HCl (3 mL, 1 M) followed by

HCl (3 mL, 0.5 M). Cd and Ag were eluted using deionised water (1 mL) followed by NH4OH (4 mL, 1 M). N.B. this step may also be achieved using the UTEVA® based method described by

Morgenstern [13] or the TEVA-TRU method described by Maxwell and co-workers [4].

Electrodeposition procedure: The solution to be electrodeposited was evaporated to dryness.

Aqueous NaHSO4 (5%, 2.5 mL), water (2 mL) and aqueous Na2SO4 (15%, 5 mL) were added to the residue and heated, the solution was transferred to an electrodeposition cell using water (3 mL) and (NH4)2C2O4 (160 mM, 1 mL) was added. The current was set to 0.5 A for 5 min, then

0.75 A for 60 min. 1 min before the end of electrodeposition, aqueous KOH (25%, 2 mL) was added. The solution was decanted off and the cell washed with aqueous NH4OH (5%, 2 mL).

The plate was rinsed with 5% NH4OH, ethanol and acetone before drying in the air.

13.46 General column Am/Eu separation method for triazine EC Resins

A known amount of polymer (0.30-1.00 g) was packed in a 2 mL column and fritted (Triskem) and conditioned with 5 M HNO3/HCl (2 x 5 mL). The acid involved in the separation was then added to further condition the column (0.1-5 M HNO3, 3 x 5 mL)

A typical complex matrix as outlined above (1 mg mL-1, all Ln (III, except Pm), Sr, Y, Ag, Cd,

Ba, Mo and tracers of 241Am (III)/152Eu (III) (10-100 Bq concentration range) and U/Pu/Np (10

Bq) was generated at 0.1, 1, 2, 3, 4 and 5 M HNO3. An AG1-X8 column separation was carried out as outlined below (or UTEVA/TEVA-TRU method for U/Pu/Np purification). The matrix was continuously loaded onto the column for the required complexation/contact time (15-90 min) with dilution (either continuously or by being inserted into a stoppered column). The

116 column was then eluted with 5 mL (2 x 2.5 mL) of 0.1 M HNO3 over 30 min. The phases were counted by gamma spectroscopy (Am/Eu). The eluted/purified phases were then diluted (1000-

10000 fold dilutions) for analysis of trace elements by ICP-MS (all Millipore grade DI and

Aristar HNO3).

13.5 Results and Discussion

Multiple soft-N donor ligands based on 1,2,4-triazines have shown selectivity for complexation with later actinides (Am, Cm) over other actinides and lanthanides [14]. This is attributed to the enhanced covalency in bonds formed by the 5f orbitals of the actinides, which confers a preference for softer ligands [15]. The triazine ligands used in this study CyMe4-BTPhen 1,

CyMe4-BTBP 2 and CA-BTPhen 3 were previously shown to separate Am/Eu with separation factors in the range of 100-400 such that a rapid selective separation can be achieved. (Figure 2)

[15] In previous work we identified these ligands for the separation of Am from aqueous matrices containing actinides, lanthanides and matrix elements. These ligands can be recycled for repeat use, adhere to the CHON principle such that they can be safely incinerated after use, avoiding the generation of toxic waste. In particular CA-BTPhen has the advantage that it is easy to synthesise (in four steps from commercially available materials).

EC resins such as TEVA, TRU, UTEVA and DGA are typically ligands adsorbed onto various forms of the epoxy resins Amberlite XAD2 to XAD7. Klug et al have a published method for the adsorption of BTP-type ligands onto Amberlite XAD-7. [16] Amberlite adsorbs and releases ions through hydrophobic and polar interaction and can readily incorporate organic molecules into its structure. Amberlite XAD-7 used in the work is 20-60 µm and is the current inert polymeric support of the UTEVA ® resin.

117 13.51 Resin Preparation

We designed our own adsorption method by dissolving known quantities of each ligand in methanol or acetonitrile (enough solvent to barely cover the resin (1-15 mL)) with a known mass of the treated uncoated resin (Amberlite XAD-7, treated as set out in the manufacturer instructions). This allows for the generation of known w/w loading resins and allows easy scale up. The mixture was slurried for at least 2 hours at room temperature. The solvent was removed under reduced pressure and dried under high vacuum for 1 hour and the resulting resin was analysed by IR spectroscopy. Phenanthrolines stretches were seen for all four resins (C=N at

~1600-1650 cm-1), elemental analysis (CHNS) was also used for analysis, (increase in nitrogen content up to 20% for highly loaded resins). The resins were then suspended in 4 M HNO3 (up to

1 month) to confirm the resins’ stability in acidic solutions then dried under high vacuum before use.

To assess the americium extraction capabilities, we adapted our liquid-liquid separation procedure by removal of the 1-octanol phase, so that EC resins 1, 2 and 3 would be suspended in an acidic matrix containing Am for a period of time, followed by filtration and stripping of the resin with low-molarity acid followed by filtration of the second phase. Firstly, the stability of these resins in 4 M HCl and 4 M HNO3 was assessed over 1 month. A comparable elemental analysis was obtained and no physical change was observed, indicating stability over a long time period and potential for re-use these resins when packed into a column. Molarities above 7 M

HCl led to discolouration of the polymer overnight which means that the molarity of phase loaded onto the column is important, also it should be noted that the ligands in solution show reduced extraction kinetics above 4 M HCl/HNO3. [15] Therefore any solution loaded to the resins is required to be adjusted to <5 M HCl/HNO3.

118 In an initial Am/Eu extraction investigation, resins 1, 2 and 3 (10% w/w, 20 mg) was suspended

241 152 in a solution of Am/ Eu (10-100 Bq each, 1 mL) in 4 M HNO3 for 1.5 h. The polymer was then back-stripped for 1.5 h with 0.1 M HCl. The resulting acidic phases were counted by gamma spectroscopy against a standard of known activity in the same geometry (Table 1). The experiment was also repeated with Amberlite XAD-7 as a control.

241Am recovery (back- 152Eu (Aq phase) recovery and error Polymer (1.5 h method) stripped) and error / % / % AMB-BTPhen 99.1 ± 1.7 94.2 ± 3.1 AMB-BTBP 93.2 ± 1.9 95.2 ± 2.8 AMB-CA-BTPhen 56.1 ± 2.1 97.2 ± 3.4 Table 1: Preliminary amberlite polymer Am/Eu separation data, determined by gamma spectroscopy against a known activity standard in the same geometry

These results show complexation time is needed to ensure that Am/Cm is separated from the mixture. AMB-CA-BTPhen showed slow americium extraction after 1.5 h. We however achieved the aim of generating EC resins from these ligands in an easier manner which allows their application to complex separations and the ability to generate these in an analytical laboratory. As long as the mixture does not come into contact with either peroxide or organic solvents then it is unaffected in aqueous media up to 7 M HCl. From this AMB-BTPhen/BTBP were shown to need >15 mins contact time to complex high percentages of americium and were taken forward for further study. We wanted to obtain weight distribution ratios/resin capacity values (Dw) as a function of 0.1-5 M HNO3 for the matrix elements of interest to complex separations (importantly Am and the lanthanides). This set of data which is usually presented with radiochemical resins such as UTEVA to develop separation methods as shown by Horwitz et al for the LN1-3 resins. We used the same methodology applied by to the LN resins to determine the resin capacity values/distribution ratios below. [17]

119 13.52 Resin loading investigation for AMB-BTPhen/BTBP

The resins with the fastest extraction kinetics (AMB-BTPhen/BTBP) from the preliminary work were generated with w/w loadings of 5, 10, 15, 20 and 50% of with the method outlined above.

These polymers (20 mg) were contacted with Am/Eu (10-100 Bq) as outlined above and back- stripped and gamma counted after 15 minutes of contact time. The phases were counted by gamma spectroscopy against a known activity standard in the same geometry (Table 2).

AMB-BTPhen loading Am recovery ± error / % Eu recovery ± error / % 5% 89.5 ± 2.1 6.6 ± 4.1 10% 89.7 ± 3.1 6.5 ± 3.6 15% 89.2 ± 2.9 6.5 ± 4.2 20% 89.0 ± 2.7 6.6 ± 3.8 50% 89.3 ± 3.0 6.5 ± 3.7

AMB-BTBP loading Am recovery ± error / % Eu recovery ± error / % 5% 76.2 ± 3.1 5.8 ± 4.2 10% 76.1 ± 2.7 5.2 ± 3.9 15% 75.9 ± 2.8 5.6 ± 3.7 20% 75.2 ± 2.8 5.5 ± 4.1 50% 76.1 ± 2.9 5.3 ± 4.2 Table 2: Am/Eu recoveries from w/w loading AMB-BTPhen/BTBP 15 min polymer separations (after back stripping) determined by gamma spectroscopy

This experiment showed that the complexation to the polymer is the limiting factor in Am extraction from solution and that 5% loading (ligand excess) is sufficient to achieve the same separation as 50% w/w with minimal Eu recovery at these activities (10-100 Bq). This allowed less ligand to be used in future experiments and shows that high ligand quantities can be accommodated onto the resin. Furthermore 0.5 h was shown to give recoveries >90% for AMB-

BTPhen/BTBP and Am/Eu separation factors comparable with those observed with the ligands in solution at 3 M HNO3.

120 13.53 Matrix elements separations using AMB-BTPhen/BTBP

EC resins 1 and 2 were tested against a complex sample matrix (outlined below) in a cartridge column that can be used in conjunction with other methods (e.g. UTEVA/TEVA-TRU/AG1-X8).

A 2-5 mL column (Triskem) was packed with 0.3 g of the three polymer samples (AMB-

BTPhen/BTBP 5% w/w) and conditioned with 5 M HNO3 (5 mL). The acid involved in the separation was then added to further condition the column (0.1-5 M HNO3, 5 mL).

A typical complex matrix as outlined previously after AG1-X8 separation (1 mg mL-1, all

Ln (III, except Pm), Sr, Y, Ag, Cd, Ba and 241Am (III)/152Eu (III) (10-100 Bq concentration range) was generated by separation of the complex matrix (containing U/Pu/Np/Mo/Zr) with

AG1-X8 and adjusted to 0.1, 1, 2, 3, 4 and 5 M HNO3. This matrix (5-10 mL) was continuously loaded onto the column over 30 min with dilution (either continuously or by being inserted into a stoppered column). This initial phase was counted by gamma spectroscopy (for Am/Eu), then diluted for analysis of trace elements by ICP-MS (all Millipore grade DI and Aristar HNO3). The column was then eluted with 5 mL (2 x 2.5 mL) of 0.01 M HNO3 over 30 min. The values obtained from gamma spectroscopy and ICP-MS were converted into KD values and plotted for the 3 polymers below. For clarity, the charts without Am are shown to display the Ln data in more detail where necessary (Figures 1-5 below).

121 13.54 KD Data for Poly-AMB-BTPhen

Acid dependencies of the uptake of Am (III), Ln (III) and FP/Matrix elements on AMB-BTPhen Sr Y Ag 1200 Cd Ba Mo 1000 La Ce 800 Pr

D Sm

K 600 Gd Tb 400 Dy Ho 200 Er Tm 0 Yb Lu 0.0 1.0 2.0 3.0 4.0 5.0 Am Eu M, HNO3

122 Acid dependencies of the uptake of Ln (III) and FP/Matrix elements on AMB-BTPhen 120 Sr Y Ag Cd Ba 80 Mo La

D Pr K Nd Sm 40 Gd Tb Dy Ho Er Tm 0 Yb Lu 0.0 1.0 2.0 3.0 4.0 5.0 Eu M, HNO3

Figure 2: Acid dependencies (KD) of the uptake of Ln (III) and FP/Matrix elements on AMB- BTPhen at 0.1-5 M HNO3, values determined by ICP-MS and gamma spectroscopy (no errors >2.0% from raw data).

The data for AMB-BTPhen shows the highest selectivity for Am (III) at 3 M HNO3 which is similar to that seen in solution, at <2 M HNO3 there is very little complexation onto the matrix apart from general adsorption of Ln ions, which allows for our designed method to strip the labile complex at <1 M HNO3 for this polymer as predicted and shown previously.

The selectivity decreases at increasing acidity to 5 M HNO3 to decomposition of the ligands at >7 M HNO3. The complexation/method time (>15 min) is the only drawback with this method but either stoppering the column with the solution or continuous addition are needed to achieve separation. The resin will separate small quantities of Cd (II) and Ln (III) but as these do not inhibit analysis and can be potentially purified as discussed in previous work with a sulfate

123 precipitation for Cd or pre-column separation. The overall method is as fast as the liquid-liquid method and on par with a DGA based separation and can achieve DF>103 for Am/Ln separation

(from multiple contact times, calculated as a ratio of separation factors calculated from ICP-MS and gamma spectroscopy). These resins are also easier as they can be disposed of as solid waste

(if necessary) and can be made in advance of a method and stored and avoid the use of organic solvents and phase separations. Also the resin is potentially more applicable at higher concentrations of Am and adsorption can be scaled easily.

13.55 KD Data for Poly-AMB-BTBP

Acid dependencies of the uptake of Am (III), Ln (III) and FP/Matrix elements on AMB-BTBP

900 Sr Y 800 Ag Cd 700 Ba Mo 600 La Ce 500

Pr D

K Nd 400 Sm Gd 300 Tb Dy 200 Ho Er 100 Tm 0 Yb Lu 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Am HNO3, M Eu

124

Acid dependencies of the uptake of Am (III), Ln (III) and FP/Matrix elements on AMB-BTBP 90 Sr Y 80 Ag Cd 70 Ba Mo 60 La Ce

D Pr K 40 Nd Sm 30 Gd Tb 20 Dy Ho 10 Er Tm 0 Yb 0 1 2 3 4 5 Lu HNO3, M Eu

Figure 4: Acid dependencies (KD) of the uptake of Ln (III) and FP/Matrix elements on AMB- BTBP at 0.1-5 M HNO3, values determined by ICP-MS and gamma spectroscopy (no errors >2.0% from raw data).

AMB-BTBP shows similar separation ability to AMB-BTPhen, Am (III) is selectively complexed at 3-4 M HNO3 with the separation factors dropping with increasing acidity but this ligand/resin has slower kinetics as seen with above and from the lower KD values calculated.

AMB-BTBP does not show the same affinity for Cd as AMB-BTPhen (as seen in solution) but shows higher complexation of Eu and Ba across all acidities. [18] Preliminary tests in HCl showed a similar trend with lower KD values but the behaviour is similar to that seen in solution and in HNO3. The presence of acid is thought to be needed to potentially form an adduct with the

125 1:1/1:2 ligand to metal complex, but more work on speciation on the solid support would need to be completed before any wider use is undertaken. Overall the adsorbed ligands may offer a good alternative to the liquid-liquid separation procedure as CyMe4BTBP is likely to be commercially available in the near future and can already be purchased from a few suppliers. Assuming pre- concentration and purification of Fe/Ca/Al and the actinides of interest in a complex sample e.g.

U/Pu/Np/Th using EC/anion exchange resins this resin offers a rapid Am/Ln separation process to augment the widely used literature methods. Further purification of the lanthanides can then be achieved using the LN resin or by ICP-MS of the Ln containing fraction allowing for purification and quantification of all elements of interest from a complex matrix to be achieved in <2 d.

13.6 Conclusions

We have synthesised three novel EC resins based on the soft-N donor BTBP/BTPhen triazine ligands by adsorption the ligands into Amberlite XAD-7. The resins were then characterised using IR spectroscopy and elemental analysis and then tested in Am/Eu separations. The complexation kinetics and polymer loading were tested and it was found that 5% w/w loading gave reliable results and that two of these polymers required between 15-30 min to achieve complexation of Am from solution. These polymers were then tested with complex matrices in conjunction with AG1-X8 chromatography to achieve complete separation and quantification of all matrix elements with a combination of ICP-MS, gamma spectroscopy and alpha spectrometry. From these results the capacity factor (KD) as a function of HNO3 concentration were calculated allowing for potential separation methods for a wide range of metal ions to be used to design separation methods. In particular these resins show excellent potential for the specific extraction of Am from complex matrices. We hope that this work can be utilised in

126 nuclear forensic investigations and in the wider field of analytical radiochemistry to provide a new rapid method for Am separation where specific EC resins can be made in advance of the required method and used in conjunction with other ion specific resins.

13.7 Acknowledgements

Funding for this project was provided by AWE via a studentship to MAH. ICP-MS analyses were carried out by Mr Paul Lythgoe (School of Earth, Atmospheric and Environmental

Sciences, The University of Manchester).

13.8 Notes and References

1. Wallenius, M., Mayer, K.: Age determination of plutonium material in nuclear forensics

by thermal ionisation mass spectrometry. Fresenius J. Anal. Chem. 366, 234 (2000).

2. Mayer, K., Wallenius, M., Ray, I.: Nuclear forensics—a methodology providing clues on

the origin of illicitly trafficked nuclear materials. Analyst 130, 433 (2005).

3. Goldstein, S. L., Hensley, C. A., Armenta, C. E., Peters, R. J.: Environmental and Human

Monitoring of Americium-241 Utilizing Extraction Chromatography and α-Spectrometry.

Anal. Chem. 69, 809 (1997).

4. Maxwell, S. L.: Rapid method for determination of plutonium, americium and curium in

large soil samples. J. Radioanal. Nucl. Chem. 275, 395 (2008).

5. Maxwell, S. L., Culligan, B. K., Kelsey-Wall, A., Shaw, P. J.: Rapid radiochemical

method for determination of actinides in emergency concrete and brick samples. Anal.

Chim. Acta. 701, 112 (2011).

6. Luisier, F., Alvarado, J. A. C., Steinmann, P., Krachler, M., Froidevaux, P. J.: A new

method for the determination of plutonium and americium using high pressure

127 microwave digestion and alpha-spectrometry or ICP-SMS. Radioanal. Nucl. Chem. 281,

425 (2009).

7. Higginson. M. A. Marsden. O. J. Thompson. P. Harwood. L. M. Hudson. M. J. Lewis. F.

W. Livens. F. R. Heath. S. L. accepted. (Radiochim Acta)

8. Higginson. M. A. Marsden. O. J. Thompson. P, Livens. F. R. Heath. S. L. submitted.

(R+F Polymers)

9. Lewis, F. W., Hudson, M. J., Harwood, L. M.: Development of Highly Selective Ligands

for Separations of Actinides from Lanthanides in the Nuclear Fuel Cycle. Synlett 2609

(2011).

10. Hudson, M. J., Harwood, L. M., Laventine, D. M., Lewis, F. W.: Use of Soft

Heterocyclic N-Donor Ligands To Separate Actinides and Lanthanides. Inorg. Chem. 52,

3414 (2013).

11. Whittaker, D. M., Griffiths, T. L., Helliwell, M., Swinburne, A. N., Natrajan, L. S.,

Lewis, F. W., Harwood, L. M., Parry, S. A., Sharrad, C. A.: Lanthanide Speciation in

Potential SANEX and GANEX Actinide/Lanthanide Separations Using Tetra-N-Donor

Extractants. Inorg. Chem. 52, 3429 (2013).

12. Trumm, S., Geist, A., Panak, P. J., Fanghänel, T.: Highly Efficient Separation of

Actinides from Lanthanides by a Phenanthroline-Derived Bis-triazine Ligand. Solvent

Extr. Ion Exch. 29, 213 (2011).

13. Morgenstern, A., Apostolidis, C., Carlos-Marquez, R., Mayer, K., Molinet, R., Single-

Column Extraction Chromatographic Separation of U, Pu, Np and Am. Radiochim Acta.

90, 81 (2002).

128 14. Hudson, M. J., Boucher, C. E., Brackers, D., Desreux, J. F., Drew, M. G. B., Foreman

M. R. S. J., Harwood, L. M., Hill C., Madic, Marken,. T. G. A. Youngs., New J. Chem.

30, 1171 (2006).

15. Lewis, F. W., Harwood, L. M., Hudson, M. J., Drew, M. G. B., Desreux, J. F., Vidick,

G., Bouslimani, N., Modolo, G., Wilden, A., Sypula, M., Vu, T.-H., Simonin, J.-P.:

Highly Efficient Separation of Actinides from Lanthanides by a Phenanthroline-Derived

Bis-triazine Ligand. J. Am. Chem. Soc. 133, 13093 (2011).

16. Adam, C., Kaden, P., Beele, B. B., Müllich, U., Trumm, S., Geist, A., Panak, P. J.,

Denecke, M. A.: Evidence for covalence in a N-donor complex of americium(III). Dalton

Trans. 42, 14068 (2013).

17. Klug, C., Sudowe, R.: A Novel Extraction Chromatography Resin for Trivalent Actinides

Using 2,6-bis(5,6-diisobutyl-1,2,4-triazine-3-yl)pyridine. Separ. Sci. Technol. 48, 2567

(2013).

18. McAlister, D, R., Horwitz, E, P,: Characterisation of Extraction of Chromatographic

Materials Containing Bis(2-ethyl-1-hexyl)Phosphoric Acid, 2-Ethyl-1-Hexyl (2-Ethyl-1-

Hexyl)Phosphonic Acid, and Bis(2,4,4-Trimethyl-1-Pentyl)Phosphinic Acid. Solvent

Extr. Ion Exch. 25, 757 (2007).

19. Cockrell, G. M., Zhang, G., VanDerveer, D. G., Thummel, R. P., Hancock, R. D.:

Enhanced Metal Ion Selectivity of 2,9-Di-(pyrid-2-yl)-1,10-phenanthroline and Its Use as

a Fluorescent Sensor for Cadmium(II). J. Am. Chem. Soc. 130, 1420 (2008).

129 14.0 Chapter 4

The material in the following section has been accepted to the journal Dalton Transactions for publication.

The author synthesised and characterised the ligands, designed and performed the radiochemical method development, analysed the data, interpreted the results and wrote first draft and the final version of the manuscript.

130

Synthesis of functionalized BTPhen derivatives -

effects on solubility and americium extraction

Matthew A. Higginson,a Nichola D. Kyle,a Olivia J. Marsden,b Paul Thompson,b Francis R.

Livensa and Sarah L. Heath.a,*

a Centre for Radiochemistry Research, School of Chemistry, The University of Manchester,

Manchester, M13 9PL, UK. b AWE, Aldermaston, Reading, RG7 4PR, UK.

* Corresponding author, [email protected], tel. +44 161 275 4696

131 14.1 Abstract

Separation of the minor actinides (Am/Cm) from spent nuclear fuel post-PUREX process is expected to play a key part in new reprocessing methodologies. To date, a number of selective americium extractants from the BTPhen ligand family have been identified. In this investigation, we synthesise 24 novel BTPhens with additional functionality to determine the effects on solubilities and americium extraction capabilities. The data obtained will allow for tuning of steric/electronic properties of BTPhens in order to assist future extractant design.

14.2 Introduction

In nuclear fuel reprocessing, separation of the minor actinides (Am/Cm) from the lanthanides potentially offers alternative waste management options. The removal of these elements, which account for ~1% by mass but ~90% of the long lived radiotoxicity of high level waste from fuel reprocessing, could reduce both the duration of the radiological hazard and the volumes of high level waste.1-3 One proposed approach partitioning of the minor actinides (and/or Np) after the PUREX process and transmutation by irradiation into short lived and/or stable isotopes.4-5 To achieve this, selective extraction methods where single/groups of actinides are separated from neutron absorbing poisons (lanthanides) are required. In addition, such methods could be utilised to obtain pure minor actinide fractions in other applications, such as environmental analysis.6-7 Such separation procedures require highly selective extractants that fulfil a wide range of criteria (e.g. resistance to radiolysis/acidity, comprising of CHON only (to reduce the generation of secondary waste streams), fast forward/reverse kinetics).

Targeted ligand design has allowed access to a handful of ligand families that meet

132 some/all of these criteria for Am/Ln separations. Currently, the main CHON-based extractant families are bis-triazinylpyridines (BTPs), bis-triazinylbipyridines (BTBPs) and bis-triazinylphen-anthrolines (BTPhens). Geist et al. provide an extensive review of the design/development of these ligand families.8 In the main, the initial ligand designs aimed at resolving practical separation issues (e.g. poor acidic stability, slow extraction kinetics, low solubility).9-10 Functionalisation of a large set of ligands with specific non

CHON moieties has not yet been fully explored, so there is potential for further functionalisation to alter the electronics of the system and affect physical properties, and hence improve Am (III)/Ln (III) extraction. As sets of groups away from CHON have not been extensively investigated previously for BTPhens, it is of interest to future extractant design to see the effect that electron withdrawing/donating groups have on solubility and extraction. We therefore aimed to synthesise sets of BTPhen ligands with additional functionality (with BTBPs, BTPhens are currently the most promising for SANEX

(Selective ActiNide EXtraction) processes)11 to explore the effects on solubility, acidic stability and Am/Ln extraction capability. This would be analysed using separation factors (SF) which are defined as the ratio of distribution co-efficient between the organic and aqueous phases for Am/Eu as representative An(III)/Ln(III). BTPhens are soft-N donor ligands which exploit the enhanced covalency of the minor actinide f-orbitals via metal-nitrogen interactions12, we were therefore interested in the effect of changing the electronics of the system on Am extraction. From a practical standpoint, we were interested in simplifying the synthesis of these functionalised BTPhens by employing commercially available starting materials and a single precursor that could be made on a large scale. This allows for the generation of numerous easily accessible extractants with

133 additional functionality without extensive synthetic chemistry. The phenanthroline moiety of the ligand is easily accessible in four synthetic steps, but functionalisation of this backbone requires harsh reaction conditions and/or protecting group chemistry as it is electron rich which can make obtaining a library of compounds for screening difficult.9

However, some previously synthesised BTPhens such as the benchmark CyMe4BTPhen require a 1,2-diketone which is not commercially available. Our strategy was therefore, a divergent functionalised BTPhen synthesis in which addition of a variety of differently functionalised commercially available diketones to a non-functionalised phenanthroline backbone would afford functionalised ligands. We designed three families of BTPhens using three families of commercially available diketones (isatin, benzil and aliphatic groups), which were firstly tested for their solubility in common nuclear reprocessing solvents and acidic stability. The ligands were then tested in Am/Eu separations to determine SFAm/Eu, with any very promising ligands subsequently tested on more complex matrices as described in our previous work reported in the supplementary information with kinetic data.13

14.3 Experimental

All reagents were purchased from Sigma Aldrich and were of analytical grade. All of the diketones used in the ligand cross coupling reactions are commercially available. Anhydrous acetonitrile was used as purchased (Sigma Aldrich) and handled under Schlenk line conditions

(under N2). All melting points were determined using a Stuart Scientific SMP10 apparatus and are uncorrected. NMR spectra were recorded for solutions in CDCl3, d4-MeOH or d6-DMSO on a

Bruker Avance III instrument (400 MHz) and were referenced to the residual solvent signal.

Compound assignments were determined using COSY and HMQC experiments. Coupling

134 constants (J values) are quoted to the nearest 0.1 Hz. IR spectra were recorded for solid samples using a Bruker Alpha-P ATR spectrometer. Preparative column chromatography was performed using Sigma-Aldrich silica gel (technical grade, 60 Å, 220–240 mesh, 35–75 μm) and the flash technique.14 Compositions of solvent mixtures are quoted as ratios of volume. Organic solutions were dried with anhydrous magnesium sulfate. Low resolution mass spectra were measured on a

Micromass Platform II instrument with electrospray ionisation (University of Manchester).

Nano-electrospray accurate mass measurements were performed by EPSRC National Mass

Spectrometry Facility (NMSF) in Swansea. All radionuclides used were provided from calibrated stocks in the School of Chemistry, University of Manchester. Micropipettes of 100 µL, 0.1–1 mL and 2–10 µL were calibrated on a 4 d.p. balance with >18 MΩ Millipore deionised water in the temperature range 18–22 °C and were found to be within their stated range. All acid solutions were made from Aristar® grade concentrated solutions and were diluted with >18 MΩ Millipore deionised water. All solutions were considered to have expired within one month of preparation.

Gamma counting was performed using a Canberra 2020 coaxial HPGe gamma spectrometer with an Ortec 919E multi-channel analyser. Gamma spectroscopy was performed against a standard of known activity counted in the same geometry and analysed using the diagnostic photon energies of 241Am (59.5 keV) and 152Eu (121.8 keV). Samples were electrodeposited and alpha spectrometry was performed on a Canberra model 7401VR detector with multi-channel analyser.

ICP-MS analysis was performed on an Agilent 7500cx spectrometer. Multiple standards for each element in the range 1-100 ppb were used for ICP-MS quantification. All reagents were of

>99.99% purity unless otherwise stated.

135 14.31 Ligand Synthesis

Amidrazone 1 was synthesised as previously reported from commercially available neocuproine.15 Amidrazone 1 was then reacted on 200 mg scale in THF (10 mL) in the presence of triethylamine (0.1 mL) with 2.1 equivalents of each of 26 diketones at reflux. The reactions were monitored by TLC (dichloromethane/MeOH) and took up to 3 d. The solvent was removed in vacuo and the solids were purified by column chromatography (dichloromethane/MeOH) to obtain compounds 2-27 (analytical data in Supplementary Information). CyMe4BTPhen and

16-17 Me4BTPhen were synthesised and purified as previously reported.

14.32 Solubility screening of BTPhen Ligands

Each of the ligands 2-27 were added to 1 mL of the three high immiscibility industrially relevant solvents 1-octanol, dodecane and cyclohexanone until the compound was at the solubility limit

(at 18-20°C). The solubility (mmol) was then calculated by dividing by the molecular weight of the compound (x 1000) to give the solubility of the compounds in mmol.

14.33 Am/Eu Separations with BTPhen Ligands

A 1 mM solution of the ligands described below in 1-octanol was contacted with 0.1-5 M

241 152 HNO3/HCl (1 mL) solution containing Am (III)/ Eu (III) (10-100 Bq each) in for 1.5 h using a vortex mixer. The ligand was then back-stripped for 1 h with 0.1 M HCl on a vortex mixer. The resulting three acidic phases were separated and counted by gamma spectroscopy against a standard of known activity in the same geometry.

136 14.34 Complex matrix separations with benzil-BTPhen ligands

Solutions of the six functionalised benzil-BTPhens in 1-octanol (1 mM, 1 mL) was contacted with the complex matrix outlined below. The solution was mixed for 1.5 h using a vortex mixer.

The ligand was then back-stripped for 1 h with 0.1 M HCl on a vortex mixer. The resulting three acidic phases were separated and counted by gamma spectroscopy against a standard of known activity in the same geometry, the aqueous phase was diluted for ICP-MS analysis against a standard matrix solution.

14.35 Generation of Complex Nuclear Matrix

Using 4 M HCl (VWR, AnalaR NORMAPUR® grade) a solution containing ~1 mg mL-1 of metal was prepared by addition of the following elements in chloride form (Sigma-Aldrich or

Fluka, >99.99% trace metal): Be, Cs, Ce, Pr, Nd, Sm, Tb, Y, Sr and Ba. Spike solutions of 241Am

(10-100 Bq range), 152Eu (10-100 Bq range) were added to generate a known molarity solution.

14.4 Results and Discussion

14.41 Ligand Synthesis

Twenty six BTPhen ligands were synthesised by reaction of amidrazone 117 with commercially available diketones from three families: isatin, benzil and aliphatic (Schemes 1-3)

Scheme 1. Synthesis of istain-BTPhens. Reagents: (i) diketone, Et3N, THF, Δ, <72 h.

137

Scheme 2. Synthesis of benzil-BTPhens. Reagents: (i) diketone, Et3N, THF, Δ, <72 h.

Scheme 3. Synthesis of aliphatic (and indan) BTPhens. Reagents: (i) diketone, Et3N, THF, Δ, <72 h.

14.42 Solubility Testing in Common Reprocessing Solvents

Increasing the solubility of BTPhens is of interest for their application to reprocessing methodology as the ligands typically have solubility in 1-octanol of <20 mM. Increased solubility would allow for smaller volumes of organic phase to be used and reduce secondary waste. To determine the effects of additional functionalisation on solubility, ligands 2-26 were dissolved in three high immiscibility industrially relevant solvents representative of three solvent groups (1-octanol (alcohol), dodecane (aliphatic),

138 cyclohexanone (cyclic ketones)) to determine their solubility (Table 1-3). We reasoned that specific functional groups such as methoxy or chloride would improve solubility of the ligands in polar solvents where bromide or iodide may potentially show the reverse trend. These alterations in the design could help improve promising extractants by tuning the solubility in process relevant solvents such as dodecane and 1-octanol. Also, cyclohexanone was reasoned to allow for stacking effects with these ligands/functional groups which we hoped would allow for trends based on functionalisation to be significant.

Isatin- 7-F-isatin- 7-Cl-isatin- 7-Br-isatin- Solubility / mM Indan- BTPhen BTPhen BTPhen BTPhen BTPhen 23 2 3 4 5 1-Octanol 8.0 8.1 2.2 8.0 6.0 Dodecane 5.0 6.2 3.8 3.0 3.0 Cyclohexan- one 22.0 98.0 50.7 100.0 42.0 5-CF3-isatin- 5-Me-isatin- 5-F-isatin- 5-Cl-isatin- 5-Br-isatin- BTPhen BTPhen BTPhen BTPhen BTPhen 12 11 6 7 8 1-Octanol 3.8 8.3 11.0 9.0 7.0 Dodecane 1.9 2.3 2.9 2.1 2.1 Cyclohexan- one 194.0 47.7 56.2 57.0 41.0 4,7-Cl2- isatin- 5-I-isatin- 5-NO2-isatin- BTPhen BTPhen BTPhen 13 9 10

1-Octanol 17.0 1.9 2.3 Dodecane 3.0 1.3 2.0 Cyclohexan- one 105.0 40.9 47.0

Table 1. Solubilities of isatin- (and indan) BTPhens in industrially relevant solvents. All isatin-BTPhen ligands had low solubility (<6.2 mM) in dodecane, and additional functionality had a detrimental effect on the solubility in straight chain solvents. This implies that functionalization in similar extractants ultimately may not improve in a process relevant non-polar solvent. Solubility was generally higher in 1-octanol for all

139 ligands which mirrors previous solubility studies and implies that perhaps functionality in certain positions could be used to improve solubility. In cyclohexanone, much higher solubilities (22–194 mM) were observed, and addition of the CF3 group markedly improved solubility, which could be useful for future ligand design if increased solubility in a cyclic solvent is required. Overall, the cyclohexanone data shows the effects of electron withdrawing groups on this general class of isatin-BTPhen ligands and shows that functionalisation to be counter-productive in most cases.

The benzil-BTPhens are generally less soluble than the isatin-BTPhens in octanol/dodecane. The amine functionalised benzil-BTPhen 21 showed higher solubility than the other ligands in 1-octanol which could be useful for ligand design (the resulting ligand would also still adhere to the CHON principle and amine/amides would potentially increase electron donation into the overall system). Solubility was highest in cyclohexanone for all ligands, with 2,2’-dichlorobenzil-BTPhen 22 showing the highest solubility (128 mM). This potentially shows that multiple chloride insertion can improve solubility in design of BTPhen ligands in specific solvents and positions.

2,2'-Cl2- 3,3'-(MeO)2- 4,4'-F2- 4,4'-Br2- Benzil- benzil- benzil- benzil- benzil- Solubility / mM BTPhen BTPhen BTPhen BTPhen BTPhen 14 22 16 18 19 1-Octanol 2.5 4.0 2.2 2.3 1.1 Dodecane 1.6 2.0 1.3 2.1 1.0 Cyclo-hexanone 50.0 128.1 9.2 42.0 11.5

4,4'-(MeO)2- 4,4'-Me2- 4,4'-Me2N- 4,4'-Cl2- benzil- benzil- benzil- benzil- Solubility / mM BTPhen BTPhen BTPhen BTPhen 17 20 21 15 2.7 1-Octanol 3.0 2.6 15.0 1.5 Dodecane 2.0 2.3 2.6 36.1 Cyclo-hexanone 78.7 40.1 63.8 Table 2: Solubilities of benzil-BTPhens in industrially relevant solvents.

140

MeEt-BTPhen MePr-BTPhen Et2-BTPhen Solubility / mM 24 25 26 1-Octanol 14.4 2.2 6.1 Dodecane 2.3 2.2 2.2 Cyclohexanone 14.9 99.8 157.6 Table 3: Solubilities of aliphatic BTPhens in industrially relevant solvents.

For the aliphatic BTPhen ligands, the solubility trends are similar to those previously reported by Asfar et al. for BTPhen/BTBP where increasing the aliphatic chain length increased the solubility in alcohol and cyclic ketone based solvents at the expense of acidic and radiolytic stability.9

14.43 Acid Stability of Functionalised BTPhens

All the BTPhens synthesised were stable when contacted with 3 M HCl/HNO3 (typically the molarity of high active raffinate) for 24 h. However, acid concentrations higher than this led to decomposition of the ligands in all cases; this is markedly different to stability of CyMe4BTPhen/BTBP which are stable in 4 M HCl/HNO3 overnight. In terms of ligand design this result shows that additional functionality appears to reduce acidic stability

141 regardless of which moieties are added to the ligand, and this would need to be considered in method design.

14.44 Am/Eu Separations by Functionalised BTPhens

Initial testing of Am (III)/Eu (III) separation for one ligand from each family was carried out prior to a full screen. The ligands (1 mM in 1-octanol) were contacted with a solution

241 152 of Am/ Eu (10 Bq each) in 0.1-4 M HCl/HNO3 for 1.5 h. The results were very similar to those of our previous work applying CyMe4BTPhen to complex matrices, therefore for separations with the functionalised BTPhens we used the same extraction procedure but with the aqueous phase as 3 M HNO3. Using this procedure, Am/Eu separation was carried out for each BTPhen ligand in triplicate (with CyMe4BTPhen as a control). This allowed for DAm/DEu and SFAm/Eu to be calculated (Tables 4-6).

Ligand SFAm/Eu CyMe4BTPhen 350 Indan-BTPhen 22 2 Isatin-BTPhen 2 4 5-Me-isatin-BTPhen 11 37

5-CF3-isatin-BTPhen 12 10 5-F-isatin-BTPhen 6 3 5-Cl-isatin-BTPhen 7 3 5-Br-isatin-BTPhen 8 1 5-I-isatin-BTPhen 9 1

5-NO2-isatin-BTPhen 10 3 7-F-isatin-BTPhen 3 8 7-Cl-isatin-BTPhen 4 8 7-Br-isatin-BTPhen 5 3

4,7-Cl2-isatin-BTPhen 13 3 Table 4: SFAm/Eu for isatin-BTPhens, determined by gamma spectroscopy.

142 All isatin-BTPhens achieved separation of Am from solution (>40% Am extracted within

1.5 h), however, they all co-extracted Eu (III) (5-27%), resulting in low separation factors. The overall observed trend is that secondary functionalisation of these extractants appears to reduce Am (III) extraction and favours Eu (III) co-extraction lowering the separation factor. However, it is interesting to note that 5-Me-isatin-BTPhen is the only ligand with an electron donating group in the family, and this shows the highest SFAm/Eu.

From this data, isatin-BTPhens do not appear to be viable ligands for Am (III) extraction due to low separation factor In addition to this it can perhaps be shown that the addition of electron withdrawing halogens to these systems reduces the metal-nitrogen interaction with Am (III) and reduces the observed selectivity.

Ligand SFAm/Eu

CyMe4BTPhen 350

Me4BTPhen 122

Et4BTPhen 25 105

Me2Et2BTPhen 23 40

Me2Pr2BTPhen 24 60 Table 5: SFAm/Eu for aliphatic BTPhens, determined by gamma spectroscopy.

The aliphatic BTPhen ligands were seen to have decreasing Am extraction capability with increasing chain length when symmetrical. It is interesting to note that the unsymmetrical ligands Me2Et2BTPhen and Me2Pr2BTPhen had lower separation factors even though there is no great increase in chain length. These unsymmetrical ligands represent a potential new area of ligand design for BTPhens, which have until now been generally symmetrical in design.

143 Ligand SFAm/Eu CyMe4BTPhen 350 Benzil-BTPhen 14 16 4,4-F2-benzil-BTPhen 18 30

4,4-Cl2-benzil-BTPhen 15 78

4,4-Br2-benzil-BTPhen 19 146

3,3-(MeO)2-benzil-BTPhen 16 71

4,4-(MeO)2-Benzil-BTPhen 17 284

2,2-Cl2-benzil-BTPhen 22 17

4,4-Me2-benzil-BTPhen 20 99 Table 6: SFAm/Eu for benzil-BTPhens, determined by gamma spectroscopy.

Overall, the benzil-BTPhens show higher SFAm/Eu than either of the other two ligand families. In addition, functionalisation appears to increase separation factors. For the 4,4- disubstituted ligands, an increase in SFAm/Eu can be seen with increasing electron donating ability of the substituent (OMe>Me>Br>Cl>F). The donation of electron density into the triazine moieties may assist with the covalent interactions required to complex Am (III) over Eu (III). Substituents at the 2- or 3-positions did not result in high separation factors, this is likely to be due to steric hindrance (2,2-Cl2-benzil-BTPhen) or due to the ortho/para directing effect of the substituents (3,3-(MeO)2-benzil-BTPhen). These results highlight the importance of electronics and sterics in the design of selective BTPhen extractants, in particular the methoxy benzil group is of interest due to its high separation factor and adherence to the CHON principle. As the benzil-BTPhens showed the best separation factors, they were then subjected to more complex Am-containing matrices to assess separation capability in the presence of competing ions, in particular the lanthanides. From these results, we think a CyMe4BTPhen extractant functionalised with electron-donating groups (e.g. ester, methyl, methoxy) would potentially improve both the solubility and observed extraction data and that from our work many groups can be

144 ignored as potential functionalisation candidates (e.g. NO2/I). Overall it is clear that functionalisation can both improve and hinder the solubility/stability in these systems and that the position and type of functionalisation is an important consideration when designing these extractants.

14.45 Complex Nuclear Matrix Separations with Benzil-BTPhens

The most promising benzil-BTPhens (omitting benzil-BTPhen and 2,2-Cl2-benzil-

BTPhen) were screened for Am extraction from more complex matrices containing competing ions expected to be encountered in reprocessing. A matrix was generated containing all lanthanides (except Pm) and Sr, Y, Cs, Ba as fission products, and extraction was carried out using each of the ligands in the same procedure as above. The

SFAm/Eu results were comparable to those obtained in the Am/Eu separations above for all ligands (see Supplementary Information). Recovery of the matrix elements was determined post-extraction by ICP-MS (Table 7).

Although recoveries for the matrix elements were typically high, the data shows most ligands are co-extracting one or more elements. 4,4-(MeO)2-Benzil-BTPhen, which has the best separation factor for Am/Eu, co-extracts ~10-15% of all matrix elements and therefore would not be a viable ligand for selective Am extraction, as high decontamination factor (>1000) could not be achieved. The 4,4-Cl2, 4,4-Br2 and 4,4-Me2- benzil-BTPhens also show co-extraction of the lanthanide ions (~10%). 3,3-(MeO)2- benzil-BTPhen shows high recoveries of the matrix elements (>93% in all cases), however the lack of Am selectivity is still a concern for separations. 4,4-F2-benzil-

BTPhen showed the lowest co-extraction of matrix elements, but this is attributed to slow

145 extraction kinetics, hence longer experiments which would be required for high Am recovery are expected to result in more co-extraction.

Ligand 88Sr 89Y 133Cs 137Ba 139La 140Ce 96.7 ± 95.6 ± 99.9 ± 99.9 ± 98.2 ± 97.6 ± 4,4-F2-benzil-BTPhen 18 1.1 0.5 1.0 0.7 0.5 0.7 91.6 ± 95.6 ± 95.5 ± 94.8 ± 93.5 ± 93.0 ± 4,4-Me2-benzil-BTPhen 20 1.3 0.8 1.0 0.8 0.7 0.8 86.2 ± 95.6 ± 89.9 ± 88.9 ± 87.7 ± 87.7 ± 4,4-(MeO)2-Benzil-BTPhen 17 1.5 0.9 0.9 0.8 0.9 0.7 88.8 ± 95.6 ± 92.5 ± 91.8 ± 90.4 ± 90.3 ± 4,4-Br2-benzil-BTPhen 19 1.7 0.9 1.2 0.8 0.9 0.7 91.8 ± 95.6 ± 95.4 ± 95.0 ± 93.5 ± 93.0 ± 4,4-Cl2-benzil-BTPhen 15 1.6 0.7 1.1 0.8 0.7 0.6 94.6 ± 95.5 ± 98.3 ± 98.3 ± 96.6 ± 95.9 ± 3,3-(MeO)2-benzil-BTPhen 16 1.3 1.4 1.2 0.8 0.7 0.6 Ligand 141Pr 146Nd 147Sm 157Gd 159Tb 163Dy 99.5 ± 99.1 ± 98.6 ± 97.8 ± 98.1 ± 97.2 ± 4,4-F2-benzil-BTPhen 18 0.7 0.8 0.9 0.9 0.9 1.2 94.5 ± 88.2 ± 94.0 ± 93.1 ± 93.0 ± 92.2 ± 4,4-Me2-benzil-BTPhen 20 0.8 0.9 0.7 0.9 0.9 1.1 89.2 ± 86..4 ± 88.6 ± 87.5 ± 87.6 ± 86.7 ± 4,4-(MeO)2-Benzil-BTPhen 17 0.9 0.8 0.8 0.9 0.9 1.0 91.9 ± 85.0 ± 90.7 ± 90.6 ± 90.3 ± 89.5 ± 4,4-Br2-benzil-BTPhen 19 1.0 1.0 0.7 0.8 0.8 1.0 95.1 ± 98.8 ± 93.7 ± 93.5 ± 93.7 ± 92.6 ± 4,4-Cl2-benzil-BTPhen 15 0.9 0.9 0.9 0.9 0.9 1.0 97.8 ± 95.5 ± 96.6 ± 96.6 ± 96.5 ± 95.6 ± 3,3-(MeO)2-benzil-BTPhen 16 0.7 0.9 0.6 0.9 0.9 1.1 Ligand 165Ho 166Er 169Tm 172Yb 175Lu 98.4 ± 97.9 ± 95.5 ± 94.7 ± 97.2 ± 4,4-F2-benzil-BTPhen 18 0.8 0.9 0.8 1.0 0.9 93.6 ± 92.7 ± 90.5 ± 90.1 ± 92.6 ± 4,4-Me2-benzil-BTPhen 20 0.8 0.9 0.9 0.9 0.8 88.0 ± 86.9 ± 84.7 ± 84.7 ± 86.7 ± 4,4-(MeO)2-Benzil-BTPhen 17 0.8 0.8 0.8 0.9 0.8 90.7 ± 89.7 ± 87.3 ± 87.4 ± 89.3 ± 4,4-Br2-benzil-BTPhen 19 0.6 0.8 0.7 0.9 0.7 93.6 ± 92.8 ± 90.2 ± 90.5 ± 92.6 ± 4,4-Cl2-benzil-BTPhen 15 0.7 0.8 0.6 0.8 0.8 96.6 ± 96.1 ± 93.2 ± 93.5 ± 95.9 ± 3,3-(MeO)2-benzil-BTPhen 16 0.8 0.8 0.8 0.9 0.8 Table 7: Percentage recovery for matrix elements from separation of complex matrix using benzil-BTPhen, determined by ICP-MS (samples run in triplicate, error calculated from instrument error plus one sigma).

146 14.5 Modelling of [AmL]3+ complexes of benzil-BTPhen ligands

The bond lengths in the [AmL]3+ complexes (minima complexes presented in the

Supplementary information) were calculated using density functional theory (DFT;

Gaussian 09)19 in the same manner as that reported by Xiao et al18 for the benzil-BTPhen ligand family. The functional used was B3LYP. With light atoms, the standard Pople- style polarized valence triple 6-311G(푑,푝) basis set was used for geometry optimization.

Small-core RECPs which replace 60 core electrons for Am and 28 electrons for Eu were also used.

All the chemical species were firstly optimized at the B3LYP/6-311G(푑,푝)/RECP level and natural bond order (NBO) analysis was undertaken. The resulting bond lengths (Am-

N2) lie in the 2.36-2.38 Å range which agrees well with previously reported calculations

19 18 for BTPhens. The shorter bond lengths when compared with (Eu-N2) No trends were observable in these data. The NBO analysis allowed a population analysis of the triazine nitrogen atoms in BTPhen ligands (used to complex Am (III)) for the additional functional groups (Table 8). The results suggest that, relative to benzil-BTPhen, all functional groups add additional electron density into the triazine system which we expected, which could potentially lead to stronger interaction when complexing with Am

(III). Overall, none of our computational analysis accounted for the observed reduction in selectivity seen when specific functionalisations were undertaken.

147 Additional population of Ligand triazine nitrogen / % of one electron benzil-BTPhen N/A

2,2-Cl2-benzil-BTPhen 5.178

3,3-(MeO)2 benzil-BTPhen 1.200

4,4-(MeO)2 benzil-BTPhen 3.174

4,4-F2-benzil-BTPhen 3.314

4,4-Cl2-benzil-BTPhen 2.124

4,4-Br2-benzil-BTPhen 2.316

4,4-Me2-benzil-BTPhen 2.124 Table 8: Calculated additional electron population on triazine nitrogens calculated by DFT (Gaussian 09, B3LYP)

14.6 Conclusions

From the three families of functionalised BTPhens synthesised, selectivity factors for Am

(III)/Eu (III) separation were obtained to determine the effects of ligand design on Am extraction. The isatin-BTPhens showed poor separation factors due to Eu co-extraction so this group of ligands would not be viable. The aliphatic BTPhen ligands showed decreasing separation factors with increasing chain length, and the unsymmetrical members of this family showed lower separation factors than symmetrical ligands. The benzil-BTPhens were the most promising group of ligands, showing the highest SFAm/Eu values, which improved when electron donating groups were present, an interesting point for future ligand design. The benzil-BTPhens were also tested for Am extraction from a more complex matrix containing lanthanides and fission products. For most benzil-

BTPhens, co-extraction of the lanthanides (10-15%) was observed, apart from the 4,4- fluoro (attributed to slow extraction kinetics). Overall, the ligands synthesised do not show a great improvement on the current benchmark ligands such as CyMe4BTPhen.

However, this study highlights how additional functionality on BTPhens can be used to

148 modify properties such as solubility and Am (III) extraction and potentially underpins the design of future Am/Cm selective extractants for reprocessing methodologies. Future work will attempt to synthesise additional diketones with electron donating groups to yield more data on the effects on selectivity to fill in the majority of potential groups.

Also work on functionalising the CyMe4BTPhen system should be attempted in light of these observations with electron donating CHON groups to attempt to resolve problems with slow phase transfer kinetics and to attempt to improve the system which could potentially be used in the proposed SANEX process.

14.7 Acknowledgements

Funding for this project was provided by AWE via a studentship to MAH. ICP-MS analyses were carried out by Mr Paul Lythgoe (School of Earth, Atmospheric and

Environmental Sciences, The University of Manchester).

14.8 Notes and references

1 Fuel Cycle Stewardship in a Nuclear Renaissance, The Royal Society Science Policy Centre Report, Oct. 2011, Chapter 4, p. 39 2 K. L. Nash, G. R.Choppin, Sep. Sci. Technol. 1997, 32, 255. 3 C. Madic, M. J. Hudson, O. J. Liljenzin, J. –P. Glatz, R. Nannicini, A. Facchini, Z. Kolarik, R. Odoj, Prog. Nucl. Energy 2002, 40, 523. 4 M. J. Hudson, L. M. Harwood, D. M. Laventine, F. W Lewis, Inorg. Chem., 2013, 52, 3414. 5 J.-P. Grouiller, S. Pillon, C. de Saint Jean, F. Varaine, L. Leyval, G. Vambenepe, B. Carlier, J. Nucl. Mater., 2003, 320, 163. 6 J. Brown, M. J. Carrott, M, O. D. Fox, C. J. Maher, C. Mason, F. McLachlan, M. J. Sarsfield,. R. J. Taylor, D. A. Woodhead, IOP Conf. Ser.: Mater. Sci. Eng., 2010, 9, 012075. 7 A. Geist, C. Hill, G. Modolo, M. R. Foreman, M. St. J. Weigl, K. Gompper, M. J. Hudson, C. Madic, Solvent Extr. Ion Exch. 2006, 24, 463. 8 P. J. Panak, A. Geist, Chem. Rev. 2013, 113, 1199. 9 D. M. Laventine, A. Afsar, M. J. Hudson, L. M. Harwood, Heterocycles, 2012, 2, 86. 10 A. Afsar, L. M. Harwood, M. J. Hudson, J. Westwood, A. Geist, Chem. Commun.2015, 51, 5860.

149 11 D. Magnusson, B. Christiansen, R. Malmbeck, J.-P. Glatz, Radiochim. Acta, 2009, 97, 497. 12 C. Adam, P. Kaden, B. B Beele, U. Müllich, S. Trumm, A. Geist, P. J. Panak, M. A. Denecke, Dalton Trans., 2013, 42, 14068. 13 M. A. Higginson, O. J. Marsden. P. Thompson. L. M. Harwood. M. J. Hudson. F. W. Lewis. F. R. Livens. S. L. Heath. submitted. (Radiochim Acta) 14 W. C. Still, M. Kahn, A. Mitra, J. Org. Chem., 1978, 43, 2923. 15 M. J. Hudson, L. M. Harwood, D. M. Laventine, F. W. Lewis, Inorg. Chem., 2013, 52, 3414. 16 D. M. Whittaker, T. L. Griffiths, M. Helliwell, A. N. Swinburne, L. S. Natrajan, F. W. Lewis, L. M. Harwood, S. A. Parry, C. A. Sharrad, Inorg. Chem., 2013, 52, 3429. 17 F. W. Lewis, L. M. Harwood, M. J. Hudson, M. G. B. Drew, J. F. Desreux, G. Vidick, N. Bouslimani, G. Modolo, A. Wilden, M. Sypula, T.-H. Vu, J.-P. Simonin, J. Am. Chem. Soc., 2011 133, 13093. 18 C.-L. Xiao, C.-Z. Wang, J-H. Lan, L.-Y. Yuan, Y.-L. Zhao, Z.-F. Chai, W.-Q. Shi, Radiochim. Acta, 2014, 102, 875. 19 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2009

150 15.0 Appendix to section 4

Synthesis of functionalised BTPhen derivatives - effects on solubility and americium extraction

Matthew A. Higginson,a Nichola D. Kyle,a Olivia J. Marsden,b Paul Thompson,b Francis R.

Livensa and Sarah L. Heath.a,*

Supplementary Information

151 15.1 Complex Matrix SFAm/Eu calculated results (by gamma spectroscopy)

Results calculated from DAm/DEu data as shown by Lewis et al in reference 17 in the manuscript

Ligand SFAm/Eu CyMe4BTPhen 350 4,4-F2-benzil-BTPhen 18 32

4,4-Cl2-benzil-BTPhen 15 90

4,4-Br2-benzil-BTPhen 19 153

3,3-(MeO)2-benzil-BTPhen 16 75

4,4-(MeO)2-Benzil-BTPhen 17 252

4,4-Me2-benzil-BTPhen 20 95 Table 1. SFAm/Eu from separation of complex matrix using benzil-BTPhen ligands

BTPhen ligand Am (III) extraction versus time

Figure 1. Extraction recovery (Am III) versus time for three BTPhen ligands from this study determined by gamma spectroscopy.

15.2 Triazine ligand experimental data

2,9-bis(9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (2) o - Bright orange solid (200 mg, 56 %); m.p.= 192 C; Rf = 0.70 (92:8 DCM: MeOH); IR (vmax / cm 1): 3441 (N-H), 2888, 2815 (C-H Ar), 1614 (C=N), 1330 (C-N); MS (ES+) found m/z 597.4 + + [M+H+DMSO] , HRMS found m/z 517.1604 [M+H] requires 517.1632; δH (400 MHz, DMSO- d6) 11.05 (2 H, s, NH), 8.81 (2 H, d, J = 7.8 Hz, ArH), 8.49 (2 H, d, J = 7.3 Hz, ArH), 8.24 (2 H, d, J = 7.2 Hz, ArH), 7.95 (2 H, s, ArH), 7.59 (2 H, t, J = 7.8 Hz, ArH), 7.51 (2 H, d, J = 8.1 Hz,

152 ArH), 7.07 (2 H, t, J = 7.9 Hz, ArH) ppm; δC (101 MHz, DMSO-d6) 184.7, 159.8, 151.2, 146.5, 142.3, 138.8, 138.4, 130.7, 125.2, 124.9, 123.2, 118.3, 112.7 ppm.

2,9-bis(8-fluoro-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (3) o -1 Red/orange solid (120 mg, 31 %); m.p.= 179 C; Rf = 0.60 (92:8 DCM: MeOH); IR (vmax / cm ): 3444 (N-H), 3093, 3061 (C-H Ar), 1637 (C=N), 1325 (C-N), 1206 (C-F); MS (ES+) found m/z + + 386.5 [M/2 + halo exchange] , HRMS found m/z 553.1065 [M+H] requires 553.1044; δH (400 MHz, DMSO-d6) 11.57 (2 H, s, NH), 8.35 (2 H, d, J = 7.3 Hz, ArH), 8.29 (2 H, d, J = 7.1 Hz, ArH), 8.26 (2 H, s, ArH), 7.80 (2 H, d, J = 7.3 Hz, ArH), 7.58‒7.50 (2 H, m, ArH), 7.38 (2 H, d,

J = 7.3 Hz, ArH) ppm. δC (101 MHz, DMSO-d6) 183.7, 159.7, 154.9, 152.8, 148.9, 146.5, 137.9, 137.8, 125.3, 125.1, 123.8, 121.1, 118.7 ppm.

2,9-bis(8-chloro-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (4) o -1 Red solid (90 mg, 21 %); m.p.= 184 C; Rf = 0.85 (92:8 DCM: MeOH); IR (vmax / cm ): 3475 (N-H), 3169, 3075 (C-H Ar), 1611 (C=N), 1317 (C-N), 775 (C-Cl); MS (ES-) found m/z 604.7 + + [M + Na – 2H] , HRMS found m/z 585.0852 [M+H] requires 585.0853; δH (400 MHz, DMSO- d6) 11.47 (2 H, s, NH), 8.41 (2 H. d, J = 8.3 Hz, ArH), 8.26 (2 H, d, J = 8.3 Hz, ArH), 8.08 (2 H, s, ArH), 7.67 (2 H, dd, J = 8.2 Hz, ArH), 7.49 (2 H, d, J = 7.3 Hz, ArH), 7.10 (2 H, t, J = 7.4 Hz,

ArH) ppm. δC (101 MHz, DMSO-d6) 183.9, 166.9, 160.1, 148.2, 142.7, 141.5, 137.9, 132.2, 124.2, 123.6, 120.4, 119.4, 116.6 ppm.

2,9-bis(8-bromo-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (5) o -1 Orange solid (240 mg, 52 %); m.p.= 177 C; Rf = 0.79 (92:8 DCM: MeOH); IR (vmax / cm ): 3455 (N-H), 3176, 3104 (C-H Ar), 1609 (C=N), 1314 (C-N), 686 (C-Br); HRMS Accurate found + m/z 687.6647 [M+NH4] requires 687.6682; δH (400 MHz, DMSO-d6) 11.30 (2H, s, NH), 8.52 (2 H, d, J = 8.0 Hz, ArH), 8.08 (2 H, s, ArH), 7.66‒7.55 (4 H, m, ArH), 7.52 (2 H, dd, J = 7.3 Hz

ArH), 7.03 (2H, t, J = 8.1 Hz, ArH) ppm δC (101 MHz, DMSO-d6) 184.1, 160.2, 149.9, 140.8, 137.2, 134.2, 124.6, 124.0, 120.6, 116.0, 105.1, 45.9, 9.0 ppm.

2,9-bis(6-fluoro-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (6) o -1 Orange solid (160 mg, 43 %); m.p.= 244 C; IR (vmax / cm ): 3308 (N-H), 3078, 2861 (C-H Ar), + 1621 (C=C), 1142 (C-F); HRMS found m/z 553.1404 [M+H] requires 553.1449; δH (400 MHz, DMSO-d6) 11.08 (2 H, s, NH), 8.85‒8.81 (2 H, m, ArH), 8.79‒8.75 (2 H, m, ArH), 8.24 (2 H, s, ArH) 7.49‒7.37 (4 H, m, ArH), 6.95‒6.90 (2 H, d, J = 8.3 Hz, ArH) ppm. δC (101 MHz, DMSO- d6) 184.4, 159.9, 159.7, 157.4, 153.6, 147.4, 137.1, 125.1, 124.8, 119.0, 113.9, 111.9, 111.7, 109.9 ppm.

2,9-bis(6-chloro-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (7) o -1 Orange solid (130 mg, 33 %); m.p.= 217 C; Rf = 0.73 (92:8 DCM: MeOH); IR (vmax / cm ): 3440 (N-H), 3091, 3065, 2986 (C-H Ar), 1615 (C=C), 1167 (C-N), 705 (C-Cl); HRMS m/z + found 585.0803 [M+H] requires 585.0858; δH (400 MHz, DMSO-d6) 11.11 (2 H, s, NH), 8.85 (2 H, d, J = 8.6 Hz, ArH), 8.72 (2 H, d, J = 8.3 Hz, ArH), 8.24 (2 H, s, ArH), 7.62 (2 H, d, J =

153 8.6 Hz, ArH), 7.56 (2 H, s, ArH), 6.93 (2 H, d, J = 8.3 Hz, ArH) ppm. δC (101 MHz, DMSO-d6) 183.8, 159.6, 159.3, 149.7, 145.7, 137.7, 131.9, 127.2, 124.6, 119.6, 119.4, 114.3, 111.55 ppm.

2,9-bis(6-bromo-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (8) o -1 Red solid (121 mg, 26 %); m.p.= 190 C; Rf = 0.63 (92:8 DCM: MeOH); IR (vmax / cm ): 3435 (N-H), 3286, 3154, 2918 (C-H Ar), 1616 (C=C), 1140 (C-N), 609 (C-Br); HRMS found m/z + 775.2536 [M+Na+DMSO] requires 775.2510; δH (400 MHz, DMSO-d6) 10.94 (2 H, s, NH), 8.79 (2 H, d, J = 5.8 Hz, ArH), 8.74‒8.70 (2 H, m ArH), 8.64 (2 H, s, ArH), 8.29 (2 H, s, ArH),

7.40 (2 H, d, J = 7.8 Hz, ArH), 7.18 (2 H, d, J = 7.8 Hz) ppm. δC (101 MHz, DMSO-d6) 185.1, 166.2, 163.3, 159.9, 149.0, 141.9, 139.2, 138.3, 132.5, 128.4, 125.3, 118.2, 112.5 ppm.

2,9-bis(6-iodo-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (9) o -1 Purple solid (180 mg, 34 %); m.p.= 155-157 C; Rf = 0.64 (92:8 DCM: MeOH); IR (vmax / cm ): 3455 (N-H), 3224, 3093 (C-H Ar), 1602 (C=N), 1382 (C-N), 595 (C-I); HRMS found m/z + 768.5368 [M+H] requires 768.5313; δH (400 MHz, DMSO-d6) 11.21 (2 H, s, NH), 7.91‒7.76 (8 H, m, ArH), 6.76 (2 H, d, J = 8.6 Hz, ArH), 5.76 (2 H, s, ArH) ppm. δC (101 MHz, DMSO-d6) 183.5, 159.2, 151.6, 150.4, 146.2, 137.1, 132.9, 125.3, 120.4, 115.1, 112.3, 96.6, 85.9 (I-C) ppm.

2,9-bis(6-nitro-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (10) -1 Brown solid (50 mg, 11 %); Rf = 0.75 (92:8 DCM: MeOH); IR (vmax / cm ): 1547 (N-O), 1363 + (N-O); HRMS found m/z 607.9985 [M+H] requires 607.9939; δH (400 MHz, MeOH-d4) 8.39‒ 8.32 (2 H, m, ArH), 8.17 (1 H, dd, J = 9.46, ArH), 8.14‒8.04 (1 H, m, ArH), 7.64‒7.53 (1 H, m, ArH), 7.42 (1 H, d, J = 8.07 Hz), 7.38‒7.28 (1 H, m, ArH), 7.17‒7.04 (1 H, m, ArH) 7.00‒6.88

(2 H, m, ArH) 6.80 (1 H, d, J = 9.33 Hz, ArH) 6.72‒6.65 (1 H, m, ArH) ppm. δC (101 MHz, MeOH-d4) 139.8, 138.3, 134.3, 130.0, 129.7, 129.1, 128.2, 126.3, 124.7, 123.1, 122.2, 117.3, 116.3 ppm.

2,9-bis(6-methyl-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (11) o -1 Orange solid (150 mg, 41 %); m.p.= 237 C; IR (vmax / cm ): 3432 (N-H), 3195, 2995 (C-H Ar), 1611 (C=C); MS (ES+) found m/z 545.6 [M+H]+, HRMS found m/z 545.1970 [M+H]+ requires

545.1951; δH (400 MHz, DMSO-d6) 11.11 (2 H, s NH), 9.11‒8.10 (4 H, m, ArH), 8.04‒6.47 (8 H, m, ArH), 3.42 (6 H, s, 2 x CH3) ppm. δC (101 MHz, DMSO-d6) 159.1, 155.8, 150.0, 140.6, 140.5, 136.2, 135.5, 127.7, 127.6, 127.4, 127.4, 120.1, 116.7, 114.7, 24.8 ppm.

2,9-bis(6-(trifluoromethoxy)-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (12) -1 Bright red solid (210 mg, 45 %); Rf = 0.61 (92:8 DCM: MeOH); IR (vmax / cm ): 3404 (N-H), 1656 (N-H), 1255 (C-O), 997 (C-F); MS (ES+) found m/z 721.3 [M+H+K]+, HRMS found m/z + 687.3612 [M + 3H] requires 687.3611; δH (400 MHz, DMSO-d6) 11.29 (2 H, s, NH), 7.59 (2 H, d, J = 8.3 Hz, ArH), 7.52 (2 H, s, ArH), 7.45 (2 H, s, ArH), 7.40 (2 H, m, ArH), 7.02 (2 H, d J =

8.3 Hz, ArH), 6.81 (2 H, d, J = 8.1 Hz, ArH) ppm. δC (101 MHz, DMSO-d6) 183.9, 159.9,

151.2, 150.0, 143.8, 136.8, 131.4, 128.0, 125.9, 124.4, 121.8, 119.1, 118.2, 113.9 ppm; δF (400 MHz, DMSO-d6) -57.6 (3 F, s, CF3) ppm.

154 2,9-bis(5,8-dichloro-9H-[1,2,4]triazino[6,5-b]indol-3-yl)-1,10-phenanthroline (13) o -1 Orange solid (210 mg, 48 %); m.p.= 232 C; Rf = 0.67 (92:8 DCM: MeOH); IR (vmax / cm ): 3449 (N-H), 1665 (N-H), 1606 (C=N), 1164 (C-N), 655 (C-Cl), 622 (C-Cl), 595 (C-Cl); MS (ES- + + ) found m/z 689.2 [M-H+Cl] , HRMS found m/z 663.4526 [M-H+Cl] requires 663.4546; δH (400 MHz, DMSO-d6) 11.37 (2 H, s, NH), 7.65 (2 H, d, J = 8.6 Hz, ArH), 7.26 (2 H, d, J = 8.3 Hz, ArH), 7.82 (2 H, d, J = 8.8 Hz, ArH), 7.09 (2 H, d, J = 8.6 Hz, ArH), 6.39 (2 H, s, ArH) ppm. δC (101 MHz, DMSO-d6) 183.7, 180.8, 159.4, 149.5, 144.9, 138.2, 131.2, 129.9, 124.8, 117.1, 115.3, 113.9, 55.4 ppm.

2,9-bis(5,6-diphenyl-1,2,4-triazin-3-yl)-1,10-phenanthroline (14)

o -1 Yellow solid (100 mg, 23 %); m.p.= 182 C; Rf 0.50 (92:8 DCM: MeOH); IR (vmax / cm ): 3057, + + + + 2927 (C-H Ar), 1618 (C=N), 1551 (C=C);MS (ES ) found m/z 662.8 [M+(K to NH4 )] , HRMS + found m/z 643.5410 [M+H] requires 643.5421; δH (400 MHz, DMSO-d6) 9.01‒8.63 (2 H, m, ArH), 8.53 (2 H, s, ArH), 8.30‒8.05 (2 H, m, ArH), 7.97‒7.92 (8 H, m, ArH), 7.78‒7.58 (8 H, m,

ArH), 7.58‒7.26 (4 H, m, ArH) ppm. δC (101 MHz, DMSO-d6) 178.7, 167.8, 161.0, 145.2, 133.1, 130.3, 129.9, 129.1, 128.8, 128.5, 126.4, 119.4 ppm.

2,9-bis(5,6-bis(4-chlorophenyl)-1,2,4-triazin-3-yl)-1,10-phenanthroline (15)

o -1 Yellow solid (100 mg, 18 %); m.p.= 239-242 C; Rf = 0.90 (92:8 DCM: MeOH); IR (vmax / cm ): 3058, 2927 (C-H Ar), 1505 (C=C), 648, 639, 619, 608 (C-Cl); HRMS found m/z 779.0768; δH (400 MHz, DMSO-d6) 7.91‒7.70 (11 H, m, ArH), 7.40‒7.28 (11 H, m, ArH) ppm; δC (101 MHz, DMSO-d6) 188.0, 162.9, 161.5, 159.6, 157.4, 153.1, 131.6, 131.5, 128.5, 115.4, 115.1 ppm.

2,9-bis(5-(3-methoxyphenyl)-6-(m-tolyl)-1,2,4-triazin-3-yl)-1,10-phenanthroline (16)

o -1 Yellow solid (130 mg, 25 %); m.p.= 192 C; IR (vmax / cm ): 3076, 3011, 2966 (C-H Ar), 1666 (C=C), 1303 (C-N), 1031 (C-O); MS (ES-) found m/z 778.5, HRMS found m/z 763.2797 + [M+H] requires 763.2781; δH (400 MHz, DMSO-d6) 8.51 (2 H, d, J = 8.6 Hz, ArH), 8.28‒8.19 (2 H, m, ArH), 7.99 (2 H, s, ArH), 7.80 (8 H, d, J = 8.1 Hz, ArH), 7.44 (8 H, d, J = 7.8 Hz, ArH),

2.42 (12 H, s, 4 x CH3O); δC (101 MHz, DMSO-d6) 195.1, 169.9, 146.9, 135.6, 130.5, 130.1, 126.6, 118.3, 109.1, 87.7, 86.0, 79.1, 21.9 ppm.

2,9-bis(5,6-bis(4-methoxyphenyl)-1,2,4-triazin-3-yl)-1,10-phenanthroline (17)

o -1 Yellow solid (85 mg, 16 %); m.p.= 166-168 C; Rf = 0.50 (92:8 DCM: MeOH); IR (vmax / cm ): 3069, 2934 (C-H Ar), 2839 (C-H), 1603 (C=C), 1017 (C-O); HRMS found m/z 763.2775 + [M+H] requires 763.2781; δH (400 MHz, DMSO-d6) 10.22 (2 H, s, ArH), 8.92 (8 H, d, J = 3.8 Hz, ArH), 8.37 (8 H, d, J = 2.0 Hz, ArH) 8.28 (2 H, d, J = 9.1 Hz, ArH), 8.10 (2 H, d, J = 9.1

Hz, ArH), 7.91 (d, J=9.1 Hz, 1 H, ArH), 7.86 (2 H, s, ArH), 3.88 (12 H, s, 4 x CH3O) ppm; δC

155 (101 MHz, DMSO-d6) 149.9, 145.4, 144.1, 134.6, 129.9, 124.2, 119.9, 108.8, 101.8, 99.8, 92.8, 91.8, 72.6 ppm.

2,9-bis(5,6-bis(4-fluorophenyl)-1,2,4-triazin-3-yl)-1,10-phenanthroline (18)

o -1 Yellow solid (95 mg, 20 %); m.p.= 246 C; IR (vmax / cm ): 3074 (C-H Ar), 1598 (C=C), 1274 (C-N), 1155 (C-F); MS (ES+) found m/z 715 [M+H]+, HRMS found 715.1961 [M+H]+ requires

715.1981; δH (400 MHz, DMSO-d6) 8.11‒8.00 (11 H, m, ArH), 7.54‒7.41 (11 H, m, ArH) ppm; δC (101 MHz, DMSO-d6) 193.0, 167.9, 165.5, 158.6, 158.0, 154.5, 133.6, 133.5, 129.5, 117.4, 117.1 ppm.

2,9-bis(5,6-bis(4-bromophenyl)-1,2,4-triazin-3-yl)-1,10-phenanthroline (19) o -1 Yellow solid (204 mg, 30 %); m.p.= 193-194 C; Rf = 0.68 (92:8 DCM: MeOH); IR (vmax / cm ):

1587 (C=C), 575, 562, 558, 552 (C-Br); HRMS found m/z 959.8737; δH (400 MHz, DMSO-d6) 8.81‒8.53 (8 H, d, J = 9.2 Hz, ArH), 8.15 (2 H, s, ArH), 7.86 (2 H, d, J = 6.8 Hz, ArH), 7.55 (2

H, s, ArH), 6.97 (8 H, s, ArH) ppm. δC (101 MHz, DMSO-d6) 155.0, 151.3, 143.2, 139.1, 133.9, 133.1, 132.1, 131.5, 131.3, 130.5, 128.5, 125.3 ppm.

2,9-bis(5,6-di-p-tolyl-1,2,4-triazin-3-yl)-1,10-phenanthroline (20)

o -1 Yellow solid (132 mg, 28 %); m.p.= 109 C; IR (vmax / cm ): 3063, 3048, 2976 (C-H Ar), 1659 + (C=C), 1329 (C-N); HRMS found m/z 699.2965 [M+H] requires 699.2984; δH (400 MHz, DMSO-d6) 8.56 (2 H, d, J = 8.3 Hz, ArH), 8.35 (2 H, d, J = 7.1 Hz, ArH), 8.29 (2 H, s, ArH), 7.47‒7.35 (16 H, m, ArH), 3.38 (12 H, s, CH3) ppm; δC (101 MHz, DMSO-d6) 194.9, 160.3, 142.5, 141.3, 133.9, 131.2, 123.4, 122.4, 116.5, 113.1, 105.6, 101.9, 55.9 ppm.

4,4',4'',4'''-((1,10-phenanthroline-2,9-diyl)bis(1,2,4-triazine-3,5,6-triyl))tetrakis(N,N- dimethylaniline) (21) o -1 Green solid (80 mg, 14 %); m.p.= 239 C; IR (vmax / cm ): 3011, 2917, 2829 (C-H Ar), 1581 + + (C=C), 1155, 1132 (C-N);MS (ES ) found m/z 833.8 [M+H/H2O] , HRMS found m/z 883.2365 + [M+MeOH+2H] require 883.2346; δH (400 MHz, DMSO-d6) 8.41‒8.37 (2 H, m, ArH), 8.30 (2 H, d, J = 8.6 Hz, ArH), 7.95 (2 H, s, ArH), 7.65 (8 H, d, J = 9.1 Hz, ArH), 6.76 (8 H, d, J = 9.3

Hz, ArH), 3.04 (24 H, s, 8 x CH2) ppm; δC (101 MHz, DMSO-d6) 193.9, 154.7, 144.2, 143.9, 136.6, 131.9, 128.7, 126.6, 120.8, 119.5, 111.6 ppm.

2,9-bis(5,6-bis(2-chlorophenyl)-1,2,4-triazin-3-yl)-1,10-phenanthroline (22)

o -1 Yellow solid (104 mg, 19 %); m.p.= 222-232 C; Rf = 0.90 (92:8 DCM: MeOH); IR (vmax / cm ): 3058, 2927 (C-H Ar), 1505 (C=C), 648, 639, 620, 611 (C-Cl); HRMS found m/z 779.1088 + [M+H] requires 779.1035; δH (400 MHz, DMSO-d6) 8.56 (2 H, d, J=8.6 Hz, ArH), 8.15 (2 H, d, J=8.3 Hz, ArH), 8.08 (2 H, s, ArH), 7.73 (8 H, d, J=7.3 Hz, ArH), 7.66‒7.38 (8 H, m, ArH) ppm;

δC (101 MHz, DMSO-d6) 193.4, 158.6, 149.0, 141.0, 132.6, 131.7, 130.1, 129.5, 127.3, 127.1, 118.7 ppm

156 2,9-bis(9H-indeno[1,2-e][1,2,4]triazin-3-yl)-1,10-phenanthroline (23) o -1 Black solid (190 mg, 53 %); m.p.= 192 C; Rf = 0.69 (92:8 DCM: MeOH); IR (vmax / cm ): 3069, 2929 (C-H Ar), 1602 (C=C);MS (ES-) found m/z 639.6 [M+I to H exchange]-, HRMS + found m/z 529.1647 [M+NH4] requires 529.1652; δH (400 MHz,CDCl3) 8.24 (2 H, d, J = 8.4 Hz, ArH), 8.04 (2 H, d, J = 8.1 Hz, ArH), 7.92 (2 H, s, ArH), 7.76‒7.58 (6 H, m, ArH), 3.42 (4

H, s, 2 x CH2) ppm; δC (101 MHz, CDCl3) 198.9, 171.6, 157.3, 153.9, 150.5, 150.4, 143.5, 136.6, 135.6, 134.4, 133.6, 125.2 ppm.

2,9-bis(6-ethyl-5-methyl-1,2,4-triazin-3-yl)-1,10-phenanthroline (24) o -1 Brown oil (60 mg, 21 %); m.p.= 159 C; Rf = 0.56 (92:8 DCM: MeOH); IR (vmax / cm ): 2971, + + 2923 (C-H Ar), 2850 (C-H), 1619 (C=C);MS (ES ) found m/z 559.7 [M+NaCl2/CO2] , HRMS found 423.2035; δH (400 MHz, MeOD-d4, mixture of regioisomers) 8.99‒7.55 (6 H, m, ArH), 1.28‒1.05 (6 H, m, 2 x CH3), 1.05‒0.99 (6 H, m, 2 x CH3), 0.84‒0.70 (4 H, m, 2 x CH2) ppm; δC (101 MHz, MeOD-d4) 136.8, 121.3, 117.6, 40.0, 31.7, 29.4, 22.3, 20.3, 13.3, 13.1, 12.8, 11.8, 7.8 ppm.

2,9-bis(5-methyl-6-propyl-1,2,4-triazin-3-yl)-1,10-phenanthroline (25) -1 Dark brown oil (190 mg, 58 %); Rf = 0.77 (92:8 DCM: MeOH); IR (vmax / cm ): 2975, 2937 (C- + + H Ar), 2878 (C-H), 1619 (C=C); MS (ES ) found m/z 243.2 [M/2+NH4] , HRMS found m/z + 451.2448 [M+H] requires 451.2410; δH (400 MHz, DMSO-d6, mixture of regioisomers) 8.80 (2 H, d, J = 8.3 Hz, ArH), 8.65 (2 H, d, J = 8.3 Hz, ArH), 7.96 (2 H, s, ArH), 4.26 (4 H, t, J = 6.9

Hz, 2 x CH2), 3.17 (6 H, s, 2 x CH3), 2.29‒2.17 (4H, m, 2 x CH2), 0.99 (6 H, t, J = 6.9 Hz, 2 x CH3) ppm; δC (101 MHz, DMSO-d6) 175.7, 144.9, 112.4, 89.4, 85.6, 27.4, 26.7, 22.9, 19.3, 12.8, 9.6 ppm.

2,9-bis(5,6-diethyl-1,2,4-triazin-3-yl)-1,10-phenanthroline (26)

-1 Dark brown oil (150 mg, 49 %); Rf = 0.50 (92:8 DCM: MeOH); IR (vmax / cm ): 2962, 2934 (C- H Ar), 2873 (C-H), 1618 (C=C); MS (ES+) found m/z 249.5 [M/2+H+Na]+, HRMS found m/z + 451.2348 [M+H] requires 451.2345; δH (400 MHz, DMSO-d6) 8.22 (2 H, d, J = 8.8 Hz, ArH), 8.19 (2 H, d, J = 8.8 Hz, ArH), 7.99 (2 H, s, ArH), 2.26‒2.15 (8 H, m, 4 x CH2), 1.09 (12 H, t, J = 7.1 Hz, 4 x CH3) ppm; δC (101 MHz, DMSO-d6) 174.8, 172.5, 169.0, 42.6, 36.0, 21.7, 21.5, 18.4, 14.6, 13.9, 13.6 ppm.

157 16.0 Conclusions and recommendations for future work

 I have achieved the aims of the research project by developing selective radiochemical

separation processes for americium and curium for application to complex nuclear

forensics samples to meet the industrial sponsor’s requirements and this work has been

demonstrated in the sponsors laboratories.

 Overall the procedures (using multiple ligands/methods with liquid-liquid separation and

EC resins) can be used alongside conventional EC resins to improve the overall

separation procedure outlined in figure 1 and they are also applicable with the current

method. A full, quantitative data set can be obtained within 2-3 days as outlined in the

brief.

In conclusion, the overall procedure utilises AG1-X8/extraction chromatography followed by triazine ligand separation in various forms. In the initial liquid-liquid procedure, purification and quantification of all major actinides can be achieved in <2 days, which would be valuable in a nuclear forensics investigation where speed of investigation is important. This first method avoids use of multiple column techniques and generation of thiocyanate waste, and can be flexible in extractant choice. All four chosen triazine ligands tested were found to extract

Am/Cm from complex matrices (e.g. containing Ln, Fe, Ca, Al) after AG1-X8 chromatography with little complexation of competing ions. Of particular note is the ability of the ligands to achieve separation of Am/Cm from all the lanthanides, which to our knowledge has not been tested before with ligands based on the triazine motif.

Work was then undertaken to find ways of implementing this work on the solid support, firstly a synthetic approach was undertaken. Of the many routes attempted, the polymer PVB-

Me4BTPhen was generated by synthesis of the C-5 brominated derivative of the ligand, which

158 allowed attachment of an aniline linker and subsequent conjugation to a PVBC solid support. I propose that this general methodology could be applied to other BTPhens and also related triazine ligands e.g. BTBPs to generate new extraction chromatography resins for use in An

(III)/Ln (III) applications.

PVB-Me4BTPhen extracted americium (>90%) from Am (III)/Eu (III) solutions with an

Am (III):Ln (III) decontamination factor of >1000 (calculated as ratio of separation factors, from percentage recoveries) over 90 min. Back-extraction of the polymer to obtain aqueous americium was possible using low molarity acid, as observed for the analogous liquid-liquid separation process. The polymer-bound nature of the ligand may enable wider use of BTPhen americium extraction in analytical radiochemistry and allow the development of new separation methods when combined with existing EC methodology.

Due to the synthetic approach being difficult for application to an analytical radiochemistry laboratory, we worked on adsorption methodologies allowing any triazine ligand to potentially be immobilised on a solid support to give a large flexible range of liquid-liquid and

EC resins for application to complex matrices. Three novel EC resins based on the soft-N donor

BTBP/BTPhen triazine ligands were generated by adsorption ligands into Amberlite XAD-7.

The resins were then characterised using IR spectroscopy and elemental analysis and then tested in Am/Eu separations. These polymers were then tested with complex matrices in conjunction with AG1-X8 chromatography to achieve complete separation and quantification of all matrix elements with a combination of ICP-MS, gamma spectroscopy and alpha spectrometry. From these results the capacity factor (KD) as a function of HNO3 concentration were calculated allowing for potential separation methods for a wide range of metal ions to be used to design

159 separation methods. In particular these resins show excellent potential for the specific extraction of Am from complex matrices.

With methods to generate separation procedure on the solid support and in solution I investigated the synthesis of functionalized BTPhen derivatives to probe effects on solubility and americium extraction. Three families of functionalised BTPhen ligands were designed and synthesised with commercially available diketones. Selectivity factors for Am/Eu separation were obtained to determine the effects of ligand design on Am extraction. The isatin-BTPhens showed poor separation factors due to Eu co-extraction and therefore this group of ligands would not be viable for selective Am extraction. The aliphatic BTPhen ligands showed decreasing separation factors with increasing chain length, and the unsymmetrical members of this family showed lower separation factors than symmetrical ligands. The benzil-BTPhens were the most promising group of ligands, showing the highest SFAm/Eu values which were seen to improve when electron donating groups were present an interesting point for future ligand design. The benzil-BTPhens were also tested for Am extraction from a more complex matrix containing lanthanides and fission products Overall, the ligands synthesised do not show a great improvement on the current benchmark ligands such as CyMe4BTPhen. However, I hope that the study highlights how additional functionality on BTPhens can be used to modify properties such as solubility and Am (III) extraction and that this work is useful in the design of future Am/Cm selective extractants for reprocessing methodologies and analytical applications.

Future work would involve testing our methods on more complex matrices and testing on more real matrices and creating new innovative separation methods for rapid extraction of An,

Ln and fission/activation products. Furthermore, I would like to address parts of the overall separation scheme for a complex nuclear forensic matrix employed by AWE and develop

160 alternative methods with selective ligands. For example, triazine ligands could be used to purify

Cd from a complex matrix or new extractants could be developed to improve Sr/Ba purification to avoid the use of fuming nitric acid and at other points that could potentially impede obtaining accurate results in a timely manner.

161 17.0 Methodology Section

17.1 Analysis Techniques Applicable to Americium Separation

17.11 Alpha Spectrometry

The two americium isotopes that can be commonly determined by alpha counting are 241Am and

243Am, gamma spectroscopy is also a widely applied technique for Am isotopes quantification.

This technique is can be used for the determination of 241Am and 243Am in the same solution.

The energy of the main alpha particles of 243Am (5.28 MeV) only differs by about 0.2 MeV from

241Am and, furthermore, several isotopes of plutonium and uranium can appear in this energy range. This means the use of alpha spectrometry requires a good radiochemical separation of an entire matrix (lanthanides also inhibit the electrodeposition), a thin layer electrodeposition of the radionuclides, long counting times (relative to liquid scintillation counting) and correction for absorption and backscattering from the plate.

17.12 Liquid Scintillation Counting

In this method the radioisotope mixture is mixed with a scintillation cocktail (usually a commercial solvent and solute mixture). Horrocks details this procedure and the benefits it offers over plate-counted analysis.70 The main benefit is the near 100% counting efficiency which allows for shorter counting times and therefore faster analysis. It also allows for continuous analysis for batch sample investigations. The scintillation cocktail chemically transfers the kinetic energy of the radiation to light energy; energy from beta, alpha and (to an extent) gamma emissions are transferred in the same way. The emitted light is received by a photomultiplier which records the energies as a spectrum. This method avoids the self-absorption which can occur with a solid sample, and less radiochemical separation is required. Liquid scintillation counting is the most applicable analysis technique for determining the presence of pure

162 americium in a solution, but in a mixture with other radionuclides, interferences and overlap of spectra due to the poorer energy resolution of liquid scintillation can cause difficulty in assignment and quantification. However a major disadvantage is the generation of mixed waste and its disposal.

17.13 Gamma Spectroscopy

Gamma spectroscopy is a non-destructive technique in which the energies of gamma photons emitted can be used to determine the radioisotopes present. Quantitative gamma analysis uses high purity germanium semiconducting detectors at cryogenic temperatures. When a gamma photon hits the semiconductor detector, an electron from the valence band is promoted to the conduction band leaving a hole. Both electron and hole migrate under an applied potential, generating a current pulse whose height is proportional to the energy dissipated in the sensitive volume of the detector. There are three modes of interaction of a gamma photon with matter

(photoelectric effect, Compton effect and pair production). The interactions of photons through the last two of these effects give rise to background features in the spectrum. The detector is coupled to a multi-channel analyser which categorises all signals in channels on the basis of pulse height, which is related to the detected energy of the photons. The resulting signal for each channel is plotted cumulatively. After background subtraction, the net count is proportional to the isotope activity while a built-in library is used to identify the radioisotopes from specific energies.

High sensitivity gamma spectroscopy is the most useful technique to quantify americium quickly in a separation. The low energy gamma rays 241Am emits (two main γ‐rays at 59.5 (36%) and 26.3 keV (2.4%)) can be quantified and the technique can identify multiple isotopes in a mixture, as well as discriminating Am from other gamma emitting isotopes. Using standard

163 counting geometry, suitable calibration sources, and sufficient count times to reduce error, a precise and accurate measurement of the amount of the radionuclide can be made. One of gamma spectroscopy’s main strengths is the ability to detect very low activity concentrations of radioactive materials (can achieve less than 1 Bq, but detection is relative to the detector calibration and count time), meaning it is applicable to environmental detection of radionuclides.71 However due to the low energies of the 241Am gammas radiochemical separations are required if there is a high background in the low energy gamma area.

17.14 UV-VIS-NIR Spectroscopy

UV–VIS–NIR absorption spectroscopy has been widely used to characterise americium solution species.72 The oxidation states of americium can be identified by their characteristic absorptions.

Furthermore UV-VIS-NIR has allowed analysis of the speciation of americium with many ligands, aiming to understand its chemistry; it is also a fast reliable technique, but requires a reasonable concentration of americium to be present.

17.15 NMR

NMR has found many uses in nuclear science; it has opened new analysis avenues particularly applicable to actinide co-ordination chemists. New variants of NMR e.g. DOSY, have allowed investigation of the co-ordination of actinide ions to ligand systems.73 It is also possible to determine stability constants and identify solution species of complexes which will not crystallise (hence solid state measurements are not possible). In terms of separations, the ability to determine the solution species and do small scale NMR titrations has provided understanding of some of the actinide species involved in radiochemical separations.

164 17.16 Mass Spectrometry

Mass spectrometry has many variants (e.g. for analysis of ions generated from surfaces) and can be used to determine the abundances of radioactive and stable isotope within a sample. The nuclides in a sample are ionised and the mass to charge ratio for each ion is measured, allowing identification.

Standard ICP-MS cannot be used when isotopes of the same nominal mass are present in the sample (e.g. 241Pu and 241Am). The technique works by introducing the sample into a plasma at around 6,000–10,000 K which is sustained within a high powered argon torch. The sample is nebulised, atomised and ionised. The positively charged ions from the sample are introduced into a very high vacuum between the torch and mass spectrometer. With the development of

Collision Cell technology the possibilities of measuring samples with interfering isobars is becoming possible.

Secondary ion mass spectrometry (SIMS) is a mass spectrometry technique that can be used with microgram sample amounts. It has high sensitivity and has a high depth resolution, when used for surface analysis. The sample can be presented as a thin film. This allows study of individual sections of an inhomogeneous sample. A primary ion beam is focused onto the particles to eject ions, which are collected in a high vacuum sample chamber. The secondary ions are collected, and passed to a mass analyser. An image of the density of the ions collected can be calculated, thus SIMS allows for single particle characterisation on the nanometre scale.74

ICP-AES/ICP-OES are alternative analytical techniques to ICP-MS but they require the elements for analysis to be present at higher levels (μg mL-1 to mg mL-1). Sample preparation is the same as for ICP-MS, and the sample is nebulised and atomised in the same manner.

165 Measurement of the atomic emission spectrum or optical emission spectrum of the analyte allows determination of the concentration.75

17.17 References

70 . D. L. Horrocks, Int. J. Appl. Radiat. Isot., 1966, 17, 441–446.

71. A. S. Murray, R. Marten, A. Johnston, P. Martin, J. Radioanal. Nucl. Chem, 2005, 115, 263–

288.

72 . C. Walther and M. A. Denecke, Chem. Rev., 2013, 113, 944–994.

73. C. Berthon, P. Delangle, P. Moisy, M. C. Charbonnel, M. Heitzmann, S. I. Nikitenko, E.

Bosse and M. S. Grigoriev, IOP Conf. Ser.: Mater. Sci. Eng,. 2010, 9, 012058.

74 . C. W. Magee, W. L. Harrington and R. E. Honig, Rev. Scient. Instrum, 1978, 49, 477–485.

75 . K. L. Linge, Geostand. Geoanal. Res., 2007, 29, 7–22.

166

167