Radionuclide Fate in Naturally Occurring Radioactive Materials (NORM) in the Oil and Gas Industry

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

2019

Faraaz Ahmad

School of Earth and Environmental Sciences

Table of Contents List of Figures…………………………….……………………………………………………………………………………….7 List of Tables…..……………………………..…………………………………………………………………………………15

List of Abbreviations ...... 17 Thesis Abstract ...... 20 Declaration ...... 22 Copyright statement ...... 22 Acknowledgements ...... 23 About The Author ...... 25

CHAPTER 1: Introduction…………………………………………..……………………….26

1.0 Project Introduction ...... 26 1.1 Aims and objectives ...... 31 1.1.1 Outline ...... 32 1.2 Thesis structure ...... 34 1.3 Paper status and author contributions ...... 36

CHAPTER 2: Literature Review…………………………………………..………………………………..38

2.0 Literature Review ...... 38 2.1 Natural uranium - 238 and thorium - 232 decay Series ...... 38 2.2 Produced Water – Liquid waste stream ...... 42 2.2.1 Characteristics ...... 42 2.2.2 Chemical incompatibility of produced waters ...... 48 2.2.3 Mineral solubility……………………………………………………………………………………… 49

2.2.4 Degree of saturation………………………………………………………………………………… 51

2.3 Reservoir processes and scale formation – Solid waste stream ...... 52 2.3.1 Pipe scale formation ...... 52 2.3.2 Reservoir processes and sulphate scales ...... 53 2.3.3 Radionuclide uptake into sulphate scales ...... 56 2.3.4 Calcium carbonate scales ...... 59 2.3.5 Radionuclide uptake into calcium-containing scales ...... 60 2.3.6 Black dust deposits...... 61 2.3.7 Radionuclide uptake into black dust deposits ...... 62 2.3.8 Sludges ...... 64 2.3.9 Radionuclide uptake into sludges ...... 64

2.3.9.1 Scale Prevention ...... 66 2.4 Operational discharge of effluent waters to surface waters and NORM formation ...... 67 2.4.1 Offshore discharges ...... 67 2.4.2 NORM formation during operational discharge ...... 68 2.4.3 Environmental characteristics ...... 70 2.5 Radioactivity of NORM ...... 73 2.5.1 Solid waste stream – Pipe scale and Sludge ...... 73 2.5.2 Liquid waste stream – Produced water ...... 77 2.6 Influence of natural processes and microorganisms on barium and radium remobilisation .. …………………………………………………………………………………………………………………78 2.6.1 Sulphate-reducing bacteria and their effect on barite ...... 79 2.7 Environmental Implications and Exposure Assessment ...... 82

CHAPTER 3: Methodology…………………………………..…………………………………………………89

3.0 Research Methods ...... 89 3.0.1 Safety ...... 89 3.1. Sample collection ...... 90 3.1.1 Marine Sediment...... 90

3.2 Precipitation Experiment (BaxSrySO4 - Inactive) ...... 91 3.2.1 Seawater and Produced Water Synthesis ...... 91 3.2.2 Precipitation Procedure ...... 93 3.3 Radium uptake experiment ...... 94

3.4 Precipitation Experiment (BaxSryRazSO4 – Active) ...... 95 3.4.1 Adjusted Brine Composition ...... 96 3.4.2 Precipitation Procedure ...... 96 3.5 Sediment Microcosm Experiment ...... 98 3.6 NORM Sample Characterisation ...... 102 3.7 Solution and Geochemical analysis ...... 103 3.7.1 Inductively-coupled Plasma Atomic Emission Spectroscopy (ICP-AES) ...... 103 3.7.2 Ion Chromatography ...... 106 3.7.3 pH and Eh ...... 108 3.7.4 UV-Vis Spectrophotometry ...... 108 3.7.5 Liquid Scintillation Counting (LSC) ...... 113 3.7.6 PHREEQC Modelling ...... 116

3.8 Chemical Extractions ...... 117 3.8.1 Heavy Liquid Extraction...... 117 3.8.2 Sequential Extraction ...... 118 3.8.3 Barite Dissolution ...... 120 3.9 Solid Phase Characterisation ...... 120 3.9.1 Environmental Scanning Electron Microscopy (ESEM) – Energy Dispersive X- Ray Analysis (EDX) ...... 120 3.9.2 Powder X-Ray Diffraction (XRD) ...... 123 3.9.3 Fourier Transform Infra-red Spectroscopy (FTIR) ...... 126 3.9.4 Raman Spectroscopy ...... 129 3.9.5 BET Surface Area Analysis ...... 130 3.9.6 X-Ray Absorption Spectroscopy (XAS) ...... 131 3.9.7 X-Ray Fluorescence (XRF) ...... 134 3.9.8 Gamma Spectroscopy ...... 136 3.9.9 Autoradiography ...... 139 4.0 Microbial Community Analysis – DNA Sequencing ...... 142 CHAPTER 4: Fate of Radium on Discharge of Oil Produced Water to the Marine Environment…………………………………..…………………………………………………………145

4.1 Abstract………………… ……………………………………………..…………………………………………………..146

4.2 Introduction…………………………………..………………………………………………………………………….147

4.3 Materials and Methods………………………………………………………………..……………………….…..152

4.3.1 The study area and experimental method…………………………………….…………….152

4.3.1.1 Marine Sediment……………………………………….…………………………………………..152

4.3.1.2 Produced water and seawater mixing experiments ...... 152 4.3.1.3 Radium uptake experiment………………………………..…………………………………..153

4.3.2 Analytical methods………………………………………..…………………………………………...153

4.3.2.1 Solid phase characterisation………………………….………………………………………..153

4.3.2.2 Radiometric analysis ...... 154 4.3.2.3 Heavy Liquid Extraction ...... 155 4.3.2.4 Autoradiography ...... 155 4.4 Results and Discussion……………………………….……………………………………………………………..155

4.4.1 Formation of NORM precipitate upon discharge of Produced water to Seawater…...... 155

4.4.1.1 Characteristics of marine sediment samples ...... 155 4.4.2 Strontiobarite formation during field and synthetic seawater and produced water mixing: morphology and composition ...... 160 4.4.2.1 Field Mixing Experiment ...... 160 4.4.2.2 Synthetic Mixing Experiment ...... 161

4.4.3 Ra uptake during strontiobarite (Ba-Sr-SO4) formation ...... 164 4.4.3.1 Synthetic seawater and produced water mixing: Radium uptake experiments ...... 164 4.5 Conclusion and Environmental Implications………………………………………………..……….……166

Acknowledgements ...... 168 Supplementary Information ...... 169 CHAPTER 5: The Effects of Bioreduction on the Fate of Ra2+(aq) and Radiobarite in marine Sediments on Discharge of Oil Produced Water……………………………………………………………………………………………………………………….... 176

5.0 Abstract ...... 177 5.1 Introduction ...... 178 5.2 Experimental Methods ...... 184

5.2.1 Radiobarite (RaBaSrSO4) and barite (BaSrSO4) formation ...... 184 5.2.1.1 Precipitation procedure and radium uptake ...... 184 5.2.2 Chemical composition and mineralogical analysis ...... 185 5.2.2.1 Dissolution of barite particles ...... 185 5.2.2.2 Scanning electron microscopy (SEM) ...... 185 5.2.3 Sediment microcosm Experiments ...... 185 5.2.3.1 Sampling ...... 185 5.2.3.2 Microcosm experiments ...... 186 5.2.3.3 Geochemical analysis ...... 187 5.2.3.4 Heavy liquid extraction ...... 187 5.2.3.5 Sequential chemical extraction ...... 187 5.2.3.6 Etch pit formation via chelating ligand ...... 188 5.2.3.7 Microbial community analysis ...... 188 5.3 Results & Discussion...... 189 5.3.1 Marine sediment and radiostrontiobarite ...... 189

5.3.2 Radiobarite (RaSrBaSO4) formation ...... 189 5.3.3 Biogeochemical reduction of sediments ...... 191

5.3.3.1 Biogeochemistry in sediment microcosms ...... 191 5.3.4 Radium speciation and fate ...... 194 5.3.5 Fate of Ra: Etch pit formation analysis ...... 197 5.3.6 Microbial community characterisation ...... 200 5.4 Environmental Implications ...... 202 Acknowledgements ...... 203 Supplementary Information ...... 204 CHAPTER 6: Chemical and Radiological Characterisation of Scales containing NORM……………………………………….……………………………………………………..……208

6.0 Abstract ...... 208 6.1 Introduction………………………………………….………….………………………………………………..209

6.2 Experimental Method…………………………………………………..…………………………………….213

6.2.1 NORM Samples & Characterisation ...... 213 6.2.1.1 Sampling ...... 213 6.2.1.2 Preparation ...... 213 6.2.1.3 Chemical composition and mineralogical analysis ...... 213 6.2.1.4 Radiological composition & distribution ...... 214 6.3 Results and Discussion ...... 215 6.3.1 Radioactivity content of scales ...... 215 6.3.1.1 North Sea Samples ...... 215 6.3.1.2 Iraq Samples ...... 216 6.3.2 Spatial distribution relationship of radionuclides with chemical and mineralogical composition of scales ...... 218 6.3.2.1 North Sea samples ...... 218 6.3.2.2 Iraq Samples ...... 229 6.4 Conclusion and Environmental Implications ...... 238 Supplementary Information ...... 242

CHAPTER 7: Conclusions and Future Work Directions………………….………...249

7.0 Conclusions…………………………………………………………….………………………………………….249

7.1 Future Work……………………………………………………………………………………………………….257 List of References…………………………………………………………………………………………………………260

Appendix 1: Conferences and Courses………………………………..…………………………………….287

List of Figures Chapter 1: Introduction

Figure 1.1: A diagram illustrating the mobilisation of certain progeny by the leaching of primordial radionuclides 238U and 232Th. Contamination via migration resulting in the precipitation or NORM in the gas, oil and water streams………………………………………………… 28

Figure 1.2: : Discharge scenario illustrating the complexity of this system and the potential scavenging mechanisms and pathways controlling the fate, speciation and mobility of radium: 1) Discharge of produced water; 2) Dispersion of Ra2+ (aq); 3) Adsorption of Ra2+ to sediment; 4) Precipitation of radiobarite; 5) Dispersion of radiobarite; 6) Deposition of particles; 7) Indigenous microbial population and biogeochemical processes in the sediment and; 8) Radium released from sediment surface or radiobarite via microbial reduction…………………………………………………………………………………………………………….…………… 29

Chapter 2: Literature Review

Figure 2.1: The full decay series of uranium (238U) and thorium (232Th), adopted radiative decay and half-life…………………………………………………………………………………………………………… 41

Figure 2.2: The decay series of uranium and thorium with the different modes of mobilisation of certain daughter radionuclides……………………………………………………….………. 42

Figure 2.3: Oil and gas accumulation and trapping schematic…………………………….…….……….44

Figure 2.4: Scale formation during enhanced oil recovery operations……………………….….…..55

Figure 2.5: A schematic diagram to show the double spiral flow in a bend……….……………….56

Figure 2.6: A diagram to illustrate the modes of uptake of radium and the three stages involved in the precipitation process……………………………………………………………………………….. 59

Figure 2.7: Equation to show the formation of calcium carbonate…………………..………………. 60

Figure 2.8: A) Illustration of the ion-exchange between iron and lead resulting in the formation of thin films consisting of metallic 210Pb or galena (PbS). Electrochemical process by which Pb2+ reduced to metallic Pb and Fe oxidised to Fe2+; B) electrochemical reactions resulting in corrosion of pipe wall via bacteria and scratched surfaces (Corg represents carbon source e.g. CH2O or CH4)………………………………………………………………………………………. 65

Figure 2.9: A diagram to show areas in which NORM can build up in the offshore production system indicated by the radioactive sign; (key: orange = oil, blue = water and purple = gas)………………………………………………………………………………………………………………………………….. 67

Figure 2.10: Hypothetical pore-water depth profile produced by successive TEAP’s during decomposition and table showing depth-related Gibbs free-energy calculations adapted from Burke et al., 2005………………………………..………………………………………………………………………………………………. 79

Figure 2.11: Illustration of the potential reduction of sulphate in barite via sulphate reducing bacteria (SRB) in sediment (Corg represents carbon source e.g. CH2O or CH4)..….. 80

Figure 2.12: Routes of exposure from NORM (internal and external sources)……………………84

Figure 2.13: Options for NORM disposal………………………………………………………..………………… 85 Chapter 3: Methodology

Figure 3.1: Schematic of the field sediment samples obtained (+) from various distances to the discharge point (•)…………………………………………………….………………………………………………. 90

Figure 3.2: (A-D); (A) Image of precipitate formation in beaker (B) BSE image of precipitate extracted from beaker; (C) EDS spectra confirming the formation of calcite with co- precipitation of Sr and Ba and; (D) XRD pattern confirming the presence of calcite.………… 93

Figure 3.3: (A-B): A) Overview of the experimental set up; (B) The undesired crystal morphology observed via BSE imaging when instantly mixing the two solutions as opposed to pumping over 6 hrs (undesired dendritic morphology = indicative of a fast growth rate)………………………………………………………………………………………………………………………………….98

Figure 3.4: Overview of the experimental setup of anaerobic sediment microcosms across all sets of microcosm experiments containing; strontiobarite, radiostrontiobarite and aqueous radium. Bottles A-C; amended with 5 mM acetate and lactate (bottle A-B amended with a further 10 mM after Day 140), D; unamended as shown by no distinct darkening of sediment, E; amended without the addition of precipitate or 226Ra (further amended after Day 140), F; autoclaved sterile control and, G; seawater containing precipitate or 226Ra only control………………………………………………………………….…………………. 101

Figure 3.5: Design of the ICP torch and temperature distribution in plasma……….…………. 106

Figure 3.6: Schematic of the energy level diagram and characteristic emission and absorption spectra for sodium (Na)……………………………………………………………………………….. 106

Figure 3.7: Schematic of the setup and components of an ion chromatograph system used for analysis…………………………………………………………….…………………………………………………….… 107

Figure 3.8: A typical Ion chromatogram produced using a Dionex instrument, showing the elution and resultant separation of anion species…………………………………………………………. 108

Figure 3.9: Beer Lambert Law and variables required to determine absorption………….…. 110

Figure 3.10: The molecular structure of the ferrozine molecule with the bidentate ligand properties of the ferroin group highlighted……………………………………………………………….……111

Figure 3.11: Reaction of dimethyl-p-phenylenediamine with iron chloride (oxidant) and hydrogen sulphide to form methylene blue under acidic conditions………………………..……..112

Figure 3.12: Illustration of the main component of a scintillator, the photomultiplier tube (PMT and the process of generating a series of amplified electrical pulses whose individual amplitudes are proportional to the quantum energies of the photons that generated them………………………………………………………………………………………………………………………….….. 114

Figure 3.13: Schematic of a LSC instrument system……………………………………………..………… 115

Figure 3.14: Example of the data feed to the PHREEQC computer geochemical modelling software to determine the saturation and speciation of barium, strontium and calcium………………………………………………………………………………………………………………..………… 116

Figure 3.15: Overview of the experimental setup of the heavy liquid density separation experiment...... 118

Figure 3.16: Schematic of the electron beam interaction volume and various signals generated………………………………………………………………………………………………………………………. 122

Figure 3.17: Schematic diagram of the various signals generated by interaction of the electron beam with the sample……………………………………………………………………………..………………………………………….… 123

Figure 3.18: X-ray diffraction by an array of atoms within a crystal lattice as described by Bragg’s Law. Arrows represent incident and diffracted x-rays, the lattice spacing (d), diffraction angle (θ) and lattice planes illustrated as horizontal lines……………………..……… 125

Figure 3.19: Image of the Perkin Elmer Spotlight 400 ATR-FTIR system. The infrared beam is generated in the Universal ATR unit from a HeNe laser and directed into the Spotlight 400 unit which contains the detector and microscope for high-precision point analyses……… 128

Figure 3.20: Schematic of a multiple reflection ATR system……………………………..……………. 128

Figure 3.21: Diagram showing resonance (elastic scattering known as Rayleigh), stokes lines (inelastic) and anti-stokes lines (inelastic)……………………………………… …………………………….. 130

Figure 3.22: (A) Image of the Horiba Xplora Raman microscope at the University of Manchester (left); B) Undesired burning of sample due to its chemical and mineralogical properties (right)……………………………………………………………………………………………………..……. 130

Figure 3.23: XAS spectrum indicating the various features and regions observed in a spectrum and resultant scattering effects of emitted photoelectrons with neighbouring scattering atoms (multiple and single events)………………………………………………………………… 133

Figure 3.24: Image of the B18 beamline set up at the Diamond Light Source, Harwell, UK (left) and; B) a schematic of a typical beam line set up for the collection of XAS data in fluorescence and transmission modes…………………………………………………………………………... 134

Figure 3.25: The band structure of the semiconductor high purity crystal, electric field (←)………………………………………………………………………………………………………………………………… 137

Figure 3.26: Schematic of the set up and process of using a storage phosphor screen (BAS- IP) and its composite structure…………………………………………………………….………………………… 141

Figure 3.27: A workflow diagram illustrating the 16s rRNA sequencing protocol….……….. 142

Chapter 4: Fate of Radium on Discharge of Oil Produced Water to the Marine Environment

Figure 4.1: Schematic of the field sediment samples obtained (+) from the offshore marine system surrounding the discharge outfall (•)…………………………………………………………………. 152

Figure 4.2: (A-B), BSE images and elemental maps of the radiostrontiobarite particles extracted from marine sediment; (C-F), corresponding elemental maps (Ba, S, Si and Sr respectively); (G), EDS spectra representative of the strontiobarite grains (bright regions) via point analysis (G); (H), EDS spectra representative of the clay rich areas (dark regions) via point analysis (H); (I), stub sample containing radiostrontiobarite particles and; J) the corresponding autoradiograph………………………………………………………………………………..……. 158

Figure 4.3: (A) BSE image showing frambodial pyrtie cyrtsals found in field sediment samples contaminatied with radiostrontiobarite and; (B) EDS spectra collected confirming the identificaation of pyrite………..………………………..……………………………………………………….. 159

Figure 4.4: (A) BSE image showing strontiobarite crystals exhibiting tabular morphology with a particle size ranging from 2-6 µm from field mixing experiments and; (B) EDS spectra representative of all grains…………………………………….……………………………………………………… 160

Figure 4.5: (A) BSE image showing the strontiobarite crystals exhibiting both tabular and rosette morphology with a particle size between 1-5 µm and; (B) EDS spectra representative of all grains………………………………………………………….………………………………… 162

Figure 4.6: A) Percentage uptake of Ba (□) and Sr (Δ) over time (1.5 mL small scale experiments); B) 226Ra uptake over time (□) (1.5ml small scale experiment; initial activity 13bq mL-1) and; C) sulphate concentration over time……………………………………….……………. 165 Supporting information

Figure S4.1: (A) image of stub containing radiostrontiobarite grain; (B) backscattered electron (BSE) image showing the morphology of the grain; (C-E) corresponding elemental maps and; (F) autoradiograph displaying activity…………………………….………………..………….. 172

Figure S4.2: (A) image of stub containing radiostrontiobarite grains and area analysed highlighted (yellow circle); (B) backscattered electron (BSE) image showing the morphology of the grain; (C-E) corresponding elemental maps and; (F) autoradiograph displaying activity…………………………………………………………………………………………………………………..………. 173

Figure S4.3: XRD of precipitate obtained from mixing of synthetic produced water and seawater representative of Ba75Sr25 (silicon (Si) incorporated for reference purposes).…. 174

Figure S4.4: FTIR spectra of precipitate obtained from mixing of synthetic produced water and seawater…………………………………………………………………………………………………………………. 174

Figure S4.5: Sr K-edge EXAFS spectra and fit table (R-factor: 0.0088). Reference (Ref.) sample corresponds to pure barite…………………………………………………………….………………..… 175

Chapter 5: The Effects of Bioreduction on the Fate of Ra2+(aq) and Radiobarite in marine Sediments on Discharge of Oil Produced Water

Figure 5.1: A) BSE image showing strontiobarite crystals exhibiting both tabular and rosette morphology with a particle size between 1-10µm and; (B) EDS spectra collected via SEM confirming the strontiobarite phase………………………………………………………………………………. 190

Figure 5.2: Results from all sets of microbial-mediated reduction experiments. Changes in pore water concentrations of Ba, Fe(II), SO42- and S2-. Error bars represent average of three replicates, error bars ±1 SD……………………………………………………………………………………………. 193

Figure 5.3: Radium concentrations in sequential extraction leachates from sediment microcosms containing Ra2+(aq) and radiobarite, following 1 and 300 days of anaerobic incubation……………………………………………………………………………………………………………………… 194

Figure 5.4: Radium concentrations in pore waters and percentage radium released during anaerobic incubation in microcosm experiments containing radiobarite and Ra2+(aq). Error bars represent average of three replicates, error bars ±1 SD………………………………..……….. 196

Figure 5.5: BSE images of the surface of barite grains after 300 days of anaerobic incubation to determine microbial-mediated etch pit formation; A) barite precipitate not exposed to anoxia (control); B-C) after anoxia (Day 300); D) exposure to seawater to account for ion- exchange/ionic-strength effect and; E) control showing the formation of elongated etch pits after exposure to an EDTA solution (5mins)…….…….…………………………………………………….… 199

Figure 5.6: Prokaryotic phylogenetic diversity of sediments from the field (Day 0) and microcosms (Day 300). Phyla/classes are illustrated if present > 0.5% of the microbial community………………………………………………………………………………………………………..…………… 200 Supporting information

Figure S5.1: XRD pattern of the sediment used in all sediment microcosm experiments.. 204

Figure S5.2: FTIR of the precipitate obtained from mixing experiments……………..…………..204

Figure S5.3: XRD pattern confirming the bulk mineralogical composition of the precipitate obtained from mixing experiments to be Ba75Sr25………………………………………………………….. 205

Figure S5.4: Barium concentrations in sequential extraction leachates from sediment microcosms containing Ra2+(aq), radiobarite, barite following 1 and 300 days of anaerobic incubation, including field sediment samples………………………………………………………………… 205

Figure S5.5: Calcium concentrations in sequential extraction leachates from sediment microcosms containing Ra2+(aq), radiobarite, barite following 1 and 300 days of anaerobic incubation, including field sediment samples………………………………………………………………… 206

Figure S5.6: BSE image of the etch pits and sizes formed on the surface of barite grains via EDTA dissolution……………………………………………………………………………………………………………. 206

Figure S5.7: Relative species diveristy of samples mesuaed by number of observed species identified vesus the total number of sequences analysed……………………………………………... 207

Chapter 6: Chemical and Radiological Characterisation of Scales containing NORM

Figure 6.1: Autoradiography map to show the areas of activity within the samples to help correlate activity distribution with mineralogy; 7470 and 7457; 2 day exposure, 7536; 17 day exposure and IRQ 1 - 4; 28 day exposure time. Dashed (white) lines (- -) dictate edge of the samples…………………………………………………………………………………………………………………… 217

Figure 6.2: XRD results obtained for North Sea scales (Brt = barite; Cal = calcite; Clt = celestite); 7470: BaSr66SO4, 7536: BaSr40SO4, 7457: CaCO3 and 7457 (corrosion material): Mag = magnetite; Lp = lepidocrocite; Gt = goethite and; Ak = akaganeite……………………… 219

Figure 6.3: (A-D); A) Backscattered electron image showing the different growth events, branching and zonation; (B and C) representative backscattered electron images of zones 2 & 4 and 1 & 3 respectively and; D) EDS spectra representative of the morphologically different zones encountered (zone 1 – 4)…………………...... ……………………………………………… 223

Figure 6.4: (A-D); A- B) Backscattered electron image showing the acicular needle-like crystal growth of the scale; (C) zoomed in image to visualise heterogeneity in morphology and; (D) corresponding EDX spectra………………………………..……………….……………………………. 225

Figure 6.5: (A-E); A- B) Backscattered electron image showing the flat smooth crystal growth and indentation created during the removal of the sample; (C) BSE image showing the two distinct areas and mineralogical phases found at the edge of the sample and; (D-E) EDX spectra of the bulk of the sample and edge respectively………………….…….………………. 228

Figure 6.6: XRD results obtained for scales (IRQ 1 - 4); Anh = anhydrite; Bsn = basanite; Gp = gypsum; Hl = halite; Hbl = hornblende and; Qz = quartz………………………………………………… 230

Figure 6.7: A) BSE image showing the heterogeneity within the sample: quartz (Point 1), anhydrite (Point 2), magnetite and pyrrhotite (Point 3) and aluminium silicate (point 4) phases in scale IRQ1; B) BSE image of celestite (Point 5); C) BSE image of a grain composed of hornblende, zinc sulphide or pyrrhotite (Point 6) and; D) Energies of the characteristic backscattered electrons detected corresponding to the elements in the scale………….….. 232

Figure 6.8: BSE image showing morphological difference within the sample (large crystals surrounded by a slurry) and; B) EDX spectra corresponding to the elements in the scale; bulk mineralogical composition of large crystal structures to be CaSO4 (Point 2 and 3) with incorporation of trace elements in surrounding slurry (Point 1) and also celestite phases being identified (Point 4) and darker regions to be resin/background (Point 5)…………….. 234

Figure 6.9: A) BSE image highlighting the heterogeneity within the sample and; B) Energies of the characteristic backscattered electrons detected corresponding to the elements in the scale confirming the bulk mineralogical composition to be CaSO4 (point 1-5) and darker regions to be resin (point 6)…………………………………………………………….…………………………….. 236

Figure 6.10: A) BSE image of IRQ 4 confirming the bulk mineralogical composition to be

CaSO4 (Point 2) and brighter grey regions containing strontium (Point 1) and; B) Energies of

12 the characteristic backscattered electrons detected corresponding to the elements in the scale……………………………………………………………………………………………………………………..……….. 238 Supporting information

Figure S6.1: Elemental map of sample 7457; A) BSE image displaying morphological differences and zonation within the scale; B) Barium elemental map; C) Sulphur elemental map and; D) strontium elemental map……………………………………………………………………..…… 242

Figure S6.2: A) the area mapped via IR spectroscopy; size: 1950 µm x 5800 µm; B) The map produced at the selected wavenumber of 1224 cm-1 showing the mineralogical distribution of strontiobarite across the mapped area and variances in crystallinity and composition and; C) Point analysis (representative spectra) of the mapped region………………..…………. 243

Figure S6.3: Zoomed in autoradiography maps to show the correlation between porous areas of the sample and activity; A) North Sea 7470 and; B) North Sea 7457…………….…… 243

Figure S6.4: Elemental map of sample 7470; A) BSE image displaying the adopted acicular needle-like morphology within the scale; B) strontium elemental map; C) barium elemental map and D) sulphur elemental map……………………………………………………….……………………… 244

Figure S6.5: A) the area mapped via IR spectroscopy; size: 1492 µm x 821 µm; B) The map produced at the selected wavenumber of 1106 cm-1 showing associated variances in crystallinity and composition and; C) Point analysis (representative spectra) of the mapped region…………………………………………………………………………………………………………………..……….. 244

Figure S6.6: Elemental map of sample 7536; A) BSE image displaying two distinct areas and mineralogical phases found at the edge of the sample; B) zinc elemental map; C) calcium elemental map and; D) lead elemental map…………………………………………………….……………. 245

Figure S6.7: A) the area mapped via IR spectroscopy; size: 300 µm x 260 µm; B) The map produced at the selected wavenumber of 1384 cm1 showing the mineralogical distribution of calcite across the mapped area; C) The map produced at wavenumber 1660 cm-1; D) Point analysis (representative spectra) of the calcite mapped region (red area) and; E) Point analysis (representative spectra) of the galena and sphalerite region……………….…………… 245

Figure S6.8: A) the area mapped via Raman spectroscopy; size: 17 µm x 15 µm; B) The map produced at the selected wavenumber of 1000 cm1 showing the mineralogical distribution of calcite across the mapped area; C) Point analysis (representative spectra) of the calcite mapped region (red area); D) The map produced at wavenumber 200 cm-1; E) Point analysis (representative spectra) of the galena region; F) The map produced at wavenumber 600 cm-1 and; G) Point analysis (representative spectra) of the resin………….………………………… 246

Figure S6.9: IRQ 1; A) the area mapped via IR spectroscopy; size: 500 µm x 1000 µm; B) The map produced at the selected wavenumber of 1160 cm1 showing the mineralogical distribution of anhydrite across the mapped area; C) The map produced at wavenumber 780 cm-1 showing the distribution of quartz; D) Point analysis (representative spectra) of the mapped region (red area); E) Point analysis (representative spectra) of the quartz region………………………………………………………………………………………………………………….………… 246

Figure S6.10: IRQ 2; A) the area mapped (red) via IR spectroscopy; size: 950 µm x 1800 µm; B) The map produced at the selected wavenumber of 1128 cm-1 showing the mineralogical distribution of anhydrite across the mapped area and; C) Point analysis (representative spectra) of the mapped region………………………………….…………………………………………………… 247

Figure S6.11: IRQ 3; A) the area mapped via IR spectroscopy; size: 950 µm x 1900 µm,; B) The map produced at the selected wavenumber of 1152 cm1 showing the mineralogical distribution of anhydrite across the mapped area; C) The map produced at wavenumber 1304 cm-1 confirming the infill of resin; D) Point analysis (representative spectra) of the mapped region; and; E) Point analysis of the resin region……………………………………………… 247

Figure S6.12: IRQ 4; A) the area mapped via IR spectroscopy; size: 900 µm x 1442 µm; B) The map produced at the selected wavenumber of 1164 cm1 showing the mineralogical distribution of anhydrite across the mapped area and; C) Point analysis (representative spectra) of the mapped region……………………………………………….……………………………………… 248

List of Tables

Chapter 1: Introduction

Table 1.1: Total alpha and beta discharges into the North Sea from 2005-2016 from the oil industry. Total alpha and beta emissions from the nuclear sector shown for comparison……………………………………………………………………………………………………………………… 30

Chapter 2: Literature Review

Table 2.1: Concentrations of primordial radionuclides in igneous, sedimentary rocks, soils (Bq kg-1), and sea water (Bq L-1), where thorium concentrations (-) are extremely low; ≥ 0.16 ng kg-1/ 4 x 10-3 dpm kg-1…………………………………………………………………….……………………………. 39

Table 2.2: Typical compositions of seawater and produced waters from the North Sea and Gulf of Arabia………………………………………………………………………………………………………………….. 46

Table 2.3: The various scales that can precipitate in oilfields including those due to the addition of chemical additives and corrosion…………………………………………………………………… 47

Table 2.4: Common mineral scales which dominantly form in the oilfield and their individual chemical properties to exemplify the difficulty of the removal of scales especially barium sulphate due to its insolubility and hardness. Sand is included for reference purposes. Scale of hardness is in accordance to HSAB theory and ranges from 1 (soft) to 10 (hard). Solubility product values are negative logarithms of activity products at 25 °C and 1 bar (pK) e.g. -10 solubility product for BaSO4 = 10 ………………………………….………………………………………………. 57

Table 2.5: The different types of NORM in oilfields and the component radionuclides within the materials and their occurrence during production……………………………………….……………. 62

Table 2.6: Radioactivity concentrations of radium isotopes within NORM taken from oilfields across the globe (n.r. = data not recorded)…………………………….………………………….. 74

Table 2.7: A comparison of the radioactivity concentrations for crude oil, NG and NGL.…. 75

Table 2.8: A comparison of the radioactivity concentrations for a range of scales (hard sulphate and carbonate scales), deposits (softer sulphate and carbonate scales) and scrapings…………………………………………………………………………………………………………………………. 76

Table 2.9: A comparison of the radioactivity concentrations for produced water……………. 78

Chapter 3: Methodology

Table 3.1: The seawater and formation water composition adopted…………………….………… 92

Table 3.2: Yield obtained from multiple mixing experiments to form strontiobarite…….…. 94

Table 3.3: The solution composition adopted to produce radium labelled strontiobarite in 40 mL………………………………………………………………………………………………………………………………. 96

Table 3.4: A summary of the sequential extraction steps adopted…………….………………….. 120 Chapter 4: Fate of Radium on Discharge of Oil Produced Water to the Marine Environment

Table 4.1: Mineralogical, radiological and chemical characteristics of marine sediment samples taken from the different locations……………………………………………………………………. 156

Table 4.2: Field and synthetic seawater and formation water compositions (Todd et al., 1992). Ionic strength and saturation index (SI) of mixtures calculated using PHREEQC modelling software with a 9:1 mixing ratio (sit.database)…………………………………………..…. 161 Supporting information

Table S4.1: XRF analysis showing major elements present in all field sediment samples, (-); not detected. Element concentrations have been deduced from the concentration percentage of compounds, e.g. Si from SiO2 and Al from Al2O3……………………………………… 169

Table S4.2: XRF analysis showing the trace elements present in sediment all field sediment samples……………………………………………………………………………………….………………………………… 171

Table S4.3: Calculated molar stoichiometry of the desired precipitate from solution and solid data analysis. Surface area analysis data via BET; (-): limited sample for analysis.…. 174

Chapter 5: The Effects of Bioreduction on the Fate of Ra2+(aq) and Radiobarite in marine Sediments on Discharge of Oil Produced Water

Table 5.1: Calculated molar stoichiometry of the desired precipitate from solution data and solid data analysis, surface area analysis via BET and precentage radium uptake in mixing experiments: initial activity of 190 Bq mL-1………………………………………………………………..…… 190

Chapter 6: Chemical and Radiological Characterisation of Scales containing NORM

Table 6.1: Gamma spectrometry analysis based on the measurement of individual radionuclides for both sets of samples (North Sea; 7457, 7470, and 7536 and Iraq: IRQ 1-4); a – externally provided results and b – internally (independently) analysed (n.d = not detected)…………………………………………………………………………………………………………………..…… 218

Table 6.2: Bulk composition determined via XRF of North Sea samples…………….…………… 220

List of Abbreviations

ATR-FTIR Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy

BET Brunauer-Emmett-Teller method of specific surface area analysis

BOEM Bureau of Ocean Energy Management

Bq Becquerel

CDT Centre for Doctoral Training

COSHH Control of Substances Hazardous to Health

DEFRA Department for Environment, Food and Rural Affairs

DIW Deionised water dpm Disintegration per minute

E Energy

EA Environmental Agency

Eb Binding energy

EDS Energy Dispersive Spectroscopy

EDTA Ethylenediaminetetraacetic Acid

EDX Energy dispersive X-ray

EOR Enhanced oil recovery

EU European Union eV Electron volts

EXAFS Extended X-ray absorption fine structure

FEDRC Federal Regulatory Commission

FT-IR Fourier Transform Infra-Red

IAP Ion activation product

ICP-AES Inductively coupled plasma atomic emission spectroscopy

ICRP International Commission on Radiation Protection

IR Infrared kBq Kilo Becquerel keV Kilo electron volts

KOH Potassium Hydroxide

Ksp Solubility product constant

MECA Ministry of the Environment and Climate Affairs

MEIM Ministry of Energy, Industry and Mineral Resources

MOG Ministry of Oil and Gas mSv milliSievert

NaOH Sodium Hydroxide n.d Not detected

NERC Natural Environment Research Council

NGL Natural gas liquids

NORM Naturally occurring radioactive material

NOR Naturally occurring radionuclides n.r Not recorded

OGA Oil and Gas Authority

OSPAR Oslo and Paris Commission

PES Polyethersulfone

PMT Photomultiplier tube

18 ppb Parts per billion ppm Parts per million

SEM Scanning Electron Microscopy

SI Saturation index

SIT Specific ion Interaction Theory

TBq Terabecquerel

TE-NORM Technologically enhanced naturally occurring radioactive material

UV-vis Ultraviolet-visible spectroscopy

XAS X-ray absorption spectroscopy

XRD X-ray diffraction

Thesis Abstract

The formation of naturally occurring radioactive materials (NORM) during oil extraction and operational marine discharges is a global ongoing problem encountered in the petroleum industry. The presence of 226Ra in effluent waters (e.g. produced water) and the mixing of incompatible waters (e.g. seawater and produced water) encountered at different stages of production drives the co- precipitation of radium into insoluble sulphate mineral phases (e.g. RaBaSO4 and

RaBaSrSO4). These materials can accumulate in large volumes across an array of components in the production system (e.g. pipe lines, valves, storage tanks) but, may also form and deposit in the environment as a result of operational effluent marine discharges. The build-up and accumulation of naturally occurring radionuclides within solid and liquid waste materials during the extraction of oil (e.g. minerals, effluent waters and sediments) has been a recognised risk in the oil and gas industry during the past century. However, there is uncertainty about the speciation and fate of radium during offshore produced water discharges, and the nature of the material formed as a result of the mixing of these waters which is poorly constrained. An understanding of the fate and uptake of key radionuclides into NORM waste streams e.g. mineral scales formed in tubulars, and from operational marine discharge of produced water as a result of mixing of incompatible waters is key in order to develop the capability to predict their fate and environmental risk.

In this study, the uptake and fate of radium in marine sediment samples obtained from a field site where produced waters are discharged were explored. A variety of techniques such as gamma spectroscopy was used to assess the radium activity concentrations in marine field samples. Radium was present in selected field samples at concentrations between 0.1 – 0.3 Bq g-1. Heavy liquid extractions were used to separate (radio)strontiobarite particles, a NORM component, from the marine sediments. The bulk chemistry, particle size (≤ 2 µm) and morphology (equant) were then characterised, and the radioactivity in these particles using autoradiography to provide validation. The uptake and fate of radium was explored

20 in model systems where synthetic (full-component) and field produced waters were mixed with synthetic and field seawaters under laboratory conditions to mimic the formation of (radio)strontiobarite particles by produced water discharge into the marine environment. Experiments showed that a significant proportion of radium (up to 48 % in 1 hour) co-precipitates with barium during mixing. Solid samples were extracted and characterised using a variety of techniques including, FTIR, XRD, BET, XAS and SEM which confirmed the mineralogical (i.e. barite particles 1 - 6 μm in size) and chemical composition (Ba~75Sr~25SO4) of precipitates obtained from laboratory mixing regimes to be consistent with the particles found in field sediments. Additionally, the long term fate, speciation and mobilisation of radium once discharged to a marine setting as the inorganic solid and aqueous ion was investigated via a series of sediment microcosm experiments where progressive anoxia was stimulated using terminal electron acceptor additions. A range of techniques including radiochemical measurements, sequential extractions, SEM, heavy liquid extractions and DNA sequencing were performed to fully quantify, understand the partitioning of radium within sediment and the resultant effect on the microbial community. Experiments showed radium remained recalcitrant to

2+ dissolution and desorption as aqueous radium (Ra ) and radiobarite (RaBaSrSO4) respectively during bioreduction and development of sulphate reducing conditions. Finally, a comprehensive investigation into the factors controlling the distribution of radionuclides in NORM samples (e.g. hard scales) obtained from tubulars from oil producing platforms (UK and Iraq) was performed to understand radionuclide attenuation within mineral phases formed under different conditions and processes. Samples comprised of anhydrite (CaSO4) and gypsum (CaSO4.2H2O) showed no levels of measurable radioactivity in comparison to scales comprised of

226 - strontiobarite (BaSrSO4), which showed measurable levels of Ra (10.7 – 18.5 Bq g 1) due to the ionic radii compatibility between radium and barium. Specific activity concentrations of 210Pb in black dust materials comprised of galena (PbS) and wurtzite (ZnS) mineral phases was found to be 30.6 Bq g-1. Here we determined the best ways of characterising NORM to aid the assessment of the fate of such materials in the environment.

Declaration

The author of this thesis declares that 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.

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, trademarks 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=24420), in any relevant Thesis restriction declarations deposited in the University Library, the University Library’s regulations (see http://www.library.manchester.ac.uk/about/regulations/) and in The University’s policy on Presentation of Theses.

Acknowledgements

First and foremost, I thank my supervisors, Sam Shaw and Kath Morris, for giving me the opportunity to work on this project in such a great research group. The completion of this PhD would not have been possible without their help thus I am appreciative of their endless support, advice and guidance throughout. I would also like to thank British Petroleum (BP) for providing the funding for my PhD, especially Peter Evans, Daniel Touzel and Oliver Pelz who provided technical support and advice throughout the duration of my PhD. The NERC Centre for Doctoral Training in Oil & Gas is acknowledged for their invaluable support in also providing funding, and the thoroughly enjoyable experience of being part of a cohort of great people, attending valuable courses and conferences.

My thanks also go to the many technical staff at Manchester who helped out over the years, including, Al Bewsher, Paul Lythgoe, John Waters, Heath Bagshaw, Cath Davies, Karen Theis, Chris Boothman, Chris Bamber and John Fellowes. I would like to thank Steve Stockley for his banter, uplifting the spirits and always believing in me. I am also in gratitude to Adrian Cleary for sharing his extensive knowledge, helping me when I initially started and answering all of my many questions. Thank you to the GeoMicro research group for being a friendly and fun bunch. A special mention to Gianni Vettese, Adrian Cleary, Lynn Foster and Naji Bassil, for being my office buddies for these past four years.

Closer to home I send my thanks to my entire family for their invaluable support. I thank my mother, father, sister, Sara and brother, Faiz for ensuring that I stayed focussed throughout. A special mention must also go to my little sister, Summah,

23 who makes me smile in the hardest times and who makes me laugh on the hardest of days. Thanks to Faiz for being the big brother that he should be, the continuous support, the togetherness, motivation and of course, the endless arguments. It’s been a rollercoaster of a journey to say the least, but we are on top! Thanks to Sara for being spoilt thus making the job as an older brother challenging. A special mention must go to my Uncle Ash for always being there for my family, being strong spirited, having a big heart and an inspiration. Thank you for all of your support throughout everything, laughter, jokes and unforgettable memories. I am now finally Dr. Saab and Gary is just being Gary! I would like to thank my friends, Arslan Amer Khalid and Yusuf Bux for listening to my regular moaning and for always believing in me. I thank Nada Al-Ghamdi for staying in contact and supporting me during this journey.

A special thanks to Emma Jayne Woolaston, who influenced and inspired me from the very beginning at Parrs Wood Sixth Form College to study Chemistry as a degree and pursue a PhD. Thanks for making the Chemistry lessons fun, interesting and easy to understand. You are hands down the best chemistry teacher. I finally made it and hope I make you proud as your first PhD graduate!

Finally, I would like to thank myself for the continued hard work and discipline over the past 8 years of higher education.

I thank everyone else who made the last four years enjoyable.

This thesis is dedicated to the memory of my grandparents, Nawab Bibi, Punu Khan, Inayat Begum and Ali Ahmad.

About The Author

The Author graduated from The School of Chemistry at The University of Manchester in 2015 with a Master of Chemistry (MChem) degree. He conducted his master’s in the field of Bioorganic Chemistry. The author then joined the Geomicrobiology group in the School of Earth and Environmental Sciences at the University of Manchester, where the work reported in this thesis was undertaken.

CHAPTER 1

Introduction

1.0 Project Introduction

Despite increasing demand for renewed focus on alternative sources of energy, the petroleum industry remains a vital contributor to the global energy demand (BP, 2019; McKinsey Global Institute, 2019; WEC, 2019).

The build-up and accumulation of naturally occurring radionuclides within solid and liquid waste materials during the extraction of oil (e.g. minerals, effluent waters and sediments) has been a recognised risk in the oil and gas industry during the past century (Doerner et al., 1925; Gordon et al., 1957; Hanan et al., 1981; Landa et al., 1983; Pardue et al., 1998; Røe Utvik, 1999; Dowdall et al., 2012; Zhang et al., 2014; Rosenberg et al., 2014; Garner et al., 2015; Van Sice et al., 2018; McDevitt et al., 2019). Materials which contain an elevated concentration of radionuclides due to technological processes are termed technologically enhanced naturally occurring radioactive materials (TE-NORM) and can pose an additional risk to the health of workers and environment. There are arrays of wider industrial activities which can lead to the formation of NORM in significant volumes. These primarily comprise of phosphate and fertiliser manufacture (Burnett et al., 1996; Keating et al., 1996), fossil fuel burning, (Hedvall et al., 1996) mining and milling of uranium ores, conventional and unconventional oil and gas production (Vidic, 2013; Kondash et al., 2014; Zhang et al., 2014).

The dominant mechanism of NORM formation during oil extraction is the mixing of incompatible waters (e.g. seawater and produced water) which can occur at particular stages during the extraction operations (e.g. water flooding and marine discharge), and is an ongoing problem encountered in the petroleum industry (Todd et al., 1990, 1992; Badr et al., 2008; Rosenberg et al., 2014; Zhang et al., 2014).

As a result of temperature and pressure conditions altering and / or injection of seawater during water flooding activities to restore formation pressure and increase oil production, carbonate and sulphate scales deposit on the inside surfaces of production equipment (Fig. 1.1) (Yuan et al., 1994; Al-Masri et al., 2005; Badr et al., 2008; Garner et al. 2015). The radioactive decay of primordial radionuclides (238U and 232Th) which naturally exist in the reservoir rock, results in the distribution of an array of naturally occurring radionuclides (e.g. isotopes of radium (224Ra, 226Ra and 228Ra) and lead (210Pb)) in geological formations, gas or oil and water streams (Fig. 1.1) (Fisher, 1998; Ojovan, 2019). Naturally occurring radionuclides (NOR’s) within extracted waters (e.g. produced water) can co- precipitate within hard scales, sludges and drilling muds (Garner et al., 2015; IOGP, 2016). In particular the ability of radium to incorporate via co-precipitation into insoluble barium and strontium sulphate mineral phases results in the formation of naturally occurring radioactive material (NORM) in the form of RaxBa1-xSO4

(radiobarite) and BaxSryRazSO4 (radiostrontiobarite) (Steffan, 2014; Al-Masri et al., 2005; Garner et al., 2015). These materials can accumulate in large volumes across an array of components in the production system, for example pipe lines, valves, storage tanks, wellheads, tubulars and may also deposit on and contaminate production equipment (Fig 1.1) (Smith, 1985; Al-Masri et al., 2005; Puntervold et al., 2008; Vearrier et al., 2009; Abdul et al., 2010; Garner et al., 2015; IOGP, 2016; Vazirian et al., 2016). The build-up of scale and sludge material impairs production leading to the need for descaling operations (Crabtree et al., 1999; Abbas, 2014; Rawahi, 2017). Material that is NORM contaminated is most difficult and costly to remove due to additional health hazards (Hamilton et al., 2004; Abdul et al., 2010; IOGP, 2016). Studies have previously focussed on the characterisation of a variety of NORM material to aid disposal and decontamination processes however, studies in relation to the radioactivity and mineralogy of such samples and its associated fate and hazard in the environment has received less attention. Understanding the factors controlling the distribution of radionuclides in NORM samples will help better understand the environmental impact and fate of such materials in the environment.

Following the precipitation of a combination of solid-NORM (e.g. scale), produced waters with lower activity concentrations of NOR’s than its initial activity are subsequently extracted (Candeias et al., 2014). Discharge of produced water effluent to the marine environment, an essential plant operation, may drive the precipitation of sulphate particles with the potential uptake of radium or lead to adsorption of radium to existing particulates in the sea column and / or to sediment (Fig. 1.2) (Neff, 2002; Fakhru’l-Razi et al., 2009; Grung et al., 2009; Abdellah et al., 2014). As radium has a long half-life and high radiotoxicity it poses a potential radiological risk (Thiry et al., 2008). There is uncertainty about the nature of these interactions and the speciation of radium during mixing of these waters which is poorly constrained (Fig. 1.2). Understanding the removal mechanisms associated with radium during operational discharges from the offshore oil and gas industry into the marine environment is essential in underpinning predictions of the fate of 226Ra in these systems and further defining its overall environmental impact (Fig. 1.2). Radium scavenging mechanisms such as adsorption and precipitation, as well as factors such as aqueous dispersion are key fundamental processes effecting the mobility and fate of radium (Fig. 1.2). It is expected that the formation of inorganic micro-particulate radiostrontiobarite (RaBaSrSO4) ternary phases during production water / seawater mixing, via the mechanism of co-precipitation is a major pathway controlling 226Ra fate in these systems but, there is a paucity of direct experimental evidence for this process (Fig. 1.2) (Landa et al., 1983; Pardue et al., 1998; Jerez

Vegueria et al., 2002; Gafvert et al., 2007; Van Sice et al., 2018; McDevitt et al., 2019). In addition, radium may also exist unassociated in the water column in the aqueous form as the more mobile Ra2+ ion, thus it is important to gain a better understanding of the fate, speciation and mobility of radium in both the solid (radiostrontiobarite) and aqueous phase (Ra2+ ion) in marine systems as a result of operational discharge (Fig. 1.2). In particular, associated with the natural biogeochemical processes (e.g. microbial reduction) in the sediment and how the speciation and fate of radium following deposition changes during microbial reduction processes (e.g. sulphate reduction) (Fig. 1.2).

Figure 1.2: Discharge scenario illustrating the complexity of this system and the potential scavenging mechanisms and pathways controlling the fate, speciation and mobility of radium: 1) Discharge of produced water; 2) Dispersion of Ra2+ (aq); 3) Adsorption of Ra2+ to sediment; 4) Precipitation of radiobarite; 5) Dispersion of radiobarite; 6) Deposition of particles; 7) Indigenous microbial population and biogeochemical processes in the sediment and; 8) Radium released from sediment surface or radiobarite via microbial reduction

NORM is one of the regulated risks that have to be accounted for in some jurisdictions, such as the North Sea where the total discharge is monitored. Oil and

29 gas production collectively account for the largest discharge of alpha emissions from industrial sources (OSPAR, 2018) (Table 1.1).

Oil & Gas Nuclear Year Total alpha 226Ra Total beta Total alpha Total beta (TBq) (TBq) (TBq) (TBq) (TBq) 2005 6.4 0.81 4.25 0.52 160 2006 6.9 0.78 4.67 0.34 58 2007 7.4 0.90 4.94 0.19 33.4 2008 6.76 0.82 4.54 0.17 27.2 2009 7.4 0.94 5.02 0.18 29.8 2010 7.6 1.03 4.94 0.18 23.1 2011 7.6 0.95 5.03 0.17 25.9 2012 8.0 1.05 5.2 0.19 20.1 2013 6.5 0.78 4.34 0.20 21 2014 6.1 0.73 4.1 0.22 21 2015 6.7 0.80 4.4 0.23 20 2016 7.1 0.85 4.7 0.29 22

Table 1.1: Total alpha and beta discharges into the North Sea from 2005-2016 from the oil industry. Total alpha and beta emissions from the nuclear sector shown for comparison (OSPAR, 2018)

HSE (Health and Safety Executive) legislation governs worker exposure to NORM and is outside the scope of this project. However, the environmental effects of NORM are not fully understood and a number of global trends are making this issue more pronounced:-

• It has been publicly reported that significant stockpiles of NORM are accumulating in countries and territories where adequate long-term disposal facilities are not available (IAEA, 2013).

• Published research has indicated an increasing discharge of water and improvements in scale control are shifting discharge of NORM into the water phase (Pardue et al., 1998; Jerez Vegueria et al., 2002; Hylland, 2003; Eriksen et al., 2006;

Gaefvert et al., 2007; Grung et al., 2009; Dowdall et al., 2012; OSPAR, 2018; Van Sice et al., 2018).

• Unconventional oil and gas reserves (e.g. shale gas) are publicly reported to be frequently associated with NORM. The location onshore and the large number of wells make this issue more apparent (Rowan, 2011; Kondash et al., 2014; Zhang et al., 2014; IOGP, 2016; Lauer et al., 2018; Van Sice et al., 2018; McDevitt et al., 2019).

• Decommissioning of ageing assets may generate large amounts of NORM scales (Al-Masri et al., 2003; Al-Masri et al., 2005; Garner et al., 2015; IOGP, 2016).

A fundamental understanding of the fate and uptake of key radionuclides into NORM waste streams e.g. mineral scales formed in tubulars, and from operational marine discharge of produced water as a result of mixing of incompatible waters (e.g. seawater and produced water), is key in order to develop the capability to predict their fate and environmental risk.

1.1 Aims and objectives

The aim of this project is to determine the speciation and fate (e.g. mobility and attenuation) of key radionuclides (e.g. 226Ra) derived from oil and gas production during two fundamental processes of NORM formation in which mixing of incompatible waters is encountered (i.e. mineral formation during water flooding operations and operational marine discharges of produced waters). To do this, primary investigations into factors controlling the distribution of radionuclides in NORM samples (e.g. sediment and hard scales) obtained from the field was performed to understand radionuclide attenuation within mineral phases formed under different conditions and processes. The secondary objective was to gain an understanding of the NORM formation process and nature of material which forms during operational discharge. This included exploring the formation, uptake and speciation of radium into fine grained sulphate solid phases (e.g. radiobarite)

31 following discharges into the marine environment via simulation experiments. The final objective was to assess the speciation and fate of radium as biological processes and reducing conditions develop in sediment, especially sulphate reduction, which has not been previously assessed.

The deliverables overall aim to contribute towards improving our understanding of the risk that NORM presents and the ability to ultimately predict the fate of radium in marine systems to further assess it environmental impact, and enable development and adoption of fit-for purpose control strategies tailored to different locations. The output can also aid processes such as decontamination and disposal.

Three main overarching areas of research were defined, which were studied using a variety of wet chemistry, geochemical, biogeochemical, radiometric, spectroscopic, state-of-the-art micro focus and mapping techniques, including synchrotron-based spectroscopy and SEM (scanning electron microscopy), to provide in-depth information of the chemical, morphological and radiological composition of NORM samples, including radionuclide distribution, speciation and uptake mechanism.

1.1.1 Outline

1. We explore the uptake and fate of radium in marine sediment samples obtained from a field site where produced waters are discharged. With marine field samples, radium activity concentrations was assessed using gamma-ray spectroscopy, and heavy liquid extractions were used to separate (radio)strontiobarite particles, a NORM component, from the marine sediments. The bulk chemistry, particle size and morphology were then characterised, and the radioactivity in these particles using autoradiography to provide validation. The process of marine discharge was mimicked in the laboratory via mixing experiments (i.e. seawater and produced water). The uptake and fate of radium was explored in model systems where synthetic (full-component) and field produced waters were mixed with synthetic and field seawaters under laboratory conditions to mimic the formation of (radio)strontiobarite particles by produced water

discharge into the marine environment. Solid samples were extracted and characterised using a variety of techniques including, Fourier transform infrared (FTIR), X-ray diffraction (XRD), Brunauer-Emmett-Teller (BET) surface area analysis and scanning electron microscopy (SEM). Other techniques such as solution analysis and geochemical modelling were also employed. The speciation of strontium associated with barite was analysed via X-ray absorption spectroscopy (XAS). This is the first paper to extract, assess and fully characterise the precipitates which form from the discharge of produced water to seawater using both natural and synthetic systems, and from field sediments providing a better understanding of the mechanism of NORM formation in these systems, and helping to predict the fate of radium (226Ra) in marine environments.

2. A comprehensive assessment of the long term fate, speciation, mobilisation and availability of radium once discharged to a marine setting as; 1) the

2+ inorganic solids (BaxSryRazSO4 and BaSrSO4) and, 2) the aqueous ion (Ra ). Aqueous radium and synthetic radiostrontiobarite chemically and morphological identical to that extracted from the field site, were deposited within marine sediment and biologically induced reducing conditions developed using microcosm experiments. Experiments were conducted under environmentally relevant conditions using seawater and sediment obtained from a produced water discharge site in the UK. A range of techniques including sequential extractions, SEM, liquid scintillation counting, ion chromatography, spectrophotometry, heavy liquid extractions and 16s rRNA sequencing were performed to fully quantify, understand the partitioning of radium within sediment and the resultant effect on the microbial community pre- and post-anoxia. This is the first study to investigate the mobility and solubility of radium in its aqueous and solid form directly applicable to discharges in the oil and gas sector, as biogeochemical conditions evolve from iron- to dominant sulphate-reducing conditions under marine field conditions.

3. Characterisation of a variety of NORM samples in the form of pipe scale extracted from tubulars from oil facilities. Development of sample handling and mounting techniques to explore the mineralogy of scales using high accuracy analytical techniques. We used a range of micro/nano focus and bulk techniques to determine the best ways of characterising NORM. We investigated the factors controlling the distribution of radionuclides in samples by identifying relationships between the bulk chemical and mineralogical composition, and radionuclide content. The spatial relationship between these factors was investigated to ultimately understand the potential environmental impact and fate of such materials in the marine environment. Links between the mineralogy and activity of the samples was achieved using a combination of techniques such as gamma spectroscopy, XRD and XRF. Mapping techniques were also used including, FTIR, Raman, SEM and autoradiography. Although previous studies have investigated the incorporation of radionuclides in a variety hard scales, a special emphasis was given to the portion of material which may remain in the produced water effluent post treatment and finds its way into surface marine waters.

1.2 Thesis structure

This thesis comprises of three primary research chapters produced for publication dedicated to the research areas described above, and has been submitted in the thesis by paper format. These research chapters are preceded by a review of the related literature within these areas of research and a thorough description of the methodologies applied. All research chapters include a statement of the contribution of the author and collaborators. A summary discussing and outlining the conclusions and future work directions is then provided. The chapters are as follows:

Chapter 2: Literature review. Provides a comprehensive review of the relevant literature and previous work to the subject area of this thesis

Chapter 3: Methodology. Details all the experimental methodologies employed throughout this research, including both practical and theoretical details.

Chapter 4: Research chapter. Contains the paper “Fate of radium on discharge of oil produced water to the marine environment’’ and associated supporting information. This paper explores the formation, uptake and fate of radium in model systems where natural and synthetic produced waters are mixed with natural and synthetic seawaters under laboratory conditions. Precipitates (e.g. strontiobarite and radiostrontiobarite) extracted from model experiments and field marine sediment samples are characterised via an array of techniques to provide validation of the NORM formation process. This paper also provides an assessment of factors which may affect the speciation and uptake of radium following marine discharges.

Chapter 5: Research chapter. Contains the paper “The effects of bioreduction on the fate of Ra2+(aq) and radiobarite in marine sediments on discharge of oil produced water’’ and associated supporting information. This paper extends the experiments of the first paper to provide a comprehensive assessment of the potential environmental impact, speciation, mobilisation and availability of radium once discharged to a marine setting as; 1) the inorganic solids (BaxSryRazSO4 and

2+ BaSrSO4) and, 2) the aqueous ion (Ra ) in a series of sediment microcosm experiments. This paper also provides an environmental impact assessment.

Chapter 6: Research chapter. Contains the paper “Chemical and radiological characterisation of scales containing NORM’’ and associated supporting information. This paper determines the best ways to characterise a variety of NORM samples obtained from well strings and assesses their potential environmental impact and fate in the marine environment.

Chapter 7: Conclusion and further work. This chapter provides a summary of the key findings and results achieved. Proposals for future work directions drawn from the thesis project are also outlined and considered.

A full list of references is delivered after Chapter 7.

Appendices include a list of attended conferences including presentations, and list of courses attended through the NERC CDT in Oil and Gas.

1.3 Paper status and author contributions

Chapter 4: “Fate of radium on discharge of oil produced water to the marine environment’’ F. Ahmad Principal author; all laboratory work; all data analysis and sample preparation and characterisation K. Morris Input to experimental concept; aided with data interpretation; extensive manuscript review pre- and post- submission G. T.W. Law Aided with gamma spectroscopy analysis and liquid scintillation counting K. Taylor Project supervisor; input into project concepts and reviewed the manuscript S. Shaw Principal supervisor; input to experimental concept; aided with data interpretation; extensive manuscript review pre- and post- submission; aided with XAS data collection

Chapter 5: “The effects of bioreduction on the fate of Ra2+(aq) and radiobarite in marine sediments on discharge of oil produced water’’ F. Ahmad Principal author; all laboratory work; all data analysis and sample preparation and characterisation K. Morris Input to experimental concept; aided with data interpretation; extensive manuscript review pre- and post- submission G. T.W. Law Aided with gamma spectroscopy analysis and liquid scintillation counting K. Taylor Project supervisor; input into project concepts and reviewed the manuscript

S. Shaw Principal supervisor; input to experimental concept; aided with data interpretation; extensive manuscript review pre- and post- submission

Chapter 6: “Chemical and radiological characterisation of scales containing NORM’’ F. Ahmad Principal author; all laboratory work; all data analysis and sample preparation and characterisation K. Morris Input to experimental concept; extensive manuscript review pre- and post-submission G. T.W. Law Aided with gamma spectroscopy analysis K. Taylor Project supervisor; input into project concepts and reviewed the manuscript S. Shaw Principal supervisor; input to experimental concept; aided with data interpretation; extensive manuscript review pre- and post- submission

CHAPTER 2

Literature Review

2.0 Literature Review

2.1 Natural uranium - 238 and thorium - 232 decay Series

Radioactive elements of differing concentrations naturally exist in soils, rocks and waters as a result of stellar processes, novae and supernovae nuclear synthesis events which have led to the incorporation and distribution of elements within the earth’s crust (Choppin, 2013). Such radionuclides are termed naturally occurring radionuclides (NOR’s) thus exist in igneous, sedimentary rocks, soils and sea water at typical concentrations shown in Table 2.1 (Fisher, 1998; Ojovan, 2019). The vast majority of naturally occurring radionuclides arise from the radiative decay of the ‘heavier’ primordial elements uranium (238U) and thorium (232Th). Some NOR’s (e.g. 14C and 3H) arise from cosmic radiation interactions in the Earth’s upper atmosphere, whereas others (e.g. 40K and 87Rb) naturally exist independent of the radiative decay of primordial radionuclides (Choppin, 2013).

Primordial radionuclides characteristically have long half-lives on the scale of

238 billions of years hence existing since the formation of earth. Uranium ( U: t1/2

9 232 10 4.470 x 10 years) and thorium ( Th: t1/2 1.405 x 10 years) are the decay chains of most importance in understanding the existence of primordial radionuclides identified in NORM in oil field liquid and solid waste streams (Choppin, 2013; Ojovan, 2019).

Rock type 238U 232Th 40K 87Rb Igneous (Bq kg-1) Mafic 7 - 10 7 – 10 70 – 400 10 – 50 Salic 7 – 10 60 – 80 1100 – 1500 170 – 200 Granite 40 70 > 1000 170 – 200 Sedimentary (Bq kg-1) Shale 40 50 800 110 Carbonate 25 8 70 8 Soil 66 37 400 50 Sea water 0.03 - 11 1 (Bq L-1)

Table 2.1: Concentrations of primordial radionuclides in igneous, sedimentary rocks, soils (Bq kg-1), and sea water (Bq L-1), where thorium concentrations (-) are extremely low; ≥ 0.16 ng kg-1/ 4 x 10-3 dpm kg-1 (Somayajulu et al., 1966; Moore, 1981; Huh et al., 1985; Fisher, 1998; Ojovan, 2019)

During the radioactive decay of primordial (‘parent’) radionuclides, such unstable elements are transformed into subsequent ‘daughter’ or ‘progeny’ radionuclides. Three types of radioactive decay are experienced with characteristic energies and mass. Ionising radiation such as, alpha particles (α), beta particles (β) and gamma rays (γ) are emitted with characteristic properties. Alpha particles consist of two neutrons and two protons (helium ion), whereas beta particles are defined as either negatrons (β-: high-speed electrons) or positrons (β+: antielectron). Gamma rays are high energy photons of electromagnetic radiation whose emission generally occurs after other forms of decay have taken place (e.g. alpha or beta decay), to transform the resulting daughter nucleus in an excited state with excess energy into to a more stable ground state of lower energy. Alpha particles are doubly charged (He2+ or α2+), poorly penetrating but highly ionising in comparison to other types of radiation. Beta particles have a moderate penetrating power and are either negatively charged (negatrons) or positively charged (positrons) with the mass of an electron. Gamma rays have no charge but, have the highest penetrating power and can therefore cause the most damage via exposure. All three types of radiation can lead to the breakage of chemical bonds and have the potential to cause damage to, or even destroy living cells or tissue. All modes of radiation emission pose the

39 greatest radiation risks and health hazards if inhaled, injected or ingested leading to elevated risks of developing various types of cancers e.g. lung cancer and bone cancer (Smith, 1992; NRC, 1999; UNSCEAR, 2000; Vearrier et al., 2009; Choppin, 2013; Carvalho et al., 2014; IOGP, 2016).

The daughter radionuclides which are most dominant and have the highest activity in NORM in the oil and gas industry are radon-222 (222Rn; α emitter), radium-228 (228Ra; β emitter), radium-226 (226Ra; α/γ emitter), polonium-210 (210Po; α emitter) and lead-210 (210Pb; β/γ emitter) which arise from the uranium and thorium decay series (Fig. 2.1). Radioactive decay of primordial radionuclides (238U and 232Th) cascade through a series of slow sequential alpha and beta decays resulting in the formation of the stable non-radioactive isotopes lead-206 (206Pb) and lead-208 (208Pb) respectively (Choppin, 2013; IOGP, 2016).

As shown from the decay series (Fig. 2.2) different modes of mobilisation exist for certain progeny during oil production. Isotopes of radium (224Ra, 226Ra and 228Ra) and lead (210Pb) are present in liquid and solid waste streams (e.g. produced waters and scales) due to leaching of primordial radionuclides situated in reservoir rock (Fig. 1.1). This leads to their mobilisation to pore waters and their resultant detection at a range of offshore and onshore installations around the world (Doyi et al., 2016; IOGP, 2016). Due to the redox potential of the surrounding hydrocarbon reservoir being strongly reducing the parent radionuclides, uranium and thorium, remain immobile, especially for the case of tetravalent uranium (U4+) (Collins, 1975; Bloch et al., 1981; Laul, 1994; Miezitis et al., 2007; Choppin, 2013). Other radionuclides such as radon gas, a decay product of radium, is mobilised via emanation and transported with gas or oil and water streams (Fig. 1.1). These are the dominant causes of increases in radioactivity observed in pipe scales and produced waters during petroleum production (Fig 1.1) (Hartog et al., 2002; Al- Masri et al., 2005; Abdellah et al., 2014; Garner et al., 2015; IOGP, 2016).

Due to the characteristic long half-lives of thorium and uranium secular equilibrium can be established for both decay chains, thus allowing an indication of their concentration from gamma ray emission intensities from their respective progeny. Radioactive equilibrium can be established when parent radionuclides decay to

40 produce progeny with much shorter half-lives. The rate of decay of the parent radionuclides thus becomes closely related (approximately constant) to the rate of production of the daughter radionuclides and radioactive equilibrium is established (Choppin, 2013). The establishment of equilibria occurs after numerous half-lives of the primordial radionuclide of interest (e.g. 238U: ~ 30 years and 232Th: ~ 106 years). Breaches in secular equilibrium can occur as a result of radon emanation which is important in determining the concentrations of 226Ra in samples, and also by leaching of radionuclides to the water column (e.g. 210Pb) (Fig. 2.2) (IOGP, 2016).

Figure 2.1: The full decay series of uranium (238U) and thorium (232Th), adopted radiative decay and half-life (IOGP, 2016)

Figure 2.2: The decay series of uranium and thorium with the different modes of mobilisation of certain daughter radionuclides (IOGP, 2016)

2.2 Produced Water – Liquid waste stream

2.2.1 Characteristics

Produced water naturally exists in geological formations within the pore and void spaces within rocks (Hunt, 1979). It is understood that rock in most hydrocarbon formations was once entirely occupied by water previous to its displacement and trapping of hydrocarbons (Amyx et al., 1960). As the density properties differ, the less dense hydrocarbons displace water from the formation leading to a hydrocarbon-containing geological formation referred to as a petroleum reservoir (Fig. 2.3). Geological formations therefore generally contain a pool of hydrocarbons in liquid and gas chemical states and water referred to as ‘formation water’ or ‘connate water’ in fractured or porous rock spaces. Once fluid is transported to the surface during production of the reservoir it is referred to as ‘produced water’. Produced water is transported to the surface with crude oil, natural gas and water from water flood activities during recovery as a complex mixture. Gas-oil, water-oil

42 separators and water treatment technologies such as membrane based desalination, thermal desalination, membrane distillation, hydrocyclones and gas floatation are implemented to separate gases, condensates, oil, waters, sand, sediments and solids producing the largest effluent by-product in the oil and gas industry (Fig. 1.1) (Arthur et al., 2005; Igunnu et al., 2014; Singh, 2015a; Jiménez et al., 2018). It estimated up to 3 – 9 barrels of water is co-produced with each barrel of oil, with over 250 million barrels per day of water produced globally, and more than 40 % discharged to the environment (Fakhru’l-Razi et al., 2009; Igunnu et al., 2014; Singh, 2015a; Oil & Gas UK, 2018; OSPAR, 2018).

Produced water NORM disposal practices and regulations vary globally however, disposal methods for NORM generated in the offshore and onshore oil and gas industry differ as marine disposal is limited onshore. Generic measures adopted include disposal at source to the marine environment or streams (e.g. produced water discharges), geological disposal (e.g. disposal/abandoned wells), well injection including salt dome disposal, storage or burial as radioactive waste (e.g. low-level radioactive waste mines) and the use of (un-) lined pits/ponds (e.g. fracking ponds and lagoons) (Roberts et al., 1998; IAEA, 2004; Al-Masri et al., 2005; Yoshida et al., 2008; OSPAR Comission 2014, 2013; UK NORM Waste Strategy, 2014; Robertson et al., 2017). As produced water is commonly treated/cleaned and discharged to the marine environment during offshore operations many challenges and factors need to be considered relating to its marine disposal (e.g. oil and radium content, and fate) (Cowan., 1976; OSPAR, 2001, 2018; Gafvert et al., 2007; Hosseini et al., 2010, 2012; Dowdall et al., 2012; Olsvik et al., 2012; Bakke et al., 2013).

Figure 2.3: Oil and gas accumulation and trapping schematic (www.geologypage.com)

The natural constituents of produced waters are mainly compounds such as organic acids, aromatic hydrocarbons, polyaromatic hydrocarbons, phenols, alkylphenols,

+ + + 2+ 2+ 2+ 2+ 2+ sand and silt. Also dissolved cations, (Na , K , NH4 , Ca , Mg , Ba , Sr and Fe ),

2- - - - 2- - - - 3- anions (SO4 , Br , Cl , HCO3 , CO3 , BO2 , NO3 , OH and PO4 ) and gases (H2S, CO2

-1 and O2) in varying amounts. Concentration of salts vary from a few mg L up to 345 000 mg L-1 (Jacobs et al., 1992; Røe Utvik, 1999; Fakhru’l-Razi et al., 2009; Kondash et al., 2014; Zhang et al., 2014; IOGP, 2016). Isotopes of radium (224Ra, 226Ra and 228Ra) and lead (210Pb) are also present in produced waters and detected at a range of offshore and onshore platforms around the world (see section 2.5) (Røe Utvik, 1999; Rowan, 2011; Kondash et al., 2014; Doyi et al., 2016; Van Sice et al., 2018). As formation waters are typically saturated with chloride, metals exist as chloride salts and as aqueous ions in solution (Steffan, 2013). Produced waters from gas production are considered more toxic as they can contain higher amounts of aromatic hydrocarbons of low molecular weight such as benzene, toluene, xylene and ethyl benzene (Jacobs et al., 1992). Other production chemicals such as corrosion products and chemical additives are present in trace amounts such as; a) equipment corrosion inhibitors; b) mineral scale inhibitors; c) biocides to reduce bacterial fouling; d) demulsifiers/emulsion breakers (water-in-oil) and reverse emulsion breakers (oil-water separators); e) flocculants and coagulants to remove solids and; f) solvents (Steffan, 2014). Chemical additives are commonly introduced as treatments to ensure safe and reliable facility operation, and to increase the oil and gas production via maintaining a constant flow of hydrocarbons, and to safeguard the quality of produced water to ensure it meets environmental and disposal guidelines (OSPAR, 2001; Her Majesty’s Stationery Office and London,

2016). Additives can also have an undesired effect on the properties of oil and water mixtures such as toxicity, increased bioavailability and reduced biodegradability (Brendehaug et al., 1992).

The chemical complexity of produced waters gives rise to its ability to be the major contributor resulting in the formation of a variety of mineral scale deposits in production equipment. There is also environmental concern related to the marine disposal of produced water. The chemical and physical compositions of produced waters globally are non-identical and alter significantly due to many factors; 1) the reservoir geology with which the produced water has been in contact for many years (hundreds of thousands); 2) the type of hydrocarbon product being produced; 3) the location of the oil field and; 4) the production life time of the reservoir as the chemical properties and volume alter over time (Li et al., 2013).

Studies have shown Gulf of Arabia oil platforms typically produce waters with increased concentrations of calcium thus the identification of dominantly calcium- containing scales (e.g. gypsum and calcite) at installations. This has been attributed to the limestone rich reservoirs present in the region (Table 2.2) (Bader, 2006; Moghadasi et al., 2007; Badr et al., 2008). In contrast, produced waters and scales rich in barium and strontium (e.g. barite and strontiobarite) are commonly encountered at offshore and onshore installations in the UK and North Sea (Table 2.2) (Mitchell et al., 1980; Todd et al., 1990, 1992; Yuan et al., 1994; Worden et al., 2000; Garner et al., 2015).

Produced water and sea water compositions Element Seawater North Sea produced Gulf of Arabia (ppm) water (ppm) produced water (ppm) Na 10890 29370 68195 K 460 372 4361 Mg 1368 504 1903 Ca 428 2809 19483 Sr 8 574 1090 Ba 0 292 0 Fe3+ 0 0 0 Cl 19700 52360 147910

SO4 2960 11 287

HCO3 124 496 256

Table 2.2: Typical compositions of seawater and produced waters from the North Sea (Todd et al., 1992) and Gulf of Arabia (Bader, 2006)

Though a correlation exists between the produced water composition and the mineralogical composition of scale produced, where variations in mineralogy is primarily affected by basin geology composition at particular geographical locations, a variety of scales are often found at installations (Table 2.3). This is because the properties of produced waters can alter during the life time of the well as on longer time scales waters existing in other formations adjacent to the hydrocarbon-bearing formation can become part of the water matrix which alter the properties of the produced water.

Name Formula

Iron carbonate (siderite) FeCO3

Iron sulphide(s) (pyrite & mackinawite) FexSy; FeS, FeS2

Iron oxide(s) (hematite & magnetite) FeO, Fe2O3, Fe3O4

Iron hydroxides Fe(OH), Fe(OH)2, Fe3O3

Magnesium Carbonate (magnesite) MgCO3

Calcium magnesium carbonate (dolomite) CaMg(CO3)2

Barium carbonate (witherite) BaCO3

Barium radium strontium sulphate BaxSryRazSO4 (radiostrontiobarite)

Strontium carbonate (strontianite) SrCO3

Silica (quartz) SiO2 Sodium chloride (halite) NaCl Organic materials Wax, bio film, asphaltenes Sand, shale Various reservoir rock grains not just quartz Chemical additives Calcium phosphate, calcium sulphite, thermally degraded polymers, quaternary ammonium salts

Barium sulphate (barite) BaSO4

Strontium sulphate SrSO4

Barium radium sulphate BaxRa1-xSO4

Calcium carbonate (calcite/aragonite) CaCO3

Calcium sulphate (anhydrite & gypsum) CaSO4, CaSO4.2H2O

Barium strontium sulphate (strontiobarite BaSrSO4 & bariocelestite) Zinc Sulphide (sphalerite) ZnS Lead sulphide (galena) PbS

Table 2.3: The various scales that can precipitate in oilfields including those due to the addition of chemical additives and corrosion (Al-Masri et al., 2005; Steffan, 2013; Garner et al., 2015; IOGP, 2016)

2.2.2 Chemical incompatibility of produced waters

During the primary extraction and recovery of oil, the natural pressure in the geological rock formation begins to deplete and resultantly the subsurface pressure becomes insufficient to force and drive oil towards the surface. Consequently, secondary and tertiary methods (e.g. from specialist floods to gas injection) are required to supply external energy to the reservoir to provide thrust to extract further oil reserves from void spaces of rocks to increase the yield of oil extracted. This is commonly accomplished offshore by injecting fluids, primarily sea water due to its abundance offshore, to maintain overall pressure and displace hydrocarbons towards the wellbore (Fig. 2.4) (Bahadori, 2018). The injection of sea water can have a detrimental impact on the productivity of the reservoir if mitigation and prediction strategies in regards to scale formation and souring of the well have not been accounted for (Ollivier, 2005). Reservoir souring occurs due to the incorporation of foreign microbes into the reservoir (e.g. sulphate-reducing bacteria) by the injection of sea water (Davies Michael, 2006). This is problematic because these bacteria have the capability to reduce sulphate to sulphide leading to the formation and accumulation of hydrogen sulphide (H2S) which can promote corrosion, scale formation, exposure and health and safety concerns (Al-Masri et al., 2005; Badr et al., 2008; IOGP, 2016). Biocides are typically used as a preventative strategy with typical levels of H2S encountered in natural gas typically ranging between 0 - 150 ppm with up to 80 000 ppm detected globally (Wang et al., 2008; Telmadarreie et al., 2012; Amosa et al., 2013; Zhu et al., 2017). Levels above 4 ppm are generally considered as ‘sour’ gas wells (EPA, 1995).

In time, injected seawater begins to form part of the fluids in the formation and are recovered, however during water-flooding operations differences in chemical properties between seawater and produced water encourages the formation of precipitates along the well-string and production system. When sea water typically

2- containing a high concentration of sulphate (SO4 ) but low concentrations of divalent cations (Ca2+, Mg2+, Ba2+ , Ra2+ and Sr2+), and formation water which is characterised by low sulphate and high divalent cation concentrations alongside elevated radionuclide levels are mixed, precipitation of a combination of sulphate

48 scales follow (including radium-containing sulphate scales) (Table 2.2) (Yuan et al., 1994; Badr et al., 2008; Candeias et al., 2014; IOGP, 2016). The key process of the formation of inorganic precipitates (e.g. RaxBa1-xSO4 and BaxSryRazSO4) is the mixing of chemically incompatible waters (e.g. produced water and sea water) and the establishment of supersaturated solutions. This process occurs during offshore marine discharges in addition to enhanced oil recovery operations (see section 2.3 and 2.4) (Zhang et al., 2014). Other types of scale can form under certain operational conditions (e.g. temperature and pressure) including the formation of sulphides and carbonates (see below) (Badr et al., 2008; Mccartney, 2008; Puntervold et al., 2008; Steffan, 2013; IOGP, 2016).

2.2.3 Mineral solubility

The solubility product constant Ksp, governs the solubility of a given mineral in solution (Petrucci et al., 2007; Clark et al., 2000; Ferguson et al.,1995). This equilibrium constant therefore represents the degree at which a solute dissolves in solution. Minerals which are more soluble have a higher Ksp in comparison to those which are less soluble (Petrucci et al., 2007; Ferguson et al., 1995). For example, the solubility product for the barite mineral (BaSO4) is given by:

Equation 2.1: A) Solubility equilibrium constant between a solid (barite) and its respective ions in solution 2+ 2- (Ba and SO4 ) and; B) The activities of ions at the state of equilibrium

Here, if barite is a pure solid phase then the activity of aBaSO4 by convention is equal to 1 as shown from the right hand side of Equation 2.1a. In dilute aqueous solutions, where a small amount of ions will be dissolved the activities of the ions

2+ 2- are described as, aBa2+ ≈ [Ba ]eq and aSO42- ≈ [SO4 )eq at the state of equilibrium

(Equation 2.1b) however, a real solution may not be in the state of equilibrium (Petrucci et al., 2007; Ferguson et al., 1995).

2+ 2- Therefore, multiplying the concentrations of the ions together ([Ba ] x [SO4 ]) will give the ion activation product (IAP) (Ferguson et al., 1995). This has the same form as the equilibrium constant (Ksp), but describes the non-equilibrium state of a

2+ 2- solution and takes into account the actual activities (e.g. IAP = [Ba ]actual [SO4 ]actual) of the mineral forming ions (Equation 2). This is because in a natural solution it is

2+ not likely that the mineral forming ion concentrations are equal (e.g. [Ba ]eq =

2- [SO4 ]eq) as there is typically more than one source of each of these ions in solution.

The saturation index (SI) is logarithmic ratio of the ion activation product (IAP) and the solubility product (Ksp) as shown by Equation 2.2.

2+ 2− [퐵푎 ]푎푐푡 [푆푂4 ]푎푐푡 퐼퐴푃 푆퐼 = 푙표푔10 = 퐾푠푝 퐾푠푝

Equation 2.2: The saturation index for determining saturation tendency

The saturation index is expressed as the log of the actual amount of mineral forming ions over the solubility product constant of that mineral. The saturation index is a beneficial measure to understand whether the solution is undersaturated, saturated or supersaturated with respect to the given mineral. An undersaturated solution will have a negative SI value (SI < 0), a saturated solution will have a SI = 0, and a supersaturated solution will have a positive SI value (SI > 0) (Ferguson et al., 1995).

SI = O IAP = Ksp Saturated (in equilibrium)

SI < 0 IAP < Ksp Undersaturated

SI > 0 IAP > Ksp Supersaturated

2.2.4 Degree of saturation

A solution is termed saturated when it is in thermodynamic equilibrium with the solid phase of its solute at a given temperature, and a solution containing a greater amount of dissolved solute than that given by the equilibrium saturation value is termed supersaturated (Harker et al., 2002). The degree of supersaturation can be expressed by:

∆푐 = 푐 − 푐∗

Equation 2.3: Degree of supersaturation expression

where, c is the solution concentration and c* the equilibrium saturation values.

The supersaturation of a solute in solution is an important determining factor that controls the crystallisation or precipitation of minerals from solution. Some minerals must precipitate from a supersaturated solution in order to reach equilibrium. Therefore, the supersaturation is generally referred to as the thermodynamic ‘driving force’ for mineral precipitation where, it is more favourable for precipitation to take place from a highly supersaturated solution in contrast to a less supersaturated solution (Krauskopf et al., 1995; Harker et al., 2002). The degree of supersaturation may change during oil production due to variations in temperature, pressure and pH typically experienced during production. The main reason for the precipitation of sulphate mineral scales in tubulars is the mixing of incompatible waters e.g. seawater and formation water. In addition, the controlling factor for the formation of carbonate scales is the reduction in pressure experienced during production leading to the loss of carbon dioxide from solution (see sections 2.3.2 and 2.3.4). The presence of common ions to solutions can decrease the solubility of minerals in solution by driving the precipitation of minerals and removal of common ions from solution (Krauskopf et al., 1995). This occurs by shifting the equilibrium towards the left (reactants) to relive the stress of the excess product (i.e. use up the excess common ions) which in turn reduces the molar solubility of the mineral (< Ksp) and thus promotes the precipitation of the mineral/solute known as the common ion

51 effect. Typically, a reaction without a common ion present has a greater Ksp in contrast to a reaction which has a common ion present (Krauskopf et al., 1995).

2.3 Reservoir processes and scale formation – Solid waste stream

2.3.1 Pipe scale formation

The formation of NORM as scale or sludge is a common problem encountered in the oilfield throughout the water production lifetime of a producing well. Under certain operating conditions and water compositions co-precipitation and accumulation of routine oilfield waste materials such as barite (BaSO4), celestite

(SrSO4), gypsum (CaSO4.2H2O), anhydrite (CaSO4) and calcite (CaCO3) deposit in production or treatment installations (Table 2.3) (Al-Masri et al., 2005; Mccartney, 2008; Abdul et al., 2010; Garner et al., 2015; IOGP, 2016). More importantly under certain conditions NOR’s (mainly isotopes of radium) can co-precipitate, incorporate and concentrate within the bulk mineral matrix as impurities in mineral scales, sludge’s, bio-film deposits, corrosion products and drilling muds classifying them radioactive (Paschoa, 1997; Al-Masri et al., 2005; Garner et al., 2015; IOGP, 2016). The most dominant mineral scales which exist are barite and calcite which form and dominate under different conditions (e.g. temperature and pressure) and at different times during production operations. Reasons for scale formation include pressure and temperature changes (e.g. experienced during transportation of oil and gas from the subsurface), water flooding operations (e.g. injection of sea water during enhanced oil recovery operations, EOR), evaporation (e.g. in fracking ponds and gas extraction piping), variations in flow (e.g. alteration among turbulent and laminar), gas expansion due to a pipeline diameter change, pH changes (e.g. due to injection of sulphate rich fluid or mixture of differing produced waters) and existence of seed crystals on the inner surface of equipment and on irregular surfaces (Vetter, 1972; Mackay, 2005; Badr et al., 2008; Steffan, 2013; IOGP, 2016). The core root for formation of sulphate scale in tubulars (e.g. barite, gypsum,

52 anhydrite and celestite) is mainly due to the mixing of incompatible waters and establishment of supersaturated solutions which is dependent upon the geological formation and its respective produced water composition (section 2.2.1) (Todd et al., 1992; Yuan et al., 1994; Houston, 2007; Rosenberg et al., 2011a, 2014; Rosenberg et al., 2011b; Zhang et al., 2014).

2.3.2 Reservoir processes and sulphate scales

The characteristics of the produced waters are an important factor in determining the type of scale which forms at an oilfield. As the age of the well begins to increase procedures such as the injection of sea water, to help improve the efficiency of oil recovery can cause a change in the composition of the formation water. Compositional change of produced water via water-flooding is caused by the fluid- rock interactions between the injected sea water and rock formation, and fluid-fluid interactions between sea water and formation water. Water injected at the injection well (e.g. seawater) is typically of lower temperature than the reservoir (Fig. 2.4). This water therefore reduces the temperature of the surrounding formation. However, as the seawater is entering the subsurface it is heated resulting in an increase in its temperature and pressure. Scale may precipitate along the well-string if the water is saturated with salts whose solubility decreases with increasing temperature (e.g. gypsum and anhydrite). Mineral precipitation can occur either forward of the mixing zone, in the mixing zone or behind the mixing zone during the injection of water (Fig. 2.4). Oil and reservoir brine are only present in the pore of rocks forward of the mixing zone. In the mixing zone precipitation may possibly occur at local temperature and pressure due to the interaction of dissolved species present in the reservoir brine and injected water. These interactions can lead to dissolution and also substitution events at the rock surface which can in turn under certain temperatures result in the precipitation of anhydrite, barite and celestite scales, and cements dependent upon the reservoir composition (Badr et al., 2008; Puntervold et al., 2008). At a different pressure the remaining clear water moves ahead of the water in the mixing zone and precipitation may occur again. This is continued until clear water reaches the

53 production well (Fig. 2.4). Behind the mixing zone, precipitation results from temperature and pressure changes typically anhydrite and gypsum (Badr et al., 2008). Changes in thermodynamic conditions and decreases in pressure and temperature experienced along the flow string up to the surface in the production well can again promote scale deposition across the production system subject to where supersaturation is established (Fig. 2.4 and 2.9) (Abdul et al., 2010; Steffan, 2013; IOGP, 2016) (Fig. 2.4 and Fig.1.1; Section 1). The content of naturally occurring radionuclides dissolved, mobilised and transported with the waters varies which in turn affects the rate of NORM scale deposition, activity and occurrence. The radionuclide content within waters differs as the mineralogy, sedimentary structure content and textural properties within geological formations vary (see section 2.5) (Todd et al., 1992; Yuan et al., 1994; Hamilton et al., 2004; Godoy et al., 2005; Houston, 2007; IOGP, 2016; Van Sice et al., 2018).

Figure 2.4: Scale formation during enhanced oil recovery operations

Solubility variations with temperature of minerals can also influence precipitation, for example temperature reductions experienced at the well bore due to pressure reduction or inhibition of cool treatment fluids can promote the leaching of calcium sulphates as its solubility increases with decreasing temperature (250 mg L-1 at 140 °C and 2.7 g L-1 at 35 °C). In contrast barite and strontium solubility increases with increasing temperature and can therefore precipitate during temperature and pressure decreases (Badr et al., 2008; Steffan, 2013; IOGP, 2016; Doubra et al., 2017). A decrease in temperature and pressure is experienced as fluid (oil, gas and formation water) migrates from the subsurface towards the surface but, can also occur due to bends in pipes. Bends in pipework can change the direction of flow of the liquid through the pipe due to its curvature which creates centrifugal force that acts upon the fluid. This in turn creates a pressure gradient which results in a spiral

55 flow of the fluid migrating from the centre of the pipe towards the outer wall and back towards the inner wall (Fig. 2.5). This results in a decrease in pressure, temperature and causes scale precipitation, and the potential for NORM deposition (Fig. 2.5) (Idelchik, 1998; Mazumder, 2012; Zardin et al., 2017).

Figure 2.5: A schematic diagram to show the double spiral flow in a bend (www.thermopedia.com/content/577/)

2.3.3 Radionuclide uptake into sulphate scales

The fundamental co-precipitation process of radium into barite has been widely studied over the past century in relation to produced water discharges, scale formation and other areas such as the nuclear sector e.g. in uranium mine tailings (Doerner et al., 1925; Gordon et al., 1957; Beneš et al., 1981; Fedorak et al., 1986; Jerez Vegueria et al., 2002; Rosenberg et al., 2011, 2014; Zhang et al., 2014). Due to presence of aqueous radium ions (Ra2+ (aq)) within the water phase possessing similar properties, such as charge state (i.e. divalent) and ionic radius (1.7 Å) to that of Ba2+ (1.61 Å) and Sr2+ (1.44 Å) ions in solution (Rumble, 1978; Housecroft, 2012; Rosenberg et al., 2014; Zhang et al., 2014), co-precipitation and preferential uptake of radium into recalcitrant barium – and / or strontium – containing scales occurs

(Table 2.4). This results in radiobarite (RaxBa1-xSO4) and radiostrontiobarite

(BaxSryRazSO4) precipitation on the surface of the pipe walls containing naturally

enhanced concentrations of radium between 0.1 to > 15000 Bq g-1 (see section 2.5) (Doerner et al., 1925; Gordon et al., 1957; USGS, 1999; Moghadasi, 2003; Al-Masri et al., 2005; Abdul et al., 2010; Garner et al., 2015; Al Attar et al., 2016; Doyi et al., 2016; IOGP, 2016).

Name Synonym Formula Molecular Specific Hardness Activity weight gravity solubility (g mol-1) (g cm-3) products (Keq) Common scales

Barium sulphate Barite BaSO4 233.4 4.5 3.3 10.0

Calcium carbonate Calcite CaCO3 100.1 2.71 3 8.35

Strontium Celestite SrSO4 183.7 3.96 3 6.5

sulphate Anhydrite CaSO4 136.14 2.96 3 4.5 Calcium sulphate

Calcium sulphate Gypsum CaSO4.2H2O 172.2 2.32 2 4.6 Sodium chloride Halite NaCl 58.4 2.17 2 - 1.58 Sand grains

Silicon Dioxide Quartz SiO2 60.1 2.65 7 3.98 (reference)

Table 2.4: Common mineral scales which dominantly form in the oilfield and their individual chemical properties to exemplify the difficulty of the removal of scales especially barium sulphate due to its insolubility and hardness. Sand is included for reference purposes. Scale of hardness is in accordance to HSAB theory and ranges from 1 (soft) to 10 (hard). Solubility product values are negative logarithms of activity -10 products at 25 °C and 1 bar (pK) e.g. solubility product for BaSO4 = 10 (Ball, 1991; Krauskopf, 1995)

Furthermore, results from many studies show that radium removal via co-

precipitation into binary (Ra-BaSO4 or Ra-SrSO4) or ternary phases (Ra-BaSrSO4) is dictated by the ionic strength of the solution and the nucleation kinetics of the barite mineral (Rosenberg et al., 2011; Rosenberg, Metz et al., 2011). The distribution model has been widely adopted to describe the co-precipitation of radium into barite through radium and barium concentration ratios in the aqueous and solid phases known as the concentration based partition coefficient (Equation 4).

Equation 4: Concentration-based effective partition coefficient; [Ra] represents the aqueous concentration of the element, and dRasolid represents the concentration of the element in the solid (e.g. Ra or Ba) (Rosenberg et al., 2014)

Rapid precipitation kinetics of insoluble barite has been shown to be the controlling factor of increased radium removal in systems (Rosenberg et al., 2014; Zhang et al., 2014). Increases in the ionic strength of solutions have also been shown to lead to increased radium removal reflected by the calculated concentration-based partition coefficient (Kd’) (Meissner et al., 1972; Bokern et al., 2003). Experimental partition coefficients reported for radium uptake into barite in simple analogous systems range between 1.07 – 1.54 in dilute solutions (I = 0 M), and up to 7.49 with increasing ionic strength (I = 3 M) (Rosenberg et al., 2014; Zhang et al., 2014). However, the partition coefficients and radium uptake process into barite under oil field relevant conditions (e.g. during produced water/seawater mixing) is unknown.

The principle mechanisms of radium attenuation into barite via co-precipitation has been previously described as; 1) inclusion; 2) occlusion and; 3) adsorption (Harvey, 2000; Zhang et al., 2014) (Fig. 2.6). Inclusion is the dominant process and refers to the substitution or lattice displacement of a lattice site in the host mineral (e.g. barite, celestite or strontiobarite) by a trace element (guest) with a similar charge state and ionic radius. Occlusion refers to the trapping of a trace element within the crystal during the stage of growth. Such mechanism is deemed to be less prevalent in the precipitation of radiobarite due to the trace concentration of radium isotopes in solution and also due to the low moisture content of barite crystals (Nichols et al., 1941). Adsorption is the mechanism by which the trace element becomes weakly bound to the surface of the host precipitate via an outer sphere adsorption mechanism (Krauskopf, 1995) (Fig. 2.6).

Figure 2.6: A diagram to illustrate the modes of uptake of radium and the three stages involved in the precipitation process (Zhang et al., 2014)

2.3.4 Calcium carbonate scales

The rate of precipitation of calcium carbonate increases with temperature increases and/or pressure decreases compared to that of barium sulphate and strontium sulphate. This is due to the hydrogen bicarbonate content within water which originates from dissolved carbon dioxide. A decrease in pressure results in a reduction in CO2 content in solution and instead favours the gas phase, however upon heating the decomposition of hydrogen bicarbonate favours the formation of carbonate ions in solution (Fig. 2.7). These carbonate ions interact with calcium in solution and precipitate (Fig. 2.7) (Burrows, 2013; Steffan, 2013). In addition the loss of CO2 in solution subsequently reduces the concentration of carbonic acid

(H2CO3) which causes an increase in brine pH (Meyers et al., 1985; Tanner et al., 1986; Payne, 1987; Wat et al., 1992). This increase in pH reduces the solubility and drives the precipitation of calcite. These equilibrium reactions occur simultaneously leading to calcite deposition. Such scale is commonly recognised at early stages of

59 field production due to reduction in pressure during production preceding seawater injection and breakthrough (Badr et al., 2008; Steffan, 2013; IOGP, 2016).

Figure 2.7: Equation to show the formation of calcium carbonate

2.3.5 Radionuclide uptake into calcium-containing scales

Studies show the existence of calcium carbonate and sulphate scales (e.g. CaSO4

226 and CaCO3) with reduced uptake of radium ( Ra) in contrast to barite due to differences in chemistry e.g. ionic radii / volumetric mismatch (see section 2.3.3 and Table 2.4) (Miyake, 1978; Al-Masri et al., 2005; Yoshida et al., 2008; Hedström et al., 2013; Garner et al., 2015). A study of Syrian scales from an oil field show that calcium carbonate and calcium sulphate form when a bicarbonate rich formation water was present (Al-Masri et al., 2005). These tend to have much lower radium activity concentrations (e.g. exempt) than those of barite scales (Al-Masri et al., 2005). Al-Masri et al., (2005) showed that 226Ra is highly correlated with barium and strontium content in scales (R2 = 0.92) than calcium in which no linear correlation was found (Al-Masri et al., 2005). The reason stated for this is due to the difference in the chemistry of the elements (Table 2.4). Radium has an ionic radius of 1.7 Å and barium has an ionic radius of 1.61 Å which is close to that of radium. However, calcium has a smaller ionic radius of 1.34 Å which is significantly different and therefore, it is very difficult for the radium to be taken up into the crystal structure (Shannon, 1976; Miyake, 1978; Al-Masri et al., 2005; Yoshida et al., 2008; Hedström et al., 2013; Garner et al., 2015). This is further reflected by the calculated partition coefficient for radium (0.15 ± 0.06) and barium (0.016 ± 0.011) uptake into calcite which confirms the reduced uptake of radium into calcite minerals (Yoshida et al., 2008).

2.3.6 Black dust deposits

Other types of scale have been identified in oil and gas production processes including lead-, zinc- and iron-containing scale deposits and corrosion products, for example lead sulphide (PbS; galena), lead-iron-sulphide (PbFeS; ‘black dust’), iron sulphide, iron oxides, iron carbonates (e.g. Fe3O4, FeS, FeS2, FeCO3 and magnetite) and zinc sulphide (ZnS) on surfaces of equipment (e.g. tubulars, inlets, gas transport systems and pumps) (Table 2.3 and 2.5) (Curti, 1999; Worden et al., 2000; Hartog et al., 2002; SNIFFER, 2003; Godoy et al., 2005; Badr et al., 2008; Trifilieff et al., 2009; Abdul et al., 2010). The mixing of lead, iron or zinc rich formation waters with hydrogen sulphide gas generally result in the formation of sulphide scales (e.g. galena, sphalerite and pyrite) (Badr et al., 2008). In contrast oxygen or bicarbonate rich waters commonly result in the formation of various types of iron oxides and carbonates (e.g. siderite, FeCO3, Fe(OH)2, Fe3O4 and Fe2O3) (Hartog et al., 2002; Steffan, 2013; Zhang et al., 2016). The formation of lead scale known as ‘black dust’ can accumulate as very thin films (composed of sub-micron size particles) to the inner surface of gas lines, natural gas processing plants and production equipment such as tanks, separators and tubes (Table 2.5 and Fig. 2.9) (Worden et al., 2000; Hartog et al., 2002; Al-Masri et al., 2005; Abdul et al., 2010; Steffan, 2013; Garner et al., 2015; IOGP, 2016).

Type Radionuclides Occurrence Ra scales 226Ra, 228Ra, 224Ra and their Wet parts of production decay products installations consisting of solid group II sulphate and carbonate scales Ra sludge 226Ra, 228Ra, 224Ra and their Separators, skimmer tanks, decay products waste pit, storage tank Pb deposits 210Pb and its decay products Wet parts of gas production installations Pb films 210Pb and its decay products Oil and gas treatment and transport Po films 210Po Condensates treatment facilities Condensates 210Po (210Pb) Gas production Natural gas 222Rn, (210Pb, 210Po as decay Gas treatment and storage products) and transport systems plated on surfaces Produced water 226Ra, 228Ra, 224Ra, 210Pb All production facilities and disposal facilities in large volumes

Table 2.5: The different types of NORM in oilfields and the component radionuclides within the materials and their occurrence during production (Abdul et al., 2010; Steffan, 2013)

2.3.7 Radionuclide uptake into black dust deposits

Black dust is a powder composed of a variety of corrosion contaminants, mostly iron oxides (hematite; Fe2O3 and magnetite; Fe3O4), iron sulphides alongside sand, silt, glycol, long hydrocarbons (asphaltenes), galena and lead (hydr)oxides (Hartog et al., 2002; Steffan, 2013). Scales and sludges containing lead (210Pb) are commonly associated with the production of gas and originate from two different sources. Firstly, from the in situ radioactive decay of 226Ra to the noble gas radon (222Rn) which is known as a ‘supported mechanism’ as it originates from the direct

62 emanation from the parent radionuclide uranium (238U) situated in the reservoir rock (Fig. 2.2 and Fig. 1.1; Section 1). Radon gas can travel in condensates to reach the gas retention equipment and reside in the installations over a long period. It is carried from the reservoir with natural gas and condensate resulting in lead (210Pb) contaminating large areas during its course (Fig. 1.1 & 2.9 and Table 2.5 &2.6) (Al- Masri et al., 2005; Badr et al., 2008; Abdul et al., 2010; Steffan, 2013). Secondly,

lead can be brought to surface as lead chloride (PbCl2) within formation water due to a separate (‘unsupported’) mechanism which commonly results in the formation of radioactive sulphide deposits (e.g. FeS, PbS or ZnS) (Hartog et al., 2002; Trifilieff et al., 2009; IOGP, 2016) most regularly found on equipment which have been in contact with produced water (Worden et al., 2000; Garner et al., 2015). This process can occur in wells rich in hydrogen sulphide (sour wells) where a sudden drop in pressure or temperature results in the establishment of a supersaturated solution. Here lead cations (Pb2+) and sulphide anions (S2-) precipitate. Deposits are also encountered in ‘dry’ gas wells suggesting the transport of lead under water- free conditions (Hartog et al., 2002). Additionally, lead (210Pb) can also originate from the radiative decay of radium in radio-barite scale precipitated on the inner surface of pipes (‘unsupported mechanism’). Activity concentrations for isotopes of bismuth (214Bi and 210Bi) and lead (214Pb and 210Pb) can also be detected as a result of the radiative decay of radon gas encountered in gas treatment and storage systems (Table 2.5 and 2.6) (Garner et al., 2015). The understanding for the formation of thin films of lead scale on the surface of equipment is described by an electrochemical process and differences in chemical properties and interactions between lead and iron coated surfaces (Hartog et al., 2002; Steffan, 2013). As the chemical properties of lead make it more resistant to corrosion and oxidation (more noble) compared to that of iron, contact of lead with iron coated surfaces along parts of the processing system (e.g. ferrous (Fe2+) steel materials) result in spontaneous and favourable ion exchange of lead and iron which leads to the formation of lead scale and corrosion of the pipe wall (e.g. lead-iron-sulphide). Lead in solution oxidises iron from the tube wall, resulting in corrosion and accumulation of lead scales greater than 10 mm (Hartog et al., 2002; SNIFFER, 2003; Jose Marcus

Godoy et al., 2005; Trifilieff and Wines, 2009; Abdul and Zaidan, 2010; Steffan, 2013) (Fig. 2.8)

Corrosion of pipe lines can also be microbial induced, where sulphate reducing bacteria colonize on pipe walls (Fig. 2.8). Here bacteria consume sulphate resulting in the production of corrosive hydrogen sulphide and carbon dioxide lowering the pH of solution resulting in iron oxidation, dissolution and production of iron sulphides aiding corrosion. Weathering of a pipe exposed to water and oxygen can also result in the corrosion of iron surfaces and formation of rust e.g. iron

2+ hydroxides. This occurs via the oxidation of iron (Fe(s) to Fe (aq)), reduction of

2+ oxygen (O2(g) to H2O(l)) and the reaction of Fe ions with oxygen resulting in hydrated iron(III)oxide (Fig. 2.8) (Trifilieff et al., 2009; Khan et al., 2015).

2.3.8 Sludges

Sludges consist of liquid and paste-like materials encompassing solid precipitates due to the high hydrocarbon content. The oil-wet fine mineral grains are commonly sulphates and carbonates originating from the same processes leading the formation of hard scales (Table 2.5). Such materials are generally less active than solid scale materials as they contain reduced amounts of solid NORM e.g. scrapings and scales (see Table 2.6 and 2.8 in section 2.5).

2.3.9 Radionuclide uptake into sludges

Due to components such as the heavier hydrocarbons and sand falling out of solution (waxy material) this prevents such material from travelling long distances in the water column reducing the incorporation of radium and total activity (see section 2.5 and Table 2.8). These are commonly encountered in waste pits, NGL storage tanks, tank bottoms, gas/oil separators and dehydration vessels (Table 2.5 and 2.6) (Steffan, 2013; Garner et al., 2015; IOGP, 2016).

B) Water

H2SO4

H2S O SO4 2 H O 2 - 2- - 2- Iron (steel pipe) SRB + Corg HS ,S + HCO3 , CO3 Cathodic site

Anodic site 2+ - Fe e FeS + FeCO3

2+ Fe Paint layer

surfaces (Corg represents carbon source e.g. CH2O or CH4) (Steffan, 2013)

2.3.9.1 Scale Prevention

Studies have shown barium to be the dominant component observed in sulphate scales obtained from oil field operations around the world despite stratigraphical differences in reservoir lithologies (Warren et al., 1993; Balson et al., 2002; Al-Masri et al., 2005; Bader, 2006; Houston, 2007; Moghadasi et al., 2007; Badr et al., 2008; Rowan, 2011; Gregory et al. 2011; Zhang et al., 2014; Garner et al., 2015; Al Attar et al., 2016; Hu et al., 2016). Due to the recalcitrant nature of barite, and resistance to acid leaching, removal of barite is most challenging and costly in the oilfield in contrast to carbonate scales (Table 2.4) (Fakhru’l-Razi et al., 2009; Antony et al., 2011; Palanisamy et al., 2017). The high thermodynamic stability of barium sulphate due to similarities in chemical properties (ionic radius and charge state)

2+ 2- between barium (Ba ) and sulphate (SO4 ), results in the characteristic hardness (3.3 Mohs’s) and low solubility leading to the difficult removal of such scales when formed and the adoption of inhibition strategies (Table 2.4) (Rumble, 1978; Rosenberg et al., 2014; Zhang et al., 2014). The most efficient and common methods adopted to prevent or delay the formation of carbonate and sulphate scales are scale inhibitors and coatings (Vazirian, 2016). Usually either, phosphonates (e.g. hexaphosphonate (PO-85), poly-phosphono carboxylic acid (PPCA) and penta-phosphonate (DETPMP)) or polymeric scale inhibitors (e.g. phosphino-polycarboxylates (PPCA), polyacrylate (PAA) and polyvinyl-sulphonates) (Bezemer et al., 1969; Vetter, 1972, 1976; Mitchell et al., 1980; Laing et al., 2003; Fried et al., 2012; Shi et al., 2013; Li et al., 2017). Characteristics of selected inhibitors must include, sufficient interaction with the reservoir thus provide long inhibition life times, be fairly stable to thermal degradation, highly soluble and compatible in brine systems to prevent premature precipitation, and achieve inhibition at low concentrations. The scale inhibitors function by sequestering ions in solution via complexation to functional groups (e.g. chelating groups) which possess variable divalent ion compatibility. (Shakkthivel et al., 2006; Amjad et al., 2014; Kelland, 2014; Li et al., 2017).

Figure 2. 9: A diagram to show areas in which NORM can build up in the offshore production system indicated by the radioactive sign; (key: orange = oil, blue = water and purple = gas) (IOGP, 2016)

2.4 Operational discharge of effluent waters to surface waters and NORM formation

2.4.1 Offshore discharges

The presence of naturally occurring radionuclides such as 226Ra and 228Ra, from the decay of naturally occurring 238U and 232Th have proven to be significant in produced water effluents from oil and gas platforms across the world (see Table 2.9 in section 2.5) (Fisher, 1998; Røe Utvik, 1999; IOGP, 2016). The release of produced water brines via discharging to surface waters from disposal sites (e.g. offshore oil and gas platforms) can result in elevated levels of Ra in sediment and surface waters to be identified in proximity to the discharge site (Landa et al., 1983; Pardue et al., 1998; Holdway, 2002; Jerez Vegueria et al., 2002; Grung et al., 2009; Dowdall et al., 2012; Bakke et al., 2013; Van Sice et al., 2018; McDevitt et al., 2019). Studies show the total activity of radium discharged globally from production sites can be

67 significant with levels up to 1200 Bq L-1 reported across installations (IOGP, 2016). Discharges from Norwegian sites are typically of the order of 306 – 480 x 109 Bq y-1 with an average radium activity concentration of 3.3 Bq L-1 amounting to 134 - 149 million m3 of produced water discharged annually. Eriksen et al., (2006) showed the concentration of 226Ra discharged from North Sea platforms varies from 0.5 - 1 Bq L- 1 to 21 Bq L-1 (Eriksen et al., 2006; Olsvik et al., 2012; Bakke et al., 2013). Jerez Vegueria et al., (2002) showed offshore discharges from platforms in Brazil are on the order of 2 - 30 m3 d-1 with a varying 226Ra concentration between 0.012 – 6 Bq L- 1 (Jerez Vegueria et al., 2002). These studies illustrate discharge volumes and levels of activity vary considerably globally with up to 37 Bq L-1 reported from American platforms (Pardue et al., 1998). Background radium concentrations within surrounding seawaters are typically around three orders of magnitude lower than that in produced waters (e.g. 0.01 – 0.03 Bq L-1) (Jerez Vegueria et al., 2002; Gafvert et al., 2007; Dowdall et al., 2012).

2.4.2 NORM formation during operational discharge

Sequestration of radium via sorption to mineral surfaces (e.g. clay minerals), existing particulates in the water column and/or to sediment, and co-precipitation within mineral phases (e.g. barite) have previously been shown to be significant mechanisms controlling its environmental speciation, mobility and bioavailability (Pardue et al., 1998; Sajih et al., 2014; Zhang et al., 2014; Siddeeg, et al., 2015; Van Sice et al., 2018; Neff, 2002; Fakhru’l-Razi et al., 2009; Grung et al., 2009). The environmental risk associated and impact posed by radium once discharged to marine waters is dependent upon its speciation in the natural environment and resultant mobility. Due to the similar chemical properties of radium with other group II alkaline earth metals (e.g. Ba and Sr) and other divalent or trivalent cations (e.g. Fe2+/3+) it is expected to: 1) adsorb onto the surfaces of mineral phases or 2) co-precipitate with sulphate and carbonate mineral phases predominantly barium- or strontium-containing minerals (e.g. BaSO4, SrSO4, BaCO3, SrCO3, CaCO3) once discharged, which can result in elevated activities of radium in receiving waters and sediments (Ames et al., 1983; Langmuir et al., 1985; Pardue et al., 1998; Gonneea

68 et al., 2006; Gonneea et al., 2008; Zhang et al., 2014). The key process of formation of such radium-containing inorganic particulates (e.g. RaxBa1-xSO4 and BaxSryRazSO4) is the mixing of chemically incompatible waters originating from the same processes leading the formation of hard scales during water flooding operations as previously described in section 2.3 (Gafvert et al., 2007; Abdul et al., 2010; Candeias et al., 2014; Zhang et al., 2014; Garner et al., 2015). An array of surfaces exist to which radium in the aqueous form (e.g. Ra2+ ion) can sorb, including clays (muscovite, illite, and kaolinite), iron oxides (goethite and ferrihydrite), manganese oxides (birnessite, manganite and todrokite) , carbonates (siderite, dolomite and magnesite) and organic matter (Taskayev et al., 1978; Hanan et al., 1981; Landa et al., 1983; Langmuir et al., 1985; Jones et al., 2011; Sajih et al., 2014; Siddeeg et al., 2015; Van Sice et al., 2018). Sajih et al., (2014) showed that radium adsorbs to iron oxide minerals such as goethite and ferrihydrite. Landa et al., (1983) showed radium sorption to clay minerals during produced water discharge was well correlated (r = 0.74) with the clay content and also the sulphate-bearing minerals. As iron phases are typically more abundant in natural sediment systems in comparison to manganese-bearing minerals, radium sorption to iron-bearing minerals generally occurs to a greater extent in sediments (Nirdosh et al., 1990; Gonneea et al., 2008). However, the tendency of sequestration via adsorption by manganese and iron oxides is highly dependent upon the environmental conditions e.g. the acidity, composition and resultant supersaturation of the receiving water in respect to solid-phases such as barite (Appelo, 2005; Van Sice et al., 2018). Carbonate minerals also have the potential to adsorb radium including siderite, dolomite and magnesite (McDevitt et al., 2019). As well as solely existing as the Ra2+

0 0 aqueous ion, radium may also form complexes in solution such as RaSO4 , RaCl2 ,

0 + RaCO3 and RaOH , and adsorb to organic matter (International Atomic Energy Agency, 2014). Studies have shown the sorption of radium by sediment and mineral surfaces is generally attributed to an outer sphere surface adsorption mechanism in which the kinetics of sorption can vary depending on the pH and salinities (ionic strength) of the receiving waters. These factors can potentially enhance or reduce the mobility of radium in surface waters thus the resultant enrichment zone as a result of surface site competition between ions in solution (Landa et al., 1983; Jones

69 et al., 2011; Zhang et al., 2014; Van Sice et al., 2018; McDevitt et al., 2019). For example, Landa et al., (1983) showed that the areal zone of sorption of radium by sediment during the discharge of produced water varied depending upon the ionic strength of the produced water. They determined that the sorption was the highest for solutions diluted with water (> 80%) in contrast to sodium chloride (> 40%). This is due to less competition between ions in solution for sorption sites which increases the probability of radium-build up at the vicinity of the discharge site which can be dictated by the characteristics of the receiving environment (e.g. surface water composition) (see below).

2.4.3 Environmental characteristics

Dependent upon the characteristics of the receiving environment (e.g. stream (fresh water), river (brackish water) or ocean (brine water)) radium can potentially incorporate into a range of mineral phases (Landa et al., 1983; Pardue et al., 1998; Jerez Vegueria et al., 2002; Van Sice et al., 2018; McDevitt et al., 2019). Which sequestration mechanism(s) dominate when produced water is released into marine or freshwater environments highly depends on the characteristics of the receiving environment, where generally adsorption and precipitation occur simultaneously to some degree. In high salinity systems (e.g. marine environments) previous studies have shown via sequential leaching of sediments surrounding a discharge point, that radium dominantly co-precipitates with barite and this controls the solubility of radium. Here the vast proportion of radium associates with the recalcitrant mineral fraction of sediments typically in the vicinity of the discharge site due to the high density of barite (4.5 g cm-3) which results in the deposition of particulates (Landa et al., 1983; Pardue et al., 1998; Jerez Vegueria et al. 2002; Appelo, 2005; Van Sice et al., 2018; McDevitt et al., 2019). Studies investigating the speciation and association of radium and barium within different geochemical fractions in sediment samples from stream systems of lower salinity (e.g. freshwater) associated with produced water discharge identified attenuation of radium in the carbonate fraction (e.g. witherite (BaCO3), strontianite (SrCO3) and calcite (CaCO3)) and also the more available labile forms (e.g. exchangeable)

70 confirming adsorption and diversity in attenuation mechanisms. This is due to the lower levels of sulphate in the receiving waters and undersaturation of the solution in respect to solid-phase barite which instead allows alternative mechanisms for radium attenuation in the environment to dominate (Hanan et al., 1981; Landa et al., 1983; Pardue et al., 1998; Van Sice et al., 2018; McDevitt et al., 2019). For example Van Sice et al. (2018), showed radium attenuation in stream sediments near the point of discharge of unconventional waste waters in Pennsylvania (USA) were up to 24.6 ± 0.74 Bq g-1 (hundreds of times higher than background). This was deemed to be associated with the solid recalcitrant fraction speculating the likely association with insoluble sulphate minerals (e.g. BaSO4 and BaSrSO4) from sequential leaching experiments and geochemical modelling. They also identified a portion of radium (~ 1.5 x background) exists in the more available forms (e.g. exchangeable) up to 31 km downstream of the discharge point. In a similar study McDevitt et al. (2019), showed radium activities of stream sediments over 30 km downstream ranged from 0.2 – 3.6 Bq g-1 with preferential association of radium with carbonate minerals (e.g. calcite) and other more labile forms/readily- exchangeable forms (e.g. adsorbed to surfaces and in interlayer sites of clays). Both studies demonstrated a greater portion of radium downstream exists in the more labile forms (e.g. as sediment-associated and aqueous species) in comparison to upstream sediment (mineral-associated). Such findings are a result of current flow effects, the low salinity of freshwater systems and the greater clay content found further downstream to the outfall leading to enhanced mobility of radium in the water column (Hanan et al., 1981; Landa et al., 1983; Pardue et al., 1998; Lauer et al., 2018; Van Sice et al., 2018; McDevitt et al., 2019). As a result, subsequent attenuation of radium via sorption dominates in contrast to precipitation and deposition which dominantly occurs at the immediate vicinity of the discharge as soon as the two waters mix. Other studies have identified the presence of radium in core samples downstream from the produced water discharge point as a result of the successive deposition of sediments and the dilution and dispersive effects of aqueous radium (Dowdall et al., 2012; Van Sice et al., 2018; McDevitt et al., 2019). Dowdall et al. (2012), showed between 22 – 35 Bq kg-1 of 226Ra in the uppermost

71 layers of sediment cores (4 – 6 cm) taken from the North Sea, UK region may result from marine discharges.

Studies related to marine discharges from offshore platforms are limited and have identified the non-existence of elevated levels of radium and barium in sediment or seawater around platforms as a result of dispersive effects by currents (Jerez Vegueria et al., 2002; Gafvert et al., 2007). Work by Jerez Vegueria et al. (2002), revealed the non-existence of significant radium contamination (above background levels) within sediment or seawater samples around platforms (250 – 1000 m) during the discharge of produced waters (12 Bq L-1) as a result of dispersive effects by currents (Jerez Vegueria et al., 2002). By contrast possible accumulation of radiobarite has been identified in other studies inferring, that under certain conditions discharge of effluents may result in barite precipitation and subsequent sedimentation. One study by Pardue et al. (1998), identified radium activities between 21.5 – 33.7 Bq g-1 in waste pit sediments at the immediate vicinity of a produced water discharge outfall in Louisiana (USA), deducing 226Ra ‘hotspots’ near the outfall are strongly connected to BaxRaySO4 minerals. Correlations between radium concentrations and the identification of barite via XRD in this study is suggestive of this mechanism however, to date NORM particles have not been extracted and directly characterised to provide evidence of radiobarite in either experimental or field based studies (Pardue et al., 1998).

Contrast in findings from different sites suggest that the environmental setting and characteristics (e.g. water depth, salinity and mineralogical distribution differences which exist between deep sea and estuarine/shallow marine settings) of the receiving environments are significant in the mechanisms of radium interactions and thus the positive identification of radium accumulation in sediments at distances from a discharge point (Landa et al., 1983; Pardue et al., 1998; Jerez Vegueria et al., 2002; Gafvert et al., 2007; Rosenberg et al., 2011, 2014; Rosenberg, Metz and Ganor, 2011; Dowdall et al., 2012; Zhang et al., 2014; Van Sice et al., 2018; McDevitt et al., 2019)

2.5 Radioactivity of NORM

In agreement with the International Atomic Energy Agency (IAEA), the International Commission on Radiological Protection (ICRP) and European Commission, any natural radioactive material containing radionuclides in the uranium and thorium decay series with an activity concentration above 1 Bq g-1 is defined as radioactive and therefore NORM (IAEA, 2004; ICRP, 2007; European Commision, 2014).

2.5.1 Solid waste stream – Pipe scale and Sludge

The radioactivity concentrations of mineral scale deposits vary between locations (Table 2.6 & 2.8) and sample analysis needs to be performed to assess the radiological risk they may present, as quantity is not an indicator of the radioactivity concentrations samples may demonstrate.

Nation Deposit 226Ra (Bq g-1) 228Ra (Bq g-1) Algeria (2001) Scale 1-950 n.r. Algeria (2001) Sludge 0.069-0.393 n.r. Australia (1998) Scale/Sludge 20-70 n.r. Brazil (2003) Sludge 0.36-367 0.25-343 Brazil (2003) Scale 19.1-323 4.21-235 Brazil (2005) Scale 109.6 – 3500 133.8 - 2195 Egypt (2001) Sludge 18 13.25 Kazakhstan (2005) Scale 5.10 – 51 0.20 - 10 Norway (1997) Scale 0.3-32.3 0.3-33.5 Norway (1997) Sludge 0.1-0.47 0.1-4.6 Syria (2005) Scale 0.3-1520 0.6-868 Syria (2003) Sludge 0.470 -1 0.359 – 0.660 UK (1987) Scale 1-1000 n.r. UK North Sea (1998 Scale 0.66-300 n.r. UK East Midlands Scale 48.5-154.6 16.4-61.6 (2015) UK East Midlands Sludge 0.39-32.6 0.32-9.01 (2015) USA (2001) Scale 15.4-76.1 n.r.

Table 2.6: Radioactivity concentrations of radium isotopes within NORM taken from oilfields across the globe (n.r. = data not recorded) (Hamlat et al., 2001; Garner et al., 2015; Doyi et al., 2016)

Generally sludge has lower activity concentrations of radium isotopes than that of scale deposits, whereas the opposite relates to lead deposits which have a higher concentration than sludge as they are found in tank bottoms, natural gas liquid (NGL) storage tanks and gas oil separators. Additionally, radioactivity can also be detected from thorium isotopes (228Th) through the radiative decay of 228Ra in aged scale and sludge (Fig. 2.2). Mercury (Hg) can also accumulate within material and is commonly found within vessels and knockout drums. Mercury contaminants are found in areas where radioactive lead (210Pb) impurities have built-up and exist, which results in higher radiation exposure levels and health risks (Abdul et al., 2010; IOGP, 2016).

The radioactive gas radon (222Rn) which has a very short half-life (~ 4 days) is found in natural gas liquid (NGL) processing installations, natural gas (NG) transmission lines or in the gas cap of crude oil storage tanks and travels across the gas-water stream (Fig. 1.1 and 2.9). However, high radiation levels are rarely measured and the radiative decay of radon to lead (210Pb) and polonium (210Po) is more problematic due to their long half-life and tendency to form thin lead scales (Table 2.5 & 2.8 and Section 2.3) (Summerlin et al., 1985; Hartog et al., 2002; Trifilieff et al., 2009; Abdul et al., 2010; Ojovan, 2019). Recorded activity concentrations of 210Pb in black dust material have been reported to be around 0.04 - 4.9 kBq kg-1 (Godoy et al., 2005).

Radionuclide Natural gas (NG) Natural gas liquid Crude oil Bq m-3 (NGL) Bq g-1 Bq L-1 238U n.r. n.r. 0.0000001 - 0.01 232Th n.r. n.r. 0.000003 - 0.002 228Th n.r. n.r. n.r. 228Ra n.r. n.r. n.r. 226Ra n.r. n.r. 0.0001 - 0.04 224Ra n.r. n.r. n.r. 222Rn 5 - 200,000 0.01 - 1,500 n.r. 210Pb 0.005 - 0.02 0.3 - 230 n.r. 210Po 0.002 - 0.08 0.3 - 100 0 - 0.01

Table 2.7: A comparison of the radioactivity concentrations for crude oil, NG and NGL (Jonkers et al., 1997; Abdul et al., 2010; Doyi et al., 2016; IOGP, 2016)

Radionuclide Scale Sludge Deposits Scrapings Bq g-1 Bq g-1 Bq g-1 Bq g-1 238U 0.001 - 0.5 0.005 - 0.01 0.001 - 0.05 n.r. 232Th 0.001 - 0.002 0.002 - 0.01 0.001 - 0.07 n.r. 228Th n.r. n.r. n.r. n.r. 228Ra 0.05 - 2,800 0.5 - 50 0.05 - 300 0.01 - 10 226Ra 0.1 - 15,000 0.05 - 800 0.8 - 400 0.01 - 75 224Ra n.r. n.r. n.r. n.r. 222Rn n.r. n.r. n.r. n.r. 210Pb 0.02 - 75 0.1 - 1,300 0.05 - 2,000 0.05 - 50 210Po 0.02 - 1.5 0.004 - 160 n.r. 0.1 - 4

Garner et al. (2018), characterised scale and sludge samples from the East Midlands, UK which contained high levels of radium and other radioactive radionuclides (214Pb, 212Pb, 210Pb and 214Bi) associated with inorganic strontiobarite. Results showed a positive correlation between the concentration of strontium and barium, and activity of radium within samples as expected. The bulk composition of sludge materials composed of baritocelestite (Ba < Sr) possessed lower activity concentrations of radium in contrast to strontiobarite (Ba > Sr) pipe scales due to the ionic radii compatibility between barium and radium compared to strontium and radium as expected (see section 2.3). In contrast calcium-rich samples showed no relationship with activity in agreeance with other studies conducted in the North Sea (Heaton et al., 1995) and other global locations which show restricted uptake of radium into calcite (Al-Masri et al., 2005). Similarly, Al-Masri et al. (2005) identified positive correlations between Ba and Ra in scale samples in contrast to Ca and Ra due to the ionic radii compatibility between ions (see section 2.3) (Al-Masri et al., 2005; Zhang et al., 2014).

2.5.2 Liquid waste stream – Produced water

NOR activity concentrations in produced waters differ between facilities due to variances in geological formations, life time of the well and operational conditions. Activity concentrations in produced water effluents are generally many orders of magnitude lower in contrast to concentrations observed in materials formed post discharge and accumulation (e.g. NORM scale) (Table 2. 9 vs. Table 2.6 and 2.8). NOR activity concentrations in waste waters from unconventional reservoirs developed by hydraulic fracturing (‘fracking’) have received great interest over the years. Reported ranges of radionuclide activities for unconventional wells are commonly lower (e.g. 226Ra = 0.2 – 660 Bq L-1) but within the ranges reported for conventional wells, though conditions leading to their accumulation are the same (Table 2.9) (Rowan, 2011; IOGP, 2016; Lauer et al., 2018). Unconventional reservoirs characteristically have greater 238U/232Th parent ratios resulting in lower 228Ra/226Ra ratios (median = 0.19) in produced waters in comparison to conventional formations and waters (median = 1.04) (Table 2.9) (Warner et al., 2013; Hladik et al., 2014; Harkness et al., 2015; Lauer et al., 2018; Van Sice et al., 2018). Characteristic differences observed in 228Ra/226Ra ratios can provide a suitable tool in determining the source of waste waters which are otherwise similar in terms of major element chemistry, thus sources of contamination at disposal sites can be determined (Rowan, 2011; Chapman, 2013; Haluszczak et al., 2013; Van Sice et al., 2018). During the initial hour to weeks of hydraulic fracturing produced water of low activity are extracted. Subsequent increases in activity, result from the equilibration between NOR’s (e.g. radium) present in the formation (e.g. in void spaces or adsorbed to surfaces of mineral) and injected water resulting in naturally enhanced concentrations of 226Ra, 228Ra and 210Pb (Robertson et al., 2017; Rowan, 2011; IOGP, 2016).

Radionuclide Produced water Produced water (conventional) (unconventional) Bq L-1 Bq L-1 238U 0.0003 - 0.001 0.2 - 1.5 232Th 0.0003 - 0.1 n.r. 228Th 0.5 - 40 0.3 - 7.8 228Ra 0.3 - 180 1.4 - 41.6 226Ra 0.002 - 1,200 0.2- 660 224Ra 0.5 - 40 0.3 - 7.8 222Rn n.r. n.r. 210Pb 0.05 - 190 0.5 - 2.6 210Po n.r. n.r.

Table 2.9: A comparison of the radioactivity concentrations for produced water (Rowan, 2011; IOGP, 2016; Lauer et al., 2018)

2.6 Influence of natural processes and microorganisms on barium and radium remobilisation

During the transportation and persistent deposition or burial of sediment containing NORM following produced water discharges, microbial activity at particular depths in sediments possess the potential to influence the mobility, bioavailability and fate of radionuclides via microbial induced reduction processes (Pardue et al., 1998; VanLoon, 2000; Phillips et al., 2001; Keith-Roach, 2002; Ouyang et al., 2017). Consumption and degradation of organic matter by microbial activity coupled to various terminal-electron-accepting-processes (TEAP’s), can result in the establishment of anoxia in sediments and development of reducing conditions. Which reduction process predominates in sediments is controlled by the thermodynamic free-energy yield of the metabolic process and the availability of the terminal electron accepting process (Fig. 2.10). As anoxia develops with depth less energy efficient biogeochemical zones of differing TEAP’s are successively encountered (e.g. nitrate, manganese, iron and sulphate reduction) which allows the speciation and mobility of radionuclides to be examined across different zones (Fig. 2.10). This is because characteristic sequential chemical changes in the sediment pore water follow as electron acceptors are consumed and reduced

78 products form (Keith-Roach, 2002; Konhauser et al., 2002; Burke et al., 2005; Newsome et al., 2017).

Biogeochemical Gibbs free-energy change per mole of acetate zone (TEAP) consumed as electron donor (KJ mol-1)

Aerobic -856 Decrease in Respiration energy yield Nitrate -806 per mole of organic Manganese -569 matter Iron -361 consumed

Sulphate -48

2.6.1 Sulphate-reducing bacteria and their effect on barite

Bioreduction processes stimulated in natural sediment systems including Fe(III)- reduction and sulphate-reduction have suggested that remobilisation and increased solubility of radium and radiobarite may occur as reducing conditions develop (Fedorak et al., 1986; Pardue et al., 1998; Phillips et al., 2001). Whilst there have been studies on terrestrial discharges of radium in engineered and natural settings (Bolze et al., 1974; Fedorak et al., 1986; Baldi et al., 1996; Phillips et al., 2001; Wilkins et al., 2007; Luptáková et al., 2015; Ouyang et al., 2017), there are limited studies that focus upon marine discharges of oil produced water (Pardue et al., 1998). For example, Baldi et al., (1974), demonstrated the release of barium (1.2 mg L-1) from barite in sewage sludge’s using cultures of Desulfovibrio desulfuricans. Similarly, Fedorak et al., (1986) showed the concurrent release of radium and sulphide from radiobarite sludge’s from uranium mine wastes (700 - 1300 Bq g-1) under sulphate reducing conditions. However, they only detected the release of

79 radium (150 - 400 Bq L-1) and barium (0.2 mmol) using concentrations of lactate (2500 and 4000 mg L-1) unlikely to be found in the environment. Luptakova et al., (2015), also confirmed the microbial reduction of barite from an acid mine drainage site by cultures of Desulfovibrio. They detected ten times the quantity of barium in inoculated experiments (2.96 mg L-1) in comparison to un-inoculated control experiments (0.35 mg L-1). The principal mechanism resulting in the potential release of radium from barite scale and sludge is thought to be sulphate reduction by sulphate reducing bacteria (SRB) from the genus Desulfovibiro sp. and Desulfobacterium (Fig. 2.11) (Robert, 1981; Fedorak et al., 1986; Phillips et al., 2001). Here, sulphate reduction is implicated in accessing sulphate within barite precipitates potentially releasing barium and analogous species to the environment (Fig. 2.11) (Carbonell et al., 1999; Konhauser et al., 2002). This has been demonstrated by the release of barium and sulphide following the dissolution of barite and decrease in sulphate under anoxia in comparison to control studies performed under aerobic (oxidised) conditions as described (Bolze et al., 1974; McCready et al., 1980; Baldi et al., 1996; Pardue et al., 1998; Luptáková et al., 2015; Ouyang et al., 2017).

BaCO BaS 3

2+ 2+ - 2- - 2- Ra + Ba + HS , S + HCO3 , CO3

Sediment-water

Interface SRB + Corg

RaxBaySrzSO4

Figure 2.11: Illustration of the potential reduction of sulphate in barite via sulphate reducing bacteria (SRB) in sediment (Corg represents carbon source e.g. CH2O or CH4)

For radium incorporated in radiobarite, and existing as Ra2+(aq), few biogeochemical studies have been reported relating to produced water discharges. Pardue et al. (1998), demonstrated sulphate reduction can potentially lead to the dissolution of radiobarite in brackish and salt water systems. They detected high levels of sulphide and barium but only trace levels of radium (0.063 ± 0.01 Bq g-1) in overlying pore waters in sediment-microcosms containing contaminated field sediment from a discharge site. Radium release was thought to be inhibited by re- sequestration via co-precipitation into newly formed barite as the radium was released from sediments, as a result of high sulphate concentrations above the sediment-water interface. Other studies for example Phillips et al. (2001), demonstrated minimal radium release (< 0.1%) from oil-field barite scale using sulphate-reducing cultures isolated from oil field brine pond sediment. Factors such as crystal size, retention mechanisms, surface area and nature of the materials (e.g. different processes of formation) were identified as possible variables effecting the amount of radium released and the non-stoichiometric release of barium, sulphide and radium to solution in agreement with other studies (Bolze et al., 1974; McCready et al., 1980; Fedorak et al., 1986; Baldi et al., 1996; Wilkins et al., 2007). In particular radium and barium retention mechanisms such as re-adsorption and inclusion into newly formed mineral phases (e.g. barium carbonate (BaCO3) and barium sulphide (BaS)) have been suggested as possible reasons for restricted release of radium and barium (Bolze et al., 1974; Baldi et al., 1996; Phillips et al., 2001; Luptáková et al., 2015). For example, Wilkins et al. (2007), did not detect desorption of Ra2+ from sediments where Ra2+ had been sorbed under oxic conditions as Fe(III)-reduction developed. This was in contrast to Landa et al. (1991) who detected ~ 3% release of radium from uranium mine tailing solids, due to the suspected re-sequestration of radium into newly formed mineral phases following

2+ Fe(III)-reduction (e.g. siderite (FeCO3) and vivianite (Fe 3(PO4)2.8H2O)). They further postulated that radium was perhaps not bound to bioavailable iron oxides in the sediments. Similarly, Baldi et al., (1996) speculated that the reduced release of barium during sulphate reduction may be due to the formation of transient species

(BaS) and barium compounds (BaCO3) however, no direct evidence was found. Other studies related to unconventional oil and gas extraction have demonstrated

81 subsurface barite dissolution during hydraulic fracturing via microbially induced bioreduction. Ouyang et al. (2017), demonstrated the potential for barite dissolution (> 10 µg/mL) by halophilic organisms in hydraulic fracturing fluids under anoxic and hypersaline conditions. Evidence of microbial induced etch pit morphology on barite grains using SEM, confirmed the microbial induced dissolution of barite under laboratory conditions. Work relating to unconventional oil and gas extraction certainly suggests barite dissolution is possible (Burgos et al., 2017; Ouyang et al., 2017). More importantly it must be noted that barite formed in the field and laboratory across these studies may have different crystal structures and grain sizes thus impacting dissolution kinetics (Chang, 1996; Phillips et al., 2001). Additionally, studies utilizing pure cultures/enrichments of sulphate reducing bacteria may not sufficiently represent the environmental systems and there is a paucity of sediment work with indigenous microbial communities. The fate of radium in sediments as a result of offshore produced water discharges thus remains poorly understood.

2.7 Environmental Implications and Exposure Assessment

NORM produced can pose health risks to occupational workers during cleaning operations, maintenance, decommissioning and transport of waste, and may pose additional environmental risks during disposal e.g. operational discharge (Hamlat et al., 2001; Chowdhury et al., 2004; Hamilton et al., 2004; Abdul and Zaidan, 2010; Ismail et al., 2011; Doyi et al., 2016; IOGP, 2016).

There are a number of national and global regulations, legislations and guidelines produced by regulatory bodies, environmental agencies, local authorities and associations comprising of experienced industrial members to control and regulate the management and disposal of NORM. Efforts to improve environmental, social and safety performances as well as health and safety are being made by the sharing of knowledge and good practices which are adopted and recognised by authorities as a source of industrial information.

The exposure of radiation to workers is an area which has been given much attention due to the effect and impact ionising radiation can impose on living cells and tissue (Paschoa, 1998; NRC, 1999; Carvalho et al., 2014). There are a number of pathways in which workers can become exposed to undesirable doses of ionising radiation divided into internal and external exposure routes (Fig. 2.12). Exposure can be achieved via inhalation of pipe scale dust particulates or accidental ingestion of airborne particles which are classed as internal routes of exposure typically encountered during cleaning operations. External exposure can originate from unclean pipes, scale on contaminated equipment or scattered on the ground during the transportation of waste or descaling and other decontamination operations of equipment (e.g. dry rattling, higher pressure water jetting, cutting and welding, hammering and wire brushing) (Hamilton et al., 2004; Abdul et al., 2010). To reduce the risk to human health effective precautions, best working practices and training programmes such as health and safety awareness are utilised. Decontamination operations of equipment is adopted by operators for mainly production reasons or for safety reasons. This is mainly due to the build-up of radioactive scale and sludge decreasing the circumference of the pipes hence reducing the flow-through or due to health risks associated with the occupational exposure to ionising radiation respectively. To reduce and control scale formation many chemical strategies are implemented. This is achieved via the use of scale inhibitors, control of pH and temperature (Section 2.3.9) (Vetter, 1972, 1976; Mitchell et al., 1980; Vetter, Kandarpa et al., 1982).

The dose rates observed internally and externally in components of the production system and facilities due to the accumulation sludge and scale within components can be significant with up to 300 µSv h-1 observed (Jonkers et al., 1997; Van Weers, 1997; Hamlat et al., 2001; Hamilton et al., 2004; Abdul et al., 2010). The dose rates experienced both internally and externally are dependent upon the radionuclide composition, the amount of radionuclides, the activity concentration of radionuclides, particle size distribution, chemical composition and also spatial distribution (Hamlat et al., 2001; Abdul et al., 2010; IOGP, 2016). In some instances contaminated equipment and apparatus can be analysed without the need of

83 opening the equipment as isotopes of radium (228Ra and 226Ra) emit high energy photons of gamma ray radiation (> 200 keV) which can permeate through the steel walls of pipes, vessels and other components but can significantly contribute to the dose rate. Scales which have aggregated for several months can also have an increased measured dose rate on the outside of contaminated equipment as a result of photon emission by radionuclides 208Tl, 212Pb and 212Bi from the 232Th series (Fig. 2.1). In comparison equipment contaminated with scales virtually solely made up of 210Pb cannot be monitored from the outside as the emitted gamma rays and beta particles are of low energy (~ 13- 47 keV and ~ 17 – 63 keV respectively) and therefore not penetrating enough to permeate the steel walls of pipes and other equipment (Abdul et al., 2010; IOGP, 2016). Studies show workers alongside NORM adopt a total year dose < 1 mSv y-1 in agreement with the International Commission on Radiological Protection (ICRP) recommended public dose of 1 mSv y-1 further to the natural background dose (~ 2.4 mSv y-1) (Wilson et al., 1992; Hamlat et al., 2001; Hamilton et al., 2004; ICRP, 2007; IOGP, 2016).

Figure 2.12: Routes of exposure from NORM (internal and external sources) (IOGP, 2016)

Exposure can be controlled by taking simple precautions such as maximising the distance from the contaminated source which follows the inverse square law,

84 reducing the amount of time exposed to the source, and the use of protective respiratory equipment, clothing (e.g. dust masks) and shielding. Such precautions should be adopted when examining the build-up of radioactive scale and sludge on equipment (Abdul et al., 2010; IAEA, 2013; IOGP, 2016).

Further risk to exposure to NORM through inadequate and improper disposal of NORM waste by-products can result in the contamination of plants, soils, waters and animals. Such risks can be reduced by controlling the disposal of NORM waste and adopting management plans to monitor, regulate and control the potential risks associated with NORM in correspondence with applicable legislation (Fig. 2.13) (Paschoa et al., 1998; O’Brien et al., 1998; IOGP, 2016).

Figure 2.13: Options for NORM disposal (IOGP, 2016)

There are many local and international regulations regarding onshore and offshore oil field emission including; (OSPAR (Oslo and Paris) Commission for EU (European Union) member states, DEFRA (Department for Environment, Food and Rural

Affairs), the Environmental Agency (EA) and the Oil and Gas Authority (OGA) for the UK overseeing industry compliance with European Union (EU) regulations. In the United States, the Environmental Protection Agency (EPA) and the Federal Regulatory Commission (FEDRC) regulate drilling and production, however offshore waters are regulated by the Bureau of Ocean Energy Management (BOEM). In Oman regulatory authorities include, the Ministry of Oil and Gas (MOG) and the Ministry of the Environment and Climate Affairs (MECA). In Saudi Arabia the Supreme Council for Petroleum and Minerals (Petroleum Council), and the Ministry of Energy, Industry and Mineral Resources (MEIM) regulates all aspects of activity (Figgins et al., 2018; Goswami, 2019). There are five main types of discharges which have a concerning environmental impact. These are; 1) drill cuttings/drilling fluids; 2) oil leaks and spills; 3) atmospheric emission; 4) produced water and; 5) NORM disposal, with the latter two of main interest in this report. Currently NORM-waste and other radioactive contaminated materials are transported onshore for thorough analysis, testing and assessment in line with nuclear industry procedures and are being stored in drums, and subsequently buried at waste disposal sites as low level radioactive waste (Fig. 2.13) (Heaton et al., 1995; IAEA, 2013; IOGP, 2016). Other common global practices include, re-injection, land irrigation, salt spreading, geological disposal in abandoned wells, salt dome disposal and well-injection (Aboud-Sayed et al., 2000; Al-Masri et al., 2003; Al-Masri et al., 2005; Van Sice et al., 2018).

Guidelines specifying the requirements of water quality for discharge to the environment in the European Union state, declare the maximum monthly average of hydrocarbon content of 30 mg L-1 for discharge directly offshore (OSPAR, 2001). Other constituents of concern are the activity concentrations of naturally occurring radionuclides (NOR’s) such as radium. The radioactivity levels of produced water is generally minimal but careful consideration and assessment is needed when production and accumulation leads to large volumes of produced water being produced, as discharge to sea can have a significant impact on the activity levels of the sea (OSPAR, 2018) (Table 1.1; Section 1).

Due to the presence of NOR’s and toxic metals (e.g. barium) in produced water, its discharge into the sea and resultant environmental, bioaccumulation and bioconcentration effects are research areas of great interest (Table 2.2 and Table 2.9).

Bioaccumulation effects in organisms classed as filter feeders such as zoo plankton, krill, mussels and fish (e.g. shell fish and cod) are of main interest due to their method of feeding (Pauwels, 2002). Filtering structures across their bodies are used to extract matter from surrounding water which can also allow pollutants to pass through their bodies. Thus this feeding process can result in radionuclides (e.g. 226Ra) entering the food chain. As filter feeders also feed upon sediments, barite and radio-barite existing in the water column and sediment post-discharge and deposition may pass through the gills of faunae resulting in the incorporation of toxic barite and radium. This can also lead to the transport of materials to deeper water and accumulation within sediments from the decomposition of deceased organisms or through the falling of faecal pellets (Waldichuk, 1971; Gafvert et al., 2007).

Radiotoxicolgy and bioaccumulation studies show radium Ra2+(aq) of up to 117 Bq L-1 (0.117 Bq mL-1), and adsorbed to sediment (6600 Bq kg-1) due to produced water discharges to marine waters display a low exposure dose effect to benthic fauna (e.g. sediment-dwelling organisms) around oil platforms (Neff, 2002; Ruus et al., 2005; Grung et al., 2009). Likewise radium incorporated into barite is considered highly recalcitrant and has a low bioavailability and low toxicity (Neff, 2002; Ruus et al., 2005; Menzie et al., 2008; Grung et al., 2009; Olsvik et al., 2012). The bioavailability and impact of radium existing in the aqueous phase during marine discharge is deemed to be negligible, due to many factors such as, the chemical characteristics of seawater in comparison to fresh water (e.g. high sulphate concentration driving the co-precipitation of sulphate minerals), the resultant tidal regimes and dilution effects (as a result of large volumes of surrounding seawater) (Betti et al., 2004; Eriksen et al., 2006; Gafvert et al., 2007; Grung et al., 2009; Olsvik et al., 2012).

Additionally, NORM-wastes deposited in landfill sites, waste pits and storage tanks are possible sources of leaks or spills thus leaching radioactive contaminants into soil and ground waters. NOR’s can be absorbed by soil, surrounding plants and vegetation leading to potential agricultural contamination and human health hazards (Štrok et al., 2013). Preventative controls by landfill operators entail the use of modelling techniques to model the transport of radionuclides to avoid contamination (Pontedeiro et al., 2007; Pontedeiro et al., 2010). Additionally the lining of waste pits and storage ponds, regular upkeep and the use of steel tanks can also reduce the risk of release (Robertson et al., 2017).

CHAPTER 3

Methodology

3.0 Research Methods

In this chapter the numerous experimental and analytical techniques used to provide data for this thesis are outlined. This chapter incudes the theory of each technique, the scientific rationale for the employed methods and experimental details specific to the approaches adopted in each research chapter (Chapters 4 - 6). All chemical reagents used were of analytical grade and solutions were prepared using deionised water (18 Ω). All glassware was washed with Decon solution (5 %) overnight and nitric acid (5 %) overnight, and successively rinsed with deionised water before and after use. All pipettes were calibrated prior to use.

3.0.1 Safety

In this thesis all laboratory safety practices and roles were conducted in compliance with relevant documentation such as Control of Substances and Hazardous to Health (COSHH), Chemical Risk Assessment (CRA), Radiological Risk Assessment (RRA) and Standard Operating Procedures (SOP), and appropriate precautions were taken. Mandatory radiological safety training, radiological risk assessments were completed prior to experimental work and manipulation of radioactive materials were conducted with trained professionals in which strict procedures and protocols were adopted throughout. All work was conducted in compliance with the institution’s guidelines, Ionising Radiation Regulations 2017, Environmental Permitting Regulations 2010, and the Health and Safety at Work Act 1974.

3.1. Sample collection

3.1.1 Marine Sediment

Sediment samples were collected in April 2017 close to a near shore discharge point (UK) for this study. A total of 5 samples (2 - 3 kg wet weight) were taken from different distances from the outfall (20 - 250 m) (Figure 3.1). Seabed sediment (10 - 15 cm) was removed using a Day grab (0.1 m2) and were placed in sterile bags in an ice box to maintain anoxic conditions during transport. Samples were stored at 4°C and prior to analysis/characterisation, homogenised and dried.

Figure 3.1: Schematic of the field sediment samples obtained (+) from various distances to the discharge point (•)

In Chapter 3 a portion of each sediment sample obtained was analysed to assess the possibility of radium (226Ra) and barium concentrations above background levels due to operational produced water discharges. The bulk mineralogical composition of the sediments were characterised via XRD (Section 3.9.2), elemental concentrations via XRF (Section 3.9.7) and radiological content via gamma spectroscopy (Section 3.9.8). Organic matter (OM) content was also deduced from loss on ignition (Section 3.9.7). After comparison of the data obtained for each sample in regards to radium and barium, heavy liquid density separation (Section

3.8.1) experiments were performed to isolate heavy mineral grains from the most contaminated sediment samples. Mineral grains extracted were subsequently morphologically and compositionally analysed via SEM-EDX (Section 3.9.1) in conjunction with autoradiography (Section 3.9.9) to provide validation that radioactivity is partitioned within the strontiobarite phase extracted. Additionally sediment microcosm experiments (Chapter 5) were conducted representative of the field conditions via the use of the least contaminated field sediment (regarded as baseline). Experiments were conducted to explore the fate of 226Ra and proxies (e.g. Ba and Sr) in the marine environment due to the process of operational discharge as biogeochemical conditions develop.

3.2 Precipitation Experiment (BaxSrySO4 - Inactive)

To investigate whether the strontiobarite phases found and extracted from field sediment samples were formed and deposited as a result of operational marine discharges, mixing experiments were conducted using field produced water and seawater samples to provide validation (Chapter 4). Analogous synthetic mixing experiments were subsequently designed to mimic the conditions adopted for field samples to further validate the methodology, approach and formation of strontiobarite micro-particulate precipitates. Synthetic mixing experiments were also conducted repeatedly to yield enough precipitate to be used in microcosm experiments.

3.2.1 Seawater and Produced Water Synthesis

It was possible during one sampling campaign to obtain field formation water from an oil producing field (UK). Seawater was obtained from Formby Beach in Merseyside (UK). Analogous synthetic North Sea seawater and produced water compositions were obtained from literature and synthesised in the laboratory accordingly (Table 3.1) (Todd et al., 1992). Nalgene wide-mouth polypropylene carboys (10 L & 20 L) were washed using nitric acid (5%) before use, rinsed using deionised water (DIW) and left to dry. When preparing a 20 L stock solution,

91 chloride salts were all initially added to deionised water (16 L) in a carboy, followed by sodium sulphate and sodium hydrogen carbonate was added to a separate beaker containing DIW (3 L). During the addition of salts, both solutions were left to stir and salts were only added once the previous salt added had dissolved. Contents of the 3L beaker were then poured into the carboy over a period of 15 minutes to prevent oversaturation when mixing. The remaining volume of DIW (1 L) was used to wash the beaker to ensure the removal of all material to the carboy. Once added together the stock solutions were left to stir until use. Carboys were also hand shaken after 24 hours to ensure salts were well mixed and fully dissolved. The pH of both formation waters and seawater stock solutions were measured directly over a course of a week using calibrated electrodes (Section 3.7.3). An amount of the stock was filtered through a 0.22 µm polyethersulfone (PES) membrane and 1 mL was added to 9 mL high purity nitric acid (2 %) for ICP-AES analysis (Ba, Sr, Ca, Mg, and K; Section 3.7.1). Sulphate and hydrogen carbonate was excluded from the formation water composition and the ionic balance of the solution was readjusted accordingly due to undesired precipitation of barium sulphate, calcium carbonate and barium carbonate upon addition of the salts (Figure 3.2).

Seawater Formation water

Ions mg L-1 mg L-1 Na 10890 29370 K 460 372 Mg 1368 504 Ca 428 2809

Sr 8 574 Ba 0 252 Cl 19729 52663

SO4 2960 0 HCO3 124 0

Table 3.1: The seawater and formation water composition adopted (Todd et al., 1992)

Figure 3.2: (A-D); (A) Image of precipitate formation in beaker (B) BSE image of precipitate extracted from beaker; (C) EDS spectra confirming the formation of calcite and; (D) XRD pattern confirming the presence of calcite

3.2.2 Precipitation Procedure

Prior to the start of the mixing experiments geochemical modelling was conducted to calculate the degrees of saturation and ionic activities using PHREEQC geochemical modelling (Section 3.7.6). All samples (field and synthetic waters) were filtered through a 0.22µm Whatman Nylon filter paper using a 2 L Buchner flask and VWR VP86 vacuum pump, to remove any solid/colloidal material prior to use e.g. sand. Seawater (4.5 L) was added to a carboy (10 L), followed by the direct addition of formation water (0.5 L) in a 9:1 ratio and left to stir overnight to allow strontiobarite precipitation (6.64 mmol L-1 strontium). After 24 hours the solution turned milky and the stirrer bar was removed using an elongated stirrer remover and switched off to allow the high-density precipitate to settle. An amount of the mixed supernatant was then syringe filtered through a 0.22 µm PES membrane and 1 mL was added to 9 mL nitric acid (2%), and 9 ml DIW for ICP-AES (Ba, Sr, Ca, Mg,

2- - and K; Section 3.7.1) and ion chromatography analysis (SO4 and Cl ; Section 3.7.2) respectively to determine the molar stoichiometry (solution) of the precipitate. The mixed water solution (5 L) was decanted down to 2 L via a peristaltic pump. The remaining 2 L was swirled to allow efficient removal of the contents and filtered using a 2 L Buchner flask, a VWR VP86 vacuum pump and 0.2 µm Whatman Nylon filter paper. The precipitate collected on the paper was then washed with DIW (~ 10 secs, to remove any salt and the paper was subsequently left to dry in a petri dish in an oven (40 °C ± 0.5°C) until dry enough to be removed from the filter paper. Vials were weighed prior the addition of precipitate and after to determine the yield accurately (Table 3.2). The dried powder was transferred to a vial, homogenised and characterised via XRD (Section 3.9.2), SEM (Section 3.9.1), XAS (Section 3.9.6), BET surface area (Section 3.9.5) and FTIR analysis (Section 3.9.3). The molar stoichiometry (solid) of strontiobarite precipitate was further determined by adding 2 mg of sample to 7 mL EDTA-KOH solution (10 mM adjusted to pH 13 using 5 M KOH solution) designed to quantitatively dissolve barite, and then measured via ICP-AES (Averyt et al., 2003).

Yield (mg) of Precipitate (4.5L : 0.5L Mixing) Exp. 1 188 Exp. 8 245 Exp. 2 184 Exp. 9 246 Exp. 3 200 Exp. 10 159 Exp. 4 219 Exp. 11 208 Exp. 5 237 Exp. 12 196 Exp. 6 227 Exp. 13 207 Exp. 7 194 Exp. 14 173

Table 3.2: Yield obtained from multiple mixing experiments to form strontiobarite

3.3 Radium uptake experiment

Further mixing experiments were conducted on a smaller scale in Chapter 4 to model and investigate the formation process of the ternary phase (RaxBaySrzSO4)

94 and the distribution of radium in the solid and aqueous phases to allow an understanding of the behaviour and fate of radium during discharge.

A synthetic formation water stock solution (500 mL) was spiked with a 226Ra stock solution (diluted in 0.001 M HCl) to give a specific activity of 100 Bq mL-1. The pH was readjusted back to the initial pH of the formation water with diluted KOH (0.1 M) solution. Synthetic seawater and spiked produced water were mixed in a 9:1 ratio (1.35 mL to 0.15 mL) in 2 mL O-ring centrifuge tubes giving a specific activity of 10 Bq mL-1 After the two solutions were mixed the lids were screwed and continuously mixed on a stirrer plate for seven hours. The time of mixing was recorded as the initial time of the experiment. After each time point the tubes were centrifuged 14800 rpm for 5 minutes to separate the supernatant from precipitate. Radium concentrations in the mixed solution were determined by liquid scintillation via the addition of 1 mL supernatant to 10 mL scintillate cocktail (Section 3.7.5). Identical inactive experiments were run alongside to allow cation and anion analysis via ICP-AES (Section 3.7.1) and Ion Chromatography (Section 3.7.2) to be conducted. Control experiments containing only (un-)spiked produced water and seawater were also run alongside to ensure the non-sticking of radium to tube walls and undesired precipitation effects. All experiments were carried out in triplicate and the results routinely reported as the mean of three measurements with one standard deviation as the error.

3.4 Precipitation Experiment (BaxSryRazSO4 – Active)

In this thesis additional mixing experiments were conducted in Chapter 5 dissimilar to the method described in Section 3.2; Chapter 4 due to the addition of increased amounts of 226Ra. As the method described in Section 3.2 would lead to a great radiological exposure risk to the researcher as reactions would be conducted on a 5 L scale, a new method was adopted to produce the desired radioactive precipitate on a smaller scale (40 mL). Enhanced amounts of radium were used to ensure potential release (1 %) of radium in microcosm experiments could be sufficiently detected in pore waters. Inactive test experiments were initially conducted to

95 ensure the desired product morphologically and chemically matched that obtained in the active experiments described in Section 3.2. Such tests were conducted prior to the incorporation of radium due to the limited analysis capabilities upon its addition.

3.4.1 Adjusted Brine Composition

A less sophisticated recipe (Table 3.3) was adopted to produce the desired radioactive strontiobarite precipitate with the desired molar stoichiometry

(Ba75Sr25), yield (120 mg) and morphology previously obtained from field and synthetic mixing experiments in Section 3.2.

Salt Mass (mg) in 40 mL BaCl2.2H2O 124.4 SrCl2.6H2O 45.3 Na2SO4 97.0

Table 3.3: The solution composition adopted to produce radium labelled strontiobarite in 40mL

3.4.2 Precipitation Procedure

Initially a 20 mL solution comprised of strontium and barium chloride and another separate 20 mL solution containing sodium sulphate were made in separate 50 mL centrifuge tubes. The sodium sulphate solution was then directly poured into the other solution for 1 minute, filtered using a 2 L Buchner flask, a VWR VP86 vacuum pump and 0.2 µm Whatman Nylon filter paper. The supernatant (1 mL) was added to 9 mL nitric acid (2%) and 9 mL DIW for ICP-AES (Ba, Sr, Ca, Mg, and K; Section

2- - 3.7.1) analysis and ion chromatography (SO4 and Cl ; Section 3.7.2) respectively to determine the molar stoichiometry (solution) of the precipitate. The precipitate collected on the paper, washed with DIW to remove excess salt and left to dry in a petri dish in an oven (40 °C ± 0.5 °C) until dry enough to be removed from the filter paper. Once dried the precipitate was analysed as described in Section 3.2 for comparison against the precipitate produced from those experiments. The SEM analysis initially provided a negative result in regards to the adopted morphology of

96 the precipitate (Figure 3.3 (B)) hence a refined method (described below) was adopted in which pumping over 6 hours (0.055 mL min-1) and continuous stirring was incorporated which proved to be successful as shown from the data presented in Chapter 5.

A beaker containing a solution of pH pre-adjusted (0.1 M KOH) barium and strontium chloride (19.6 mL) was spiked with 7.6 kBq of radium (0.38 mL). Subsequently 20 mL of sodium sulphate was placed in a syringe and pumped over 6 hours (0.055 ml min-1) into the beaker containing the radioactive solution and left to stir for 24 hours (Figure 3.3 (A)). After 24 hours the content of the beaker was slowly swirled, decanted into a 50 mL centrifuge tube and left to settle for 4 days to allow radium uptake to occur. The beaker was washed with 0.1 M HCl to determine whether radium was sticking to the walls. Radium concentrations in the acid wash were determined by adding 1 mL of the solution to 10 mL of scintillation fluid (Ultima Gold, Perkin Elmer) in a scintillation tube. Tubes were shaken before analysis and immediately measured in a Quantulus scintillation counter (Perkin Elmer; limit of detection approx. 0.2 Bq) for 20 minutes per sample (repeated measurement). Samples were then left for 1 month to allow the Ra progeny to come into equilibrium and recounted. Liquid scintillation was conducted on the supernatant to determine the uptake of radium over the period of settling (3 days) and then decanted down to 2 mL ready to be used in microcosm experiments. Upon its addition to serum bottles the beaker was washed with seawater to ensure the complete removal of precipitate. Identical inactive experiments were previously run alongside, in which precipitate was obtained via vacuum filtration (0.22 µm Whatman PES filter) and dried (40 °C ± 0.5 °C) for characterisation. The syringe pump was calibrated to ensure the withdrawal of 20 mL solution using DIW prior to experimentation.

A) B)

Figure 3.3 (A-B): A) Overview of the experimental set up; (B) The undesired crystal morphology observed via BSE imaging when instantly mixing the two solutions as opposed to pumping over 6 hrs (undesired dendritic morphology = indicative of a fast growth rate)

3.5 Sediment Microcosm Experiment

In Chapter 5 of this thesis microcosm experiments were utilised to investigate the potential long term environmental impact, speciation, mobilisation and availability of radium once discharged to a marine setting as the; 1) inorganic radium-

containing precipitate (BaxSryRazSO4); 2) synthetic inactive precipitate (BaSrSO4) and; 3) aqueous ion (Ra2+), which are the main forms in which radium exists in the marine environment during discharge. Sediment microcosm experiments containing aqueous radium were developed to assess radium sorption and scavenging. Biological processes and reducing conditions were developed to assess the long term fate of radium in both solid and aqueous forms under conditions relevant to discharges in shallow marine environments to develop a predictive capability of radium behaviour and mobility.

Microcosms containing sediment (10 ± 0.1 g) and synthetic seawater (100 ± 1 mL) typically with a 1:10 solid solution ratio were set up using an aseptic technique in sealed glass serum bottles similar to past work from our laboratories (Burke et al., 2005; Newsome et al., 2017). Sediment obtained from the baseline location of the field site (Point 6; Section 3.1.1) was used (within 6 months of sampling) in all sediment microcosms and chemically confirmed prior to its use via characterisation

(e.g. XRF and gamma spectroscopy) detailed in Section 3.1.1. To ensure weight consistency across the experiments large material such as stones, rocks, gravel, shells and plant roots were hand-picked from the sediment. Dry-weight analysis was then conducted in triplicate by weighing 15 g of wet sediment in pre-weighed centrifuge tubes, subsequently dried in the oven (40 °C ± 0.5 °C) until dry (3 - 5 days) before re-weighing and determining the ratio. Sediment was also stored at - 80 °C to conduct microbial community analysis via 16s rRNA sequencing (Section 4.0) at the end of the experiments (Day 300) to represent the field conditions. The artificial seawater (Section 3.2.1) was used for all sediment microcosms and consisted of (mg L-1): Ca, 420 ± 6; K, 467 ± 4; Mg, 1300 ± 18; Sr, 8 ± 0.1; Ba, 0; Na, 10100 ± 121;

Cl, 19200 ± 268; SO4, 2930 ± 23; HCO3, 111 ± 10 (Todd et al., 1992). The synthetic seawater was made anoxic via flushing with a N2:CO2 gas mixture (80:20), prior to sterilising in an autoclave (120 °C, 20 min) before use. Synthetic seawater was then added to serum bottles containing pre-weighed sediment in an anaerobic cabinet before sealing with a butyl rubber stopper and aluminium crimps. In Chapter 5 sets of microcosm experiments contained either; 1) the inorganic radium-containing precipitate (BaxSryRazSO4; Section 3.4.2); 2) the synthetic inactive precipitate

2+ (BaSrSO4; Section 3.2.2) and; 3) the aqueous ion (Ra ). Therefore the bottles were removed from the anaerobic chamber and placed on the work bench or fume hood where the seals of six microcosms were de-crimped in the presence of a stream of argon to maintain anoxia and subsequently spiked accordingly with the radium- containing phases (strontiobarite precipitate (120 mg); radiostrontiobarite precipitate solution (4 mL/120 mg); or 2 kBq aqueous 226Ra). Triplicate microcosms across all three experiments and a sterile control bottle (one replicate) were inoculated aseptically with filter-sterilised electron donor as 5 mM acetate and 5 mM lactate. Control bottles (one replicate) containing: seawater, sediment and precipitate or aqueous 226Ra; seawater and precipitate or aqueous 226Ra; and sediment and seawater only, contained no electron donor (Figure 3.4). Results from amended microcosms carried out in triplicate were reported as the mean of three measurements with one standard deviation as the error. The headspace was purged with N2 prior to sealing with butyl rubber stoppers and aluminium crimps. The bottles were then incubated in the dark at ambient temperature (20 °C) and

99 aseptically sampled every 7 - 56 days under anaerobic conditions over 300 days to monitor changes in biogeochemistry. After 140 days microcosms initially amended were spiked with a second addition of electron donor (2 out of 3 bottles; 10 mM acetate and 10 mM lactate) to give a final concentration of 15 mM acetate and 15 mM lactate, to further promote the biological evolution of sediments. Alternative sediment microcosms across all sets of experiments (one replicate) included sediment and seawater only (no electron donor) were run to investigate any shifts in the microbial community after the addition of radium as the solid precipitate or aqueous solution. A seawater and precipitate only (no electron donor) microcosm was also used to account for any surface dissolution due to the ionic strength effects (Risthaus et al., 2001; Kowacz et al., 2008; Kowacz et al., 2009), and accountancy of radium sticking to the wall of the serum bottle. An unamended microcosm (containing seawater, sediment and precipitate or aqueous 226Ra) representative of the natural system was utilised to explore the natural bioreduction process in sediments. Finally a sterile control amended with precipitate or aqueous 226Ra was created for all three sets of experiments by autoclaving additional microcosms (3 cycles, 20 minutes at 120 °C) prior to the addition of electron donor to confirm sediments were microbially active (Burke et al., 2005; Wilkins et al., 2007; Thorpe et al., 2012; Newsome et al., 2014, 2017; Newsome, et al., 2014).

In all experiments microcosms were shaken to ensure constituents were in suspension and rubber tops sterilised with ethanol prior to being de-crimped for sampling. Typically two consecutive aliquots (1.6 mL and 0.4 mL) of sediment slurry was removed aseptically using a degassed syringe with argon and in the presence of a stream of N2 to preserve the samples. At each time point the total bioavailable iron and Fe (II) concentrations in sediment slurries (0.1 mL) were analysed using the ferrozine assay (Section 3.7.4.1). The remainder of the aqueous phase was separated by centrifugation (14 800 rpm, 5 minutes) and aqueous Ra was measured via liquid scintillation counting (1 mL), strontium, barium and calcium analysis by

ICP-AES (0.1 mL acidified in 2% HNO3), and sulphate and chloride analysis via ion chromatography (0.1 mL). Calibrated electrodes were used to measure the, pH and

100

Eh (on the remaining 0.3 mL solution). Hydrogen sulphide (HS-) was detected in pore waters from a second aliquot of sediment slurry (0.4 mL; see above) which was centrifuged as described above and analysed using the methylene blue assay (Section 3.7.4.2). After Day 300 sediment was removed from amended and non- amended microcosms across all sets of experiments (e.g. amended with electron donor and non-amended) to investigate the composition of the microbial community pre- and post-anoxia via 16s rRNA sequencing, and to validate the establishment of dominant sulphate reducing conditions (Section 4.0). A portion of sediment was also extracted and oven dried from an inactive (amended) microcosm bottle to isolate strontiobarite mineral grains via heavy liquid density separation (Section 3.8.1) for etch pit analysis via SEM (Section 3.9.1) to further validate the results.

Sediment microcosms are deemed to be a suitable technique to investigate the long term fate of radium in its many forms (e.g. solid and solution phase) as biological processes and reducing conditions develop under controlled anaerobic conditions representative to the field. Similarly radium uptake experiments are suitable techniques to determine the uptake of radium in respect to operational

101 discharges. However, limitations arise as sediment microcosm systems and uptake experiments represent a closed system or a system where there is maximum uptake respectively (Section 3.5 and 3.3). In addition, there are many challenges with mimicking an open system due to the complex variables of a natural system (e.g. current flow effects and associated tidal regimes. Discharge times and volumes are other important factors in determining for example the mass of precipitate generated at a particular discharge site and have not been fully assessed in this thesis but have been described. Nonetheless such data can be utilised in discharge modelling systems as feed data to develop a predictive capability of radium behaviour and mobility in these systems and enable development and adoption of fit-for purpose control strategies tailored to different locations.

3.6 NORM Sample Characterisation

In this thesis NORM scale samples from oil field tubulars where characterised in Chapter 6 to understand the factors controlling the distribution of radionuclides in NORM samples and fate in the marine environment. The materials analysed for study within this project were supplied by industrial partners. Two different sets of samples were analysed which originated from separate locations, those from the North Sea (7536, 7470 & 7457) and those from Iraq (IRQ 1-4). At least 5 - 7 g and 30 - 140 g of each sample were taken from Iraq and North Sea tubulars respectively. A portion of the samples were set into an epoxy resin (a mixture of bisphenol a- (epichlorohydrin) and hardener (amines, polyethylene, polytriethylenetetramine fraction; 10:1) and prepared as sections and polished to give a flat surface for analysis using a range of micro-focus techniques. Dependent upon the sample availability a variety of techniques were used to characterise the bulk mineralogical composition of the samples via XRD (Section 3.9.2), elemental concentrations via XRF (Section 3.9.7), SEM-EDX (Section 3.9.1), FTIR spectroscopy mapping (Section 3.9.3) and Raman spectroscopy (Section 3.9.4). Radiological content and distribution was assessed via gamma spectroscopy (Section 3.9.8) and autoradiography (Section 3.9.9) to determine the total radium in the sample and

102 spatial distribution relationships between the bulk mineralogical composition and radionuclide content.

Sectioning and polishing of the resin embedded samples was carried out by Mr. Steve Stockley, School of Earth & Environmental Sciences, The University of Manchester

3.7 Solution and Geochemical analysis

3.7.1 Inductively-coupled Plasma Atomic Emission Spectroscopy (ICP- AES)

ICP-AES measures the atomic spectra of elements and was used to determine the total elemental composition of solutions. A series of elements were determined such as barium, calcium, strontium, magnesium, potassium and sodium (Chapters 4 and 5). Additionally total concentrations of sulphate (measured as sulphur), manganese and iron were also measured in sequential extraction experiments (Chapter 5).

Fundamentally the distinguishing capability of ICP-AES as a spectroscopic source is its ability to produce linear calibration lines over several orders of magnitude thus allowing the routine analysis and detection of major and trace components (detection limit of < 10 ppb). The typical components of an ICP-AES instrument include the: source unit (ICP torch); spectrometer; and computer. The source unit is the component which generates the characteristic elemental emission spectral lines by providing an energy source. The spectrometer in turn resolves and separates the emission lines generated and measures the strength of the recorded signal. The computer performs a conversion in which the signal is converted into a numerical measurement of the total concentration of the elements in solution (generally as ppm or mg L-1). Upon addition of the sample to the instrument via pumping through a capillary tube a flow of argon is introduced into the nebuliser resulting in the formation of an aerosol. The aerosol with argon gas then tunnels up the central injector tube of the ICP torch in which evaporation, atomisation and ionisation

103 occurs. The torch is the component in which the energy source in the form of a plasma is generated and powered by a radiofrequency generator connected to multiple hollow copper coils (load coil; cooled via internal water flow) surrounding the torch. The plasma is generated in the vitreous silicon tube by the passing of current through the coil inducing an oscillating magnetic field and spark resulting in the ionisation of argon flowing through the silica tube (Figure 3.5). A cascade of energetic collisions between free electrons and argon atoms precede resulting in temperatures of up to 10 000 K being reached inside the plasma. A temperature gradient typically exists between the load coil (cooler region of argon gas) and central part of the plasma in aid to prevent the melting of the quartz tubing (Figure 3.5). At such high temperatures thermal excitation of sample atoms results in the promotion of valence electrons to higher ‘excited’ energy orbitals (Gill, 2014; Matthews, 2014). When the electron returns to the ground state configuration upon relaxation, a photon of energy is subsequently emitted at characteristic wavelengths relating to the energy difference between the excited and ground state of individual elements (Figure 3.6). The comparison of standards of known concentration to the measured emission intensity allows the concentration of the element to be deduced (Burrows, 2013; Matthews, 2014).

Supernatant samples were removed from experiments and solids separated via filtration methods (0.22 µm (PES) membrane (Chapter 4), 14 800 rpm for 5 minutes (Chapter 5) or 5000 rpm for 20 minutes (Chapter 5)). Aqueous samples (0.1 and 1 mL) were then made up to 10 mL with 2% HNO3 such that the analytes concentration was between 0.1 and 10 mg L-1 and radionuclide concentrations below the 4 Bq mL-1 threshold. Analyses was performed using a Perkin Elmer Optima 5300 DV (dual view) ICP-AES system. Samples were analysed in triplicate and the results routinely reported as the mean of three measurements with one standard deviation as the error.

Sample analysis was performed by Mr. Paul Lythgoe, and standards of elements of interest were also prepared from VWR certified standard solutions and were run after every 10 samples (Boss et al., 2004; Burrows, 2013; Gill, 2014).

104

Figure 3.5: Design of the ICP torch and temperature distribution in plasma (Gill, 2014)

Figure 3.6: Schematic of the energy level diagram and characteristic emission and absorption spectra for sodium (Na)

105

3.7.2 Ion Chromatography

In this project ion chromatography (Dionex ICS 5000) was exploited to determine

2- - the solution concentration of SO4 , Cl and a variety of volatile fatty acids (VFA’s) such as acetate and lactate. Ion chromatography is an application of high- performance liquid chromatography (HPLC) which is commercially adopted to determine concentrations of metals, organics, cations and anions in solution (Woods et al., 1997; López-Ruiz, 2000; Ohta et al., 2000). The separation of ions is achieved by the differences in affinity and size of analytes to an ion exchange resin packed column termed the ‘stationary phase’ (AS11HC and AS18). A typical ion chromatograph consists of a: mobile phase (eluent; 34 mM KOH); pump (to pump liquid through the column); injection valve (to inject the analyte-containing solution); stationary phase (analytical column to adsorb analytes); suppressor column (to improve separation and detection sensitivity limits); detector (detects eluted ions) and computer/software (integrates the peak detected and compares against calibration standards) (Gill, 2014) (Figure 3.7). Typically when a liquid sample is injected into the carrier stream (mobile phase/eluent) and the resultant solution pumped through the chromatography column, competition between ions exist for exchange sites on the separation column. Separation of the components in solution occurs as a result of dissimilar interactions with the ion exchange sites on the resin principally based on size and ionic charge. The eluent solvent (mobile phase) flows through the column (stationary phase) and competes with the absorbed analytes resulting in the elution of analytes at different characteristic times known as the retention time into the mobile phase with smaller ions generally eluting earlier (e.g. Cl-) and detected as they emerge producing a signal (Burrows, 2013) (Figure 3.8). The attached computer identifies each compound eluted at different times by comparison to a series of peaks from calibration standards and the peak is integrated in which the area under the chromatogram peak recorded is proportional to the ion concentration (López-Ruiz, 2000; Gill, 2014).

In this thesis all sample preparation was performed by the author. All samples were either centrifuged at 14 800 rpm for 5 minutes or syringe filtered using a 0.22 µm

106

(PES) membrane, (Chapters 4 and 5) and typically 1 mL or 0.1 mL of supernatant was diluted with deionised water. Samples were made up to 10 mL to ensure total dissolved solid (TDS) values did not exceed 100 mg L-1 and radionuclide concentrations were below the 4 Bq mL-1 threshold (laboratory safe limit). Samples were analysed in triplicate and the results routinely reported as the mean of three measurements with one standard deviation as the error.

All ion chromatography analysis and further dilutions to samples prior to analysis was performed by Mr. Alastair Bewsher, School of Earth & Environmental Sciences, The University of Manchester.

Figure 3.7: Schematic of the setup and components of an ion chromatograph system used for analysis

107

Figure 3.8: A typical Ion chromatogram produced using a Dionex instrument, showing the elution and resultant separation of anion species (Gill, 2014)

3.7.3 pH and Eh

Solution pH was measured using a Jenway 3520 pH/Eh meter with a Fisherbrand FB68801 electrode. Prior to analysis the probe was immersed into Fluka buffer solutions (pH 4, 7 and 10) for calibration which was conducted routinely. Post calibration, the probe was immersed into experimental solutions and a reading was recorded once stabilised. Solution Eh was also measured using a Jenway 3520 pH/Eh meter with a Mettler Toledo InLab Redox Mirco electrode attached. Prior to use it as immersed in a Mettler Toledo redox buffer solution (electrode potential of + 220 mV). Post calibration the probe was immersed into experimental solutions and a reading was recorded once stabilised.

3.7.4 UV-Vis Spectrophotometry

Spectrophotometric methods were used to measure and monitor the concentrations of aqueous Fe (II) and HS- (hydrogen sulphide) in sediment and pore water solutions respectively, to asses Fe (III)-reduction and sulphate-reduction in

108

Chapter 5. Ultraviolet-visible spectrophotometry is a method widely used to quantitatively determine the concentrations of analytes in solution based on the interaction of electromagnetic radiation with matter initiating the excitation of an electron to an excited state (e.g. higher energy level). This practise utilises the intensity of light at one particular wavelength (UV-Vis region) and measures the absorbance of the compound in solution which in turn determines the solution concentration as described by the Beer-Lambert Law (Figure 3.9). This law states, as molecules absorb light at characteristic wavelengths the degree of absorption (A, dimensionless) is proportional to the concentration of absorbing molecules (c, mol L-1) for a fixed path length (l, cm). The molar absorption coefficient (ε, L mol-1 cm-1) indicates the absorption capability of the species. The law is also defined in terms of the intensity of the incident radiation (Io) on a sample solution and the transmitted radiation (It) which emerges with a lower intensity (Perkampus, 1992; Burrows, 2013). The law predicts a linear relationship, however at higher concentrations the plot digresses (curves) which is not useful for analysis. Therefore increased dilution or reductions in the path length are means to rectify non-linearity when utilising this method.

In this thesis the ferrozine (Lovley et al., 1987) and methylene blue assay (Cline, 1969) were used to specifically produce distinct coloured solutions based on the presence of iron or hydrogen sulphide complexing to chelating compounds. Calibration standards of known concentrations were measured to generate a calibration curve where the straight line equation (y = mx + c) and regression from known standards (R2 values > 0.99) were used to calculate the concentration of species in solutions.

109

I c I

퐼0 푙표푔 = 푐 푥 푙 푥 휀 = 퐴 10 퐼푡

Figure 3.9: Beer Lambert Law and variables required to determine absorption

3.7.4.1 Ferrozine Assay

The ferrozine assay was used to measure the total bioavailable iron and Fe (II) concentrations in sediment slurry suspensions from sediment microcosm experiments (Chapter 5) (Burke et al., 2005; Wilkins et al., 2007; Newsome et al., 2017). The method is based on the use of a chelating ferrozine bidentate ligand, 3- (2-pyridyl)-5,6-bis(4-phenylsufonic acid)-1,2,4-triazine (Figure 3.10) which produces a highly soluble, stable and intense magenta-coloured solution upon complexation with aqueous Fe (II). Absorbance was measured at 562 nm (Stookey, 1970; Viollier et al., 2000).

Specifically at each time point Fe (II) concentrations in sediment slurries were analysed via the immediate digestion of slurry (0.1 mL) in 0.5 M HCl (4.9 mL) for one hour in falcon tubes. Subsequently after mixing, 200 µL was then added to 2.3 mL ferrozine solution (1 g ferrozine and 11 g HEPES in 1 L deionised water adjusted to pH 7) in a macro-cuvette, measured at 562 nm (Lovley et al., 1987). Additionally total bioavailable iron in sediments was assessed via the addition of 200 µL hydroxylamine hydrochloride (6.25 M; 0.43 g in 1 mL of 2 M HCl) to the 0.5 M HCl digest for a further one hour to reduce Fe (III) to Fe (II) and measured again at 562 nm as described above. Standards with concentrations of 1, 5, 10 and 20 mM

110

FeSO4.H2O in 0.5 M HCl (0.1 mL in 4.9 mL 0.5M HCl; 1:50 dilution) and a blank (0.2 mL 0.5 M HCl in 2.3 mL ferrozine solution) were prepared to produce a linear standards curve to convert absorbance measurements to a Fe concentration. Uncertainties were measured by calculating the standard deviation (1σ) of triplicate measurements. In this thesis ratios of Fe (II)/Fe (III) in sediments provided an indicator for whether the cascade of biogeochemical terminal-electron processes progressed and evolved over time as anoxia developed despite the fact the total iron present in sediment may not be completely extracted by both the reducing agent and extractant (Burke et al., 2005; Wilkins et al., 2007; Newsome et al., 2017).

Figure 3.10: The molecular structure of the ferrozine molecule with the bidentate ligand properties of the ferroin group highlighted (Stookey, 1970)

3.7.4.2 Methylene Blue Assay

- 2- Hydrogen sulphide (H2S, HS , and S ) concentrations in pore waters were determined to act as an additional indicator for the onset of sulphate-reduction in microcosm experiments (Chapter 5) (Cline, 1969). The method is based on the formation of a methylene blue compound from the reaction of dimethyl-p- phenylenediamine with hydrogen sulphide in the presence of an oxidant (iron (III) chloride) (Figure 3.11). The formation of methylene blue produces a characteristic blue coloured solution. Briefly, an aliquot of sediment slurry was removed aseptically using a degassed syringe with argon in the presence of a stream of N2 to preserve the samples. The sample was centrifuged (14 800 rpm for 5 minutes) and 0.2 mL of supernatant was added to a degassed O-ring tube (2 mL) in the presence of a stream of N2, followed by a quick addition of 16 µL anaerobic mixed diamine

111 reagent previously synthesised in the anaerobic chamber (20 g of N, N-dimethyl-p- phenylenediamine dihydrochloride and 30 g of iron (III) chloride in 500 mL HCl (6 M)). The tubes were sealed and agitated to mix the sample and reagent, stored up to 1 hour and then immediately diluted in deionised water (4.9 mL) prior to analysis in which 0.5 mL was removed and added to a further 0.5 mL of deionised water in a 1 mL micro-cuvette. Standards were made in compliance to radiello (RAD171) methylene blue concentrate calibration standard (Sigma Aldrich) including a blank (deionised water only). Absorbance was measured at 670 nm to produce a linear standards curve to convert absorbance measurements to a sulphide concentration (Cline, 1969; Wangersky, 2003). Uncertainties were measured by calculating the standard deviation (1σ) of triplicate measurements. Although hydrogen sulphide is volatile, easily oxidised by oxygen and scavenged from solution as pyrite (FeS), this operationally defined parameter provides a useful indicator of the onset of sulphate reduction in conjunction to sulphate depletion (monitored via ICP-AES) and Fe (II) concentrations (monitored via ferrozine assay).

Figure 3.11: Reaction of dimethyl-p-phenylenediamine with iron chloride (oxidant) and hydrogen sulphide to form methylene blue under acidic conditions (Cline et al., 1969)

112

3.7.5 Liquid Scintillation Counting (LSC)

Liquid scintillation counting was used to detect the activity of radium in solution (Passo et al., 1994; Horrocks, 1974). This technique entails the dissolution of sample in a liquid scintillation cocktail (scintillator solution) in which the matrix is mainly comprised of phosphorescent molecules dissolved in a solvent (generally aromatic compounds). During this process energy emitted from radioisotopes (alpha or beta radiation) is absorbed by the solvent π-cloud orbitals and transferred to the dissolved phosphorescent molecules resulting in the formation of excited state scintillate molecules thus a countable solution. Occasionally scintillation cocktails comprise of primary and secondary phosphorescent compounds because the initial transfer of energy to the primary scintillate molecules generally results in the emission of a photon of smaller wavelength following absorption, which is insufficient for detection. Therefore secondary scintillate molecules absorb the photons emitted by the primary molecules and re-emit characteristic photons at a longer wavelength. The characteristic photons of low energy (> wavelength) are re- emitted upon de-excitation and detected by the photocathode of the photomultiplier tube (PMT) (Gill, 2014). The photocathode emits electrons due to the photoelectric effect which produces primary electrons which are accelerated electrostatically via an electrical potential to ensure contact with the dynodes (metal surfaces) situated in the PMT. Successive contact results in the production of secondary electrons amplifying the current at each dynode at a higher potential to accelerate electrons and produce an amplified output signal sufficient for detection at the anode (Figure 3.12). The output signal at the anode is a measurable electrical pulse from each group of photons (alpha or beta) from the original decay events in the scintillation cocktail that arrives at the photocathode and thus conveys specific and characteristic information relating to the energy of the original incident radiation in the sample. The pulse is passed to amplifiers which amplify the pulse to shape the pulse within the instrument allowing the activity of the sample to be calculated (Passo et al., 1994; Horrocks, 1974; Gill, 2014). The intensity of each luminescence measured is therefore proportional to the emission energy of the photons from the original ionizing event in the scintillator that arrived at the

113 photocathode. Information about the energy of the original incident radiation is carried in the output signal, and the number of decays/pulses per second refers to the number of radioactive emissions in the sample (Neame, 1974). The pulse of certain amplitude produced by the photomultiplier is characteristic for an alpha or beta decay event thus the energy of photon or particle can be deduced from the size of the generated pulse.

Figure 3.12: Illustration of the main component of a scintillator, the photomultiplier tube (PMT and the process of generating a series of amplified electrical pulses whose individual amplitudes are proportional to the quantum energies of the photons that generated them (https://en.wikipedia.org/wiki/Scintillation_counter)

Some instruments have the ability to distinguish between alpha and beta decay events via the adoption of a Pulse Shape Analyser (PSA) circuit. This effectively amplifies and integrates the tail of input pulses and compares it with the overall pulse length, and derives an amplitude independent value to describe the character of the pulse proportional to the energy of the particle (alpha or beta) exciting the scintillate molecules. The integration of the input pulse generally generates an output voltage or pulse which can be measured and compared for each input pulse detected. This additional information allows overlapping peaks produced by alpha and beta decays to be separated and provides a much more accurate measurement of radioactivity (Figure 3.13).

In this thesis a Wallac Quantulus (Perkin Elmer) 1220 ultra-low level scintillation counter was used to measure radium activity via the sum of 226Ra, 222Rn, 218Po and 214Po alpha emissions. Typically supernatant (1 mL) extracted via centrifugation or syringe filtration in Chapters 4 and 5 respectively was added to 10 mL of Scintsafe 3

114 scintillation cocktail and 0.1 mL HCl (pH 3 analytical reagent grade) in a scintillation vial. Samples were left for 30 days prior to analysis to allow the radium daughter progeny to equilibrate and reach secular equilibrium (226Ra, 222Rn, 214Bi and 214Pb) in which the activity concentration of these radionuclides is uniform (Al-Masri et al., 1996). Blank samples and standards containing HCl, radium and cocktail were routinely used to test the accuracy of measurement and the limit of detection. Blank measurements (n = 12) indicated a background activity of 0.0545 ± 0.0175 Bq mL-1. The limit of detection, 3 standard deviations, was therefore 0.107 Bq mL-1. The pulse shape analyser setting was optimised prior to the use of the instrument using a radium standard (2 Bq mL-1), in which a PSA setting of 85, alpha/beta discrimination, and count time of 240 minutes per sample was adopted. Colour quenching was investigated and accounted for in sequential extraction experiments (Chapter 5) by making standards from the aqueous phases extracted from uncontaminated field sediment extractions and spiking with aqueous radium (100 Bq mL-1).

Figure 3.13: Schematic of a LSC instrument system (Passo, C.J., Cook, G.T., 1994; Horrocks, 1974)

115

3.7.6 PHREEQC Modelling

Geochemical modelling was conducted to perform a variety of aqueous geochemical calculations to determine aqueous speciation, degrees of saturation and ionic activities using a geochemical software package, PHREEQC version 3.3.9 (Parkhurst et al., 2013) (Chapter 4). Specifically the synthetic and field mixing experiments were modelled prior to experimentation in which input data included water compositions, temperature, pH and volumes to mimic the conditions undertaken in the experimental method to investigate strontiobarite precipitation (Figure 3.14). Calculations performed via the software corresponded to the SIT (specific ion interaction theory) database in order to determine solution chemistry at equilibrium (e.g. saturation and speciation) from a library of known solubility products and stability constants.

Figure 3.14: Example of the data feed to the PHREEQC computer geochemical modelling software to determine the saturation and speciation of barium, strontium and calcium

116

3.8 Chemical Extractions

3.8.1 Heavy Liquid Extraction

Heavy liquid density separation experiments were carried out in Chapter 4 to isolate radiostrontiobarite minerals from field marine sediment for characterisation purposes to provide evidence of NORM formation processes. In Chapter 5 barite minerals exposed to anoxia (oxygen-free conditions) were isolated from sediment microcosms to provide further clarification whether potential microbial-mediated barite dissolution under anoxic conditions occurred by examining the surface of the mineral for etch-pits. Heavy liquids are typically dense chemical fluids or solutions which are often commonly used to separate mixtures based on mineralogical density differences or similarly for density-based gradient centrifugation. Typically material possessing a density greater than the heavy liquid will sink (sink fraction), whereas material adopted a lower density than the heavy liquid will float and suspend on the surface (float fraction) providing a method for separation and preconcentration (van Beek et al., 2002; Proske et al., 2015).

Oven dried sediment sample (5 - 10 g) was slightly disaggregated and sieved through a 215 µm sieve to remove large particles to obtain finer grains. A heavy liquid (diiodomethane) at a density of 3.3 g cm-3 was used to separate the barite- containing (4.5 g cm-3) dense fraction of the sediment (~ 2.5 g cm-3). 5 - 10 g of sample was added to a separating funnel containing 80 mL heavy liquid. The sediment-fluid suspension was stirred with a stirring rod to immerse the sediment and left to separate for 10 minutes forming the float and sink fraction (Figure 3.15). The dense fraction (> 3.3 g cm-3) was filtered under gravity through a funnel containing a 2.5 µm Whatman (ashless, grade 42) filter paper. The lighter fraction (< 3.3 g cm-3) was subsequently filtered separately. Both fractions were rinsed with deionised water and acetone then dried under the fume hood. Both fractions were examined under a microscope and barite particles extracted and analysed via SEM- EDS to identify etch pitch formation on the surface of barite crystals in Chapter 5. Additionally autoradiography was used to radiologically characterise barite minerals in Chapter 4.

117

Figure 3.15: Overview of the experimental setup of the heavy liquid density separation experiment

3.8.2 Sequential Extraction

In Chapter 5 sequential leaching experiments were performed to determine the speciation and phase association of radium (Ra), barium (Ba) and calcium (Ca) in sediments pre- and post-anoxia in operationally defined fractions of sediment samples. Typically a variety of lixiviants (leachates) are progressively applied to the sample of increasing strength to fractionate and release elements associated with various solid mineral phases into solution to study the distribution and estimate the potential mobility of selective metals of interest under various environmental conditions in the various fractions. A variety of publications exist in literature in regards to sequential extraction procedures however the protocol adopted in this study was adapted from Tessier et al., (1979) and Aguado et al., (2004). (Tessier et al., 1979; Aguado et al., 2004). Rather than using a HF/HClO4 mixture, aqua regia

(3:1; HCl/HNO3) was used as the leachate to extract the residual fraction. An additional modification was also applied which required using a solution of 0.1 M

Na2EDTA as the penultimate solvent to release radium and barium from radiobarite or strontiobarite (Beneš et al., 1981). After 300 days of anaerobic incubation

118 microcosms from all experiments were shaken and the resultant slurry was decanted under anaerobic conditions via the use of an argon box and separated via centrifugation (5000 rpm, 20 minutes). The resultant supernatant was decanted to allow the sediment to be accurately weighed into centrifuge tubes. Triplicate samples of wet sediments (1 g) from microcosms amended with acetate and lactate (spiked with an additional 10 mM) were accurately weighed and added to pre- weighed 50 mL centrifuge tubes under anoxic conditions. Contaminated field sediment and sediment from microcosms amended with precipitate (BaSO4 and

RaBaSrSO4), and aqueous radium, were placed in sediment microcosms for a period of 24 hours to identify any possible changes in the partitioning and mineral association of radium, barium and calcium as anoxia developed. Progressive extractions were conducted under anaerobic conditions via the use of an argon box.

This involved 1 M anaerobic MgCl2 (1 hour, pH 7, “exchangeable fraction”), 1 M anaerobic sodium acetate (5 hours, pH 5, “carbonate fraction”), 0.4 M anaerobic hydroxylamine hydrochloride (16 hours, pH 3, “reducible fraction”), 0.02M nitric acid, 30% hydrogen peroxide (2 hours, pH 2) and 3.2 M ammonium acetate (16 hours, “oxidisable fraction”), 0.1 M disodium salt of EDTA (24 hours, “barite dissolution”) and finally aqua regia (5 hour, “residual fraction”) (Summarised in Table 3.4) (Beneš et al., 1981; Nixon et al., 1983; Jerez Vegueria et al., 2002; Aguado et al., 2004). Separation of solids was achieved after agitation at each step by centrifugation (5000 rpm, 20 minutes). In the case of heating, sample was transferred to 50 mL beakers and heated on a hot plate. The leached aqueous phase was diluted and acidified (2% nitric acid) and analysed for Ba, Sr and Ca using ICP-AES. Radium concentrations were also determined in the leachate solutions via liquid scintillation counting. The remaining leachate was preserved. All experiments were carried out in triplicate and the results routinely reported as the mean of three measurements with one standard deviation as the error.

119

Fraction Lixivant Time Temp. °C Exchangeable 1 M anaerobic MgCl2 (pH 7) 1 hour RT Carbonate 1 M anaerobic sodium acetate 5 hours RT (pH 5) Reducible 0.4 M anaerobic 16 hours RT hydroxylamine hydrochloride (pH 3) Oxidisable 0.02M nitric acid, 30% 2 hours 85°C hydrogen peroxide (pH 2)

3.2 M ammonium acetate 85°C - RT Barite Dissolution 0.1 M disodium salt of EDTA 24 hours 70°C - RT Residual Aqua regia 5 hours 85°C

Table 3.4: A summary of the sequential extraction steps adopted

3.8.3 Barite Dissolution

In Chapters 4 and 5 the molar stoichiometry of strontiobarite precipitate was determined by adding 2 mg of sample to 7 mL EDTA-KOH solution (10 mM adjusted to pH 13 using 5 M KOH solution) in 15 mL centrifuge tubes designed to quantitatively dissolve barite. The solution was sonicated for 24 – 48 hours prior to being measured via ICP-AES as previously described (Averyt et al., 2003).

3.9 Solid Phase Characterisation

3.9.1 Environmental Scanning Electron Microscopy (ESEM) – Energy Dispersive X-Ray Analysis (EDX)

ESEM was used to non-destructively image, characterise (chemically and morphologically) the composition and surface structure of solid scale samples (Chapter 6) and strontiobarite minerals produced (Chapter 4) and extracted (Chapter 5) from field sediment and sediment microcosm experiments. The basic principle of ESEM involves focusing a high-energy electron beam on the sample which results in two modes of electron emission and detection (Egerton et al., 2005; Gill et al., 2014). During the bombardment of the sample with a focused

120 primary electron beam, electrons and X-rays can be emitted and scattered back from the sample specimen at various depths and detected to provide a variety of information about the sample (Figures 3.16 & 3.17). This includes; 1) secondary electrons (SE) which emerge with lower energies (≤ 50 eV) generated from close to surface (~ 10 nm) of the sample as a result of beam-induced inner shell ionisation and the repulsive force between the primary electrons and specimen atomic electrons resulting in the ejection of electrons to the surface (inelastic scattering – transfer of energy); 2) back-scattered electrons (BSE) which are of higher energy and generated from deeper/shallower regions (~ 1 µm) and occur due to primary electron interactions with the nucleus of atoms by which the incident electron is reflected from (rather than attracted to) the nucleus and remerges (elastic scattering – no loss of energy) and; 3) characteristic X-rays specific to an element in the sample generated deeper (~ 2-5 µm) resulting from electrons from higher energy shells falling to fill gaps in lower shells emitting characteristic energy X-rays which are then detected by an Energy Dispersive X-Ray Spectroscopy (EDS) system (Figures 3.16 & 3.17). All types of signal provide a variety of information and sample preparation such as coating of the sample (e.g. carbon or gold coating) was kept to a minimal as coating would prove to be problematic on other micro-focus instruments making this instrument very versatile. High resolution topographic imaging is typically produced from secondary electrons where troughs appear darker and peaks brighter. Back-scattered electrons allow the differentiation between elements within a sample as the intensity of back-scattering linearly increases with the average atomic number (Z) of the element being irradiated. Therefore intensity contrasts observed in the BSE image relates to compositional differences between elements where a high atomic number phase appears brighter. X-ray detection via an EDS system allows the investigation of the elemental distribution via spot chemical analysis and elemental mapping of samples (Danilatos, 1993; Reed, 2005; De Mello Donegá, 2014; Gill, 2014).

121

Figure 3.16: Schematic of the electron beam interaction volume and various signals generated

In this thesis a Philips XL30 FEG ESEM and FEI Quanta 650 FEG ESEM (Field Emission Gun Environmental Scanning Electron Microscope) equipped with an EDAX Gemini running Genesis software or Bruker Quantaz EDS (Energy Dispersive Spectroscopy) system with an XFlash detector running Bruker Esprit V2 software, was used for imaging and chemical analysis to obtain information on the morphology, size and chemical composition of samples. Solid scale samples were analysed in polished epoxy resin blocks (Section 3.6) attached to sticky carbon pads on aluminium stubs (diameter - 12 mm) and were uncoated as coating effected FTIR analysis. Particles obtained from heavy liquid extractions and mixing experiments were fixed to sticky carbon pads on aluminium stubs and analysed uncoated to identify the high atomic number element barium in BSE mode. Generally an operating voltage of 15 keV, high vacuum mode and working distance of 10 - 15 mm was adopted using the backscattered electron (BSE) detector, and EDS for elemental mapping and analysis.

122

Figure 3.17: Schematic diagram of the various signals generated by interaction of the electron beam with the sample

3.9.2 Powder X-Ray Diffraction (XRD)

As the wavelength of X-rays (10-10 m) is of the same order of magnitude to the internuclear distance between the atoms in solids (d-spacing), X-ray diffraction is used to determine the atomic structure of crystalline or ionic solids. Principally a finely focused beam of incident X-rays are directed at the sample and X-rays are diffracted upon interaction with the atoms in successive planes of atoms that make up the molecular structure of a solid. The detector detects the X-rays that are scattered / reflected from the atoms in the crystal planes / lattice in different directions and at different angles relating to the specific positions of the atoms known as the ‘scattering angle’. The scattering ‘power’ of an atom is therefore proportional to its atomic number (Z). The X-rays are scattered by the electron cloud (electrons) surrounding the nuclei from which a characteristic three- dimensional scattering pattern of the electron density within the crystal can be produced known as a diffraction pattern (Figure 3.18).

123

Wulf Bragg demonstrated the diffraction of X-rays via crystals can be described by a simple equation known as the ‘Bragg equation’:

2푑 sin 휃 = 푛휆

Equation 3.1- Bragg equation

Bragg showed that waves of X-ray radiation are incident on the array of atoms in the crystal plane at angle theta (θ) and diffracted X-rays will leave at an angle equal to the incident beam when the waves are in phase (constructive) (Equation 3.1). One wave is reflected from an atom in the first lattice plane and the second is reflected by an atom in the second lattice plane (Figure 3.18). The two scattered waves will only be in phase (constructive) if the additional distance (calculated via trigonometry) travelled by the second wave is equal to a multiple (n) of the incident wavelength (λ). Therefore the Bragg equation relates wavelength of incident X-ray radiation to the lattice spacing (d) as this value determines the extra distance travelled by the second wave of radiation (Figure 3.18). Thus the scattered waves can provide information about the lattice spacing and structural information as the inter-plane distance (d) is characteristic of the diffracting crystal providing a method for mineral identification (Burrows, 2013; Gill, 2014).

124

Measurements were performed on a Bruker D8 Advance diffractometer, equipped with a Göbel Mirror and Lynxeye detector to determine the bulk mineralogical composition of samples throughout the thesis (Chapters 4, 5 and 6). The X-ray tube had a copper source, providing CuKα1 X-rays with a wavelength of 1.5406 Å. Sample preparation involved grinding ~ 0.1 g of sample material using a pestle and mortar into a fine powder and then mixed with ~ 1 mL of amyl acetate. The resultant slurries were transferred to glass microscope slides and air dried. Samples were

o o scanned from 5 to 70 , with a step size of 0.02 and a count time of 0.2s per step. Resultant patterns were assessed via EVA version 4 software, to compare experimental patterns to standards from the ICDD (International Centre for Diffraction Data) database. A semi-quantitative analysis of relative mineral amounts was conducted by Topas version 4.2 software (Coelho, 2009; Cullity, 2013).

All XRD analysis was performed by Dr. John Waters, School of Earth & Environmental Sciences, The University of Manchester.

125

3.9.3 Fourier Transform Infra-red Spectroscopy (FTIR)

Infrared radiation has the energy equivalent to the transitions involved in chemical bond vibrations (Burrows, 2013). IR spectroscopy uses the interaction of infrared radiation with matter to obtain information about molecules (e.g. solid, liquid or gas) via transitions between discrete quantized vibrational energy levels to produce characteristic spectra for each molecule. A spectrum consists of many absorption bands at various intensities which signify the absorption of infrared radiation due to the different bond vibrations that take place in the molecule (e.g. stretching, bending and compressing of bonds). Vibration of bonds is eluded to the motion of atoms in individual bonds thus related to the characteristic bond angle and bond length of molecules which have specific associated energies. Therefore interaction of IR radiation with a molecule of the same energy results in the absorption of IR radiation and a peak at a characteristic frequency or wavenumber. IR spectroscopy is therefore used for functional group identification in molecules (e.g. organic chemistry) to determine and confirm compound structures, measure concentrations, vibrational frequencies and force constants of chemical bonds within compounds.

The frequency (V) at which particular bond vibrations occur depends on many factors such as the mass of the atoms in the bond (µ = reduced mass), and the strength of the bond (k = force constant) as these variables dictate the velocity of the vibration of the particular bond of interest depicted by Hooke’s Law (Equation 3.2).

1 푘 푉 = √ 2휋 µ

Equation 3.2 - Hooke’s law

In addition to identifying particular bonds and functional groups present within a sample, information in regards to a molecules bonding environment can also be deduced. Due to symmetry rules only certain vibrational modes of a molecule will

126 be IR active hence absorbs IR radiation. As the bonding environment of a molecule alters its symmetry and therefore number of IR active vibrational modes change. This can be observed by looking at the number and position (shifts) of absorption peaks in the spectrum recorded. Alterations in molecular symmetry and bonding environment can lead to the splitting of degenerate vibrational modes. Thus IR spectroscopy can provide information in regards to the extent of metal substitution, crystal morphology, ion adsorption and degree of crystallinity (Burrows, 2013). A typical FTIR spectrometer consists of a: mid-infrared beam (source of radiation directed at the beam splitter); beam splitter (equally splits the beam which is reflected by the mirrors); fixed mirror and oscillating mirror (reflects the beams back to the beam splitter constructively (in phase) and destructively (out of phase), and the beam is focussed on the sample); and a mercury-cadmium- telluride detector (beam reflected from the sample reaches the nitrogen cooled detector where an interferogram is Fourier transformed into a spectrum) (Smith, 1996; Skoog, 2007)

In this thesis (Chapters 4, 5 & 6) a Perkin Elmer Spotlight-400 ATR-FTIR imaging microscopy system (Figure 3.19) was used (minimum peak-position resolution of 0.1 cm-1) to collect reflected light photomicrographs, mid-IR spectra and infrared maps to identify minerals and metal substitution in scale samples (Chapter 6). In this instance the infrared beam is directed using a series of mirrors from the Universal ATR sampling system into this microscope unit and focussed to a spot for high-precision point analyses between 6 – 1000 µm using a series of aperture windows. Maps in regards to the spatial infrared intensity thus mineralogical distribution was created using HyperView software and spectra were analysed using Spectrum_Image software. The Universal ATR sampling system (Figure 3.19) was used to measure changes in the totally internally reflected infrared beam after contact with precipitate samples (Chapter 4 & 5). The incident IR beam is directed into and protrudes beyond (0.5 – 5 µm) the optically dense diamond crystal in which the evanescent wave is either absorbed (transmission) or altered, thus reflected (reflectance) by the sample. In this case the reflected wave (Figure 3.20) is passed to the detector situated on the opposite end of the crystal in the Spotlight

127

400 system unit in which an IR spectrum presented as a plot of percentage radiation transmitted vs. frequency of the incident radiation is represented as wavenumbers (cm-1). Polished resin blocks and precipitates were scanned at 50-15 µm spatial resolution to produce maps and spectra respectively. Scan intervals of 4 cm-1, scan number of 20 and wavenumber detection range of 500-4000 cm-1 was adopted.

Spotlight 400 Universal ATR Sampling Accessory

Figure 3.20: Schematic of a multiple reflection ATR system

128

3.9.4 Raman Spectroscopy

Raman microscopy is a spectroscopic technique complimentary to infrared spectroscopy which also provides information associated to characteristic molecular vibrations which allows molecules to be identified and quantified. This technique is based on the inelastic scattering of a beam of monochromatic light in the visible, near infrared and near ultraviolet range interacting with the sample. Inelastic scattering occurs when the energy of the incident photons of monochromatic light change upon interaction with the sample. Photons of energy from the laser are absorbed by the sample which induces an electric dipole moment and deformation of the molecules, and a higher energy excited vibrational state is populated. Upon vibrational relaxation characteristic photons of radiation are emitted. The energy of the remitted photons is either higher in energy (anti- stokes), or lower in energy (stokes) in relation to the original incident energy of monochromatic light (‘Raman Effect’) (Figure 3.21). Raman microscopy consists of the accumulation of unique quantized vibrational energy levels and probing the vibrational frequency shifts to identify molecules from these characteristic shifts (Burrows, 2013).

In this thesis a Horiba Scientific, Xplora plus Raman microscope was used (Figure 3.22). Polished blocks were analysed by a wide range of lasers 532 nm, 638 nm and 785 nm with a spatial resolution of 0.5 µm to produce maps. In this project Raman spectroscopy had proven to be an ineffective technique during method development for solid scale characterisation due to the tendency of burning samples due to the differing structural and compositional properties of each sample (Figure 3.22).

129

Fig 3.21: Diagram showing resonance (elastic scattering known as Rayleigh), Stokes lines (inelastic) and anti- Stokes lines (inelastic)

A) B)

Figure 3.22: (A) Image of the Horiba Xplora Raman microscope at the University of Manchester (left); B) Undesired burning of sample due to its chemical and mineralogical properties (right)

3.9.5 BET Surface Area Analysis

The specific surface area of strontiobarite precipitates formed from mixing experiments (Chapters 4 and 5) were examined using the Brunauer-Emmett-Teller (BET) method for comparison (Brunauer et al., 1938). The technique entails the use

130 of adsorbates in the form of gases e.g. nitrogen to adsorb to the surface of the sample in which a BET isotherm is constructed (Atkins et al., 1989). BET surface area measurements were performed using a Micromeritics Gemini V, model 2365. Sample vials were purged using a Micromeritics Flowprep 060 prior to analysis under a flow of CP grade nitrogen gas while heated to 80 - 100oC for two hours (due to the nature of the samples). After air cooling, the purged vials were weighed, before the addition of sample materials (typically 0.1 - 0.2 g), which are then purged under similar conditions for 18 hours. Once cool, the sample vials are again weighed and the mass of the empty vial is subtracted to give the mass of the purged samples within. The number of nitrogen molecules required to form a monolayer on the substrate (sample) surface is determined, as the molecular dimensions of nitrogen is known and the surface area (m2 g-1) of the sample is calculated.

3.9.6 X-Ray Absorption Spectroscopy (XAS)

X-ray absorption spectroscopy (XAS) is a non-destructive quantitative technique typically used to determine the local coordination environment of an element and the oxidation state in crystalline or amorphous samples (Penner-Hahn, 2001; Gill, 2014). Advantages of this technique include the small amount of sample required for analysis, rapid data collection time and variety of conditions e.g. temperature and pressure at which samples can be studied. Briefly, a monochromatic beam of x- rays produced by a synchrotron is scanned and directed across a sample in which the incident x-ray energy can be defined by the monochromator. The interaction of the x-rays possessing energies at specific incident energies (E) equal to the binding energy (Eb) of core electrons results in excitation and ejection of electrons into the continuum state, and formation of inner core holes (see also section 3.9.1 and 3.9.7). The vacancy is consequently occupied by electrons from higher levels (M or N shell) leading to the emission of a photon of energy (secondary fluorescent X- rays) or, Auger electron via the transfer of energy to an outer electron during this process (Figure 3.17). Most importantly, the specific energy that is required to excite an inner electron (i.e. the binding energy) is characteristic to individual

131 elements and dependent upon the oxidation state of the absorbing atom thus allowing for elemental specific information. The absorption of energy is generally represented by an abrupt increase in absorption cross-section representing the ejection of electrons from inner core shells known as the absorption edge in which the edges are labelled K (1s), L (2s, 2p (j=1/2), 2p (j=3/2)) and M (3s, 3p, 3d) in accordance to the electron being excited (Parsons et al., 2002; Newville, 2004; Gill, 2014) (Figure 3.23). Absorption edges exist if the energy of the incident photon is equal or larger than the binding energy of a core electron and the position of the edge is dependent upon the oxidation state of the element (Newville, 2004, 2014). A spectrum is generally divided into four sections (Figure 3.23); 1) The information collected before the absorption edge is referred to as the pre-edge region determined by the improbable transition of core electrons to the lowest unoccupied states as the incident x-ray energy is lower than the electron biding energy; 2) the X-ray absorption near edge structure (XANES) which extends out to ~ 10 eV from the absorption edge, and is where an electron is excited and ejected (photoelectrons) shown by the sharp increase in absorption as the incident x-ray energy is greater than the electron binding energy; 3) the near edge x-ray absorption fine structure (NEAXFS) which extends out to ~ 50 eV relating to multiple scattering events of photoelectrons with energies just above the electron binding energy with multiple neighbouring atoms. This feature is usually categorised with the XANES region, and sensitive to oxidation state and geometry, and is usually difficult to model and used as a ‘fingerprint region’ in comparison to the EXAFS region; 4) the extended X-ray absorption fine structure (EXAFS) region which extends to 1000 eV relating to the oscillations generated in the final region of spectra due to single scattering of high energy photoelectrons (well above the electron binding energy) with nearest neighbouring atoms, used for quantitative determination of bond length, coordination number, interatomic distances and bonding environment (local order). As a photoelectron is ejected from the absorbing atom it is scattered by the nearest neighbouring atoms in which the resulting wave interactions (constructive and destructive) produce oscillations characteristic to the location and identity of the nearest neighbouring atoms and

132 the amplitude related to the electron density and number of backscattering neighbouring atoms (Figure 3.23) (Newville, 2004; Willmott, 2011).

Figure 3.23: XAS spectrum indicating the various features and regions observed in a spectrum and resultant scattering effects of emitted photoelectrons with neighbouring scattering atoms (multiple and single events) (http://www.chem.ucalgary.ca/research/groups/faridehj/xas.pdf)

X-ray absorption involves either the measurement of the incident and the transmitted X-rays that pass though the sample (transmission mode) or secondary fluorescent X-rays emitted at near right angles (45 °) (fluorescence mode) (Dent et al., 2009) (Figure 3.24). The measurement of the x-ray absorption coefficient of a sample as a function of energy (µ(E)) is determined from expressions of the Beer- Lambert law (Teo, 1986; Denecke, 2006; Gill, 2014).

In this thesis the speciation of strontium associated with barite was analysed via X- ray absorption spectroscopy (XAS). A sample (~ 1.2 and ~ 2.4 wt. %) was prepared as a pellet using cellulose as the binder and data was then collected at the Diamond Light Source, Harwell, UK on B18 beamline (Figures 3.24). Sr K-edge spectra were collected in transmission mode at liquid nitrogen temperature. Background subtraction and improvement of signal to noise ratio of the data was obtained via averaging multiple scans for the sample using the Demeter software package Athena and Artemis FEFF6 (Ravel et al., 2005). Modelling was performed with the support of Mr Gianni Vetesse.

133

A) B)

Figure 3.24: A) Image of the B18 beamline set up at the Diamond Light Source, Harwell, UK (left) and; B) a schematic of a typical beam line set up for the collection of XAS data in fluorescence and transmission modes (Parsons, Aldrich and Gardea-Torresdey, 2002)

3.9.7 X-Ray Fluorescence (XRF)

X-ray fluorescence spectrometry is a non-destructive analytical technique which entails the use of a radiation source usually an X-ray tube (encompassing a tungsten filament, metal anode and transparent beryllium window) in which X-ray radiation is used to excite individual atoms within samples. The bombardment of high-energy x-rays leads to ionisation thus the generation and emission of secondary X-rays (fluorescent) due to the formation of ‘holes’ in inner orbital electron shells (K or L shell) of atoms as electrons are ejected and subsequently occupied by electrons from higher levels (M or N shell) during this process (Figure 3.17). The X-rays are of characteristic energies related to the different major and trace elements present for a given sample and detected at various wavelengths according to the conditions described by Bragg’s law (see also Section 3.9.1 and 3.9.2) (Jenkins, 1999).

The height intensity of each characteristic X-ray peak corresponds to the concentration of the given element in the sample allowing quantitative analysis (1 ppm to 100%). The characteristic energy of the X-rays emitted allows the identity of the specific individual elements to be deduced and the discrete electronic transition from which the X-ray has been generated (Gill, 2014).

134

In this thesis XRF was used to characterise the bulk elemental concentration of field sediments and solid scale samples (Chapters 4 & 6) using a PANalytical Axios Sequential Spectrometer. Samples were dried at 40 °C, ground to a fine powder using a pestle and mortar (12 g), mixed with 3 g wax binder (Hoechst wax C micro powder) and then pressed into a homogenous pellet.

3.9.7.1 Loss on ignition

Organic matter (OM) content of sediment can be crudely determined from loss on ignition which entails strongly heating a sample at selected temperatures inducing the evaporation of moisture and volatile substances (e.g. organics and carbonates), and calculating the percentage weight lost (Sutherland, 1998). A porcelain crucible was ignited at 110 ⁰C for 30 minutes in a muffle furnace, then allowed to cool in a desiccator and accurately weighed. Sediment (1.0 - 2.0 g) was placed in the clean pre-ignited crucible and weighed accurately prior to analysis. The crucible was then transferred to a muffle furnace, heated to 110 ⁰C for 1 hour and then cooled to ambient temperature in a desiccator prior to reweighing to calculate the moisture (water) loss (dry weight). The crucible was then further heated to 1000 ⁰C for 1 hour to ash the sediment to combust organic carbon (OC) including carbonates and then cooled in a desiccator to ambient temperature prior to being reweighed to obtain ignition loss data as a percentage. The following equation is used to calculate loss on ignition:

푀푐 − 푀푑 퐿표푠푠 표푛 푖푔푛푖푡푖표푛 = × 100 푀푏 − 푀푎 Equation 3.3: Loss on ignition calculation where,

Ma is the mass of the weight of the empty pre-ignited crucible in grams

Mb is the weight of the crucible containing the sediment sample in grams

Mc is the mass of the crucible containing sediment after heating to 110°C in grams

135

Md is the mass of the crucible containing sediment after heating at 1000°C containing the ashed sediment in grams

3.9.8 Gamma Spectroscopy

Gamma spectroscopy is a technique used to determine the radiological composition of samples via the identification and quantification of radionuclides. Gamma ray emission is not spontaneous but occurs as a result of an alpha or beta decay process which leaves the daughter nucleus in an excited state. The particular energy levels associated and populated by each nuclei is characteristic of each individual atom. Therefore the gamma rays emitted are of characteristic energies corresponding to well-defined energy differences between the excited energy level (ground state parent) and the de-excited energy level (ground state product) of particular atoms which allows the identity of elements and isotopes to be achieved. Gamma rays can be counted with high precision using semiconductor detectors composed of Ge(Li) (germanium-lithium) and Si(Li) (silicon-lithium), and to also provide high energy resolution (Eberth et al., 2008). The former has a higher atomic number and density thus suited for high energy gamma rays, whereas the latter is generally suited for lower energy x-rays and gamma rays (< 100 keV). Gamma rays are typically absorbed by the detector crystal principally via pair production, photoelectric and Compton Effect. The preferential mechanism is the photoelectric effect as the detector absorbs all of the energy of the incident gamma ray energy. In the semiconductor crystal a band structure exists, with the band gap being larger for Si (1.1 eV) in comparison to Ge (0.66 eV). In the event of emission of gamma ray radiation sufficient energy is supplied to the semiconductor to promote electrons from the valence band to the conduction band thus forming electron-hole pairs (Figure 3.25) where the number of promotions to the conduction band is proportional to the energy of the incident photon. Subsequently in the presence of an electric field or voltage applied to the detector, electrons and holes migrate to the cathode (positive contact) and anode (negative contact) respectively, generating a current (Gilmore, 2008). The current is then amplified and transported to the ancillary counting equipment. The signal is digitised and stored in a Multi-

136

Channel Analyser (MCA) to generate a spectrum composed of characteristic gamma lines. Determination of the photon energy is calculated from total number of electron-hole pairs generated which is proportional to the size of the signal produced from the detector. A variety of radionuclides can be identified and measured from a single spectrum as it spans a wide energy range. Typically the operation conditions entails the solid-state detector to be constantly cooled (77 K) with liquid nitrogen to prevent thermal excitation of electrons, overheating, and damage to the crystal detector and to increase resolution. The detection limit (50 mBq) and counting efficiency of such instruments greatly depend on interfering radionuclides, shielding (lead enclosure to reduce background radiation), the size of the crystal, gamma ray energy, and sample distance from the detector. In order to quantitatively determine activity concentrations of radionuclides in samples, standards are required to calibrate and determine detector efficiency matching the volume and matrix (standard geometry) of the samples of interest (Ehmann et al., 1994; Gilmore, 2008).

In this thesis gamma spectroscopy was used to determine radium concentrations in field sediment samples and solid scales (Chapters 4 & 6). The main advantage of this technique is the little need for chemical alteration via methods such as separation prior to analysis which additionally minimises cross-contamination.

Figure 3.25: The band structure of the semiconductor high purity crystal, electric field (←)

137

3.9.8.1 Detection Calibration & Counting Geometry Procedure

A standard counting geometry was determined prior to the analysis of sediment samples. This consisted of sealing sediment (50 g) in a polypropylene container to provide a tight seal and prevent loss of 222Rn gas. Samples were left for 30 days to allow the radium daughter progeny to equilibrate and reach secular equilibrium (226Ra, 222Rn, 214Bi and 214Pb) in which the activity concentration of these radionuclides is uniform. Height of the sediment in the container were checked and corrected prior to sealing samples. A calibration standard was made prior to counting the samples. The standard was made by adding homogenised soil sample (50 g) to a beaker, adding ethanol (200 mL) and stirring prior to the addition of aqueous radium spike (200 Bq) to the top layer of ethanol. The slurry was further stirred to homogenise the sample and then left to dry in the fume hood until all ethanol evaporated. The dry sample was then transferred into a polypropylene container in standard geometry and sealed for 30 days prior to counting. Both the standard and samples were counted for a period of 24 hours and the primary quantification of radium was estimated from measurements of the 214Bi daughter product gamma line at 609 keV and 214Pb gamma line at 352 keV. These peaks were analysed to isolate photo peaks in areas less obscured by the Compton distribution/background (produced as a result of scattering gamma rays; Compton effect) and to obtain a mean estimate for the 226Ra content from averaging these gamma lines. The increase in peak areas and net area counts (number of counts) were measured at these respective energies thus the detector efficiency and quantification of radionuclides determined. The activity (Bq g-1) is calculated from the count rate (counts per time) generated from the detector upon interaction with gamma radiation along with the efficiency of the detector as shown from equations:

푁푢푚푏푒푟 표푓 푑푒푡푒푐푡표푟 푐표푢푛푡푠 퐶표푢푛푡 푅푎푡푒 (푐표푢푛푡푠 푠−1) = 퐷푒푡푒푐푡푖표푛 푡푖푚푒 (푠)

Equation 3.4: Count rate calculation

138

퐶표푢푛푡 푟푎푡푒 (푠) × 퐴푐푡푖푣푖푡푦 (퐵푞) 퐴푐푡푖푣푖푡푦 (푢푛푘푛표푤푛) (퐵푞 푔−1) = (푢푛푘표푤푛) (푡푟푎푐푒푟/푠푡푎푛푑푎푟푑) 푀푎푠푠 표푓 푠푎푚푝푙푒 (푔)× 퐶표푢푛푡 푟푎푡푒 (푠)(푡푟푎푐푒푟/푠푡푎푛푑푎푟푑)

Equation 3.5: Sample activity calculation

A high-purity Ge(Li) Canberra gamma spectrometer was used to analyse samples. Routinely, standards were re-analysed to evaluate the stability of the gamma spectrometer and background measurements were taken to provide information on the detector efficiency.

3.9.9 Autoradiography

Autoradiography is a technique used to visualise and determine the spatial distribution of radionuclides in a sample via the use of a recording medium which is most commonly a photographic film which acts as a detector. There are many different forms of solid state nuclear track detectors (SSNTD) such as, photographic emulsions, glasses, plastics and crystals. A latent image is produced via the exposure of a radioactive sample by the absorption of ionising radiation emitted from the radioisotopes (e.g. x-ray, beta and gamma radiation) in the sample by a storage phosphor screen. Following the exposure (generally longer exposure times result in higher image resolution), the screen which has stored the radiation emitted by the sample is scanned with a helium-neon (HeNe) laser beam. The latent image is made visible after the development stage in which a map of the radionuclide distribution is produced.

Storage phosphor screens generally contain a phosphorescent substance in the form of a transition-metal lanthanide complex such as barium fluorobromide (BaFBrEu2+) and comprise of three layers (Figure 3.26). The process of emission of radiation upon loss of excitation exhibited by phosphorescent substances is termed luminescence, however aspects of this phenomenon are defined as fluorescence or phosphorescence. The crystal lattice of the BaFBrEu2+ (5 µm) are located in the central layer of the phosphor screen regarded as the ‘luminescent centre’ as the bivalent europium cation (Eu2+) adopts the luminescent properties. The phosphor

139 layer is shielded from damage by a protective layer and base layer which comprises of a bendable or inflexible plastic for support (Figure 3.26). Incident radiation absorbed by the lanthanide compounds results in photoionization excitation of the Eu2+ resulting in photo stimulated luminescence (PSL). Many theories exist proposing mechanisms for energy storage and release. Generally Eu2+ is oxidised in which the electron excites into the vacant hole in the conduction band of BaFBr forming Eu3+ and BaFBr-, a metastable state, in which the energy emitted by the sample is stored. The BaFBr- complex strongly absorbs at ~ 590 nm and irradiation with laser light (~ 633 nm) results in either; 1) the excited BaFBr- complex recombining with the Eu3+ ion resulting in the formation of an excited Eu2+ (4f6 5d1 state) which decays to the ground state emitting energy at 390 nm or; 2) a recombination mechanism in which BaFBr and BaFBr- centres combine resulting in the excitation of Eu2+, thus resulting in photo stimulated luminescence (Leblans et al., 2011). When the screen is scanned the luminescence is quantified, measured and digitized using a imaging system in which a spatial distribution map is produced on screen as an image (Johnston et al., 1990). The image produced allows the activity of particles to be estimated as the degree of luminescence measured is proportional to the activity in the sample. This also allows the spatial distribution to be determined accurately.

In this thesis naturally occurring radioactive materials in the form of solid pipe scale and individual grains isolated from field sediment samples were analysed via autoradiography (Chapters 4 & 6). Samples were placed on the storage phosphor screen BAS-IP SR, super resolution (G.E Healthcare) in a dark cupboard for 6 hrs up to 4 weeks. Samples were either contained in an epoxy resin block (Chapter 6) or fixed to sticky carbon pads on aluminium stubs and placed onto the screen (Chapter 4). The activity on the surface of the sample was imaged using a Typhoon 9410 variable mode imager where the screen was then scanned and imaged using a HeNe laser (l ¼ 633 nm) with a pixel size of 10 - 25 µm. The extent of darkening recorded is quantitatively proportional to the activity on the sample surface displayed using ImageQuant_software.

140

Figure 3.26: Schematic of the set up and process of using a storage phosphor screen (BAS-IP) and its composite structure (https://www.nationaldiagnostics.com/electrophoresis/article/autoradiography)

141

4.0 Microbial Community Analysis – DNA Sequencing

DNA sequencing is used to compare sample phylogeny and taxonomy composition from complex environmental samples by the sequencing of 16S ribosomal RNA genes. This technique is inexpensive and fast allowing microbe identification via the determination of microbial sequences. Briefly this method entails the following steps: 1) extraction of DNA from sediment; 2) amplification (replication) of 16s rRNA gene via a polymerase chain reaction (PCR); 3) separation of DNA via gel electrophoresis to target the specific base pair product; 4) sequencing of the PCR products (amplicons) using region of interest-specific forward and reverse primers and; 5) purification, cleaning, normalisation and pooling of PCR products, followed by characterization via comparison to known 16s rRNA gene sequences (Figure 3.27).

Figure 3.27: A workflow diagram illustrating the 16s rRNA sequencing protocol (https://infravec2.eu/product/metagenomic-illumina-sequencing-of-16s-rrna-gene-amplicons)

142

In this thesis (Chapter 5) sediment was extracted aseptically under anaerobic conditions from all sets of biogeochemical microcosm experiments at Day 1 and Day 300, specifically from amended and non-amended microcosm bottles (e.g. treated and non-treated with electron donor). Investigation into the compositional change in the microbial community biodiversity pre- and post-anoxia and potential effects of radium on the microbial community, and associated risk was conducted similar to other studies (Burke et al., 2005; Wilkins et al., 2007; Newsome et al., 2017). A frozen (- 80 °C) anaerobic back sample of sediment representative of that used in microcosm experiments was also analysed to represent the natural field conditions (e.g. Day 0).

DNA was extracted from 200 µL of sediment slurry using a DNeasy PowerLyzer PowerSoil Kit (Qiagen, Manchester, U.K). Amplification of the 16S rDNA via PCR entailed; strand separation via heating at (94-96 °C) to separate the strands; hybridization of 8F (5’-AGAGTTTGATCCTGGCTCAG-3’), and 1492R (5’- TACGGYTACCTTGTTACGACTT-3’) primers at (55-65 °C); and DNA synthesis in which Taq polymerase extends the primers to complete the replication of the original strand of DNA at 72 °C. The 16S rRNA gene was chosen as it is very well conserved among most bacterial species. Following amplification via PCR, the DNA was stained before placement in an agarose gel, where it was subsequently separated using electrophoresis to determine the purity of the products (amplicons) formed in the first amplification (PCR) step. The stained DNA was viewed under UV light, to identify the 16S rRNA products (~ 1500 base pair) by comparison to a ladder of DNA fragments of varying lengths.

Sequencing of PCR amplicons of 16S rRNA was performed using the Illumina MiSeq platform (Illumina, San Diego, CA, USA) specifically targeting the V4 hyper variable region using forward primer, 515F, 5′-GTGYCAGCMGCCGCGGTAA-3′; and reverse primer, 806R, 5′-GGACTACHVGGGTWTCTAAT-3′, for 2 × 250-bp paired-end sequencing (Illumina) (Caporaso et al., 2011, 2012). PCR amplification (second step) was performed using Roche FastStart High Fidelity PCR System (Roche Diagnostics Ltd, Burgess Hill, UK) in 50 μl reactions. This entailed: 1) initial denaturation and separation at 95°C for 2 min, followed by 36 cycles of 95°C for 30 s; 2) hybridisation

143 of primers at 55°C for 30 s; 3) synthesis at 72°C for 1 min and; 4) a final extension step of 5 min at 72°C. Purification and normalisation of the PCR products to ~ 20 ng each was achieved using the SequalPrep Normalization Kit (Fisher Scientific, Loughborough, UK). The PCR amplicons from all samples were pooled in equimolar ratios. The run was performed using a 4 pM sample library spiked with 4 pM PhiX to a final concentration of 10 % following the method of Schloss and Kozich (Kozich et al., 2013).

Raw sequences were divided into samples by barcodes (up to one mismatch was permitted) using a sequencing pipeline. Quality control and trimming was performed using FastQC Cutadapt, and Sickle. SPADes (Nurk et al., 2013) was used to perform MiSeq error correction. Forward and reverse reads were incorporated into full-length sequences with Pandaseq (Masella et al., 2012). Chimeras were screened and removed using ChimeraSlayer (Birren et al., 2011). Subsequently Operational taxonomic units (OTU) were generated with UPARSE (Edgar, 2013) and classified by Usearch (Edgar, 2010) at the 97% similarity level, and singletons were removed. Rarefaction analysis was conducted using the original detected OTUs in Qiime to describe microbial diversity (Caporaso et al., 2010). The taxonomic assignment and phylogenetic examination was performed by the RDP (Wang et al., 2007) classifier and using the Blastn nucleotide search (http://blast.ncbi.nlm.nih.gov) as described previously (Lloyd et al., 2019).

All DNA extractions and analysis was performed by Mr. Christopher Boothman, School of Earth & Environmental Sciences, The University of Manchester

144

CHAPTER 4

Fate of Radium on Discharge of Oil Produced Water to the Marine Environment

This chapter is a manuscript prepared for submission in the journal Marine Pollution Bulletin. Supporting Information provided with this manuscript is included in the following manuscript

145

Faraaz Ahmad1, Katherine Morris1, Gareth T.W. Law2, Kevin Taylor1 and Samuel Shaw1* 1Research Centre for Radwaste Disposal and Williamson Research Centre, School of Earth & Environmental Sciences, Williamson Building, The University of Manchester, M13 9PL, United Kingdom;

2Radiochemistry Unit, Department of Chemistry, University of Helsinki, P.O. BOX 33 (Yliopistonkatu 4), 00014, Finland

*Corresponding Author ([email protected])

Keywords: Radium, Produced water, Precipitation, Barite, Offshore discharges, NORM

4.1 Abstract

Understanding the speciation and fate of radium during operational discharge from the offshore oil & gas industry into the marine environment is important in assessing its long term environmental impact. In the current work, we explore the behaviour of radium in experimental systems where synthetic produced water and seawater were mixed under laboratory conditions. Experiments showed that a significant proportion of radium (up to 48 % in 1 hour) co-precipitates with barium during mixing. Additionally, barite precipitation occurred during mixing of seawater and produced waters from an oil field similar in character to the synthetic water experiments. Radium concentrations in sediments from a discharge site were assessed using gamma spectroscopy. Radium was present in selected field samples (0.1 - 0.3 Bq g-1), and heavy liquid extractions were used to separate barite particles from the marine sediments. Barite particles were then characterised using SEM and autoradiography confirming the radiobarite fate pathway.

146

4.2 Introduction

The presence of radionuclides, such as 226Ra and 228Ra, from the decay of naturally occurring 238U and 232Th have proven to be significant in produced water effluents from oil and gas platforms across the world (Fisher, 1998; Røe Utvik, 1999). During extraction and production of oil and gas, radionuclides can be transported from the subsurface to the produced waters dependent upon the chemical conditions of the reservoir (IOGP, 2016). Discharges of offshore effluents can thus result in the release of naturally occurring radionuclides into the marine environment (Holdway, 2002; Grung et al., 2009; Dowdall et al., 2012; Bakke et al., 2013). In addition the release of produced water into the marine environment can result in the formation of inorganic particles e.g. barite (BaSO4) due to the mixing of incompatible waters (e.g. produced water and seawater) (Zhang et al., 2014). These particles can contain radium (e.g. 226Ra) due to the ability of this radionuclide to co-precipitate into binary phases e.g. celestine (SrSO4) and barite (BaSO4) and in ternary phases e.g. strontiobarite (BaSrSO4) (Al-Masri et al., 2005; Rosenberg et al., 2014; Zhang et al., 2014). To understand the environmental fate of radium from produced waters it is critical to gain an understanding of radionuclide incorporation during the particle formation process and nature of the material which forms upon discharge of produced water into the marine environment. This will help to underpin assessments of the environmental risk and fate of radium in these systems.

Bedrock in most hydrocarbon formations is typically accompanied by formation water. Formation water is transported to the surface with crude oil, natural gas and sometimes sea water during extraction as a complex mixture (Hunt, 1979; Holdway, 2002; Jerez Vegueria et al., 2002). This mixture subsequently undergoes industrial separation and leads to the creation of produced water, the largest effluent in the petroleum industry (Grung et al., 2009). The chemical and physical compositions of produced waters differ due to factors such as differing reservoir geology and the stage of oil and gas production. Produced waters comprise of dissolved cations

+ + 2+ 2+ 2+ 2+ 2- - - 2- (Na , K , Ca , Mg , Ba , Sr ), anions (SO4 , Cl , HCO3 , CO3 ) and dissolved gases. Concentration of salts vary from a few mg L-1 up to 345 000 mg L-1 (Jacobs et al.,

147

1992; Røe Utvik, 1999; Fakhru’l-Razi et al., 2009; Zhang et al., 2014). Isotopes of radium (224Ra, 226Ra and 228Ra) and lead (210Pb) are also present in produced waters due to leaching of primordial radionuclides (uranium and thorium) situated in reservoir rock, which leads to their mobilisation to pore waters and resultant detection at a range of offshore installations around the world (Doyi et al., 2016).

As a result of temperature and pressure conditions altering and/or injection of seawater to maintain reservoir pressure during oil and gas extraction, carbonate and sulphate scales deposit on the inside surfaces of production equipment (Todd et al., 1990; Yuan et al., 1994; Al-Masri et al., 2005; Badr et al., 2008; Garner et al., 2015). The ability of radium to incorporate via co-precipitation into insoluble barium and strontium sulphate mineral phases results in the formation of naturally occurring radioactive material (NORM) in the form of RaxBa1-xSO4 (radiobarite) and

BaxSryRazSO4 (radiostrontiobarite) in production equipment (Garner et al., 2015; Doyi et al., 2016). The dominant formation mechanism of this radium-containing scale is the mixing of chemically incompatible waters thus establishment of supersaturated solutions during oil extraction operations, where injection of fluids primarily sea water, interact with produced water to form precipitates. Due to the ionic radii compatibility between ions in solution (e.g. radium and barium) co- precipitation of radium in barite occurs (Zhang et al., 2014). When sea water

2- containing a high concentration of sulphate (SO4 ) but low concentrations of divalent cations (Ca2+, Mg2+, Ba2+ , Ra2+ and Sr2+), and produced water which is characterised by low sulphate and high divalent cation concentrations alongside elevated radionuclide levels are mixed, precipitation of radium-containing sulphate particles follows, and water lower than its initial radioactivity is subsequently produced (Candeias et al., 2014). Discharge of produced water effluent to the marine environment, an essential plant operation, may drive the precipitation of sulphate particles with the potential uptake of radium, or may lead to the sequestration of radium via adsorption to existing particulates in the sea column and/or to sediment (Neff, 2002; Fakhru’l-Razi et al., 2009; Grung et al., 2009). There is uncertainty about the nature of these interactions and the speciation of radium during mixing of these waters which is poorly constrained. Understanding the

148 removal mechanisms associated with radium during operational discharges from the offshore oil and gas industry into the marine environment is essential in underpinning predictions of the fate of 226Ra in these systems and further defining its overall environmental impact. Radium scavenging mechanisms such as adsorption, precipitation and aqueous dispersion are key fundamental processes effecting the mobility and fate of radium. Overall it is expected that the formation of inorganic micro-particulate radiostrontiobarite (RaBaSrSO4) ternary phases during production water / seawater mixing, via the mechanism of co-precipitation is a major pathway controlling 226Ra fate in these systems, but there is a paucity of direct experimental evidence for this process (Gafvert et al., 2007).

Studies show the total activity of radium discharged globally from production sites can be significant with levels up to 1200 Bq L-1 reported across installations (IOGP, 2016). Dowdall et al., (2012) reported discharges from Norwegian sites are typically of the order of 306 – 480 x 109 Bq y-1 with an average radium activity concentration of 3.3 Bq L-1. Eriksen et al., (2006) showed the concentration of 226Ra discharged from North Sea platforms varies from below limit of detection (0.5 - 1 Bq L-1) to 21 Bq L-1 (Eriksen et al., 2006; Olsvik et al., 2012; Bakke et al., 2013). Jerez Vegueria et al., (2002) showed offshore discharges from platforms in Brazil are on the order of 2 - 30 m3 d-1 with a varying 226Ra concentration between 0.012 – 6 Bq L-1 (Jerez Vegueria et al., 2002). These studies illustrate discharge volumes and levels of activity vary considerably globally with up to 37 Bq L-1 reported from American platforms (Pardue et al., 1998). Background radium concentrations within surrounding seawaters are typically around three orders of magnitude lower than that in produced waters (e.g. 0.01 – 0.03 Bq L-1) (Jerez Vegueria et al., 2002; Gafvert et al., 2007; Dowdall et al., 2012). This suggests that discharge of produced waters provides a point source elevation in radium discharges. In terms of dilution and dispersion, research by Jerez Vegueria et al., (2002) revealed the absence of significant radium contamination (above background levels) within sediment or seawater samples around platforms as a result of dispersive effects by currents (Jerez Vegueria et al., 2002). By contrast possible accumulation of radiobarite has been identified in other studies inferring, that under certain conditions discharge of

149 effluents may result in barite precipitation and subsequent sedimentation. For example, Pardue et al., (1998) identified contaminated oilfield sediment samples contained barite from bulk chemical analysis (e.g. X-ray diffraction), thus postulated radium solubility is most likely controlled by co-precipitation with barite. Correlations between radium concentrations and the identification of barite in this study is suggestive of this mechanism however to date, NORM particles have not been extracted and directly characterised to provide evidence of radiobarite in either experimental or field based studies (Pardue et al., 1998). Contrast in findings from different sites suggest that the setting and characteristics (e.g. water depth, salinity and mineralogical distribution differences which exist between deep sea and estuarine/shallow marine settings) of the receiving environments is key to controlling the mechanisms of radium interactions, and thus the positive identification of radium accumulation in sediments at distances from a discharge point (Landa et al., 1983; Pardue et al., 1998; Gafvert et al., 2007; Van Sice et al., 2018; McDevitt et al., 2019).

The fundamental co-precipitation process of radium uptake into barite has been widely studied experimentally over the past century in relation to produced water discharges and other areas such as the nuclear sector i.e. in uranium mine tailings (Doerner et al., 1925; Gordon et al., 1957; Beneš et al., 1981; Fedorak et al., 1986; Jerez Vegueria et al., 2002; Rosenberg et al., 2011(b), 2014; Zhang et al., 2014). Results from numerous studies show that radium removal via co-precipitation into binary (Ra-BaSO4 or Ra-SrSO4) or ternary phases (Ra-BaSrSO4) is dictated by the ionic strength of the solution and the nucleation kinetics of the barite mineral (Rosenberg et al., 2011(a); Rosenberg et al., 2011(b)). The distribution model has been widely adopted to describe the co-precipitation of radium into barite through radium and barium concentration ratios in the aqueous and solid phases known as the concentration-based effective partition coefficient (Equation 4.1).

150

Equation 4.1: Concentration-based effective partition coefficient; [Ra] represents the aqueous concentration of the element, and dRasolid represents the concentration of the element in the solid (e.g. Ra or Ba) (Rosenberg et al., 2014)

Experimental partition coefficients reported in simple analogous systems range between 1.07 – 1.54 in dilute solutions (0 M) and up to 7.49 with increasing ionic strength (3 M) (Rosenberg et al., 2014; Zhang et al., 2014). To date work on the co- precipitation of radium during barium uptake has been based on simple synthetic solutions rather than field samples or synthetic (full-component) brines. Therefore the formation process, composition and morphology of the NORM which is produced during discharge to the marine environment is poorly constrained under field relevant environmental conditions.

In this study, firstly we explored the uptake and fate of radium in sediment samples obtained from a field site where produced waters are discharged into the marine environment. The bulk field sediment samples were assessed using gamma-ray spectroscopy, X-ray diffraction (XRD) and X-ray fluorescence (XRF). In addition, heavy liquid extractions were utilised to separate barite particles from marine sediments, and to characterise their bulk chemistry, morphology/particle size (Scanning electron microscopy, SEM) and radioactivity (autoradiography). Secondly, we extend these observations to model systems where synthetic (full-component) and field produced waters were mixed with synthetic and field seawaters under laboratory conditions to mimic the formation process of (radio)barite particles by discharge into the marine system. By combining the field and laboratory based components of the study we were able to assess the fate of radium (226Ra) within produced water when discharged into the marine environment.

151

4.3 Materials and Methods

4.3.1 The study area and experimental method 4.3.1.1 Marine Sediment

Sediment samples were collected in April 2017 close to a near shore produced water discharge point. A total of 5 samples were taken from different distances from the outfall (20 - 250 m) (Fig. 4.1). Seabed sediment (10 - 15 cm) was removed using a Day grab (0.1 m2) and were placed in sterile bags in an ice box to maintain the redox conditions during transport. Samples were stored at 4°C and prior to characterisation homogenised and dried.

Figure 4.1: Schematic of the field sediment samples obtained (+) from the offshore marine system surrounding the discharge outfall (•)

4.3.1.2 Produced water and seawater mixing experiments

Produced water from an active oil field in the North Sea and seawater from Formby Beach in Merseyside, UK was obtained. Synthetic North Sea seawater and produced water compositions were also synthesised (Table 2) (Todd et al., 1992). Both field and synthetic seawaters and produced waters were mixed in large scale (5 L) precipitation experiments in a 9:1 ratio and stirred (180 rpm) for 24 hrs. This was

152 followed by filtration (0.22 µm PES filter) and drying (40°C ± 0.5°C) to harvest the resultant precipitates for characterisation. Modelling using PHREEQC 3.3.9 was used to assess the saturation index of the solid phase (e.g. barite) for these experimental systems.

4.3.1.3 Radium uptake experiment

A synthetic produced water stock solution (Table 2) was spiked with 226Ra (100 Bq mL-1) and pH was readjusted with dilute KOH solution to allow the detection of radium over the course of the experiment. Small scale synthetic seawater and spiked produced water experiments were conducted by mixing the solution in a 9:1 ratio. Each experiment had a total volume of 2 mL and several experiments were run in series to allow sacrificial sampling. The samples were continuously mixed (180 rpm) for seven hours with periodic sampling. After each time point samples were centrifuged at 14800 rpm for five minutes and radium concentrations were determined by adding 1 mL of the solution to 10 mL of scintillation cocktail in a sealed tube. Samples were then analysed in a Quantalus scintillation counter (Perkin Elmer) after 1 month to allow the Ra progeny to equilibrate. Parallel inactive experiments were run to allow cation and anion analysis using Inductively Coupled Plasma Atomic Emission Spectroscopy (Perkin Elmer Optima 5300 dual view system) and Ion Chromatography (Dionex ICS 5000).

4.3.2 Analytical methods

4.3.2.1 Solid phase characterisation

XRD (Bruker D8 Advance diffractometer) and XRF (PANalytical Axios) were conducted to determine bulk mineralogical and chemical compositions of inactive precipitates and sediments. Sediment was dried in an oven (40°C ± 0.5°C), disaggregated using a pestle and mortar and homogenised for XRD and XRF analysis. Organic matter content of the sediments was estimated from loss on ignition (Siddeeg et al., 2015). Fourier-transform infrared spectroscopy (Perkin Elmer Spotlight 400 FTIR imaging system Universal ATR) was used to characterise

153 the inactive precipitate obtained from parallel inactive mixing experiments. Brunauer- Emmett – Teller (BET) surface area analysis of the inactive precipitates was performed (Micromeritics Gemini V Surface Area Analyser 2365). The molar stoichiometry of the strontiobarite precipitates formed from large mixing experiments was determined by dissolution into pH 13 EDTA to totally dissolve barite. The Sr an Ba concentration were then measured using ICP-AES (Averyt et al., 2003). A FEI QUANTA 650 FEG ESEM (Field Emission Gun, Environmental Scanning Electron Microscope) equipped with Bruker Quantaz Energy Dispersive Spectroscopy system with an XFlash detector was used for imaging and analysis of the chemical composition of the particles. Here the backscattered electron (BSE) detector was used to image the samples and allow the differentiation between elements atomic number (Z) contrast. EDS was used for chemical analysis of the samples. The speciation of strontium associated with barite was analysed via X-ray absorption spectroscopy (XAS). A sample (approximately 2 wt. %) was mounted in a cryo vial and analysed at the Diamond Light Source, Harwell, UK on beam line B18. Sr K-edge spectra were collected in transmission mode at liquid nitrogen temperature (77 K). Background subtraction and improvement of signal to noise ratio of the data was obtained via averaging multiple scans for the sample using the Demeter software package Athena and Artemis, FEFF6 (Ravel et al., 2005).

4.3.2.2 Radiometric analysis

Radium concentrations in sediments and produced waters were determined using gamma spectrometry (Canberra gamma spectrometer). Samples were sealed in a double polypropylene container in a standard geometry for 1 month prior to analysis to allow the Ra progeny to equilibrate. The gamma spectrometer was calibrated using known 226Ra standards in the standard geometries for sediment and water. Samples were counted for 24 hours (sediment) or 7 days (produced water) and the radium concentration was calculated from measurements of the 214Bi and 214Pb daughter products and comparing these to known standards (Siddeeg et al., 2015).

154

4.3.2.3 Heavy Liquid Extraction

Oven dried sediment were disaggregated and sieved through a 215 µm sieve. Heavy liquid (di-iodomethane; 3.3 g cm-3) was then used to separate the barite-containing (4.5 g cm-3) dense fraction of the sediment. Here an accurately weighed, sieved solid sample between 10 and 20 g was added to a separating funnel containing 80 mL heavy liquid and left to separate for 10 minutes. Both the dense fraction (> 3.3 g cm-3) and light fraction (< 3.3 g cm-3) were then separated and filtered. Both fractions were then rinsed with deionised water and acetone then dried. The samples were then examined under a microscope and suspected white barite particles handpicked for SEM analysis.

4.3.2.4 Autoradiography

Autoradiography was used to determine the spatial distribution of radionuclides on selected samples from field site studies. Handpicked barite samples were placed on the storage phosphor screen BAS-IP SR, super resolution (G.E Healthcare) in a dark cupboard for 4 weeks. The activity on the surface of the sample was imaged using a Typhoon 9410 variable mode imager where the screen was then scanned and imaged using a HeNe laser (633 nm) with a pixel size of 10-25µm. The extent of darkening recorded is quantitatively proportional to the activity on the sample surface (Zeissler et al., 2001).

4.4 Results and Discussion

4.4.1 Formation of NORM precipitate upon discharge of Produced water to Seawater 4.4.1.1 Characteristics of marine sediment samples

The bulk mineralogical composition of the five sediment samples collected mainly comprised of silicate and clay minerals such as quartz (SiO2), mica (muscovite;

KAl3Si3O10(OH)2), feldspar (albite; NaAlSi3O8) and chlorite (clinochlore;

(Mg,Fe,Al)3(Al,Si)4O10(OH)8). All samples contained carbonate minerals (aragonite

155 and calcite; CaCO3) and halite (NaCl). Three samples contained kaolinite

(Al4(SiO10)(OH)8), ankerite (Ca(Fe,Mg,Mn)(CO3)2) or microcline (KAlSi3O8) (Table 4.1 & Table S4.1-4.2). The concentrations of elemental Ba and Sr were measured by XRF and are summarised in Table 4.1.

Sediment Mineralogy Chemical Composition γ- Sample (XRD) (XRF) spectrometry Sr (ppm) Ba (ppm) Ra226 (Bq g-1) A: 250m W Quartz, Muscovite, Albite, 870 241 0.04 ± 0.004 Clinochlore, Calcite, Aragonite, Halite B: 100m W Quartz, Analcime, 558 370 0.06 ± 0.01 Muscovite, Albite, Microline, Clinochlore, Kaolinite Calcite, Aragonite, Ankerite, Halite C: 20m W Quartz, Muscovite, Albite, 375 310 0.05 ± 0.01 Clinochlore, Calcite, Aragonite, Halite D: 100m E Quartz, Muscovite, Albite, 552 1176 0.32 ± 0.05 Microline, Clinochlore, Kaolinite, Calcite, Aragonite, Halite E: 250m E Quartz, Muscovite, Albite, 252 345 0.11 ± 0.01 Microline, Clinochlore, Kaolinite Calcite, Aragonite, Ankerite, Halite

Table 4.1: Mineralogical, radiological and chemical characteristics of marine sediment samples taken from the different locations

In terms of radium analyses samples displayed a range of concentrations from background levels (0.04 – 0.06 Bq g-1) to modestly elevated (up to 0.3 Bq g-1). Here there was a spatial trend with samples west of the discharge outfall (A – C; Table 4.1) at or near background, and samples east of the outfall (D and E; Table 4.1) elevated above background (0.1 and 0.3 Bq g-1). Interestingly, the elevated Ra concentrations were associated with the Ba concentrations in the sediments (Table 4.1) suggesting co-enrichment of Ra within Ba bearing sediments due to Ra incorporation into barite (Jerez Vegueria et al., 2002; Rosenberg et al., 2014; Zhang et al., 2014; Garner et al., 2015). The spatial distribution is presumably due to

156 current dispersion effects as there is a north-east outflow direction at the site (Fig. 4.1). This clearly affects the distribution of Ba and Ra in the sediments with enrichment to the east. As discussed above, samples A – C display Ra concentrations at background levels (0.01 – 0.05 Bq g-1) found within marine sediments (Landa et al., 1983; Jerez Vegueria et al., 2002; Hosseini et al., 2010; Dowdall et al., 2012; Environment Agency, 2015). These are consistent with both Ba and Sr levels (300 – 400 and 200 – 400 ppm respectively) which again are considered to be at background concentrations (Stevenson et al., 1995; Jerez Vegueria et al., 2002). For samples D and E the Ra levels are elevated compared to background and this coincides with higher Ba and Sr levels in the sample (345 – 1170 ppm). This strongly suggests Samples D and E have technologically enhanced levels of Ra that is co-associated with barite in the sediment.

Further analysis of the mineralogy of the sediment was performed to explore the relationship between Ra and Ba observed at bulk levels. Here, heavy liquid separation allowed separation and isolation of minerals grains from the dense mineral fraction in Sample E. The isolated grains were typically between 243-426 μm and analysis of these particles using SEM showed the grains were irregular aggregates a few 100 µm in size, which consist of individual equant particles ≤ 2 µm in size (Fig. 4.2). EDS spectra and elemental mapping indicates these particles are rich in Ba, Sr and S indicative of strontiobarite. In addition, certain areas were abundant in Al, Ca and Si reflecting the presence of clay particles (Fig. 4.2).

The isolated agglomerates of strontiobarite were analysed using autoradiography where it was clear that the isolated particles contained measurable activity most likely corresponding to radium (226Ra) due to its identification in gamma spectroscopy data (Fig. 4.2 and S4.1-4.2). Here, the discreet nature of the radioactivity detected on the autoradiograph allows the individual strontiobarite particles to be imaged in terms of their activity concentration. Overall, these indicate that marine sediment samples east of the discharge outfall contained enhanced levels of Ba and Ra in all probability as a result of radiostrontiobarite co- precipitation and deposition.

157

158

Field sediment containing elevated levels of radium (226Ra) showed evidence of biogeochemical processes occurring (e.g. microbial reduction) below the sediment- water interface due to the identification of framboidal pyrite (FeS2) particles via SEM (Fig. 4.3) (Folk, 2005). EDS spectra indicates these particles are rich in Fe and S which is indicative of the pyrite, Ba and Sr as a result of radiostrontiobarite, and Al, Mg and Si from clay particles also within the analysed aggregate (Fig. 4.3b). The size of the mineral grains (< 10 µm) and morphology (framboidal) indicates this pyrite has formed in dysoxic (waters containing low concentrations of oxygen) to euxinic waters (waters containing no oxygen but concentrations of hydrogen sulphide

(H2S)) (Roychoudhury et al., 2003). Roychoudary et al., (2003) and Proske et al., (2015) have shown that the formation of pyrite with this morphology in marine sediments is due to the redox conditions below the sediment-water interface, where microbial activity results in the consumption of organic matter during sulphate reduction and formation of pyrite as shown (Fig. 4.3) (Roychoudhury et al., 2003; Folk, 2005; Proske et al., 2015).

A) B)

Interestingly this indicates that sulphate reduction is occurring within the sediment in which barite is deposited. Microbial induced sulphate reduction causes a decrease in sulphate which may induce barite dissolution as shown from previous studies (Pardue et al., 1998; VanLoon, 2000; Phillips et al., 2001; Keith-Roach, 2002;

159

Ouyang et al., 2017). This therefore proposes that the long term environmental fate of radiobarite as biological processes and reducing conditions develop merits further investigation.

4.4.2 Strontiobarite formation during field and synthetic seawater and produced water mixing: morphology and composition 4.4.2.1 Field Mixing Experiment

To further investigate strontiobarite precipitation produced water mixing experiments were conducted using the produced water sample from an active platform and seawater sampled from the shore (Table 4.2). Produced water and seawater were mixed in a 1:9 ratio which results in the formation of a white precipitate. After separation, SEM analysis revealed mineral grains with an equant morphology (Fig. 4.4a).

A) B)

EDX analysis showed that the composition was consistent with strontiobarite (Fig. 4.4b). Here, particle size distribution determined by SEM was between 1 - 6 µm which is typical of natural barite found in sediment (Phillips et al., 2001; Gonneea et al., 2006). Aqueous analysis of the experimental solutions before and after mixing,

show that the composition of the precipitate was (Ba73.1Sr26.3SO4) (Table S4.3),

160 which is consistent with strontiobarite (Todd et al., 1990, 1992). Further analysis of the composition of the precipitate via XRD and EDTA dissolution was not possible as the volume of sample produced was very small (see below). Other minor phases such as sodium chloride (NaCl) and magnesium sulphate (MgSO4) were also identified by SEM. Overall, these particles confirm that mixing of production water and seawater produce strontiobarite particles 2- 6 µm in diameter which are consistent with the strontiobarite particle aggregates from the marine samples (Fig. 4.2).

Ions Produced Seawater – Produced Seawater – water – field field (mg L-1) water – synthetic (mg L-1) synthetic (mg L-1) (mg L-1) Ca 7860 ± 220 336 ± 8 2260 ± 41 420 ± 6 K 1840 ± 30 381 ± 61 380 ± 4 467 ± 4 Mg 4190 ± 100 972 ± 26 367 ± 9 1300 ± 18 Sr 323 ± 9 6 ± 1 537 ± 6 8 ± 0.1 Ba 13 ± 1 0.3 ± 0.3 209 ± 3 0 Na 80500 ± 1900 8290 ± 196 23900 ± 356 10100 ± 121 Cl 15700 ± 182 7690 ± 1710 50700 ± 68 19200 ± 268

SO4 278 ± 8 838 ± 228 0 2930 ± 23 HCO3 39 18 ± 0.5 0 111 ± 10 Initial pH 8.27 ± 0.13 8.13 ± 0.02 5.45 ± 0.08 8.12 ± 0.02 Initial Ionic 2.69M 0.41M 1.38M 0.62M Strength Final Ionic 0.64M 0.69M Strength

SIBaSO4 1.22 2.94

4.4.2.2 Synthetic Mixing Experiment

As well as the field production water and seawater sample experiment, we also performed mixing experiments with synthetic production water and seawaters synthesised (Table 4.2) in the laboratory to further validate our results,

161 methodology and approach, and to obtain an ample yield for additional characterisation and experimentation.

A) B)

Figure 4.5: (A) BSE image showing the strontiobarite crystals exhibiting both tabular and rosette morphology with a particle size between 1-5 µm and; (B) EDS spectra representative of all grains

The synthetic produced waters and seawaters contained barium and strontium at concentrations of the same order of magnitude as those found in field brines from the North Sea (Mitchell et al., 1980; Yuan et al., 1994; Røe Utvik, 1999). After separation of the precipitate following the mixing of the waters (1 : 9 ratio), SEM analysis revealed mineral grains exhibiting an equant morphology (Fig. 4.5a). Here, the particle size distribution was identified to be between 1 – 5 µm. The mass of precipitate produced in these experiments was amenable to XRD analysis which indicated the solid comprised of strontiobarite (Fig. S4.3). EDTA dissolution experiments further confirmed the composition of the precipitate to be

Ba76.4Sr23.8SO4 consistent with that obtained from solution analysis before and after mixing (Ba75.7Sr24.3SO4) (Table S4.3). Indeed, EDX confirmed the precipitates contained Sr and Ba in similar ratios to the sediment samples and active produced water and seawater mixing experiments. EDX analysis also indicates there were minor amounts of calcite also present in the sample (Fig. 4.5b). FTIR analysis showed intense peaks at 1084 cm-1 (with a distinct shoulder at 988 cm-1) corresponding to the sulphur-oxygen (S-O) stretch and at 609 cm-1 (with a shoulder at 639 cm-1) corresponding to the bending motion of the sulphur-oxygen bond

162 within sulphate characteristic of barite (Fig. S4.4) (Adler et al., 1965; Ramaswamy et al., 2010). Finally, the best fit to the Sr EXAFS analysis data was consistent with Sr substituted within the structure of barite (Fig. S4.5). The coordination numbers and interatomic distances in respect to strontium closely correspond to the local

2+ environment of Ba in pure barite (Fig. S4.5). The distance of the first-shell Sr-O1

(2.63 Å) in the sample is shorter than the Ba-O1 (2.81 Å) distance in barite, which is consistent with the larger size of the Ba2+ (1.68 Å) ion in relation to Sr2+ (1.48 Å) (Tokunaga et al., 2018). This confirms Sr2+ precipitates with barite and substitutes for Ba2+ into the crystal lattice via co-precipitation as previously reported by Tokunga et al., (2018) (Hedström et al., 2013; Tokunaga et al., 2018).

Despite the differences in composition of the synthetic waters used in the experiments in respect to the field waters, the composition and phase produced is uniform in relation to that extracted from marine sediment samples and field mixing experiments. Small differences in the adopted crystal morphology and size across the different mixing experiments (e.g. rosette, equant), and other minor phases identified maybe due to variations in the concentration of scaling ions (e.g. Sr2+ and Ba2+) present in the produced waters and seawaters and their resultant supersaturation in solution as shown by the saturation indices (SI) of mineral phases calculated via geochemical speciation modelling (PHREEQC) (Table 4.2). Despite the higher difference in supersaturation possibly leading to the smaller crystals sizes and aggregation observed in synthetic mixing experiments, compositions between the different experiments are very similar, and the crystal morphologies identified across these experiments are typical of barite and have been widely reported in literature (Todd et al., 1990, 1992; Phillips et al., 2001; Badr et al., 2008). Overall this shows that the precipitate that forms from the synthetic fluids used to mimic the product produced in the field closely resemble those formed from field fluids, and matches chemically and morphologically to those found in field sediment. This laboratory based synthetic method can therefore be used to mimic the uptake of radium, and formation of the precipitate formed in the field in a marine setting (see section 4.4.3).

163

4.4.3 Ra uptake during strontiobarite (Ba-Sr-SO4) formation 4.4.3.1 Synthetic seawater and produced water mixing: Radium uptake experiments

After confirmation that strontiobarite, with similar characteristics to those found in the field samples precipitated upon mixing of synthetic production waters and seawaters, further experiments were undertaken to investigate radium uptake during this process. Experiments were performed which determined the distribution of radium between the solid and aqueous phases following formation of the ternary phase (RaBaSrSO4). The strontium, barium and radium concentrations in solution with time are shown in Fig. 4.6a-b. After mixing, removal of Ba and Ra from solution occurred rapidly (1 - 3 hours) presumably due to precipitation of radiostrontiobarite (Fig. 4.6a-b). This confirms rapid barite precipitation kinetics dominantly controls the uptake of radium in this system. As reported by Zhang et al., (2014). The rapid nucleation of BaSO4 due to the supersaturation of the solution in respect to barite (Table 2) leads to an increase in precipitation rate as sufficient sites for crystal growth are quickly established during the initial mixing of the brines. During this period of crystal growth (3 – 24 hours) a portion of radium is incorporated into the lattice of the carrier mineral (i.e. barite) via direct substitution of Ra2+ for Ba2+ (Fig. 4.6a-b) (Zhang et al., 2014). After this period (24 – 1000 hours) the activities of electrolytes and nucleation rate decreases as the supersaturation of the solution decreases due to fewer ions present in solution and equilibrium is established (Fig. 4.6a-b) (Todd et al., 1990; Zhang et al., 2014). Further to this due to the volumetric mismatch/ionic radii differences between radium (1. 7 Å) and strontium (1. 44 Å) in relation to barium (1. 61 Å), lattice replacement via strontium is considered less likely and unfavourable reflected by the smaller uptake of strontium over the course of the experiment (Fig. 4.6a) (Todd et al., 1990; Zhang et al., 2014; Vinograd et al., 2018). As shown from Fig. 4.6a-b radium removal occurred coincident with barium removal and strontium removal was not directly coupled to either Ba or Ra. In addition, this suggests that barite is the main carrier mineral for radium consistent with past work (Zhang et al., 2014). Radium uptake increases over time, from 48 % to 79 % between 1 - 7 hours

164

followed by a further increase up to 97 % by 24 hours. Equilibrium was then established with a maximum radium uptake of 97.5 %. An effective partition coefficient (Kd’) for Ra2+ uptake into barite of 1.14 ± 0.1 was calculated which is in agreeance with results reported by other studies in this area of 1.07 – 1.54 under similar ionic strengths (NaCl concentration: 0 - 3M) (Ceccarello et al., 2004; Rosenberg et al., 2014; Zhang et al., 2014). Overall this method effectively mimics the formation of precipitate in field samples (Section 4.4.2), and also produces uptake coefficients similar to those reported in literature (Rosenberg et al., 2014; Zhang et al., 2014).

A) B)

165

4.5 Conclusion and Environmental Implications

Elevated levels of radium were identified in field marine sediment samples (D & E) associated with radiostrontiobarite at sites distant from the point of produced water discharge proving with certainty, that these discharges could be associated with detectable, but non-hazardous levels of radiation. Sediment extractions from field marine sediments in this study show radium is attenuated within the sediment via co-precipitation into strontiobarite. The mechanism of strontiobarite formation was further demonstrated from field and synthetic mixing experiments confirming the formation process. Micro-particulate strontiobarite precipitate phases with uniform composition, characteristic morphologies (equant) and particle sizes (1 - 6 μm) were identified in field samples and different experimental regimes (e.g. field and synthetic mixing experiments). This study of produced water discharges to the marine environment thus shows a direct link between the morphology and composition of the radiostrontiobarite precipitate extracted from field marine sediment samples, and the mixing of full- component synthetic and field produced waters and seawaters. This compares with other studies where related morphologies and compositions are seen for similar systems (e.g. water flooding studies) producing strontiobarite during the mixing of incompatible waters (Todd et al., 1990, 1992). Insight into radium solid-solution chemistry and radiostrontiobarite formation via this mechanism of mixing incompatible waters, as a result of operational discharges, can lead to significant barium and radium uptake helping to predict and model the fate and behaviour of 226Ra in this system. Radium uptake experiments show that a portion of radium (48 -79%) is incorporated into strontiobarite between 1 – 7 hours and 97% over 24 hours, suggesting that on discharge some radium will be dispersed into solution as the aqueous Ra2+ ion and be subject to sorption, dependent upon the characteristics of the receiving environment. A value of 1.14 ± 0.1 was calculated for the effective partition coefficient (Kd’) for Ra2+ incorporation into strontiobarite which is comparable to other studies in this area investigating binary and ternary phases e.g. in fracking systems (or other controlled studies) under controlled conditions (Rosenberg et al., 2014; Zhang et

166 al., 2014). This study has successfully identified the existence, isolated and mimicked the formation of the radiostrontiobarite phase which forms as a result of operational discharges from the offshore industry. Results obtained in this study in respect to radium uptake are representative of a closed-system, whereas marine discharges in reality occur in an open system (e.g. of infinite seawater). This suggests radium which is not incorporated into barite will be diluted and dispersed due to current flow effects. This may increase the mobility of barium and radium once discharged, thus lead to a greater portion of the radium (and barium) dispersing and existing in the aqueous phase (as the mobile Ra2+ ion), or sorbed phase (e.g. adsorbed to sediment and particulates) relative to that incorporated into the barite solid phase, and that predicted by the closed-system precipitation experiments and other studies (Landa et al., 1983; Van Sice et al., 2018; McDevitt et al., 2019). The amount of radium uptake into the strontiobarite phase can vary with mixing times and discharge rates, however the exact nature of this relationship is not clear from this study. In this shallow water environment there is clear evidence for radiostrontiobarite precipitation and deposition due to the lack of dispersion and/or rapid deposition of radiostrontiobarite to the sediment. This is in contrast to deep water environments where production waters are discharged near surface and presumably dispersion occurs, resulting in the non- detection of contamination in the surrounding sediments and waters (Jerez Vegueria et al., 2002; Eriksen et al., 2006; Gafvert et al., 2007; Olsvik et al., 2012). The environmental setting of the receiving environment (e.g. shallow marine or deep sea), its characteristics (e.g. fresh or salt water), and its corresponding tidal regimes are therefore also fundamental factors which need to be considered to understand the attenuation, impact and fate of radium in the water column. This is key as such factors vary globally between installations as shown from other studies (Landa et al., 1983; Pardue et al., 1998; Jerez Vegueria et al., 2002; Van Sice et al., 2018; McDevitt et al., 2019). Due to radiostrontiobarite particulates settling to the sea bed via deposition into shallow water depths, further research into the cumulative effect and fate of such solid phase in a marine environment needs to be conducted to consider the long term environmental impact of radium in barite of this nature. The identification of

167 pyrite promotes studies to underpin assessments of the environmental risk and fate of radium in these systems via natural biogeochemical processes. In addition similar studies should focus on the fate of radium existing in the aqueous phase in a marine environment as sorption to sediment (e.g. clay minerals) can also be an effective sequestration mechanism.

Acknowledgements

The work contained in this paper contains work conducted during a PhD study undertaken as part of the Natural Environment Research Council (NERC) Centre for Doctoral Training (CDT) in Oil & Gas [grant number [NE/M00578X/1] under its Environmental Impact and Regulation research theme. It is sponsored via a scholarship provided by University of Manchester (UoM) and the NERC CDT studentship whose support is gratefully acknowledged.

168

Supporting information for Chapter 4: Fate of Radium on Discharge of Oil Produced Water to the Marine Environment

Additional Data

Major Elements Sample A Sample B Element Conc. Element Conc. L.O.I. Element Conc. Element Conc. L.O.I. (%) (%) (%) (%) Si 24.7 Ce - 26.5 Si 37.8 Ce 0.017 17.6 Al 7.33 Sr 0.127 Al 11.5 Sr 0.08 Ca 28.9 Zn 0.011 Ca 14.7 Zn 0.013 Fe 3.79 Cr 0.012 Fe 6.59 Cr 0.014 C 26.5 Br 0.014 C 17.6 Br 0.013 K 1.37 W 0.02 K 1.98 W 0.023 Mg 1.50 Pb 0.006 Mg 2.50 Pb 0.008 Na 1.96 Rb 0.006 Na 2.38 Rb 0.008 Cl 1.42 Co 0.003 Cl 1.62 Co 0.009 Ti 0.497 Ni 0.006 Ti 0.877 Ni 0.01 S 0.74 Cu 0.004 S 0.991 Cu 0.005 H 0.82 Zr - H 0.88 Zr - P 0.246 Ga - P 0.291 Ga - Mn 0.094 Y - Mn 0.09 Y 0.002

Sample C Sample D Element Conc. Element Conc. L.O.I. Element Conc. Element Conc. L.O.I. (%) (%) (%) (%) Si 47.6 Ce 0.017 13.9 Si 43.5 Ce 0.082 13.9 Al 11.6 Sr 0.054 Al 10.7 Sr 0.085 Ca 9.73 Zn 0.01 Ca 13.3 Zn 0.01 Fe 5.81 Cr 0.017 Fe 7.82 Cr 0.011 C 13.9 Br 0.014 C 13.9 Br 0.01 K 1.94 W 0.036 K 1.95 W 0.036 Mg 2.12 Pb 0.007 Mg 2.14 Pb 0.016 Na 2.53 Rb 0.007 Na 2.53 Rb 0.007 Cl 1.71 Co 0.009 Cl 1.36 Co 0.012 Ti 0.762 Ni 0.006 Ti 0.755 Ni 0.008 S 0.936 Cu 0.006 S 1.15 Cu 0.005 H 0.82 Zr - H 0.62 Zr - P 0.289 Ga - P 0.301 Ga -

169

Mn 0.078 Y - Mn 0.098 Y 0.002

Sample E Element Conc. Element Conc. L.O.I. (%) (%) Si 53.1 Ce 0.021 10.8 Al 12.2 Sr 0.037 Ca 5.24 Zn 0.012 Fe 5.84 Cr 0.013 C 10.8 Br 0.015 K 2.18 W 0.037 Mg 2.23 Pb 0.008 Na 2.53 Rb 0.008 Cl 2.05 Co 0.008 Ti 0.793 Ni 0.008 S 1.49 Cu 0.004 H 1.01 Zr 0.016 P 0.281 Ga 0.002 Mn 0.07 Y 0.002

170

Trace Elements (ppm) V Cr Mn Co Ni Cu Zn Ga Ge

Sample 62.2 67.1 684.9 16.1 20.9 14.4 69.2 7.5 1.3 A B 102.3 93.5 590.4 22.9 31.8 21.3 95.1 12.1 2 C 87.7 90.1 517.3 27.3 28 19.2 88 10.9 2.1 D 85.6 75.6 624.5 27.6 24.7 15 74.9 9.8 1.5 E 90.2 94.2 481.5 26.4 28.8 17.8 87.7 11.1 2.2 Br Rb Sr Y Zr Nb Mo Cd As A 82.4 40.8 869.8 12.8 112.3 6.3 1 0.9 3.2 B 83.6 51.1 557.5 18.9 178.3 13.8 0.9 1.2 0.7 C 79.5 48.4 375.1 17.8 214.3 10.9 0.5 -0.2 9.4 D 55.6 48 551.5 16.9 172.9 10 1.7 4.9 9.5 E 80.2 50.7 252.1 17.6 232.9 11.1 0.9 1.8 5.4 Sn Sb Te I Cs Ba La Ce Nd A 6.4 3.1 51.2 50.4 7.2 241.2 29 45.7 25.5 B 10.9 0.9 50.6 39.3 4.7 370.4 33.3 61.4 29.3 C 9 0.7 50.6 41.9 5.3 309.5 27.9 55.5 24.3 D 10.5 0.7 54.6 30.7 4.5 1175.8 28.3 45.8 22.4 E 8.9 2.2 51.2 32.8 6.1 345 30.1 50 22.9 Yb Hf Ta W Hg Pb Th U Sm A 1.9 6 1.3 126 6.5 24.5 7.7 8.4 4.4 B 0.4 6.1 1.7 151.1 6.1 32.9 7.1 7.7 5.8 C 0.8 6.2 0.3 252.2 6.1 30.3 6.5 7.3 3.4 D 0.2 6.6 -0.4 223.2 4.8 21.8 6.7 7 2.7 E 0.9 6 1.2 235.6 6.7 30.8 5.5 6.1 1

Table S4.2: XRF analysis showing the trace elements present in sediment all field sediment samples

171

0.5cm

172

173

Calculated molar stoichiometry BET Analysis m2/g Ion Solution Solid Field Ba 73.1 ± 0.1 - - Mixing Sr 26.9 ± 0.1 - Synthetic Ba 75.7 ± 0.1 76.4 ± 1.5 2.93 ± 0.07 Mixing Sr 24.3 ± 0.1 23.8 ± 1.3

Table S4.3: Calculated molar stoichiometry of the desired precipitate from solution and solid data analysis. Surface area analysis data via BET; (-): limited sample for analysis

Figure S4.3: XRD of precipitate obtained from mixing of synthetic produced water and seawater representative of Ba75Sr25 (silicon (Si) incorporated for reference purposes)

Figure S4.4: FTIR spectra of precipitate obtained from mixing of synthetic produced water and seawater

174

2 2 Sample Shell CN ΔE0 (eV) 2σ (Å ) R (Å) Ref. BaSrSO4 Sr-03-Ba 7 -0.186 0.0085 2.63 2.81 Sr-01-S/Ba 2 0.157 0.0075 3.23 3.08 Sr-02-S/Ba 2 -0.183 0.0075 3.14 3.32

Figure S4.5: Sr K-edge EXAFS spectra and fit table (R-factor: 0.0088). Reference (Ref.) sample corresponds to pure barite

175

CHAPTER 5

The Effects of Bioreduction on the Fate of Ra2+(aq) and Radiobarite in Marine Sediments on Discharge of Oil Produced Water

This chapter is a manuscript prepared for submission in the journal Applied Geochemistry or Environmental Science & Technology. Supporting Information provided with this manuscript is included in the following manuscript

176

Faraaz Ahmad1, Katherine Morris1, Gareth T.W. Law2, Kevin Taylor1 and Samuel Shaw1*

1Research Centre for Radwaste Disposal and Williamson Research Centre, School of Earth & Environmental Sciences, Williamson Building, The University of Manchester, M13 9PL;

2Radiochemistry Unit, Department of Chemistry, University of Helsinki, P.O. BOX 33 (Yliopistonkatu 4), 00014, Finland

*Corresponding Author ([email protected])

Key words: Bioreduction, Barite, Radiobarite, Offshore discharges, Oil and Gas industry, NORM, Produced water, Radium, Barium, Sulphate-reducing bacteria

5.0 Abstract

Sequestration of radium via sorption to mineral phases in sediments and co- precipitation within recalcitrant mineral phases (e.g. barite) are fundamental mechanisms controlling the environmental fate of 226Ra during operational discharges of produced water to the marine environment. The environmental impact of 226Ra in the marine system, as bioreduction develops, is important in predicting its long term fate in marine systems. Here, we studied the fate of radium

2+ as aqueous radium (Ra ) and in radiobarite (RaBaSrSO4) during bioreduction and development of sulphate reducing conditions in sediment microcosm experiments. Sediment obtained from a produced water discharge site in the UK was amended with Ra2+(aq) or radiobarite, and bioreduction of the indigenous microbial population was stimulated with the addition of acetate and lactate as an electron donor. Experiments containing barite (BaSrSO4) were also conducted as controls. The microcosms were monitored for changes in geochemistry and molecular ecology at selected time points. For the Ra2+ system up to 90% of radium was adsorbed to sediment with no remobilisation of radium to solution observed during progressive anoxia to sulfidic conditions. In Ra2+(aq) experiments, sequential

177 extraction experiments pre- and post-anoxia showed radium was largely retained via adsorption before and after bioreduction. Results from the radiobarite system confirmed there was no release of radium to solution throughout progressive anoxia. Here, sequential extractions showed that ~ 94% of the radium remained in the radiobarite fraction before and after bioreduction. Electron donor amended control experiments containing inactive barite showed similar results to the radiobarite systems with no release of barium to solution during bioreduction and with ~ 84% of barium attenuated in barite pre- and post- anoxia. In the barite only experiments, barite grains were separated from sediments using heavy liquid extractions. The barite grains were then analysed using SEM which showed no evidence for any etch pits on the barite and further confirmed the indigenous microbes did not consume barium sulphate as a terminal electron acceptor. In addition, in sediments stimulated with electron donor, the microbial community also reflected the onset of bio-reduction with an increased relative abundance of Fe(III) and sulphate-reducing bacteria such as Ferrimonas pelagia, Desulfuromonas svalbardensis, Desulfobacter curvatus, and Desulfotalea sp.

5.1 Introduction

Isotopes of naturally-occurring radium (226Ra and 228Ra) exist in most conventional and unconventional oil and gas waste water effluents (e.g. produced water) as a result of the decay of primordial radionuclides (238U and 232Th) (IOGP, 2016). The release of produced water brines via authorised release to surface waters from discharge sites (e.g. offshore oil and gas platforms) may result in elevated levels of Ra in sediment and surface waters close to the discharge site (Landa et al., 1983; Pardue et al., 1998; Jerez Vegueria et al., 2002; Van Sice et al., 2018; McDevitt et al., 2019). Sequestration of radium via sorption to mineral surfaces (e.g. clay minerals / iron oxides) and co-precipitation within mineral phases (e.g. barite) are key controls on its environmental speciation, mobility and bioavailability (Pardue et al., 1998; Sajih et al., 2014; Zhang et al., 2014; Siddeeg et al., 2015; Van Sice et al., 2018). In addition, radium may also exist as the Ra2+(aq) species in the environment. As 226Ra has a long half-life (1600 years) and high radiotoxicity it

178 poses a potential radiological risk (Thiry et al., 2008). It is therefore important to gain a better understanding of the fate, speciation and mobility of radium during its discharge to the environment as radiobarite and as Ra2+(aq) in marine systems. This study focusses on the fate of Ra2+ and radiobarite during microbial reduction in marine sediments. Understanding radium speciation and fate is important in developing predictive models for the fate and environmental risk of radium in these systems.

Due to the mixing of incompatible waters (e.g. produced water and seawater) and the ionic radii compatibility between radium and barium ions in solution, it is expected that, when discharged, a portion of Ra2+ will dilute and disperse due to current flow effects, adsorb to, and/or co-precipitate with sulphate and carbonate mineral phases (e.g. BaSO4, SrSO4, BaCO3, SrCO3, CaCO3) due to the establishment of supersaturation (Al-Masri et al., 2005; Rosenberg et al., 2014; Zhang et al., 2014). This can result in elevated activities of radium in receiving waters and sediments due to dispersion and deposition of Ra2+(aq) and radium-containing particulates (Ahmad et al., 2019 (Ames et al., 1983; Langmuir et al., 1985; Pardue et al., 1998; Gonneea et al., 2006; Gonneea et al., 2008; Zhang et al., 2014). Ra2+(aq) can sorb to a range of surfaces, including clays (muscovite, illite, and kaolinite), iron oxides (goethite and ferrihydrite), manganese oxides (birnessite, manganite and todrokite) , carbonates (siderite, dolomite and magnesite) and organic matter (Taskayev et al., 1978; Hanan et al., 1981; Landa et al., 1983; Langmuir et al., 1985; Jones et al., 2011; Sajih et al., 2014; Siddeeg et al., 2015; Van Sice et al., 2018).

Studies have shown the sorption of radium by sediment and mineral surfaces is generally an outer sphere surface adsorption mechanism in which the kinetics of sorption can vary depending on the pH and salinity of the receiving waters (Landa et al., 1983; Jones et al., 2011; Zhang et al., 2014; Van Sice et al., 2018; McDevitt et al., 2019). As well as surface complexation, radium co-precipitation into highly insoluble sulphate mineral phases is a significant process. Studies focussed on mixing of high salinity synthetic waters show radium co-precipitation into sulphate binary phases (e.g. barite and celestine) and ternary phases (e.g. strontiobarite). Indeed, radium precipitation in sulphate phases is the key process controlling

179 radium mobility in marine systems and can result in the formation of a concentration of naturally occurring radioactive materials (NORM) (Ahmad at al., 2019; Gafvert et al., 2007; Rosenberg et al., 2014; Zhang et al., 2014). The key process of formation of such radium-containing inorganic particulates (e.g. RaxBa1- xSO4 and BaxSryRazSO4) is the mixing of chemically incompatible waters which occurs during offshore discharges and water flooding operations (Ahmad et al., 2019; Zhang et al., 2014). Sea water contains a high concentration of sulphate

2- 2+ 2+ 2+ 2+ 2+ (SO4 ) but lower concentrations of divalent cations (Ca , Mg , Ba , Ra and Sr ), in comparison to formation water with high divalent cations, low sulphate and elevated concentrations of radium. The mixing of these two waters results in the establishment of supersaturation thus precipitation of radium-containing sulphate micro-particulates and scales due to the ionic radii compatibility between radium and barium (Ahmad et al., 2019; Gafvert et al., 2007; Abdul et al., 2010; Candeias et al., 2014; Zhang et al., 2014; Garner et al., 2015). In recent work, heavy liquid extractions on marine sediments obtained from a field site where produced waters are discharged confirmed radiostrontiobarite with a crystal size between ~ 1-2 µm forms upon release of produced water to the marine environment (Ahmad et al., 2019). In this study, model systems using field and synthetic produced waters and seawaters further confirmed the formation of (radio)strontiobarite particles upon release of production water to the marine environment. Ra2+(aq) in synthetic produced water was mixed with synthetic seawater confirming radium uptake into strontiobarite over 24 hours, suggesting that on discharge a portion of radium will also be dispersed and subject to sorption.

There is a paucity of information on the fate of Ra2+ upon its discharge to the marine environment, however discharges to freshwater settings have been studied. Studies have shown discharges to surface waters typically with low sulphate concentrations and/or barium (e.g. freshwater or coastal marsh systems), result in the incorporation of radium into carbonate phases, and enhanced adsorption to sediment surfaces downstream (McDevitt et al., 2019). This is due to the depleted levels of sulphate in the receiving waters and undersaturation of the solution in respect to solid-phase barite. This allows alternative mechanisms for radium

180 attenuation in the environment to dominate (Landa et al., 1983; Langmuir et al., 1985; Jones et al., 2011; Van Sice et al., 2018; McDevitt et al., 2019). In high salinity, high sulphate marine systems radium co-precipitation with barite controls its solubility. Barite precipitation occurs within the local vicinity of the discharge site due the rapid establishment of supersaturation, and high density of barite (4.5 g cm-3) shown from sequential leaching studies (Landa et al., 1983; Pardue et al., 1998; Jerez Vegueria et al., 2002; Appelo, 2005; Van Sice et al., 2018; McDevitt et al., 2019). Which sequestration mechanism(s) dominate, sorption of Ra2+ or precipitation of radiobarite phases, depends on the characteristics of the receiving environment where release of Ra2+ and subsequent adsorption and/or precipitation occur simultaneously to some degree (Landa et al., 1983; Pardue et al., 1998; Van Sice et al., 2018; McDevitt et al., 2019). Marine discharges from offshore platforms have identified modestly elevated levels of radium and barium in sediment at distances from the outfall due to dilution and dispersion, and the characteristics of the receiving environment (Jerez Vegueria et al., 2002) (Ahmad et al., 2019). Furthermore, radiotoxicology and bioaccumulation studies show radium Ra2+(aq) of up to 117 Bq L-1 (0.117 Bq mL-1), and adsorbed to sediment (6600 Bq kg-1) due to produced water discharges display a low exposure dose effect to benthic fauna around oil platforms (Neff, 2002; Ruus et al., 2005; Grung et al., 2009). Likewise radium incorporated into barite is considered highly recalcitrant and has a low bioavailability and low toxicity (Neff, 2002; Ruus et al., 2005; Menzie et al., 2008; Grung et al., 2009; Olsvik et al., 2012). Despite a body of work exploring radium speciation and fate, there is a lack of work examining the long term fate and mobility of radium in environmental systems as progressive anoxia develops in both the solid and aqueous phase, and this is the focus of this study.

181 there are limited studies that focus upon marine discharges of oil produced water (Pardue et al., 1998). The principal mechanism resulting in the potential release of radium from barite scale and sludge is thought to be sulphate reduction by sulphate reducing bacteria (SRB) from the genus Desulfovibiro sp. and Desulfobacterium (Robert, 1981; Fedorak et al., 1986; Phillips et al., 2001). Here, sulphate reduction is implicated in accessing sulphate within barite precipitates potentially releasing barium and analogous species to the environment (Carbonell et al., 1999; Konhauser et al., 2002). This has been demonstrated by the concurrent release of barium and sulphide following the dissolution of barite and decrease in sulphate under anoxia in comparison to control studies performed under aerobic (oxidised) conditions (Bolze et al., 1974; McCready et al., 1980; Baldi et al., 1996; Pardue et al., 1998; Luptáková et al., 2015; Ouyang et al., 2017). For radium incorporated in radiobarite, and existing as Ra2+(aq), few biogeochemical studies have been reported relating to produced water discharges. Pardue et al. (1998), demonstrated sulphate reduction can potentially lead to the dissolution of radiobarite in brackish and salt water systems. They detected high levels of sulphide and barium but only trace levels of radium (0.063 ± 0.01) in overlying pore waters in sediment-microcosms containing contaminated field sediment from a discharge site. Radium release was thought to be inhibited by re- sequestration via co-precipitation into newly formed barite as the radium was released from sediments, as a result of high sulphate concentrations above the sediment-water interface. Other mechanisms of remobilisation have been reported including desorption of radium from sediment surfaces attributed to an ion- exchange reaction (Landa et al., 1983; Ouyang et al., 2017). Other studies for example Phillips et al. (2001), demonstrated minimal radium release (< 0.1%) from oil-field barite scale using sulphate-reducing cultures isolated from oil field brine pond sediment. Factors such as crystal size, retention mechanisms, surface area and nature of the materials (e.g. different processes of formation) were identified as possible variables effecting the amount of radium released and the non- stoichiometric release of barium, sulphide and radium to solution in agreement with other studies (Bolze et al., 1974; McCready et al., 1980; Fedorak et al., 1986; Baldi et al., 1996; Wilkins et al., 2007). In particular radium and barium retention

182 mechanisms such as re-adsorption and inclusion into newly formed mineral phases (e.g. barium carbonate and barium sulphide) have been suggested as possible reasons for restricted release of radium and barium (Bolze et al., 1974; Landa et al., 1991; Baldi et al., 1996; Phillips et al., 2001; Luptáková et al., 2015). Wilkins et al. (2007), did not detect desorption of Ra2+ from sediments where Ra2+ had been sorbed under oxic conditions as Fe(III)-reduction developed. This was in contrast to Landa et al. (1991), who did detect radium release from sediments, due to the suspected re-sequestration of radium into newly formed mineral phases. Other studies related to unconventional oil and gas extraction have demonstrated subsurface barite dissolution during hydraulic fracturing via microbially induced bioreduction. Ouyang et al. (2017), demonstrated the potential for barite dissolution via ionic-exchange and by halophilic organisms in hydraulic fracturing fluids under anoxic and hypersaline conditions. Evidence of microbial induced etch pit morphology on barite grains using scanning electron microscopy (SEM), confirmed the microbial induced dissolution of barite under laboratory conditions. Work relating to unconventional oil and gas extraction certainly suggests barite dissolution is possible (Burgos et al., 2017; Ouyang et al., 2017). More importantly it must be noted that barite formed in the field and laboratory across these studies may have different crystal structure and grain size thus impacting dissolution kinetics (Chang, 1996; Phillips et al., 2001). Additionally studies utilizing pure cultures/enrichments of sulphate reducing bacteria may not sufficiently represent the environmental systems and there is a paucity of sediment work with indigenous microbial communities. The fate of radium in sediments as a result of offshore produced water discharges thus remains poorly understood.

In this study we used a multi-technique approach to provide a comprehensive assessment of the potential environmental impact, speciation, and fate of radium once discharged to a marine setting as, the aqueous ion (Ra2+) and inorganic solid

(BaxSryRazSO4). Control experiments containing barite (BaSrSO4) were also used to investigate any shifts in the microbial community due to the addition of radium. Radiobarite and barite were synthesised using a methodology that produced particles of a similar chemical composition and morphology as the field site. These

183 representative barite precipitates and aqueous radium were then spiked into microcosms containing marine surface sediments from a representative discharge site. The sediments (with their indigenous microbial populations) were then stimulated to bioreduction using terminal electron acceptor additions. A range of techniques including radiochemical measurements, sequential extractions, SEM, heavy liquid extractions and DNA sequencing were performed to explore the partitioning of radium within sediments and the impact of radium on the biodiversity of the sediments. Overall, we investigate the mobility and fate of both aqueous radium, and radiobarite as biogeochemical conditions evolve through to sulphate-reducing under marine conditions.

5.2 Experimental Methods

5.2.1 Radiobarite (RaBaSrSO4) and barite (BaSrSO4) formation

Radiobarite precipitate was synthesised chemically (Ba75Sr25) and morphologically (equant) similar to the inactive analogue of barite synthesised using full component brines, and that extracted from field marine sediment described in our previous study (Chapter 4).

5.2.1.1 Precipitation procedure and radium uptake

A solution of barium and strontium chloride (19.6 mL) was spiked with 7.6 kBq of radium (226Ra). Sodium sulphate (20 mL) was then added to the solution and left to stir for 24 hours (Ahmad et al., 2019). The solution was then decanted and left to settle for 4 days. Radium concentrations in the supernatant were periodically determined by adding 1 mL of the solution to 10 mL of scintillation fluid (Ultima Gold, Perkin Elmer) in a sealed tube. Samples were then analysed after 1 month to allow the Ra progeny to come into equilibrium using liquid scintillation analysis. Parallel inactive experiments (without 226Ra) were run to allow full characterisation of the barite precipitate using SEM.

184

5.2.2 Chemical composition and mineralogical analysis

X-ray diffraction (Bruker D8 Advance diffractometer) was conducted to determine the bulk mineralogical composition of the parallel inactive materials. Fourier- transform infrared spectroscopy (FTIR) was used to determine the chemical composition of the inactive barite precipitates (Perkin Elmer Spotlight 400 FTIR imaging system Universal ATR). Surface area analysis of the inactive barite was performed using a Micromeritics Gemini V Surface Area Analyser (model 2365).

5.2.2.1 Dissolution of barite particles

The molar stoichiometry of barite precipitate was determined by dissolving barite in EDTA-KOH to totally dissolve barite (Averyt et al., 2003). Barium and strontium in the sample was then measured via ICP-AES (Perkin Elmer Optima 5300 dual view system). All experiments were carried out in triplicate.

5.2.2.2 Scanning electron microscopy (SEM)

A FEI QUANTA 650 FEG ESEM (Field Emission Gun, Environmental Scanning Electron Microscope) equipped with Bruker Quantaz Energy Dispersive Spectroscopy system with an XFlash detector was used for imaging and analysis of the chemical composition of the particles. Here the backscattered electron (BSE) detector was used to image the samples and allow the differentiation between elements atomic number (Z) contrast. EDS was used for chemical analysis of the samples. Precipitate was imaged uncoated under high vacuum (15 keV).

5.2.3 Sediment microcosm Experiments 5.2.3.1 Sampling

Sediment samples collected in close proximity to a known marine discharge point were collected in April 2017. Samples were stored in sterile bags at 4°C in the dark prior to use in microcosm experiments. XRD (Bruker D8 Advance diffractometer)

185 and XRF (PANalytical Axios) were conducted to determine bulk mineralogical and chemical compositions of sediments. Sediment was dried in an oven (40°C ± 0.5°C), disaggregated using a pestle and mortar and homogenised for XRD and XRF analysis (Ahmad et al., 2019).

5.2.3.2 Microcosm experiments

Microcosms containing sediment and synthetic seawater (1:10 ratio) were set up using an aseptic technique in 100 mL serum bottles. The artificial seawater consisted of (mg L-1): Ca, 420 ± 6; K, 467 ± 4; Mg, 1300 ± 18; Sr, 8 ± 0.1; Ba, 0; Na,

10100 ± 121; Cl, 19200 ± 268; SO4, 2930 ± 23; HCO3, 111 ± 10 (Todd et al., 1992). Three sets of microcosms were amended with one the following: barite precipitate (120 mg); radiobarite precipitate (120 mg); or 20 Bq mL-1 226Ra. The synthetic seawater was flushed with a N2:CO2 gas mixture (80:20), prior to its use. Triplicate microcosms across all experiments, a sterile control (one replicate) and control containing seawater and sediment (one replicate) were inoculated with 5 mM acetate and 5 mM lactate (< 0.2 µm, filter sterilised) as electron donor, while other controls containing no electron donor contained: seawater, sediment and precipitate/226Ra; and only seawater and precipitate/226Ra. The headspace was purged with N2 prior to sealing with butyl rubber stoppers and aluminium crimps. The bottles were then incubated in the dark at ambient temperature (20 ± 2 °C) and sampled periodically over 300 days using aseptic technique. After 140 days selected electron donor amended microcosms were re-spiked with 10 mM acetate and 10 mM lactate. Controls included sediment and seawater only (no electron donor), and were used to explore biodiversity in experiments with and without radium. A seawater and (radio)barite only (no electron donor) microcosm was also set up to monitor release of radium and barium from (radio)barite (Risthaus et al., 2001; Kowacz et al., 2008; Kowacz et al., 2009). Radium (Ra2+ and radiobarite) spiked microcosms with no acetate/lactate amendment (unamended) were also prepared to explore the natural bioreduction process in sediments. Finally sterile controls amended with either radiobarite or Ra2+ were created for all three sets of

186 experiments by autoclaving (3 cycles, 20 minutes at 120°C) prior to the addition of spike and 0.22 µm filter sterilised electron donor (Burke et al., 2005).

5.2.3.3 Geochemical analysis

Periodically, microcosms were sampled by removing sediment slurry aseptically at selected time points under argon. Samples were analysed for total bioavailable iron and Fe(II) concentrations using the ferrozine assay (Lovley et al., 1987). The aqueous phase was then separated by centrifugation (14 800 rpm, 5 minutes) and aqueous Ra was measured via liquid scintillation counting (1 mL), strontium, barium and calcium analysis by ICP-AES (acidified in 2% HNO3) (Perkin Elmer Optima 5300) and sulphate, chloride and volatile fatty acids (VFAs) by ion chromatography (Dionex ICS 5000). Samples for radium analysis were left for 1 month to allow the Ra progeny to equilibrate, and analysed in a Quantulus scintillation counter (Perkin Elmer). The pH and Eh were measured using calibrated electrodes. Hydrogen sulphide was analysed in microcosm waters using a colorimetric technique (Cline, 1969; Wangersky, 2003).

5.2.3.4 Heavy liquid extraction

In parallel inactive microcosm experiments heavy liquid (diiodomethane) extraction was used to separate the barite-containing (4.5 g cm-3) dense fraction of the sediment adapted from Proske et al., 2015, Korozinkova et al., 2007 and Van Beek et al., 2002. Both light and heavy separated fractions were rinsed with deionised water and acetone, and then dried. Both fractions were examined under a microscope and barite particles hand-picked and analysed via SEM-EDS to explore whether etch pitch formation had occurred in any of the samples.

5.2.3.5 Sequential chemical extraction

Sequential leaching was performed to provide an operationally defined assessment of the elemental associations of Ra, Ba and Ca in sediments (Tessier et al., 1979; Rauret et al., 1999; Aguado et al., 2004). The procedure was modified by including a 0.1 M EDTA leach to target barite dissolution (Beneš et al., 1981). Sediments from

187 both fresh (1 day incubations) and 300 day end member incubations, amended with electron donor, were separated by centrifugation (5000 rpm, 20 minutes) and then harvested for sequential extractions. Triplicate samples of field sediment (uncontaminated and contaminated) from the discharge site described in our previous study (Ahmad et al., 2019), were also subjected to sequential extraction.

This involved exposure sequentially to 1 M anaerobic MgCl2 (1 hour, pH 7, “exchangeable fraction”), 1 M anaerobic sodium acetate (5 hours, pH 5, “carbonate fraction”), 0.4 M anaerobic hydroxylamine hydrochloride (16 hours, pH 3, “reducible fraction”), 0.02 M nitric acid, 30% hydrogen peroxide (2 hours, pH 2) and 3.2 M ammonium acetate (16 hours, “oxidisable fraction”), 0.1 M EDTA (24 hours, “barite dissolution”) and finally aqua regia (5 hour, “residual fraction”) (Beneš et al., 1981; Nixon et al., 1983; Jerez Vegueria et al., 2002; Aguado et al., 2004). Extractions were conducted anaerobically and separation of solids between lixiviants was achieved by centrifugation (5000 rpm, 20 minutes). Aqueous samples were diluted and acidified (2% nitric acid) for Ba and Ca analysis by ICP-AES, and Ra was measured using liquid scintillation counting after 1 month equilibration.

5.2.3.6 Etch pit formation via chelating ligand

Abiotic dissolution of barite was conducted as a control study. A solution of EDTA (18 M) was prepared and the pH was adjusted to 12 using NaOH (5 M). Inactive synthetic strontiobarite crystals were then placed into the etching solution at room temperature for 5 minutes with gentle agitation. The solution was then decanted and filtered (0.22 µm Whatman PES filter) and the resultant precipitate rinsed with 0.1 M HCl to remove any absorbed EDTA clusters on the mineral surface before SEM analysis (Dunn and Yen, 1999; Wang et al., 1999, 2002).

5.2.3.7 Microbial community analysis

At experiment end points (Day 300), bioreducing sediments were sampled for microbial community analysis. Sediment at Day 0 was also analysed. DNA was extracted from 200 µL of sediment slurry using a DNeasy PowerLyzer PowerSoil Kit (Qiagen, Manchester, U.K), amplified via PCR (Roche Diagnostics Ltd, Burgess Hill,

188

UK), sequenced (Illumina, San Diego, CA, USA) and taxonomy assigned using the Blastn nucleotide search (http://blast.ncbi.nlm.nih.gov) as described previously (Lloyd et al., 2019).

5.3 Results & Discussion

5.3.1 Marine sediment and radiostrontiobarite

The bulk mineralogical and chemical composition of marine sediment had been characterised prior to use (Ahmad et al., 2019). Briefly, XRD of the sediment showed it contained a mixture of silicates such as muscovite, kaolinite, microcline, albite as well as quartz (Fig.S5.1). The concentrations of radium (0.04 Bq g-1) and barium (241 ppm) in the sediment were at background concentrations (Stevenson et al., 1995; Jerez Vegueria et al., 2002; Landa et al., 1983; Jerez Vegueria et al., 2002; Hosseini et al., 2010; Dowdall et al., 2012; Gafvert et al., 2007).

5.3.2 Radiobarite (RaSrBaSO4) formation

5.3.2.1 Synthetic mixing experiments: Morphology, Composition and Radium Uptake

Radiobarite was precipitated under conditions representative of produced water seawater mixing conditions (Ahmad et al., 2019) (Todd et al., 1990, 1992; Phillips et al., 2001; Badr Merdhah et al., 2008). BSE imaging showed the precipitate exhibited dense tabular/rhombic and rosette morphology (Fig. 5.1a-b). The particle size distribution was between 1 - 8 µm with the most dominant crystal sizes ranging between 2 - 5 μm. XRD analysis indicated the solid comprised of strontiobarite

(Ba75Sr25SO4) (Fig. S3). EDTA dissolution experiments on the inactive barite from parallel inactive experiments, further confirmed the composition of the precipitate to be (Ba79.5Sr22.6SO4) consistent with that obtained from solution analyses before and after mixing (Ba75.2Sr24.8SO4), EDX analyses and past work (Fig. 5.1 & Table 5.1) (Ahmad et al., 2019). FTIR analysis showed intense peaks at 1058 cm-1 (with a distinct shoulder at 986 cm-1) corresponding to the sulphur-oxygen (S-O) stretch and at 603 cm-1 (with a shoulder at 637 cm-1) corresponding to the bending motion

189 of the sulphur-oxygen bond within sulphate characteristic of barite (Fig. S5.2) (Adler et al., 1965; Ramaswam, et al., 2010).

A) B)

Radium uptake into this phase was consistent with previous synthetic brine mixing experiments allowing up to 1% release of radium in microcosm water to be detected in the microcosm experiments (Table 5.1) (Ahmad et al., 2019).

BET Analysis m2 g-1 Ra uptake Calculated molar stoichiometry (%) Ion Solution Solid Synthetic Ba 75.2 ± 0.8 79.5 ± 3.2 5.55 ± 0.18 94.3 ± 1.19 Mixing Sr 24.8 ± 0.2 22.6 ± 1.2

190

5.3.3 Biogeochemical reduction of sediments

Sediment microcosm experiments containing either Ra2+(aq) radiobarite and

(RaSrBaSO4) were conducted using marine sediment and synthetic seawater. Here, radium in solution was tracked through the development of progressive anoxia through iron- and sulphate-reduction. Parallel non-active controls containing barite

(BaSrSO4) were also run to investigate any shifts in the microbial community due to the addition of radium and changes in solubility against radiobarite. Microcosms without added electron donor were set up to explore the bioreduction potential of unamended marine sediments. In addition microcosms containing seawater and radium only, sediment and seawater (with no radium) and autoclaved sediment (sterile controls) were run. At selected time points geochemical and microbiological characterisation was performed.

5.3.3.1 Biogeochemistry in sediment microcosms

Geochemical results across all groups of experiments containing sediment with and without electron donor, revealed the presence of Fe(II) in pore-water at the beginning of the experiments (Fig. 5.2.). This indicates the field sediment sample was anoxic and suggests microbial Fe(III)-reducing conditions had already been established from consumption of natural organic matter (NOM). The electron donor amended sediments, showed an immediate increase in Fe(II) in pore water when compared to the unamended experiments, with substantial amounts of Fe(III) reduction between Day 0 and Day 28 above unamended sediments. Analysis of the extractable Fe(III) in sediments showed Fe(III) was consumed by Day 28 and levels remained at around 10 - 12 mM compared to unamended microcosms which contained 8 mM of Fe(II) in solution (Fig. 5.2). Analysis of inorganic carbon and volatile fatty acids in solution showed the rapid consumption of acetate and lactate, and the formation of a mixture of hydrolysis and fermentation products as anoxia progressed.

Sulphate was present in solution at initial time points and that there was a significant decrease by Day 14 and a further decrease by Day 28, indicative of the development of sulphate reduction. In addition, sulphide ingrowth to solution was

191 detected between Days 14 - 28 confirming the onset of sulphate reduction (Fig. 5.2) (Fedorak et al., 1986; Baldi et al., 1996; Pardue et al., 1998; Phillips et al., 2001). A reduction in aqueous sulphide concentration in microcosm waters was observed from Day 28 - 140 most likely due to secondary mineral formation (e.g. iron sulphides) and the sediments progressively turned black. Previous studies have suggested a range of metastable monosulfide minerals such as mackinawite, greigite and pyrite can precipitate out of solution as a result of increased concentrations of Fe(II) produced from Fe(III)-reduction reacting with sulphide formed via sulphate-reducing bacteria (Berner, 1970; Postma, 1983). A second addition of electron donor (10 mM acetate and 10 mM lactate) at Day 140 was necessary to further stimulate bioreduction as microcosms were limited stoichiometrically (~ 168 e- eq. per mol) as a result of the inadequate electron donor availability introduced by the initial acetate and lactate amendment (Fig. 5.2). On restimulation with 10 mM acetate and 10 mM lactate a substantial decrease in sulphate and Eh from Day 140 occurred in microcosms amended with electron donor (Fig. 5.2). In addition, an increase in sulphide was seen in pore- waters from 196 days after restimulation, and again at Day 300 confirming dominant sulphate-reducing conditions. Control experiments (sterile and seawater only) showed no changes in sulphate and Fe(II) concentrations (Fig. 5.2).

192

193

5.3.4 Radium speciation and fate

To explore the impact of bioreducton on the speciation of Ra and Ba in these microcosm systems geochemical, microbiological and sequential extraction approaches were adopted. Sequential extractions were performed after 1 day and 300 days on electron donor amended (15 mM acetate and 15 mM lactate) sediments from all sets of microcosm experiments (e.g. radiobarite, barite and Ra2+(aq) systems) (Fig. 5.3, Fig. S5.4 & S5.5).

Figure 5.3: Radium concentrations in sequential extraction leachates from sediment microcosms containing Ra2+(aq) and radiobarite, following 1 and 300 days of anaerobic incubation

5.3.4.1 Ra2+(aq) microcosm experiments

Up to 75 - 80% of the Ra2+ spike was adsorbed onto unamended and amended sediments by 24 hours (Fig. 5.4). Following this initial fast uptake, adsorption continued at a slower rate until Day 140 where around 90% of radium was associated with the sediments (Fig. 5.4). As reported by Landa et al. (1982), the degree of radium attenuation during oil field brine discharges via sorption to sediment mineral surfaces is dependent upon the salinity of the receiving waters. This is to some extent reflected in this study by the time taken for equilibration to

194 be established between radium in solution and the sediment (Day 140 – 300). The high salinity of the seawater suggests competition for sorption sites on the sediments exist between ions in solution which explains the fast and slow uptake of radium observed. This indicates that upon production water discharge a small fraction of Ra2+ may well be diluted and dispersed in the marine environment or sequestered further downstream. Interestingly, no release of adsorbed Ra2+ was detected in microcosm waters during Fe(III) and sulphate reduction, suggesting radium remains adsorbed to mineral surfaces in agreement with past work (Wilkins et al., 2007). In addition, sequential extraction results (Fig. 5.3) showed that prior to progressive anoxia the radium was associated with the carbonate (57%), reducible (17%), exchangeable (10%) and organic (9%) phases. This is in stark contrast to the radiobarite experiments and confirms as expected that the Ra2+ is bound in a more readily available form in the sediment compared to the (radio)barite. Adsorption and accumulation of radium to clay minerals such as muscovite and kaolinite is expected due to its capability to sorb considerable portions of radium and quartz to a lesser extent (Ames et al., 1983; Landa and Reid, 1983; Beneš et al., 1984; Beneš et al., 1986; Siddeeg et al., 2015; Van Sice et al., 2018; McDevitt et al., 2019). Additionally, radium sequestration via adsorption or partitioning into carbonate phases is known (e.g. calcite, siderite, ankerite, aragonite and dolomite) to occur and provide effective sites for sorption, or drive the precipitation/co-precipitation of mineral phases (Landa et al., 1983; Jones et al., 2011; Van Sice et al., 2018; McDevitt et al., 2019). The observed radium fractionation profiles strongly suggest attenuation in the carbonate, exchangeable and organic phases (Fig. 5.3). After anoxic conditions there was no change in the radium associations in the sediment between Day 1 and 300 and the bulk of radium and barium remained in the carbonate, exchangeable and organic fractions (Fig. 5.4). This confirms that radium is not remobilised, re-sequestered into newly formed mineral phases or re- adsorbed to sediment. Microcosms containing solely seawater and radium provided confirmation that radium remained undersaturated as expected.

195

Pore water (Bq mL-1):

Release (%):

5.3.4.2 Radiobarite and barite microcosm experiments

As progressive anoxia developed no release of radium or barium was detected in microcosm waters in electron donor amended barite and radiobarite microcosms (Fig. 5.4). This confirms the recalcitrance of radium to remobilisation during bioreduction. In addition, sequential extraction results (Fig. 5.3) indicate that there was no change in the radium associations in the sediment between Day 1 and 300 and the bulk of radium and barium remained in the barite fraction even after progressive anoxia. This confirmed radium and barium was recalcitrant to remobilisation from the radiobarite and barite minerals (Fig. 5.3 and S5.4).

In summary microcosm experiments showed no release of radium as biogeochemical processes developed in the sediment via microbial dissolution or desorption. No release of radium or barium occurred during both Fe(III)-reduction

196 and sulphate-reducing conditions, strongly indicating radium and barium remain sequestered as the recalcitrant barite mineral and/or adsorbed to mineral surfaces recalcitrant to remobilisation. Trends between barium and radium fractionation profiles were very similar for sediment-microcosms containing radiobarite and barite pre- and post-anoxia (Fig. 5.3 & Fig. S5.4). This confirms attenuation via adsorption as well as precipitation are key sequestration processes in the environment during discharge. However more importantly radium fractionation profiles across all experiments overall remained uniform pre- and post-anoxia confirming that radium is not remobilised as progressive anoxia developed. The possible formation of transient species such as BaS and BaCO3 as reducing conditions progress and re-sequestration of radium into secondary iron, sulphate and carbonate mineral phases or re-adsorption onto mineral surfaces following the release of radium or barium from barite and sediment of this nature, is highly unlikely as illustrated from leaching experiments in contrast to other studies under differing field conditions (Baldi et al., 1996; Konhauser et al., 2002; Luptáková et al., 2015). To fully investigate whether microbial reduction impacted barite dissolution, etch pit analysis in addition to the sediment leaching experiments were conducted pre- and post-anoxia (Section 5.3.5).

5.3.5 Fate of Ra: Etch pit formation analysis

5.3.5.1 Spectroscopic analysis of barite

Barite minerals exposed to anoxic conditions in microcosm experiments were extracted via the use of heavy liquid. The surfaces of barite grains were investigated to further clarify whether microbial-mediated barite dissolution occurred. BSE imaging and EDS spectra confirmed the presence of Ba and S peaks in the EDS spectra, and the presence of the characteristic barite morphology (Fig. 5.5b & c). There was no evidence for etch pit formation on any of the barite surfaces, even those exposed to sulfidic conditions for months, suggesting no microbial induced dissolution occurred (Fig. 5.5a). This confirms radium is retained in the barite phase under sulphate reducing conditions in contrast to results from other studies under differing conditions (Risthaus et al., 2001; Becker et al., 2005; Ouyang et al., 2017).

197

Variances in results across studies can be attributed to the method of formation and nature of the material, as this can result in crystals adopting different grain sizes, morphologies and solubilities unrepresentative of the field conditions and materials of interest. This is the first study to synthesise and utilise barite chemically and morphologically identical to that formed and deposited in marine sediments due to produced water discharges in such experiments with an indigenous microbial community directly representative of the field. Barite harvested from control microcosm experiments containing only precipitate and seawater were also examined to consider potential etch pit formation via the ionic strength effect by background ions in solution (Fig. 5.5d) (Buhmann et al., 1987; Kowacz et al., 2008; Putnis et al., 2008). BSE imaging showed no indication of surface dissolution due to the effects of electrolytes on water structure dynamics and solute hydration (i.e. ionic strength effect) consistent with the absence of barium in pore waters in microcosm data (Section 5.3.3) (Kowacz et al., 2008). Abiotic dissolution of barite using EDTA was conducted as a control to demonstrate etch pit morphologies similar to microbial induced etching can form on the surface of barite (Wang et al., 1999, 2000; Kowacz et al., 2009; Ouyang et al., 2017) (Fig. 5.5e & Fig. S5.6). Here, EDTA leached samples showed etch pit formation on the surface (topmost), and increased dissolution adjacent to the surface (edges) similar to other studies (Wang et al., 1999, 2000; Kowacz et al., 2009) (Fig. 5.5e & Fig. S5.6).

198

A) BB))

C) D)

Figure 5.5: BSE images of the surface of barite grains after 300 days of anaerobic incubation to determine microbial-mediated etch pit formation; A) barite precipitate not exposed to anoxia (control); B-C) after anoxia (Day 300); D) exposure to seawater to account for ion-exchange/ionic-strength effect and; E) control showing the formation of elongated etch pits after exposure to an EDTA solution (5mins)

199

5.3.6 Microbial community characterisation

5.3.6.1 Evidence of progressive anoxia and radiological effects

Figure 5.6: Prokaryotic phylogenetic diveristy of sediments from the field (Day 0) and microcosms (Day 300). Phyla/classes are illustrated if present > 0.5% of the microbial community

Sequencing of 16S rRNA was conducted to investigate compositional changes in the microbial community after progressive anoxia and in the presence of radium. Bioreducing sediments were extracted at Day 1 and Day 300 from microcosms. DNA was extracted from electron donor amended, and unamended sediments from radiobarite, barite and Ra2+(aq) microcosms. Field sediment (e.g. Day 0) was also analysed. In the field sediment, a diverse population of soil bacteria was present (> 1700 species), with sequences associated with Fe(III)-reducing and sulphate-reducing

200 bacteria (SRB) (Fig. S5.7). The microbial community comprised mainly of Gammaproteobacteria (22%), Deltaproteobacteria (21%), Campylobacteria (12%), Bacteroidia (10%), Oxyphotobacteria (6%), Alphaproteobacteria (4%) and Anaerolineae (3%) confirming the presence of anaerobic species. Sequences closely related to known Fe(III)-reducing species such as Pelobacter sp. (2%), and sulphate- reducers such as Desulfobulbus sp.(3.9%), Desulfocapsa Sulfexigen (1.4%), Desulfosacrina (0.3%) and Desulfotalea (0.5%) close relatives of Deltaproteobacteria (> 97% match) were identified (Haas, 2008; Punjungsari, 2017). Sediments without electron donor additions across all sets of experiments showed enrichments of fewer anaerobic bacteria in comparison to amended microcosm systems (Fig. 5.6) (Lovley et al., 1995; Burke et al., 2005). Results showed a reduction in some species such as Gammaproteobacteria (15%) and Bacteroidia (7%) between field sediment (Day 0) and unamended microcosms (Day 300). However, an increase in Deltaproteobacteria (28%), Spirochaetia (0.6%), Thermoplasmata (3%), Nitrososphaeria (3%), Calditrichia (2%), Thermodesulovibrionia (1%), Woesearchaeia (3%), Anaerolineae (6%), Gemmatimonadetes (2%) and Ignavibacteria (1%) confirmed the ingrowth of anaerobes in the sediment. An enrichment was observed in Delta proteobacterium (2.7%), Desulfobulbus sp. (6%) and Uncultured sulphate-reducing bacterium (0.3%) reflecting the development of anoxic conditions by Day 300. Sediments amended with electron donor across all sets of experiments showed clear shifts in the community structure in comparison to Day 0, with noticeable enrichments of additional anaerobes. Results showed similar reductions in the taxonomic units observed in unamended sediments (i.e. no electron donor). However an increased presence of mainly Clostridia (13%), Spirochaetia (4.8%), Thermococci (1.8%), and others such as Methanomicrobia (1.4%), Atribacteria (1.8%), Thermoplasmata (1.5%), Woesearchaeia (0.7%), Cloacimonadia (0.9%), Synergistia (0.9%), Calditrichia (0.5%), Aegiribacteria (1.4%) and Mollicutes (0.7%) was detected. An increased presence of sequences affiliated with bacteria capable of both Fe(III)-reduction and sulphate-reduction were accounted for such as, Desulfuromonas svalbardensis (~ 2%), Desulfobacter curvatus (> 4%), Desulfotalea sp (> 2%) and Ferrimonas pelagia (> 3%) close relatives of Deltaproteobacteria and

201

Gammabacteria (> 97% sequence homology) (Haas, 2008; Punjungsari, 2017). This data also reflects the progressive development of anoxic conditions by the presence of methanogenic archaea such as thermoplasmata and methanmicrobia. At 300 Days enrichments of methanosaeta pelagica (1%) and methanosaeta harundinacea (> 2%) close relatives to known methanmicrobia (> 97% sequence match), were detected in electron donor amended sediments. Additionally, data show the microbial community structure was unperturbed by the presence of radium and barium suggesting low toxicity in agreement with other studies (Neff, 2002; Ruus et al., 2005; Grung et al., 2009). The amendment of field sediment with electron donor show we examined and modelled the fate of radium as anoxia developed if such conditions were established in the natural environmental setting. This also confirms radium remains recalcitrant to remobilisation as sulphate-reducing conditions dominate.

5.4 Environmental Implications

During produced water discharge to shallow marine waters, sequestration of Ra2+(aq) to sediment by surface adsorption occurs. No release of adsorbed Ra2+ was detected in microcosm waters from electron donor amended, and unamended sediments during the development of Fe(III)- and sulphate-reduction. Radiobarite experiments did not show radium remobilisation during bioreduction under these conditions. In the natural field sediment a diverse microbial consortium was present including anaerobes capable of sulphate-reduction. Progressive anoxia rapidly developed in microcosms amended with electron donor shown by the increased sulphate-reducing and methanogenic species. This implies radium released to shallow marine systems will be attenuated in sediment by sorption and precipitation, and remain recalcitrant to remobilisation even when levels of bioavailable carbon mediate bioreduction in sediments. Biodiversity was high and similar across all treatments and no impact from the experimentally enhanced radium concentrations was observed suggesting sediment biodiversity is not affected by produced water discharge. The impact of radium discharges in these systems is negligible due to dilution and dispersion effects,

202 sorption and barite formation. (Betti et al., 2004; Eriksen et al., 2006; Gafvert et al., 2007; Grung et al., 2009; Olsvik et al., 2012). These data can aid the management of radium in the environment and inform the prediction of radium mobility during discharge or other similar disposal scenarios.

Acknowledgements

The work contained in this paper was conducted during a PhD study undertaken as part of the Natural Environment Research Council (NERC) Centre for Doctoral Training (CDT) in Oil & Gas [grant number NE/M00578X/1], under its Environmental Impact and Regulation research theme. It is sponsored via a scholarship provided by University of Manchester (UoM) and the NERC CDT studentship whose support is gratefully acknowledged. We thank Paul Lythgoe, Alastair Bewsher and Christopher Boothman for analytical support at the University of Manchester.

203

Supporting information for Chapter 5: The Effects of Bioreduction on the Fate of Ra2+(aq) and Radiobarite in Marine Sediments on Discharge of Oil Produced Water

Additional figures

Figure S5.1: XRD pattern of the sediment used in all sediment microcosm experiments

Figure S5.2: FTIR of the precipitate obtained from mixing experiments

204

Figure S5.3: XRD pattern confirming the bulk mineralogical composition of the precipitate obtained from mixing experiments to be Ba75Sr25

205

Figure S5.6: BSE image of the etch pits and sizes formed on the surface of barite grains via EDTA dissolution

206

Figure S5.7: Relative species diveristy of samples mesuaed by number of observed species identified vesus the total number of sequences analysed

207

CHAPTER 6

Chemical and Radiological Characterisation of Scales containing NORM

Faraaz Ahmad1, Katherine Morris1, Gareth T.W. Law2, Kevin Taylor1 and Samuel Shaw1*

1Research Centre for Radwaste Disposal and Williamson Research Centre, School of Earth & Environmental Sciences, Williamson Building, The University of Manchester, M13 9PL;

2Radiochemistry Unit, Department of Chemistry, University of Helsinki, P.O. BOX 33 (Yliopistonkatu 4), 00014, Finland

*Corresponding Author ([email protected])

Key words: Radiobarite, Oil and Gas industry, NORM, Produced water, Radium, Barium, Scale

6.0 Abstract

Hard scales containing naturally occurring radioactive materials (NORM) from tubulars from oil producing platforms in Iraq and the North Sea have been comprehensively investigated to determine the best ways of characterising NORM to aid the assessment of the fate of such materials in the environment. In current work, we explore the relationship between the bulk composition and radionuclide content, and the spatial relationship between these factors to understand the radionuclide uptake within these materials. Samples from Iraq were mainly comprised of anhydrite (CaSO4) and gypsum (CaSO4.2H2O), and were all exempt of radioactivity due to the ionic radii incompatibility between calcium and radium. In contrast, North Sea samples containing varying amounts of strontium and barium

208

226 (SrBaSO4 and BaSrSO4) provided evidence of measurable activity of Ra (10.7 – 18.5 Bq g-1). Specific activity concentrations of 210Pb in black dust materials comprised of galena (PbS) and wurtzite (ZnS) mineral phases was found to be 30.6 Bq g-1.

6.1 Introduction

During extraction and production of oil and gas naturally occurring radionuclides (NOR’s), can be transported from the subsurface to the produced waters which can lead to the co-precipitation and accumulation of radioactive-containing inorganic solids such as barite (BaSO4), celestite (SrSO4), gypsum (CaSO4.2H2O) and calcite

(CaCO3) (IOGP, 2016). Such minerals can accumulate within pipe lines, valves, storage tanks, wellheads, tubulars and may also deposit on and contaminate production equipment (Garner et al., 2015; Puntervold et al., 2008; Vearrier et al., 2009). NOR’s (mainly isotopes of radium e.g. 226Ra) can co-precipitate, incorporate and concentrate, not only in inorganic mineral scales, but also within sludge’s and drilling muds classifying them radioactive (Al-Masri et al., 2005; Garner et al., 2015; IOGP, 2016; Paschoa, 1997). Therefore, such materials have attracted increasing attention globally over the past decades due to potential environmental and health exposure risks imposed. Henceforth it is important to characterize and analyse such materials to aid processes such as decontamination and disposal, and to also understand the associated fate and hazard in the environment.

Reasons for scale formation include pressure and temperature changes (e.g. experienced during transportation of oil and gas from the subsurface), water flooding operations (e.g. injection of incompatible waters during enhanced oil recovery operations (EOR) such as seawater to maintain reservoir pressure) and evaporation (e.g. in fracking ponds and gas extraction piping) (Badr et al., 2008; Mackay, 2005; Vetter, 1972). The core root for the formation of scale in tubulars in the petroleum industry is mainly due to the mixing of incompatible waters (e.g. seawater and produced water), and the establishment of supersaturated solutions (Zhang et al., 2014). In such instance the richness of ions within produced waters

209 such as group II (alkaline earth) metals (e.g. Ca2+, Sr2+, Ba2+, Mg2+ and Ra2+) and their

2- 2- resultant interaction with dissolved sulphate and carbonate anions (SO4 and CO3 ) introduced from injected waters, coupled with temperature and pressure changes, result in the precipitation of a vast array of sulphate and carbonate scales (El- Hattab, 1985; IOGP, 2016; Todd et al., 1990; Vetter, 1976; Vetter et al., 1982; Yuan et al., 1994; Zhang et al., 2014). Additionally, due to the decay of primordial radionuclides 238U and 232Th situated within reservoir rock and the highly reducing conditions established within the reservoir, daughter progeny for example radon (222Rn), radium (228Ra, 226Ra), polonium (210Po) and lead (210Pb) within formation waters can incorporate into solid mineral phases via processes such as co- precipitation (e.g. radium co-precipitation into barite) and/or radiative decay (ingrowth) (Al-Masri et al., 2003; Fisher, 1998; Heaton et al., 1995; Røe Utvik, 1999; Zhang et al., 2014). The precipitation of mineral phases generally depends on the point at which supersaturation is established and is therefore typically encountered at various locations across the production system (e.g. inside gas/oil separators or near the wellhead) (IOGP, 2016; Steffan, 2013). The mineralogical compositions of precipitates encountered is controlled by the composition of the formation waters which in turn can be influenced by the basin geology formation composition and its respective location (Bader, 2006; Badr et al., 2008; Garner et al., 2015; Mitchell et al., 1980; Moghadasi et al., 2007; Todd et al., 1990; Worden et al., 2000). Other factors include the life time of the well as on longer time scales waters existing in other formations adjacent to the hydrocarbon-bearing formation can become part of the water matrix which alter the properties of the produced water (Li et al., 2013; Rosenberg et al., 2014; Y O Rosenberg et al., 2011a; Rosenberg et al., 2011b; Todd et al., 1992; Yuan et al., 1994; Zhang et al., 2014).

Among the most common scales that form in oil and gas process systems are radiobarite and/or strontium-containing scales (e.g. RaxBa1-xSO4, RaxSr1-xSO4 and

BaxSryRazSO4) which form by mixing processes and changes in thermodynamic conditions (e.g. pressure and temperature) (Al-Masri et al., 2005; Doyi et al., 2016; Garner et al., 2015; IOGP, 2016). During the formation process as the barite crystals precipitate on the surface of the pipes, radium and/or strontium is taken up into

210 these particles as they grow via co-precipitation. Other studies show the existence of calcium carbonate and sulphate scales (e.g. CaSO4 and CaCO3) with reduced uptake of radium (226Ra) in contrast to barite due to differences in chemistry (e.g. ionic radii / volumetric mismatch) (Al-Masri et al., 2005; Garner et al., 2015; Hedström et al., 2013; Miyake, 1978; Yoshida et al., 2008). A study of Syrian scales from an oil field show that calcium carbonate and calcium sulphate scales form when a bicarbonate rich formation water was present (Al-Masri et al., 2005). These tend to have much lower radium concentrations than those of barite scales. The reason stated for this is due to the difference in the chemistry of the elements. Radium has an ionic radius of 1.7 Å and barium has an ionic radius of 1.61 Å which is close to that of radium. However, calcium has an ionic radius of 1.34 Å which is significantly different and therefore, it is very difficult for the radium to be taken up into the crystal structure (Al-Masri et al., 2005; Garner et al., 2015; Hedström et al., 2013; Miyake, 1978; Shannon, 1976; Yoshida et al., 2008).

Further studies have identified lead- and iron-containing NORM scale deposits for example, lead sulphide (PbS; galena), lead-iron-sulphide (PbFeS; ‘black dust’), iron sulphide and iron oxides (e.g. Fe3O4, FeS and magentite) on surfaces of equipment e.g. tubulars, inlets, gas transport systems and pumps (Abdul a et al., 2010; Badr et al., 2008; Curti, 1999; Hartog et al., 2002; SNIFFER, 2003; Trifilieff et al., 2009; Worden et al., 2000). Differences in the mineralogical composition of scales and amounts globally found at installations is influenced by factors such as differences in reservoir geology, stage of production, method of production and associated temperature and pressure changes encountered during production. The formation and accumulation of scales and sludge can cause blockage of pipe lines and tubulars restricting the flow of fluids ensuing the need for descaling operations which engender radiation exposure and protection issues to workers, and the environment due to the potential risk of contamination (Hamilton et al., 2004; IOGP, 2016; Ismail et al., 2011; O’Brien et al., 1998; Paschoa, 1998). Descaling and decontamination methods correspondingly result in the formation of large amounts of potentially hazardous material containing an array of toxic heavy metals, as well as radioactive pipe scale and sludge material (NORM) (Wilson et al., 1992).

211

Therefore, safety measures and disposal methods need to be carefully considered and determined prior to decontamination and/or decommissioning of wells and oil producing assets based upon the nature of materials formed (Pontedeiro et al., 2010, 2007).

NORM disposal practices and regulations vary globally, however disposal methods for NORM waste generated in the offshore and onshore oil and gas industry differ. Generic measures adopted offshore and onshore include disposal at source to the marine environment (e.g. offshore produced water discharges), geological disposal (e.g. disposal/abandoned wells), well injection including salt dome disposal, storage or burial as radioactive waste (e.g. low-level radioactive waste mines) and the use of (un-) lined pits/ponds (e.g. fracking ponds and lagoons) (IAEA, 2004; OSPAR Comission 2014, 2013; Roberts et al., 1998; UK NORM Waste Strategy, 2014; Yoshida et al., 2008). Estimations of the amount of scale produced annually have been previously estimated to be around 100 tonnes per producing well (American Petroleum Institue (API), 1992). Other forms of NORM waste which exists in the oil and gas industry includes sludge materials, typically of lower activity in contrast to solid scale and contaminated soils (Abdellah et al., 2014; International Atomic Energy Agency, 2014; Landa et al., 1983; McDevitt et al., 2019; Van Sice et al., 2018). Studies have previously focussed on the characterisation of a variety NORM material to aid disposal and decontamination processes onshore, for example Al- Masri et al., (2005) correlated the mineralogical compositions, elemental and radioactivity concentrations of a variety of Syrian scales to identify the age and activity of samples. Similarly, Garner et al., (2015) evaluated the activity concentrations of radiostrontiobarite scales from the North Sea by deducing correlations between barium-strontium concentrations and radium activity concentrations. However, extensive characterisation of NORM thus studies in relation to the radionuclide content, mineralogy and the associated fate and hazard of such materials in the environment has received less attention.

In this study we characterise the mineralogical, morphological and radiological composition of a variety of NORM samples taken from tubulars from oil producing facilities. We aim to understand the factors controlling the distribution of

212 radionuclides in NORM samples. This includes the relationship between the bulk composition and radionuclide content, and the spatial relationship between these factors to understand the radionuclide uptake within these materials. We use a range of microfocus (e.g. SEM, IR and Raman) and bulk techniques (XRF and XRD) to determine the best ways of characterising NORM to aid the assessment of the fate of such materials in the environment.

6.2 Experimental Method

6.2.1 NORM Samples & Characterisation 6.2.1.1 Sampling

Two different sets of scale samples from tubulars were analysed which originated from separate locations, the North Sea (North Sea 7536, 7470 & 7457) and Iraq (IRQ 1 - 4). Up to 5 - 7 g and 30 - 140 g of each sample were taken from Iraq and North Sea tubulars respectively. On the North Sea 7457 sample, it was observed that it contained a component which seemed to be corroded metal. This was separated individually and is henceforth referred to as North Sea 7457 corrosion. No additional information was provided on the samples.

6.2.1.2 Preparation

Samples were set into an epoxy resin (a mixture of bisphenol a-(epichlorohydrin) and hardener (amines, polyethylene, polytriethylenetetramine fraction; 10:1) and prepared as sections and polished to give a flat surface.

6.2.1.3 Chemical composition and mineralogical analysis

X-ray fluorescence spectrometry (PANalytical Axios) and XRD (Bruker D8 Advance diffractometer) were conducted to determine the bulk mineralogical and chemical compositions of the scale samples. Samples were dried in an oven (40°C ± 0.5°C), ground by hand using a pestle and mortar and homogenised for XRD and XRF analysis. Analysis of lattice parameters from peak positions of strontium and barium

213 end members were analysed to determine the percentage shift in peak data and composition. A Philips XL30 FEG and FEI Quanta 650 FEG ESEM (Field Emission Gun Environmental Scanning Electron Microscope) equipped with an EDAX Gemini or Bruker Quantaz EDS (Energy Dispersive Spectroscopy) system with an XFlash detector, was used for imaging and chemical analysis. EDS was used for chemical analysis of the samples. Precipitates were imaged uncoated under high vacuum (15 keV).

Fourier-transform infrared spectroscopy (Perkin Elmer Spotlight 400 FTIR imaging system Universal ATR) was used to map and determine the chemical composition of scale samples. Intensity maps were created using HyperView software and spectra were analysed using Spectrum_Image software. A scan range of 500 - 4000 wavenumbers was used (20 scans, 4 cm-1 intervals and 15 - 30 µm aperture). Raman spectroscopy (Horiba Scientific; Xplora Plus; spatial resolution of 0.5 µm) was also used to create chemical maps. Maps were created using LabSpec6 – Horiba Scientific software. Analysis was performed using a 785 nm laser, filter 25 %, acquisition time 5, accumulation 5, RTD 0.5, 50 x scope, 100 µm slit and 100 µm hole.

6.2.1.4 Radiological composition & distribution

Radium concentrations in selected scale samples (Iraq) were determined using gamma spectrometry (Canberra gamma spectrometer). Samples were sealed in a double polypropylene container in a standard geometry for 1 month prior to analysis to allow the Ra progeny to equilibrate. The gamma spectrometer was calibrated using known 226Ra standards in the standard geometries. Samples were counted for 24 hours and the radium concentration was calculated from measurements of the 214Bi and 214Pb daughter products and comparing these to known standards (Siddeeg et al., 2015). Scale samples from the North Sea were analysed externally via gamma spectrometry prior to receiving the samples.

Autoradiography was used to determine the spatial distribution of radionuclides on all samples. Scale samples were placed on the storage phosphor screen BAS-IP SR,

214 super resolution (G.E Healthcare) in a dark cupboard for 2 - 28 days. The activity on the surface of the sample was imaged using a Typhoon 9410 variable mode imager where the screen was then scanned and imaged using a HeNe laser (633 nm) with a pixel size of 10 – 25 µm. The extent of darkening recorded is quantitatively proportional to the activity on the sample surface (Zeissler et al., 2001).

6.3 Results and Discussion

6.3.1 Radioactivity content of scales 6.3.1.1 North Sea Samples

Gamma spectrometry and autoradiography of the North Sea samples show they contain measurable levels of activity (Table 6.1 and Fig. 6.1). The gamma spectrometry data shows two of the samples (North Sea 7457 & 7470) from the North Sea contain activity concentrations in respect to radium (226Ra) of 18.5 and 10.7 Bq g-1, respectively. In addition, activity corresponding to its progeny (210Pb and 210Po), and thorium decay (228Th and 228Ac) can also be detected (Table 6.1). The lower measured activity concentration of 210Pb in these two samples is attributed to the in situ decay and resultant ingrowth of 226Ra (‘supported mechanism’) typically observed in offshore North Sea scale materials (Garner et al., 2015; Heaton et al., 1995). In contrast, sample North Sea 7536 showed trivial recorded activity in respect to 226Ra but greater activity of its progeny 210Pb and 210Po (Table 6.1). The greater observed concentration of 210Pb in comparison to that of 226Ra in the sample indicates its presence is not a result of the in situ decay of 226Ra but due to a separate (‘unsupported’) mechanism such as the presence of lead rich water which commonly results in the formation of radioactive sulphide deposits (e.g. FeS, PbS or ZnS) (Hartog et al., 2002; IOGP, 2016; Trifilieff et al., 2009). Autoradiography shows radioactivity is homogenously partitioned across samples North Sea 7457 and 7470 (Fig. 6.1). Differences in the recorded image intensities further confirm sample North Sea 7457 adopts the highest measurable activity of 226Ra and 228Ac in comparison to sample North Sea 7470. In contrast radioactivity is not evenly distributed across the bulk of sample North Sea 7536 but, is partitioned and associated with the edges of the sample strongly inferring the

215 formation of a thin film radioactive lead deposit (e.g. black dust) disconnected from the bulk of sample (Fig. 6.1).

Additionally, North Sea 7457 corrosion exhibited measurable concentrations of radioactivity as shown from autoradiography (Fig. 6.1). Activity was assumed to be associated with the decay of 226Ra due to the nature of sample North Sea 7457, however this was not confirmed via gamma spectrometry due to limited sample availability.

6.3.1.2 Iraq Samples

Table 1 confirms the non-existence of radioactivity in these samples in respect to both 238U and 232Th decay in contrast to North Sea samples. Autoradiography provides further evidence of exempt activity across the entirety of these samples as expected (Fig. 6.1).

216

Figure 6.1: Autoradiography map to show the areas of activity within the samples to help correlate activity distribution with mineralogy (Blue = no recorded measurable activity; Green & Red = Increased measurable activity respectively); 7470 and 7457; 2 day exposure, 7536; 17 day exposure and IRQ 1 - 4; 28 day exposure time. Dashed (white) lines (- -) dictate edge of the samples

217

Sample location Isotope Decay Series Activity (Bq g-1) and number

North Sea 7457a Pb-210 U-238 (Ra-226) 0.14 Po-210 U-238 (Ra-226) 0.99 Th-228 Th-232 7.63 Ra-226 U-238 (Ra-226) 18.5 Ac-228 Th-232 10.7

North Sea 7470a Pb-210 U-238 (Ra-226) 0.12 Po-210 U-238 (Ra-226) 0.10 Th-228 Th-232 1.79 Ra-226 U-238 (Ra-226) 10.7 Ac-228 Th-232 9.89

North Sea 7536a Pb-210 U-238 (Ra-226) 30.6 Po-210 U-238 (Ra-226) 29.00 Th-228 Th-232 0.02 Ra-226 U-238 (Ra-226) 0.09 Ac-228 Th-232 0.04

IRQ 1-4b Ra-226 Ra-226 n.d Bi-214 Th-232 n.d Pb-214 Ra-226 n.d Pb-212 Th-232 n.d Pb-210 Ra-226 n.d Th-234 U-238 n.d

6.3.2 Spatial distribution relationship of radionuclides with chemical and mineralogical composition of scales 6.3.2.1 North Sea samples

Figure 6.2 shows the XRD patterns for North Sea samples which display peaks primarily for calcium carbonate minerals (e.g. North Sea 7536), strontiobarite (e.g. North Sea 7457) and bariocelestite minerals (e.g. North Sea 7470) encompassing different amounts of strontium and barium from peak position analysis, typical of reported North Sea scale deposits (Garner et al., 2015; Heaton et al., 1995; IOGP,

218

2016). Additionally, North Sea 7457 corrosion displayed peaks predominantly for iron-bearing minerals (e.g. goethite, akaganeite, magnetite and lepidocrocite) which is indicative of iron corrosion products. Backscattered electron (BSE) imaging and EDX in conjunction with XRF data (Table 6.2) was further used to assess inhomogeneity in morphology and composition of the scales.

219

Major Elements

CO2 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl CaO Cr2O5 MnO (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) Sample 7536 42.6 0.07 0.407 0.089 0.139 0.009 0.19 0.052 47.6 0.011 1.06 7470 0.064 1.37 0.138 0.302 0.672 0.012 37.2 1.55 1.91 0.028 0 7457 1.00 0.206 0.027 0.124 0.136 0.081 31.9 0.186 2.21 0.01 0 Fe2O3 NiO CuO ZnO SrO Y2O3 BaO Br Ga2O3 K2O WO3 (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) 7536 6.93 0.005 0.002 0.149 0.45 0.007 0.226 0 0 0 0 7470 0.276 0.013 0 0.006 27.7 0 28.6 0.003 0.015 0.092 0.014 7457 0.154 0.017 0 0 18 0 45.9 0 0 0.086 0

Table 6.2: Bulk composition determined via XRF of North Sea samples

6.3.2.1.1 North Sea 7457

The bulk of the scale sample was found to be composed of strontiobarite

(Ba~75Sr~25SO4) shown from the XRD (Fig. 6.2) data containing around 40.8 mol. % strontium deduced from peak position analysis of the XRD pattern. The XRF (Table 6.2) shows the barium concentration (BaO: 45.9 wt. % and Ba: 53.8 mol. %) is greater than the mole percent of strontium (SrO: 18 wt. % and Sr: 46.2 mol. %) confirming the strontiobarite phase in agreeance with the XRD data. BSE imaging shows distinct growth events and zones/layers in which acicular (fibrous and needle like) and partially dendritic crystal habits are adopted (Fig. 6.3). This texture is likely formed by a series of growth events as the scale formed on the pipe wall (Fig. 6.3 (A-C)). Elemental mapping shows the co-location of barium, strontium and sulphur with noticeable regions and zones comprised of reduced amounts of strontium in comparison to barium (Fig. S6.1). In addition, EDX indicates zones 2 and 4, where the crystals have adopted an acicular morphology, contain greater concentrations of strontium in contrast to zones 1 and 3 which adopt a dendritic morphology and have a lower Sr concentration (Fig. 6.3 (D) and Fig. S6.1). This suggests the formation of strontiobarite is the dominant process associated with this scale. Produced water data from the oil producing wells were not made available to decisively establish the cause of the observed inhomogeneity in the chemical and morphological composition. However, the variation in strontium concentration in

220 the scale is likely due to changes in the composition of the fluids (e.g. produced waters of greater salinity, differing compositions and/or other production/injection fluids) flowing through the pipes during production. These compositional changes may have also led to the observed changes in crystalline morphology within different zones (Fig. 6.3 and Fig. S6.1). The observed dense, compact dendritic structures in zones 1 and 3 in conjunction with an increased concentration of barium infers waters containing greater amounts of barium could have been introduced (e.g. pumped or injected) resulting in the rapid establishment of supersaturated waters in respect to barite. Thus due to the fast precipitation kinetics of barite and limited space for crystal growth the observed dendritic morphology is adopted. In zones 2 and 3 waters containing greater amounts of contaminants, or other competing divalent ions (i.e. strontium) could have been episodically injected or extracted altering the properties of the produced water, resulting in the co-precipitation of more strontium into the barite and disturbance of the crystallisation kinetics. This leads to the adopted fibrous acicular and subhedral morphology as described in other studies (Dana, 1985; Todd et al., 1990, 1992; Ceccarello et al., 2004; Garner et al., 2015).

FTIR analysis showed peaks at 1058 cm-1 (with a distinct shoulder at 986 cm-1) corresponding to the sulphur-oxygen (S-O) stretch and at 603 cm-1 (with a shoulder at 637 cm-1) corresponding to the bending motion of the sulphur-oxygen bond within sulphate characteristic of barite (Fig. S6.2) (Adler et al., 1965; Ramaswamy et al., 2010). Broadening of peaks between 1030 cm-1 and 1232 cm-1 corresponding to sulphate stretching modes, and increased peak variations above 1200 – 1300 cm-1 in zones 2 and 4 reflect mineral compositional, structural and crystallinity changes (Adler et al., 1965; Ramaswamy et al., 2010; Aroke et al., 2013). Variations in crystallinity and inhomogeneity in composition across the scale is further reflected in infrared mapping at 1224 cm-1 where these zones are clearly distinguished (Fig. S6.2).

The spatial distribution of radionuclides shown from autoradiography is inhomogeneous across the entire sample with a clear favourable distribution of radionuclides associated with the more dense zones richer in barium (zones 1 & 3)

221 in comparison to the less compact and dense zones rich in strontium (zones 2 & 4) (Fig. 6.1 and Fig. S6.3). The strong correlation between barium and radium is due to the ionic radii compatibility between these elements (Al-Masri et al., 2005; Brown et al., 2015; Ceccarello et al., 2004; Garner et al., 2015; Hedström et al., 2013; Miyake, 1978; Monnin et al., 1988; Zhang et al., 2014). Zhang et al., (2014) showed that the volumetric/ionic radii match between radium (1.7 Å) and barium (1.61 Å) and the faster precipitation kinetics of barite in contrast to celestite, results in greater radium attenuation during precipitation. Vinograd et al.,( 2018) also showed that radium mixes favourably with barium in contrast to strontium and that at increasing strontium concentrations in barite, radium uptake decreases. Garner et al., (2015), showed a positive correlation between the concentration of strontium and barium, and activity of radium within scale and sludge samples from the East Midlands, UK. The bulk composition of sludge materials composed of baritocelestite (Ba < Sr) possessed lower activity concentrations of radium in contrast to strontiobarite (Ba > Sr) pipe scales due to the ionic radii compatibility between barium and radium compared to strontium and radium. Variations in the concentrations of ions in produced water, ionic strength and temperature have also been shown to affect the uptake of radium in barite (Brown et al., 2015; Rosenberg et al., 2014; Rosenberg et al., 2011a; Rosenberg et al., 2011b; Vinograd et al., 2013; Zhang et al., 2014; Zhu, 2004).

XRD further indicated that the North Sea 7457 corrosion material obtained from this sample consisted of a mixture of iron oxide minerals (goethite, akaganeite, magnetite and lepidocrocite) (Fig. 6.2). Radioactivity was shown to be partitioned homogenously throughout the sample as shown from autoradiography images (Fig. 1). This suggests the possible attenuation of radium, via co-precipitation or adsorption to the iron oxide minerals. Sajih et al., (2014) showed that radium adsorbs to iron oxide minerals (e.g. goethite and ferrihydrite). Beck et al., (2013) also showed that radium can sorb to other iron oxides such lepidocrocite (174 ± 21 L kg-1) which has a higher Ra sorption coefficient than goethite (20 ± 8 L kg-1). Radium attenuation into iron oxides has been previously recognised at offshore

222 installations in the North Sea (Garner et al., 2015; Okonkwo et al., 2014; Read et al., 2013).

6.3.2.1.2 North Sea 7470

The scale was found to be predominantly composed of bariocelestite

(Sr~75Ba~25SO4) shown from XRD (Fig. 6.2) data containing 66.8 mol. % of strontium deduced from peak position analysis of the XRD pattern. The XRF (Table 6.2) shows the barium concentration (BaO: 28.6 wt. % and Ba: 32 mol. %) is lower than the

223 mole percent of strontium (27.7 wt. % and Sr: 68 mol. %) confirming the bariocelestite phase in agreeance with the XRD data. BSE imaging and EDX confirms inhomogeneity in morphology and composition of the scale sample (Fig. 6.4 (A-C). BSE images show the bulk of the sample is mainly comprised of overlapping interlocked fine needle (acicular) crystal morphologies with moderate porosity and permeability further visualised in the autoradiography image (Fig. 6.1). EDX indicates the crystals are rich in sulphur, strontium with minor barium corresponding to bariocelestite. XRD and EDX data both indicate barium co- precipitation into celestite (i.e. bariocelestite) in contrast to sample North Sea 7457, which is strontiobarite due to the larger amounts of barium present in this sample. Other crystal phases were identified such as sodium chloride (NaCl) in much lower amounts. Elemental maps additionally show the co-location of strontium, barium and sulphur in the sample and the greater abundance of strontium and sulphur across the sample in comparison to barium in agreeance with XRD data (Fig. S6.4). Heterogeneity in the mineral composition of the scale was not observed via SEM hence activity was uniformly distributed across the sample with no spatial distribution relationships between these factors identified (Fig. 6.1, 6.4 & Table 6.1). Variations in crystallinity and morphology were further reflected in infrared mapping (Fig. S6.5). It is widely known that the width of infrared peaks broaden with a decrease in crystallinity and sharpen with an increase in crystallinity (Hagemann et al., 1989). This is observed at the edge of the sample (Fig. 6.4C), which adopts a laminar morphology composed of large cemented crystals adopting high crystallinity, typical of slow precipitation kinetics. Therefore, sharp peaks between 1024 cm-1 and 1184 cm-1 corresponding to sulphate stretching modes were detected (Fig. S6.5) (Adler et al., 1965; Ramaswamy et al., 2010; Aroke et al., 2013). In comparison the bulk (middle) of the sample which adopts an acicular/dendritic and needle-like morphology is composed of smaller crystallites and less crystalline (Fig. 6.4B). Therefore, broadening of the sulphate stretching peaks were detected reflecting a faster process of precipitation and changes in crystallinity (Fig. S6.5). Though changes in chemical composition were not identified via elemental mapping and via additional peaks in infrared spectra. Observed changes in crystallinity may reflect aging and episodic build-up of layers in the scale.

224

In this case factors such flow rates of waters through well-strings may influence precipitation kinetics rather than abrupt changes in water composition as shown in sample North Sea 7457. Areas with no activity were identified in autoradiography corresponding to the porous areas within the sample (Fig. 6.1 and Fig. S6.3). The lower activity of the sample shown from autoradiography imaging (Fig. 1) in relation to North Sea 7457, and the lower activity concentration of radium from gamma spectroscopy analysis (Table 6.2) is reflected by the greater concentration of strontium and celestite in the scale sample which has a lower overall capacity of radium uptake (see Section 6.3.2.1.1).

225

6.3.2.1.3 North Sea 7536

The sample was found to be composed of calcite (CaCO3) shown from the XRF (Table 6.2) and XRD data (Fig. 6.2) comprising of 96.4 % calcite and 3.6 % magnesite from peak analysis of the XRD pattern. BSE imaging shows the presence of two distinct phases in the sample (Fig. 6.5 (C)). The bulk of the sample is comprised of a flat smooth well-developed crystals with minimal porosity. A second phase was identified on the edge of the sample appearing brighter in the BSE image indicative of elements of higher atomic numbers. BSE image showed the edge of the sample was composed of two phases reflected by the atomic number contrast (Fig. 6.5 (C)). Elemental mapping (Fig. S6.6) and EDX analysis (Fig. 6.5) showed the edge of the sample mainly comprised of lead (Pb), sulphur (S) and lower concentrations of zinc (Zn) corresponding to a thin layer likely comprised of a mixture of galena (PbS) and sphalerite (ZnS) (Fig. 6.5 (E)). In comparison, the bulk region (darker cemented region) which consisted of calcium, oxygen and carbon representative of calcium carbonate (calcite Fig. 6.5 (D)). This was further reflected by infrared and Raman spectroscopic analysis and mapping (Fig. S6.7 and S6.8). The presence of peaks at 876 and 1512 cm-1 were detected via infrared spectroscopy corresponding to the

CO3 group of calcite and subsequently mapped (Huang et al., 1960) (Fig. S6.7 (B and D)). The non-detection of the thin layer (~ 10 µm) of galena and sphalerite is attributed to the spatial resolution limits of the spectromicroscope (50 µm) (Fig. S6.7 (C)). Similarly, peaks recorded at 159, 282 and 1087 cm-1 was detected via Raman spectroscopy corresponding to calcite and was subsequently mapped (Fig. S6.8 (B and C)) (Gunasekaran et al., 2006; Rutt et al., 1974). Peaks in relation to sphalerite were not detected, however peaks relating to galena (PbS) were detected at 133, 171 and 970 cm-1 and were mapped (Fig. S6.8 (C and E)) (Smith et al., 2002). The detection of galena reflects the higher portion of such phase in the edge of the sample shown via BSE imaging. Therefore, the non-detection of sphalerite may be due to the spatial resolution limits of the spectrometer (~ 5 µm), or the low concentration of this mineral. The depression observed on the surface of the sample relates to the detachment of the scale from the pipe wall further confirmed by the presence of iron most likely removed from the pipe surface

226 shown in the XRF data (Fig. 6.5 (B) and Table 6.2). BSE imaging in combination with autoradiography and gamma spectrometry data show that the measurable radioactivity of 210Pb and 210Po is clearly associated with the edge of the sample consisting of a thin layer of galena in the form of (lead sulphide) and sphalerite (zinc sulphide) indicative of ‘black dust’ materials reported at installations (Table 6.1 & Fig. 6.1)) (Badr et al.., 2008; Hartog et al., 2002; IOGP, 2016; Trifilieff et al., 2009; Worden et al., 2000). The non-correlation of radioactivity associated with calcium is in agreeance with other studies in the North Sea and other global regions. This is due to the ionic radii incompatibility and crystallographic differences between radium (1.7 Å) and calcium (1.34 Å) (Al-Masri et al., 2005; Curti, 1999; Garner et al., 2015; Heaton et al., 1995; Shannon, 1976; Yoshida et al., 2008). Yoshida et al., (2008) showed that the smaller ionic radius of calcium leads to a reduced uptake of barium and radium into calcite. This was further reflected by the calculated partition coefficients for radium (0.15 ± 0.06) and barium (0.016 ± 0.011) uptake into calcite. This is in contrast to the calculated partition coefficient for radium into barite (1.07 – 7.49: NaCl 0-3M) which confirms the reduced uptake of radium into calcite than into barite (Rosenberg et al., 2014; Zhang et al., 2014). The greater activity of 210Pb in contrast to 226Ra in the sample and the presence of two distinct phases infers lead and zinc rich formation waters may have resulted in lead (210Pb)

being brought to the surface as lead chloride (PbCl2) which resultantly mixes with hydrogen sulphide to form a mixture of sulphide scales such as galena and sphalerite or wurtzite as previously described and reported (Badr et al., 2008; Worden et al., 2000). The formation of calcite can be attributed to changes in water composition where inter-connecting strata waters rich in calcium are introduced with subsequent changes in thermodynamic conditions in the production system e.g. increase in temperature and/or decrease in pressure (Abdul et al., 2010; Badr et al., 2008; IOGP, 2016; Mitchell et al., 1980; Moghadasi et al., 2007; Vazirian et al., 2016; Vetter, 1976).

227

10 µm

228

6.3.2.2 Iraq Samples

Figure 6 shows the XRD patterns for all the samples obtained from Iraq (IRQ 1 - 4). Peaks displayed primarily coincide with calcium sulphate minerals, anhydrite

(CaSO4) and gypsum (CaSO4.2H2O). Other minor mineral phases such as quartz

(SiO2) (IRQ 1 & 3), celestite (SrSO4) (IRQ 1, 2, & 4), basanite (2CaSO4.H2O) (IRQ 2 & 4) and halite (NaCl) (IRQ 3 & 4) were also detected. All samples did contain measurable radioactivity particularly in respect to 226Ra as observed for sample North Sea 7536. The non-correlation between radium and calcium observed is in agreement with other studies primarily due to the incompatibility between the ionic radii of calcium and radium as described in Section 6.3.2.1.3. The difference in mineralogical composition of these samples in comparison to those obtained from the North Sea is undoubtedly due to geographical related salinity and compositional differences of the waters extracted (e.g. formation waters). This is fundamentally influenced by the differing reservoir basin geology encountered at particular geographical locations as shown from studies conducted on NORM formation and produced water compositions in the Gulf of Arabia and North Sea (Bader, 2006; Badr et al., 2008; Moghadasi, 2003; Todd et al., 1990, 1992; Yuan et al., 1994).

229

Figure 6.6: XRD results obtained for scales (IRQ 1 - 4); Anh = anhydrite; Bsn = basanite; Gp = gypsum; Hl = halite; Hbl = hornblende and; Qz = quartz

6.3.2.2.1 IRQ 1

BSE imaging (Fig. 6.7) shows inhomogeneity in the morphology and composition of the scale due to incorporation of a variety of mineralogical phases. The images show the bulk of the material is comprised of an interlocking, fine needle (acicular) crystal structure of low porosity similar to sample North Sea 7470 with the addition of other phases present. EDS spot analysis of the acicular crystals (Fig. 6.7(D) – Point 2) produced specra that contained peaks for Ca, S and O, which is likely to be the anhydrite phase identified in the XRD patterns collected from the same sample. IR mapping further confirmed the composition due to the presence of the distinct peaks at 1160 - 1170 cm-1, 680 cm-1 and 590 cm-1 characteristic of anhydrite corresponding to the antisymmetric stretching and asymmetric bending of the SO4 groups respectively (Fig. S6.9) (Periasamy et al., 2009; Liu et al., 2009). High

230 intensity peaks for strontium, oxygen, sulphur and low intensity peaks corresponding to calcium were confirmed from EDX analysis indicating the presence of celestite (SrSO4) with co-precipitation of calcium (Fig. 6.7 (D) - Point 5). The presence of silicon and oxygen peaks in EDX spectra were identified representative of quartz grains as expected as a result of sand being drawn up the tubular during extraction of oil (Fig. 6.7 - Point 1). Characteristic doublet peaks observed in infrared spectroscopy data between 900 – 1300 cm-1 (Si-O stretch mainly involving the motion related to the oxygen atom) and at 780 – 800 cm-1 (Si-O stretch linked with the motion of the silicon atom) further confirm the existence of quartz grains (Fig. S6.9) (Lippincott et al., 2012). IR mapping at 1600 cm-1 and 780 cm-1 further illustrates the heterogeneity in mineral distribution in the sample in respect to anhydrite and quartz respectively (Fig. S6.9). Concentrations of various trace elements such as iron, aluminium, sodium, silicon, zinc, and calcium were identified in the EDX spectra attributed to iron oxide and aluminium silicate minerals distributed within the scale (e.g. magnetite, pyrrhotite, silicate minerals such as hornblende and wurtzite). Iron incorporation due to the scrapping of steel pipes during the process of scale removal from (corroded) pipe walls and/or injection of production fluids are expected sources resulting in the distribution of fine grained iron phases across the sample.

231

6.3.2.2.2 IRQ 2

BSE imaging (Fig. 6.8 (A)) shows the surface of the scale is heterogeneous composed of large euhedral crystal structures enclosed by much smaller cemented crystallite structures. EDS analysis of the large crystal structures (Fig 6.8 (A) - Points 2 & 3) produced spectra that contained peaks for calcium, sulphur and oxygen which is likely to be anhydrite identified in the XRD pattern from the sample (Fig. 6.8 (B)). Material surrounding these crystals additionally contained low

232 concentrations of impurities (trace elements) such as aluminium and strontium displaying the occurrence of co-precipitation into anhydrite (Fig 6.8 (A) – Point 1). Other phases and impurities were also identified such as celestite (Fig 6.8 (A-B) - point 4). Infrared mapping further confirmed the composition of the sample and porosity features similar to IRQ 1 (Fig. S6.10). Peaks at 1136 cm-1, 664 cm-1 and 600 cm-1 were recorded corresponding to anhydrite as expected with no distinct peak shifts observed as shown in Figure S6.10 (Periasamy et al., 2009; Liu et al., 2009). This data indicates that the difference in morphology and size of the crystals observed is likely governed by the fluid flow properties and rate of crystallisation. This suggests that the larger crystals observed may have experienced longer growth times and may have formed earlier downhole in the core matrix in the mixing zone during water flooding reflected by the adopted morphology and particle size. Subsequent enclosure and entrapment by a cemented slurry of faster growing anhydrite crystallites is likely, as shown by the incorporation of various trace elements (e.g. aluminium and strontium) into this less well-established phase indicative of a competitive environment with minimal space for crystal growth hence the adopted cemented anhedral morphology (Hagemann et al., 1989; Todd et al., 1992).

233

Figure 6.8: A) BSE image showing morphological difference within the sample (large crystals surrounded by a slurry) and; B) EDX spectra corresponding to the elements in the scale; bulk mineralogical composition of large crystal structures to be CaSO4 (Point 2 and 3) with incorporation of trace elements in surrounding slurry (Point 1) and also celestite phases being identified (Point 4) and darker regions to be resin/background (Point 5)

234

6.3.2.2.3 IRQ 3

The surface of the scale is inhomogeneous and irregular with two distinct phases identified, a well-structured cemented/laminar phase (brighter phase) and a disordered phase composed of a number of smaller crystals adopting a need-like morphology (darker phase) fused to some degree (e.g. subhedral morphology) as shown from BSE imaging (Fig. 6.9 (A)). Close examination of the BSE image shows the amalgamation of two bright phases (Fig. 6.9 (A); Point 1; and Points 2 & 4). The EDX spectrum (Fig. 6.9 (B)) indicates that both of these phases despite morphological differences comprise of calcium, sulphur and oxygen corresponding to anhydrite with the brighter cemented areas integrating greater portions of strontium via co-precipitation (CaxSrySO4). This is further reflected by the atomic number contrast (brightness) between Points 1 and Points 2 and 4 in the BSE image (Fig. 6.9). EDX spectra indicates the darker finer grained phase surrounding the brighter phase is dominantly composed of anhydrite with a significantly reduced presence of strontium reflected by the presence of only fragments of brighter phases (Fig. 6.9; Point 3 & 5). This was further reflected by infrared analysis and mapping (Fig. S6.11). The presence of peaks at 612, 668 and 1152 cm-1 corresponding to the SO4 group confirm the presence of anhydrite with no distinct peak shifts observed as shown in Figure S6.11 (Periasamy et al., 2009; Liu et al., 2009). This suggests either a single brine or multiple brines of different compositions may have been pumped/injected or extracted at different intervals resulting in two phases of precipitation. The adopted cemented laminar/anhedral morphology results from the rapid establishment of supersaturation and precipitation of crystals in limited spaces (e.g. narrow pipes or pore spaces in the core matrix) resulting in them fusing/merging. The co-precipitation of strontium is likely to result from the injection of highly saline brine with high concentrations of calcium and strontium. The subsequent passing of dilute waters or a decrease in supersaturation of the solution results in the precipitation of reduced amounts of finer anhydrite crystallites adopting subhedral needle-like morphology. The increased porosity of the sample in comparison to IRQ 1 and IRQ 2, but similar to

235

IRQ 4, can be visualised from the IR map illustrating the distribution and presence of resin within the pores of the sample at 1304 cm-1 (Fig. S6.11).

236

6.3.2.2.4 IRQ 4

Heterogeneity in morphology and composition of the sample displayed in the BSE images is similar to that observed in sample IRQ 3 (Fig. 6.9 and 6.10). BSE imaging and EDX spectra both indicate the presence of both well-established laminar (e.g. larger and smoother crystals) and subhedral crystal forms (e.g. smaller acicular crystallites) of anhydrite with co-precipitation and uptake of strontium analogous to sample IRQ 3 (Fig. 6.10 (B)). Adopted morphologies and cementation of anhydrite is in agreement with reasons described in Section 6.3.2.2.3. IR mapping also shows porosity features similar to IRQ 3 and further confirms the bulk composition from peaks recorded at 604, 668 and 1168 cm-1 representative of anhydrite (Fig. S6.12) (Periasamy et al., 2009; Liu et al., 2009).

237

Figure 6.10: A) BSE image of IRQ 4 confirming the bulk mineralogical composition to be CaSO4 (Point 2) and brighter grey regions containing strontium (Point 1) and; B) Energies of the characteristic backscattered electrons detected corresponding to the elements in the scale

6.4 Conclusion and Environmental Implications

Two sets of solid scale samples from different geological locations were fully characterised. The dominance of a preferred scale forming at a particular oil field is shown to be influenced by a combination of factors (e.g. temperature, pressure, method of operation, basin geology, salinity of waters and site of extraction in the production system etc.) as shown from this study. Variations in mineralogical

238 composition between locations is likely due to differences in basin geology governing the composition of the formation waters within reservoirs globally, thus controlling the resultant mineral precipitated during water flooding operations.

Mineralogical compositions of North Sea samples (BaSrSO4, SrBaSO4 and CaCO3) and those from Iraq (CaSO4) were similar to those reported at offshore and onshore installations in the North Sea and Gulf of Arabia (Al-Masri et al., 2005; Badr et al., 2008; Garner et al., 2015). Additionally, the radiological content of scale samples strongly correlates with its mineralogical composition as shown from autoradiography imaging. Barite- containing phases demonstrated greater radium attenuation in comparison to calcium-containing minerals predominantly due to differences in ionic radii compatibility reflected in reported structural parameters, partition coefficients and studies conducted on similar solid scale (Al-Masri et al., 2005; Rosenberg et al., 2014; Shannon, 1976; Vinograd et al., 2018; Yoshida et al., 2008; Zhang et al., 2014). Variations in barium and strontium concentrations in strontiobarite and bariocelestite scales was also shown to affect the degree of radium attenuation within samples (Vinograd et al., 2018; Garner et al., 2015). The environmental and exposure risks posed by hard scales therefore dominantly influenced by the composition of the scale. During marine disposal of waste water effluents, solid scale deposits are separated however, some deposits may enter the water column. Additionally, deposits may dislodge from pipes and accidently enter the environment. Due to solubility variances of minerals, more readily available mineral forms (e.g. calcium carbonate) pose greater environmental hazard in relation to radionuclide release in the marine environment in contrast to recalcitrant barite minerals ( Al- Marsi et al., 1998; Landa et al., 1983; Menzie et al., 2008; Van Sice et al., 2018; McDevitt et al., 2019). However, it is globally recognised radioactivity (e.g. 226Ra) favourably attenuates in strontium and barium rich scales during production as shown from this study and others (Al-Masri et al., 2005; Garner et al., 2015; IOGP, 2016). Therefore, it is important to note that the composition of the scale which forms at a particular field is significantly influenced and dependent upon the oil field characteristics which in turn determines the associated risk it may pose.

239

Characteristics of the scale such as porosity and the form in which the scale exists (e.g. solid or dust) may also influence the hazard imposed (Chang, 1996; Phillips et al., 2001), for example scale consisting of a thin labile radioactive film (e.g. ‘black dust’) will most certainly pose a greater risk relative to radium attenuated in recalcitrant barite scales. Due to the observed grain boundary between the bulk calcite mineral and radioactive galena exhibited in the North Sea 7537 sample, this postulates easy detachment thus increased risk of contamination. Furthermore, dissolution of NORM deposits via microbial mediated dissolution, ionic strength effects and natural events (e.g. episodic dilution or acidification) can result in the release of radioactivity from these scales to the water column (Fedorak et al., 1986; Pardue et al., 1998; Phillips et al., 2001; Raahauge et al., 2016; Risthaus et al., 2001). Factors such as the environmental setting (e.g. deep sea or shallow marine), environmental characteristics (e.g. tidal regimes and mineralogical distribution), re- sequestration mechanisms of radionuclides and surface area of the scale deposit have been shown to significantly influence the rate of dissolution, the mobility, availability and hazard associated with radionuclides in the water column (Gafvert et al., 2007; Landa et al., 1983; Phillips et al., 2001; Van Sice et al., 2018; Wilkins et al., 2007). Phillips et al., (2010) showed that the microbial mediated reduction and release of radionuclides from solid scale deposits are deemed to be insignificant due to the lack of radium (0.03 % / 0.8 Bq L-1) detected in microcosm waters during sulphate reduction. The low radium release measured was attributed to the small surface area of the scales and large grain sizes (Phillips et al., 2001). However, radiological exposure effects to biota and benthic fauna may prove to be significant dependent upon the mass of the scale deposited due to the substantial activity concentrations typically found in NORM (Abdul et al., 2010; Grung et al., 2009; IOGP, 2016; Olsvik et al., 2012). Such values are considered to be minor if best practice is executed in relation to NORM disposal (e.g. following legislation and regulation) and as the vast portion of solid ought to be separated prior to operational effluent discharges. Material exceeding levels of exemption for NORM disposal in legislation would require adequate disposal and infers sediments containing improper levels of radium due to discharge of solid scale material to

240 marine environment during produced water discharge and other methods, should be addressed in accordance to legislative guidelines and regulatory requirements.

241

Supporting information for Chapter 6: Chemical and Radiological Characterisation of Scales containing NORM

Additional figures

242

Figure S6.2: A) the area mapped via IR spectroscopy; size: 1950 µm x 5800 µm; B) The intensity map produced at the selected wavenumber of 1224 cm-1 showing the mineralogical distribution of strontiobarite across the mapped area, variances in crystallinity and composition (Blue = no measurable intensity; Green/Yellow/Red = increase in measurable intensity respectively) and; C) Point analysis (representative spectra) of the mapped region

Figure S6.3: Zoomed in autoradiography maps to show the porosity features (porous areas) of the samples exempt of activity; A) North Sea 7470 and; B) North Sea 7457 (Blue = no measurable activity Green/Red = increase in measurable activity respectively)

243

Figure S6.5: A) the area mapped via IR spectroscopy; size: 1492 µm x 821 µm; B) The map produced at the selected wavenumber of 1106 cm-1 showing associated variances in crystallinity and composition (Blue = no measurable intensity; Green/Yellow/Red = increase in measurable intensity respectively) and; C) Point analysis (representative spectra) of the mapped region

244

Figure S6.7: A) the area mapped via IR spectroscopy; size: 300 µm x 260 µm; B) The map produced at the selected wavenumber of 1384 cm1 showing the mineralogical distribution of calcite across the mapped area (Blue = no measurable intensity; Green/Yellow/Red = increase in measurable intensity respectively); C) The map produced at wavenumber 1660 cm-1; D) Point analysis (representative spectra) of the calcite mapped region (red area) and; E) Point analysis (representative spectra) of the galena and sphalerite region

245

Figure S6.8: A) the area mapped via Raman spectroscopy; size: 17 µm x 15 µm; B) The map produced at the selected wavenumber of 1000 cm1 showing the mineralogical distribution of calcite across the mapped area (Blue = no measurable intensity; Green/Yellow/Red = increase in measurable intensity respectively); C) Point analysis (representative spectra) of the calcite mapped region (red area); D) The map produced at wavenumber 200 cm-1; E) Point analysis (representative spectra) of the galena region; F) The map produced at wavenumber 600 cm-1 and; G) Point analysis (representative spectra) of the resin

Figure S6.9: IRQ 1; A) the area mapped via IR spectroscopy; size: 500 µm x 1000 µm; B) The map produced at the selected wavenumber of 1160 cm1 showing the mineralogical distribution of anhydrite across the mapped area (Blue = no measurable intensity; Green/Yellow/Red = increase in measurable intensity respectively); C) The map produced at wavenumber 780 cm-1 showing the distribution of quartz; D) Point analysis (representative spectra) of the mapped region (red area); E) Point analysis (representative spectra) of the quartz region

246

Figure S6.10: IRQ 2; A) the area mapped (red) via IR spectroscopy; size: 950 µm x 1800 µm; B) The map produced at the selected wavenumber of 1128 cm-1 showing the mineralogical distribution of anhydrite across the mapped area (Blue = no measurable intensity; Green/Yellow/Red = increase in measurable intensity respectively) and; C) Point analysis (representative spectra) of the mapped region

Figure S6.11: IRQ 3; A) the area mapped via IR spectroscopy; size: 950 µm x 1900 µm,; B) The map produced at the selected wavenumber of 1152 cm1 showing the mineralogical distribution of anhydrite across the mapped area (Blue = no measurable intensity; Green/Yellow/Red = increase in measurable intensity respectively); C) The map produced at wavenumber 1304 cm-1 confirming the infill of resin; D) Point analysis (representative spectra) of the mapped region; and; E) Point analysis of the resin region

247

Figure S6.12: IRQ 4; A) the area mapped via IR spectroscopy; size: 900 µm x 1442 µm; B) The map produced at the selected wavenumber of 1164 cm1 showing the mineralogical distribution of anhydrite across the mapped area (Blue = no measurable intensity; Green/Yellow/Red = increase in measurable intensity respectively) and; C) Point analysis (representative spectra) of the mapped region

248

CHAPTER 7

Conclusions and Future Work Directions

7.0 Conclusions

The aim of the project was to determine the speciation and fate of 226Ra derived from oil and gas production during two fundamental processes of NORM formation in which mixing of incompatible waters is encountered:

 Operational marine discharges of produced waters and  Mineral formation during water flooding operations

The aim was to determine the fate of radium in NORM in marine sediment and in solid scales formed during the mixing of incompatible waters encountered at different stages of oil and production. The factors controlling the distribution of radionuclides in a variety of NORM samples from the field (e.g. hard scales and sediment) was determined to understand the radionuclide attenuation within mineral phase’s formed under different conditions and processes. The principle aim of this project was to gain an understanding of the NORM formation process and nature of material which forms during operational discharge to improve our understanding of the risk that NORM presents and the fate of radionuclides in the environment. Prior to this work, research on the formation of (radio)barite during offshore discharge has been primarily focused on the mixing of simple synthetic waters (Zhang et al., 2014; Rosenberg et al., 2014). Other studies focus on the mixing of incompatible waters during water flooding operations dissimilar to the process of marine discharge (Todd et al., 1990, 1992). Most recent research has examined the discharge of produced water to freshwater settings (Landa et al., 1983; Langmuir et al., 1985; Jones et al., 2011; Van Sice et al., 2018; McDevitt et al., 2019), but a paucity of information exists on marine settings. The fundamental co-precipitation mechanism of radium into barite using full-component and field brines remains unresolved, including the long term fate of aqueous radium and radiobarite as

249 biogeochemical reducing conditions develop in marine settings. The work in this thesis contributes to these areas.

Here radium concentrations in sediments from a discharge site in the UK were assessed using gamma spectroscopy, and heavy liquid extractions were used to separate TE-NORM barite particles from marine sediments. The discharge and formation process of TE-NORM barite was simulated in the laboratory via mixing experiments entailing both field and synthetic produced waters and sea waters. This was conducted to produce particulates morphologically and chemically similar to those extracted from marine sediments, and to also validate our methodology and approach (Chapter 4). Radiobarite formation was further investigated to produce precipitates chemically and morphologically identical to that extracted from the field site to assess the long term fate of radiobarite during progressive anoxia in marine sediments (Chapter 5), and to explore the uptake and speciation of radium in radiobarite. Additionally, the long term fate of radium once discharged to a marine setting as the inorganic solid and aqueous ion was investigated via a series of sediment microcosm experiments where progressive anoxia was stimulated using terminal electron acceptor additions. A range of techniques including radiochemical measurements, sequential extractions, SEM, heavy liquid extractions and DNA sequencing were performed to fully quantify, understand the partitioning of radium within sediment and the resultant effect on the microbial community (Chapter 5). Finally, a comprehensive investigation into the factors controlling the distribution of radionuclides in NORM samples (e.g. hard scales) obtained from tubulars from oil producing platforms (UK and Iraq) was performed to understand radionuclide attenuation within mineral phases formed under different conditions and processes (Chapter 6). Here we determined the best ways of characterising NORM to aid the assessment of the fate of such materials in the environment.

Fate of radium during produced water marine discharges (Chapter 4), investigated the discharge process and formation of strontiobarite under marine conditions via the characterisation of marine sediment samples from a field discharge site, and seawater and produced water mixing experiments. XRD showed the bulk

250 mineralogical composition of marine sediments comprised mainly of silicate, carbonate, halite and clay minerals. Gamma spectroscopy in combination with XRF showed marine sediment samples from East of the outfall (D & E) displayed modestly elevated levels of radium (0.11 - 0.32 Bq g-1) and barium (345 – 1176 ppm) providing evidence of technologically enhanced levels of radium in the sediment that is co-associated with barite. Mineral grains typically between 245- 426 µm were isolated via the use of heavy liquid and analysed via SEM which displayed the presence of irregular aggregates, consisting of individual equant particles ≤ 2µm in size. EDS and elemental mapping confirmed these particles were rich in Ba, Sr and S indicative of strontiobarite. Additionally, radiometric analysis of strontiobarite particles via autoradiography showed they possessed clear measurable activity concentrations confirming radiostrontiobarite incorporation in marine sediments via barite co-precipitation and deposition. SEM further confirmed the presence of framboidal pyrite (FeS2) in field sediment samples indicative of highly reducing conditions below the sediment-water interface such as sulphate reduction promoting studies to underpin assessments of the environmental risk and fate of radium in these systems via natural biogeochemical processes (Chapter 5) (Roychoudary et al., 2003, Folk., 2005, Proske et al., 2015). Model systems where field produced water was mixed with field seawater under laboratory conditions to mimic the formation process of (radio)barite particles by discharge into the marine system were subsequently conducted. These experiments produced particles with similar morphology, composition (Ba73.1Sr26.3SO4) and particle size (2-6 µm) to those extracted from marine sediments. Similarly, mixing experiments using synthetic (full-component) produced water and synthetic seawater generated particles consistent with those obtained from marine sediments and field mixing experiments. XRD in combination with IR, EDS, XAS and EDTA dissolution data confirmed the mineralogy and composition of the precipitate to be strontiobarite

(Ba76.4Sr23.8SO4) with a particle size of 1-5 µm. Parallel experiments investigating radium uptake upon mixing of synthetic produced water and seawater were undertaken to assess the fate of radium within produced water when discharged into the marine environment. The rapid removal of Ba and Ra from solution (1 -3 hours) confirms the uptake of radium in this system is dominantly controlled by the

251 rapid precipitation kinetics of barite in agreeance with similar studies (Zhang et al., 2014). Radium uptake increases over time from 48% to 79% between 1 -7 hours followed by a further increase up to 97% by 24 hours, with a maximum radium uptake of 97.5%. Therefore, these results suggest a portion of radium will exist as the inorganic solid and aqueous ion in solution upon discharge. Dependent upon the characteristics of the receiving environment radium which is not incorporated into barite will be diluted and dispersed due to current flow effects and be subject to sorption further downstream. A value of 1.14 ± 0.1 was calculated for the effective partition coefficient (Kd’) for Ra2+ incorporation into strontiobarite which is comparable to other studies in this area investigating binary and ternary phases e.g. in fracking systems (or other controlled studies) under controlled conditions (Rosenberg et al., 2014; Zhang et al., 2014). Factors such as volumes discharged, discharge times and environmental setting can govern the amount of radium uptake into the strontiobarite phase thus the associated environmental impact.

Overall, this research presents the first extraction and characterisation of the precipitate which forms during the discharge of produced water into marine systems. This work provides a comprehensive assessment of (radio)strontiobarite formation using field and synthetic fluids consistent with marine field observations thus underlining the mechanism of radium incorporation into strontiobarite, and prediction of the fate of radium in these systems. The findings of this work conclude:

1. That strontiobarite is the main phase that forms during produced water discharge to the marine environment, with particle size and shape characteristics described above 2. It is likely that a portion of radium is taken up into barite as well as being dispersed and adsorbed to marine sediment, therefore these are the key phases which need to be assessed to understand the fate of radium from oil production in the marine environment

The effects of bioreduction on the fate radium in marine sediments (Chapter 5), extended the experiments of Chapter 4 where the formation of radiobarite was investigated under conditions representative of mixing conditions in Chapter 4.

252

Specifically, a simplified solution composition containing BaCl2, SrCl2 and Na2SO4 was adopted. XRD in combination with IR, EDS, and EDTA dissolution data confirmed particles from parallel inactive experiments adopted similar morphology, composition (Ba75Sr25SO4) and particle size (2-5 µm) to particles extracted from marine sediments, field and synthetic mixing experiments in Chapter 4. Secondly, a series of sediment microcosm experiments were conducted to provide a comprehensive assessment of the potential environmental impact, speciation and fate of radium once discharged to a marine setting. Three sets of microcosms containing field marine surface sediments from the discharge site (with their indigenous microbial populations) were doped with radiobarite (BaxSryRazSO4),

2+ barite (BaSrSO4) and aqueous radium (Ra ), and reducing conditions stimulated using electron donor additions (acetate/lactate). Triplicate microcosms across all sets of experiments, a sterile control (one replicate) and control containing seawater and sediment (one replicate) were inoculated with 5 mM acetate and 5 mM lactate as electron donor, while other controls containing no electron donor contained: seawater, sediment and precipitate/226Ra; and only seawater and precipitate/226Ra. Geochemical monitoring across all sets of experiments revealed the presence of Fe(II) in pore-water at the beginning of experiments indicating the highly anoxic nature of the field sediments. Electron donor amended sediment microcosms, progressed from Fe(III)-reduction to sulphate-reducing conditions across all sets of experiments. This was detected by a substantial decrease in sulphate concentration in microcosm waters and Eh recordings from Day 14-140 in electron donor amended microcosms. Control experiments (sterile and seawater only) across all sets of microcosm experiments showed no changes in sulphate and Fe(II) concentrations as expected. As progressive anoxia developed no release of radium or barium was detected in pore waters in electron donor amended barite and radiobarite microcosms. This confirms the recalcitrance of radium to remobilisation during bioreduction. Sequential extraction of electron donor amended sediments pre- and post-anoxia showed no changes in speciation of radium or barium before and after bioreduction. In contrast, up to 75-80% of Ra2+ was sorbed onto sediment mineral surfaces by 24 hours in aqueous radium microcosm experiments, and continued at

253 a slower rate until Day 140 where ~ 90% of radium was adsorbed. Sequential extraction results showed no changes in speciation of radium as anoxia developed, with dominant attenuation in the more readily available forms in stark contrast to (radio)barite experiments. Electron donor amended control experiments containing inactive barite showed similar results to the radiobarite systems with no release of barium to solution during bioreduction and with ~ 84% of barium attenuated in barite pre- and post- anoxia. In the barite only experiments, barite grains were separated from sediments using heavy liquid extractions. SEM analysis showed no evidence for any etch pits on the barite and further confirmed the indigenous microbes did not consume barium sulphate as a terminal electron acceptor. In addition, in sediments stimulated with electron donor, the microbial community also reflected the onset of bio-reduction with an increased relative abundance of Fe(III) and sulphate-reducing bacteria such as Ferrimonas pelagia, Desulfuromonas svalbardensis, Desulfobacter curvatus, and Desulfotalea sp. This confirms anoxic conditions (e.g. sulphate- reduction) were established in the experiments and radium remains in the recalcitrant barite mineral phase and adsorbed to mineral surfaces. Additionally, data show the microbial community structure was unperturbed by the presence of radium and barium suggesting low toxicity in agreement with other studies (Neff, 2002; Ruus et al., 2005; Grung et al., 2009).

Overall, this is the first comprehensive study to investigate the mobility and fate of both aqueous radium, and synthetic (radio)barite directly related to oil field discharges, as biogeochemical conditions evolve from iron- to sulphate-reducing conditions under marine field conditions. These data can aid the management of radium in the environment and inform the prediction of radium mobility during discharge or other similar disposal scenarios. The findings from this work conclude:

1. That radium will be attenuated into recalcitrant barite via co-precipitation in the long term within marine sediments, and assessments of risk should focus on these particulates when assessing the environmental effect of NORM in marine systems

254

2. Aqueous radium remain as an adsorbed phase at distances from the discharge point due to dispersion therefore, models of the fate of radium will need to take this into account

The chemical and radiological characterisation of scales containing NORM (Chapter 6), focused on the characterisation of a variety of solid scale samples containing NORM formed under different conditions (e.g. water flooding) to those described in Chapter 4 & 5. Here, a range of microfocus (e.g. SEM, IR and Raman) and bulk techniques (XRF and XRD) were used to determine the best ways of characterising NORM to aid the assessment of the fate of such materials in the environment. Samples were extracted from tubulars from two global locations (Iraq and North Sea). Factors controlling the distribution of radionuclides in these materials was investigated by identifying global relationships between the bulk chemical and mineralogical composition, and radionuclide content via a range of microfocus (e.g. SEM, IR and Raman) and bulk techniques (e.g. gamma spectroscopy, XRD and XRF). The spatial relationships between these factors was further assessed to understand the radionuclide uptake within these materials thus the environmental impact and fate of such materials in the environment. Samples from Iraq were mainly comprised of anhydrite (CaSO4) and gypsum (CaSO4.2H2O), and were all exempt of radioactivity due to the ionic radii incompatibility between calcium and radium reflected in reported structural parameters, partition coefficients and studies conducted on similar solid scale (Al-Masri et al., 2005; Rosenberg et al., 2014; Shannon, 1976; Vinograd et al., 2018; Yoshida et al., 2008; Zhang et al., 2014). In contrast, the mineralogical composition of the North Sea samples varied with calcium carbonate and strontiobarite minerals detected in the XRD and XRF data.

However, the mineralogical compositions of the North Sea (BaSrSO4, SrBaSO4 and

CaCO3) and Iraq (CaSO4) samples were similar to those reported at offshore and onshore installations in the North Sea and Gulf of Arabia (Al-Masri et al., 2005; Badr et al., 2008; Garner et al., 2015; Heaton et al., 1995; IOGP, 2016). Radiometric analysis via gamma spectroscopy provided evidence of increased measurable activity of 226Ra in North Sea samples containing greater concentrations of barium, due to the ionic radii compatibility between barium and radium (Vinograd et al.,

255

2018; Garner et al., 2015). Radioactivity was also associated with galena (PbS) and wurtzite (ZnS) mineral phases indicative of ‘black dust’ materials reported at installations (Badr et al., 2008; Hartog et al., 2002; IOGP, 2016; Trifilieff et al., 2009; Worden et al., 2000). XRD in combination with autoradiography showed corrosion material dominantly comprised of goethite, akaganeite, magnetite and lepidocrocite possessing measurable radioactivity (Beck et al., 2013; Sajih et al., 2014; Garner et al., 2015; Okonkwo et al., 2014; Read et al., 2013). Overall, variations in mineralogical composition between the two sets of samples is primarily affected by basin geology which governs the composition of the formation waters within reservoirs globally. It was also reasoned that NORM samples containing radium within recalcitrant barite phases possess a reduced environmental risk, in comparison to thin radioactive films of galena (e.g. black dust) associated with more easily leachable carbonate mineral phases (e.g. calcite) due to solubility variances (Al-Marsi et al., 1999; Landa et al., 1983; Menzie et al., 2008; Van Sice et al., 2018; McDevitt et al., 2019). These data are useful to predict the fate and hazard of such materials in a marine setting, as there is a lack of studies in this area. This work can also aid processes such as decontamination and disposal. SEM in combination with bulk techniques (e.g. gamma spectrometry, XRF and XRD) had proven to be most useful. Raman and IR analysis was restricted due to the spatial resolution limits of both instruments, and burning of the samples by Raman spectroscopy. Autoradiography was an important tool in assessing the distribution of radioactivity across samples. Gamma spectroscopy and XRF were also beneficial however, analysis can be restricted by sample availability.

256

7.1 Future Work

i) Barite formation using more complex brines e.g. scale inhibitors & production chemicals

In this project experiments utilising full-component brine compositions similar to those reported from North Sea installations, were essential in characterising the co-precipitation process and formation of (radio)strontiobarite during offshore marine disposal. Understanding the co-precipitation process provides strong scientific underpinning to the mechanism of radium incorporation into strontiobarite and prediction of the fate of radium in these systems relevant to various global installations. This provides a firm foundation from which to build up complexity in the effluent compositions. Additional experiments utilising more complex produced water compositions which contain multiple added components such as production chemicals (e.g. commercial scale inhibitors) would be beneficial to gain a further understanding of the co- precipitation mechanism. Also, site specific produced waters from other global locations can be used in mixing experiments to examine the material formed during discharge and to assess its fate and environmental impact.

ii) Model effects of dilution and dispersion in an open system

Radium uptake experiments conducted in Chapter 4, determined the mechanism of radiostrontiobarite formation and considered the resultant environmental implications of radium during offshore discharges in the marine environment. Factors such as dilution and dispersion were addressed in the study however, the scenario where there is maximum uptake was examined due to the complexity of modelling an open system. Therefore, future work, if possible, could

257

utilise software models (e.g. ERICA or R&D 128) and include parameters such as water depth, current velocities and data from such laboratory uptake experiments representative of a specific site of interest as means to predict and model the fate of radium in these systems and the distribution between radium in barite and aqueous radium in real plant and environmental scenarios.

iii) Effects of microbial induced anoxia on hard scales - Subject to receiving scales with data

A comprehensive assessment of the speciation and fate of radium in the forms which exist during marine discharges was successfully assessed via sediment microcosm experiments and sequential extractions in Chapter 5. To extend this work, experiments could be carried out which investigate the long term fate of hard scales (e.g. pipe scale) which enter the environment during operational discharges or burial of NORM under reducing conditions in the marine environment. Here, NORM of differing composition and types (e.g. sludge) can be assessed. This was not assessed in this research due to the insufficient levels of activity in the scale samples provided.

iv) Information on the operational conditions

Temperature, pressure, location data and water samples passing through the tubulars in which scale was extracted would provide a greater in-depth understanding of the formation process of the materials in Chapter 6. This would also help understand the variations in morphology and composition across the samples. Additionally, extra data such as discharge times and volumes would allow a deeper insight into the environmental implications of radium in produced water during marine discharges in Chapter 4. Such data could

258

be included in software models to provide a detailed assessment of fate of radium during discharges at various installations of interest.

259

List of References

Abbas, A. J. (2014) ‘Descaling of Petroleum Production Tubing utilising Aerated High Pressure Flat Fan Water Sprays’. Abdellah, W. M. and Al-Masri, M. S. (2014) ‘Treatment of norm contaminated soil from the oilfields’, Journal of Environmental Radioactivity. Elsevier Ltd, 129, pp. 63– 67. doi: 10.1016/j.jenvrad.2013.12.001. Abdul, J. and Zaidan, R. (2010) ‘Natural Occurring Radioactive Materials (NORM) in the oil and gas industry', Journal of Petroleum Researches and Studies, pp. 4-21. Aboud-Sayed, A. et al. (2000) ‘Guidelines for produced water injection’, OGP - International Association of Oil & Gas Producers, (2.80/302). Adler, H. H. and Kerr, P. F. (1965) ‘Variations in infrared spectra, molecular symmetry and site symmetry of sulfate minerals’, The American Mineralogist, 50, pp. 132–147. doi: 10.3389/fnhum.2015.00216. Aguado, J. L., Bolívar, J. P. and García-Tenorio, R. (2004) ‘Sequential extraction of226Ra in sediments from an estuary affected historically by anthropogenic inputs of natural radionuclides’, Journal of Environmental Radioactivity, 74(1–3), pp. 117– 126. doi: 10.1016/j.jenvrad.2004.01.016. Al Attar, L., Safia, B. and Abdul Ghani, B. (2016) ‘Chemical composition of scales generated from oil industry and correlation to radionuclide contents and gamma- ray measurements of 210Pb’, Environmental Pollution, 210, pp. 246–252. doi: 10.1016/j.envpol.2015.12.007. Al-Masri, M. S. and Aba, A. (2005) ‘Distribution of scales containing NORM in different oilfields equipment’, Applied Radiation and Isotopes. United Kingdom, 63(4), pp. 457–463. doi: DOI:101016/japradiso200505050. Al-Masri, M. S. and Blackburn, R. (1996) ‘Radioanalytical methods for determination of alpha emitters in the environment’, Radiation Physics and Chemistry, 47(2), pp. 171–175. doi: 10.1016/0969-806X(95)00030-2.

Al-Masri, M. S. and Suman, H. (2003) ‘NORM waste management in the oil and gas industry: The Syrian experience’, Journal of Radioanalytical and Nuclear Chemistry, 256(1), pp. 159–162. doi: 10.1023/A:1023380932621. Al-Marsi, M.S., Ali, A. I., Khietou, M. and Al-Hares, Z. (1998) 'Leaching of Ra-226 from Syrian phosphogypsum', Environmental Radiochemcial Analysis, Department of Protextion and Saftey, Atomic Enegry Commission of Syria, Damascus

Ames, L. L., McGarrah, J. E., Walker, B. A. and Salter, P. F. (1983) ‘Uranium and radium sorption on amorphous ferric oxyhydroxide’, Chemical Geology, 40(1), pp. 135–148. doi: https://doi.org/10.1016/0009-2541(83)90095-5. American Petroleum Institue (API). (1992). 'Bulletin on management of naturally occurring radioactive materials (NORM) in oil & gas production'. 1st

260 edition. Bulletin E2.

Amjad, Z. and Koutsoukos, P. G. (2014) ‘Evaluation of maleic acid based polymers as scale inhibitors and dispersants for industrial water applications’, Desalination, 335(1), pp. 55–63. doi: https://doi.org/10.1016/j.desal.2013.12.012. Amosa, M., Mohammed-Dabo, I. and Yaro, S. (2013) ‘Environmental and Corrosion Control Effects of Hydrogen Sulfide Scavengers in Oil and Gas Exploration’, Corrosion and Materials in the Oil and Gas Industries, (July 2015), pp. 453–488. doi: 10.1201/b14703-22. Amyx, J. W., Bass, D. M. and Whiting, R. L. (1960) Petroleum Reservoir Engineering. Physical properties. New York: McGraw-Hill Book Co. Antony, A., Low, J. H., Gray, S., Childress, A. E., Le-Clech, P. and Leslie, G. (2011) ‘Scale formation and control in high pressure membrane water treatment systems : A review’, Journal of Membrane Science. Elsevier B.V., 383(1–2), pp. 1–16. doi: 10.1016/j.memsci.2011.08.054. Appelo, C.A. J. and Postma, D. (2005) ‘Geochemistry, Groundwater and Pollution’, 2nd Edition, Balkema, Rotterdam pp. 1-649. http://dx.doi.org/10.1201/9781439833 544 Arthur, J. D., Langhus, B. G. and Ph, D. (2005) ‘Water Treatment Technologies’, (March 2005), pp. 1–53. Aroke, U. O., Abdulkarim, A. and Ogubunka, R. O. (2013). 'Fourier-transform Infrared Characterization of Kaolin, Granite, Bentonite and Barite'. ATBU J. Environ. Technol. 6, 42 53.

Atkins, P. and De Paula, J. (1989) ‘Atkins Physical Chemistry’, Journal of Chemical Information and Modeling, pp. 798–825. doi: 10.1017/CBO9781107415324.004.

Averyt, K. B., Paytan, A. and Li, G. (2003) ‘A precise, high-throughput method for determining Sr/Ca, Sr/Ba, and Ca/Ba ratios in marine barite’, Geochemistry, Geophysics, Geosystems, 4(4), pp. 1–10. doi: 10.1029/2002GC000467. Bader, M. S. H. (2006) ‘Sulfate scale problems in oil fields water injection operations’, Desalination, 201(1–3), pp. 100–105. doi: 10.1016/j.desal.2005.09.039. Badr Merdhah, A., Azam, A. and Yassin, M. (2008) ‘Study of Scale Formation Due To Incompatible Water’, Jurnal Teknologi, 49, pp. 9–26. Bahadori, A. (2018) ‘Enhanced Oil Recovery’, Fundamentals of Enhanced Oil and Gas Recovery from Conventional and Unconventional Reservoirs. Elsevier. Bakke, T., Klungsøyr, J. and Sanni, S. (2013) ‘Environmental impacts of produced water and drilling waste discharges from the Norwegian offshore petroleum industry’, Marine Environmental Research, 92, pp. 154–169. doi: https://doi.org/10.1016/j.marenvres.2013.09.012. Baldi, F., Pepi, M., Burrini, D., Kniewald, G., Scali, D and Lanciotti, E.. (1996) ‘Dissolution of barium from barite in sewage sludges and cultures of Desulfovibrio

261 desulfuricans’, Applied and Environmental Microbiology, 62(7), pp. 2398–2404. Ball, J. W. (1991) ‘WATEQ4F Database, The United States Geological Survey (USGS), Open-File Report 91–183 (2001).’, U.S. Geological Survey, Open-File. doi: 10.1107/S 1600536811054717. Balson, P., Butcher, A., Holmes, R., Johnson, H., Lewis, M. and Musson, R. (2002) ‘North Sea Geology : British Geological Survey (BGS)’, Techincal report (August), pp - 1-49. Beck, A.J., Cochran, M.A., (2013) 'Controls on solid-solution partitioning of radium in saturated marine sands'. Mar. Chem. 156, 38–48. https://doi.org/https://doi.org/10.1016/j.marchem.2013.01.008

Becker, U., Biswas, S., Kendall, T., Risthus, P., Putnis, C. V. and Pina, C. M. (2005) ‘Interactions between mineral surfaces and dissolved species: From monovalent ions to complex organic molecules’, American Journal of Science, 305(6-8 SPEC. ISS.), pp. 791–825. doi: 10.2475/ajs.305.6-8.791. Beneš, P., Sedlacek, J., Sebesta, F., Sandrik, R. and John, J. (1981) ‘Method of selective dissolution for characterization of particulate forms of Radium and Barium in natural and waste waters’, Water Research, 15(12), pp. 1299–1304. doi: https://doi.org/10.1016/0043-1354(81)90002-6. Beneš, P. and Strejc, P. (1986) ‘Interaction of radium with freshwater sediments and their mineral components’, Journal of Radioanalytical and Nuclear Chemistry, 99(2), pp. 407–422. doi: 10.1007/BF02037602. Beneš, P., Strejc, P. and Lukavec, Z. (1984) ‘Interaction of radium with freshwater sediments and their mineral components. I.’, Journal of Radioanalytical and Nuclear Chemistry, 82(2), pp. 275–285. doi: 10.1007/BF02037050. Berner, R. A. (1970) ‘Sedimentary pyrite formation’, American Journal of Science, 268(1), pp. 1–23. doi: 10.2475/ajs.268.1.1. Betti, M. et al. (2004) ‘Results of the European Commission MARINA II study: part I—general information and effects of discharges by the nuclear industry’, Journal of Environmental Radioactivity, 74(1), pp. 243–254. doi: https://doi.org/10.1016/j.jenv rad.2004.01.022. Bezemer, C. and Bauer, K. A. (1969) ‘Prevention of Carbonate Scale Deposition: A Well-Packing Technique with Controlled Solubility Phosphates’, Journal of Petroleum Technology. Society of Petroleum Engineers, 21(04), pp. 505–514. doi: 10.2118/2176-PA. Birren, B. W., DeSantis, T. and Petrosino, J. (2011) ‘Chimeric 16S rRNA sequence formation and detection in Sanger and 454-pyrosequenced PCR amplicons’, Genome Research, 21(3), pp. 494–504. doi: 10.1101/gr.112730.110.

Bloch, S. and Key, R. M. (1981) ‘Modes of formation of anomalously high radioactivity in oil-field brines’, American Association of Petroleum Geologists Bulletin. United States, 65(1), pp. 154–159.

262

Bokern, D. G., Hunter, K. A. and McGrath, K. M. (2003) ‘Charged Barite−Aqueous Solution Interface: Surface Potential and Atomically Resolved Visualization’, Langmuir. American Chemical Society, 19(24), pp. 10019–10027. doi: 10.1021/la0269255. Bolze, C. E., Malone, P. G. and Smith, M. J. (1974) ‘Microbial mobilization of barite’, Chemical Geology, 13(2), pp. 141–143. doi: https://doi.org/10.1016/0009- 2541(74)90006-0. Boss, C. B. and Fredeen, K. J. (2004) ‘Concepts, Instrumentation and Techniques in Inductively Coupled Plasma Optical Emission Spectrometry Charles’, p. 120.

BP (2019) ‘BP Energy Outlook 2019 edition The Energy Outlook explores the forces shaping the global energy transition out to 2040 and the key uncertainties surrounding that’, BP Energy Outlook 2019. Brendehaug, J. et al. (1992) ‘Toxicity Testing and Chemical Characterization of Produced Water --- A Preliminary Study’, in Ray, J. P. and Engelhardt, F. R. (eds) Produced Water: Technological/Environmental Issues and Solutions. Boston, MA: Springer US, pp. 245–256. doi: 10.1007/978-1-4615-2902-6_20. Brown, P.L., Ekberg, C., Ramebäck, H., Hedström, H. and Matyskin, A., (2015) 'Solubility of Radium and Strontium Sulfate across the Temperature Range of 0 to 300°C'. Uranium - Past and Future Challenges. Springer International Publishing, Cham, pp. 553–564. doi: 10.1007/978-3-319-110592_63

Brunauer, S., Emmett, P. H. and Teller, E. (1938) ‘Adsorption of Gases in Multimolecular Layers’, Journal of the American Chemical Society. American Chemical Society, 60(2), pp. 309–319. doi: 10.1021/ja01269a023.

Buhmann, D. and Dreybrodt, W. (1987) ‘Calcite dissolution kinetics in the system H2O-CO2-CaCO3 with participation of foreign ions’, Chemical Geology, 64(1), pp. 89– 102. doi: https://doi.org/10.1016/0009-2541(87)90154-9. Burgos, W. D., Castillo-Meza, L., Tasker, T. L., Geeza, T. J., Drohan, P. J., Lui, X., Landis, J. D., Blotevogel, J., McLaughlin, M., Borch, T. and Warner, N. R. (2017) ‘Watershed-Scale Impacts from Surface Water Disposal of Oil and Gas Wastewater in Western Pennsylvania’, Environmental Science and Technology, 51(15), pp. 8851– 8860. doi: 10.1021/acs.est.7b01696. Burke, I. T., Boothman, C., Lloyd, J. R., Mortimer, R. J. G., Livens, F. R. and Morris, K. (2005) ‘Effects of Progressive Anoxia on the Solubility of Technetium in Sediments’, Environmental Science & Technology. American Chemical Society, 39(11), pp. 4109– 4116. doi: 10.1021/es048124p. Burnett, W. C. and Schultz, M. K. (1996) ‘Radionuclide flow during the conversion of phosphogypsum to ammonium sulfate’, Journal of Environmental Radioactivity. Burrows, A. (2013) Chemistry3 : introducing inorganic, organic and physical chemistry. 2nd ed. Edited by Andrew. Burrows. Oxford: Oxford University Press. Candeias, J. P., de Oliveira, D.F., dos Anjos, M.J. and Lopes R.T. (2014) ‘Scale analysis

263 using X-ray microfluorescence and computed radiography’, Radiation Physics and Chemistry, 95, pp. 408–411. doi: https://doi.org/10.1016/j.radphyschem.2013.03. 007. Caporaso, J. G., Kuczynski, J. and Stombaugh, J. (2010) ‘QIIME allows analysis of high-throughput community sequencing data’, Nature methods. 2010/04/11, 7(5), pp. 335–336. doi: 10.1038/nmeth.f.303.

Caporaso, J. G., Lauber, C. and Walters, W. (2011) ‘Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample’, Proceedings of the National Academy of Sciences, 108(Supplement 1), pp. 4516 LP – 4522. doi: 10.1073/pnas.1000080107.

Caporaso, J. G., lauber, C. and Walters W. (2012) ‘Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms’, The ISME journal. 2012/03/08. Nature Publishing Group, 6(8), pp. 1621–1624. doi: 10.1038/ismej.2012.8.

Carbonell, A. A., Pulido, R., DeLaune, R. D. and Patrick Jr, W. H. (1999) ‘Soluble barium in barite and phosphogypsum amended Mississippi River alluvial sediment’, Journal of Environmental Quality, 28(1), pp. 316–321. doi: 10.2134/jeq1999.00472425002800010039x. Carvalho, F., Chambers, D., Fesenko, S., Moore, W. S. and Porcelli, D. (2014) ‘Environmental pathways and corresponding models’, IAEA - International Atomic Energy Agency (Technical reports), pp. 106-172 BT-The Environmental Behaviour of Radium. Ceccarello, S., Black, S., Read, D. and Hodson, M.E. (2004) ‘Industrial radioactive barite scale: Suppression of radium uptake by introduction of competing ions’, Minerals Engineering, 17(2), pp. 323–330. doi: 10.1016/j.mineng.2003.11.007. Chang, L. L. Y., Howie, R. A. and Zussman J. D. (1996) ‘Rock Forming Minerals. Volume 5B: Non-silicates; sulphates, carbonates., phosphates and halides’, Geology, pp. 3–29. doi: 10.1201/9781315275376-5.ISBN 0-582-30093-2 Chapman, M. (2013) ‘Naturally Occurring Contaminants in the Piedmont and Blue Ridge Crystalline-Rock Aquifers and Piedmont Early Mesozoic Basin Siliciclastic-Rock Aquifers , Eastern United States , 1994 – 2008 Scientific Investigations Report 2013 – 5072’, pp. 1994–2008. Choppin, G. R. (2013) Radiochemistry and nuclear chemistry. 4th edition. Edited by J.-O. Liljenzin, J. Rydberg, and C. Ekberg. Amsterdam: Academic Press. Chowdhury, S. et al. (2004) ‘Human health risk assessment of naturally occurring radioactive materials in produced water - A case study’, Human and Ecological Risk Assessment, 10(6), pp. 1155–1171. doi: 10.1080/10807030490887203. Clark, R. W. (2000) 'Solubilty and Solubilty Products'. J. Chem. Educ. 1998, 75, pp. 1179-1881 Cline, J. D. (1969) ‘Spectrophotometric Determination of Hydrogen Sulfide in

264

Natural Waters’, pp. 454–458. doi: 10.4319/lo.1969.14.3.0454. Coelho, A. (2009) ‘TOPAS 4.2 User Manual’, Bruker AXS GmbH. doi: 10.1038/scientificamerican0610-4.

Collins, A. G. (1975) ‘Geochemistry of oil-field waters. Devel’, Petroleum Sc, (1). Cowan Jack C., W. D. J. (1976) Water Formed Scale Deposits, Water-Formed Scale Deposits. Knovel. Crabtree, M. and Johnson, A. (no date) ‘Fighting Scale — Removal and Prevention’, pp. 30–45. Cullity, B. D. (2013) Elements of X-Ray Diffraction. 3rd ed. Edited by S. R. Stock and S. R. Stock. Harlow: Pearson Education UK.

Curti, E. (1999) ‘Coprecipitation of radionuclides with calcite: estimation of partition coefficients based on a review of laboratory investigations and geochemical data’, Applied Geochemistry, 14(4), pp. 433–445. doi: https://doi.org/10.1016/S0883- 2927(98)00065-1. Dana, J.D., Klein, C. and Hurlbut, C. S. (1985) 'Manual of Mineralogy'. 20th ed. New York; Wiley.

Danilatos, G. D. (1993) ‘Introduction to the ESEM instrument’, Microscopy Research and Technique, 25(5–6), pp. 354–361. doi: 10.1002/jemt.1070250503.

Davies Michael, S. P. J. B. (2006) ‘Sulfate-Reducing Bacteria (SRB) and Sulfide- Producing Bacteria (SPB)’, Oilfield Water Technology. NACE International. De Mello Donegá, C. (2014) 'Nanoparticles: Workhorses of nanoscience', pp. 191- 255. doi: 10.1007 /978-3-662-44823-6.

Denecke, M. A. (2006) ‘Actinide speciation using X-ray absorption fine structure spectroscopy’, Coordination Chemistry Reviews, 250(7–8), pp. 730–754. doi: 10.1016/j.ccr.2005.09.004.

Dent, A. J. et al. (2009) ‘B18: A core XAS spectroscopy beamline for Diamond’, Journal of Physics: Conference Series. IOP Publishing, 190, p. 12039. doi: 10.1088/1742-6596/190/1/012039.

Doerner, H. A. and Hoskins, W. M. (1925) ‘Co-precipitation of radium and barium sulphates’, Journal of the American Chemical Society. American Chemical Society, 47(3), pp. 662–675. doi: 10.1021/ja01680a010. Doubra, P. et al. (2017) ‘Thermodynamic modelling of scale (Calcite, Barite, Anhydrite and Gypsum) deposition from brine’, Journal of Molecular Liquids, 230, pp. 96–103. doi: https://doi.org/10.1016/j.molliq.2016.11.135. Dowdall, M. and Lepland, A. (2012) ‘Elevated levels of radium-226 and radium-228 in marine sediments of the Norwegian Trench (“Norskrenna”) and Skagerrak’, Marine Pollution Bulletin, 64(10), pp. 2069–2076. doi: https://doi.org/10.1016/j.marpolbul.2012.07.022.

265

Doyi, I., Essumang, D.K., Dampare, S. and Glover, E.T. (2016) ‘Technologically Enhanced Naturally Occurring Radioactive Materials (TENORM) in the Oil and Gas Industry: A Review’, Reviews of environmental contamination and toxicology. National Radioactive Waste Management Centre, National Nuclear Research Institute, Ghana Atomic Energy Commission , 238, pp. 107–119. doi: 10.1007/398_ 2015_5005. Dunn, K. and Yen, T. F. (1999) ‘Dissolution of barium sulfate scale deposits by chelating agents’, Environmental Science and Technology, 33(16), pp. 2821–2824. doi: 10.1021/es980968j. Eberth, J. and Simpson, J. (2008) ‘From Ge(Li) detectors to gamma-ray tracking arrays–50 years of gamma spectroscopy with germanium detectors’, Progress in Particle and Nuclear Physics, 60(2), pp. 283–337. doi: https://doi.org/10.1016/j. ppnp.2007.09.001.

Edgar, R. C. (2010) ‘Search and clustering orders of magnitude faster than BLAST’, Bioinformatics, 26(19), pp. 2460–2461. doi: 10.1093/bioinformatics/btq461.

Edgar, R. C. (2013) ‘UPARSE: highly accurate OTU sequences from microbial amplicon reads’, Nature Methods. Nature Publishing Group, a division of Macmillan Publishers Limited. All Rights Reserved., 10, p. 996. Available at: https://doi.org/ 10.1038/nmeth.2604.

Egerton, R. F. (2005) 'Physical Principles of Electron Microscopy, Physical Principles of Electron Microscopy'.

Ehmann, W. D., Robertson, J. D. and Yates, S. W. (1994) ‘Nuclear and radiochemical analysis’, Analytical chemistry, 66(12), p. 229R. doi: 10.1021/ac00084a011.

El-Hattab, M.I. (1985) 'Scale Deposition in Surface and Subsurface Production Equipment in the Gulf of Suez'. J. Pet. Technol. 37, 1640–1652. https://doi.org/10.2118/11449-PA

Environment Agency, Food Standards Agency, Food Standards Scotlands, Natural Resources Wales, Northern Ireland Environment Agency and Scottish Environment Protection Agency. (2015) ‘Radioactivity in Food and the Environment, 2015’, RIFE 21. EPA (1995) ‘Envrionmental Protection Agency, AP 45, Fifth Edition, Volume 1, Chapter 5: The petroleum industry’, (3), pp. 1–7. Eriksen, D., Sidhu, R., Stralberg, E., Iden, K. I., Hylland, K., Ruus, A., Royset, O., Royset, M. H. G. and Rye, H. (2006) ‘Radionuclides in produced water from Norwegian oil and gas installations - Concentrations and bioavailability’, Czechoslovak Journal of Physics, 56(SUPPL. 4), pp. 43–48. doi: 10.1007/s10582-006- 0486-7. EuropeanCommision (2014) ‘European Commission, Council Directive 2013/59/Euratom of 5 December 2013 laying down basic safety standards for protection against the dangers arising from exposure to ionising radiation, and repealing Directives 89/618/Euratom, 90/641/Euratom, 96/29/Eur’, (December

266

1990), pp. 1–73. Fakhru’l-Razi, A., Pendashteh, A., Abdullah, L. Q., Biak, R. A. W., Madeaeni, S. S. and Abidin, S. S. (2009) ‘Review of technologies for oil and gas produced water treatment’, Journal of hazardous materials. Department of Chemical and Environmental Engineering, Faculty of Engineering, Universiti Putra Malaysia, 170(2–3), pp. 530–551. doi: 10.1016/j.jhazmat.2009.05.044. Fedorak, P. M., Westlake, D. W., Anders, C., Kratochvil, B., Motkosky, N., Anderson, 226 W. B. and Huck, P. W. (1986) ‘Microbial release of Ra from (Ba,Ra)SO4 sludges from uranium mine wastes’, Applied and environmental microbiology, 52(2), pp. 262–268. Figgins, A., Lansdell, J. and Taqi, Y. (2018) ‘Oil and gas regulation in Oman : overview’, 1725(December 2016), pp. 1–21. Fisher, R. S. (1998) ‘Geologic and geochemical controls on naturally occurring radioactive materials (NORM) in produced water from oil, gas, and geothermal operations’, Environmental Geosciences, 5(3), pp. 139–150. Fried, R. and Mastai, Y. (2012) ‘The effect of sulfated polysaccharides on the crystallization of calcite superstructures’, Journal of Crystal Growth, 338(1), pp. 147–151. doi: https://doi.org/10.1016/j.jcrysgro.2011.09.044. Folk, R. L. (2005) ‘Nannobacteria and the formation of framboidal pyrite: Textural evidence’, Journal of Earth System Science, 114(3), pp. 369–374. doi: 10.1007/BF02702955. Ferguson, R. J., Freedman, G. F., Kulik, A. J., Robson, J. and Weintritt, D. J. (1995) 'The practical application of ion association model level indices to commerical water treatment problem solving'. ISBN 978-1-4899-1400-2 Gafvert, T., Svaeren, I., Gywnn, J., Brungot, A. L., Kolstad, A. K., Lind, B., Alvestad, P., Heldal, H. E., Stralberg, E., Christensen, G. C., Drefvelin, J., Dowdall, M. and Rudjord, A. L. (2007) ‘Radioactivity in the Marine Environment 2005’, Results from the Norwegian Marine Monitoring Programme (RAME). StralevernRapport 2007:10. Ostera Norwegian Radiation Protection Authority, 2007, pp. 42. ISBN 0804-4910. Garner, J., Cairns, J. and Read, D. (2015) ‘NORM in the East Midlands’ oil and gas producing region of the UK’, Journal of Environmental Radioactivity, 150, pp. 49–56. doi: https://doi.org/10.1016/j.jenvrad.2015.07.016. Gill, R. (ed.) (2014) 'Modern analytical geochemistry : an introduction to quantitative chemical analysis techniques for earth, environmental and materials scientists'. London: Routledge (Geochemistry Series).

Gilmore, G. (2008) 'Practical gamma-ray spectrometry'. 2nd ed. Chichester, England ; Wiley. doi: 10.1002/9780470861981.

Godoy, José Marcus et al. (2005) ‘210Pb content in natural gas pipeline residues (“black-powder”) and its correlation with the chemical composition’, Journal of Environmental Radioactivity, 83(1), pp. 101–111. doi: https://doi.org/10.1016/j.jenvrad.2005.02.018.

267

Gonneea, M. E. et al. (2008) ‘New perspectives on radium behavior within a subterranean estuary’, Marine Chemistry, 109(3), pp. 250–267. doi: https://doi.org/10.1016/j.marchem.2007.12.002. Gonneea, M. E. and Paytan, A. (2006) ‘Phase associations of barium in marine sediments’, Marine Chemistry, 100(1–2), pp. 124–135. doi: 10.1016/j.marchem.2005.12.003. Gordon, L. and Rowley, K. (1957) ‘Coprecipitation of Radium with Barium Sulfate’, Analytical Chemistry. American Chemical Society, 29(1), pp. 34–37. doi: 10.1021/ac60121a012. Goswami, R. (2019) ‘Oil and gas regulation in Saudi Arabia : overview’, pp. 1–20. Gregory, K. B., Vidic, R. D. and Dzombak, D. A. (2011) ‘Associated with the Production of Shale Gas by Hydraulic Fracturing Shale gas development’, 2030. doi: 10.2113/gselements.7.3.181. Grung, M., Ruus, A., Holth, T. F., Sidhu, R. S., Eriksen, D. O. and Hylland, K. (2009) ‘Bioaccumulation and lack of oxidative stress response in the ragworm H. diversicolor following exposure to 226Ra in sediment’, Journal of Environmental Radioactivity, 100(5), pp. 429–434. doi: https://doi.org/10.1016/j.jenvrad.2009 .03.001. Gunasekaran, S., Anbalagan, G. and Pandi, S. (2006) 'Raman and infrared spectra of carbonates of calcite structure'. J. Raman Spectrosc. 37, 892–899. https://doi.org/10.1002/jrs.1518

Haas, A. (2008) Carbonate authigenesis and worm tube mineralization- biogeochemical and geobiological processes at methane seeps on the Congo deep- sea fan. Thesis (PhD). Hagemann, H., Snyder, R.G., Peacock, A.J. and Mandelkern, L. (1989) 'Quantitative infrared methods for the measurement of crystallinity and its temperature dependence: polyethylene'. Macromolecules. 22, 3600–3606. https://doi.org/10.1021/ma00199a017

Haluszczak, L. O., Rose, A. W. and Kump, L. R. (2013) ‘Geochemical evaluation of flowback brine from Marcellus gas wells in Pennsylvania, USA’, Applied Geochemistry, 28, pp. 55–61. doi: https://doi.org/10.1016/j.apgeochem. 2012.10.002. Hamilton, I. S., Arno, M.G., Rock, J.C., Berry, R.O., Poston, J.W., Cezeauz, J.R. and Park, J.-M. (2004) ‘Radiological assessment of petroleum pipescale from pipe- rattling operations’, Health physics : the radiation safety journal. Baltimore, MD : Lippincott Williams & Wilkins, pp. 382–397. doi: 10.1097/01.HP.0000128580.09991.05. Hamlat, M. S., Djeffal, S. and Kadi, H. (2001) ‘Assessment of radiation exposures from naturally occurring radioactive materials in the oil and gas industry’, Applied Radiation and Isotopes, 55(1), pp. 141–146. doi: https://doi.org/10.1016/S0969- 8043(01)00042-2.

268

Hanan, M. A. (1981), 'Geochemistry and Mobility in Sediments of Radium from Oil- field Brines, Grand Bay, Plaquemines Parish, Louisiana'. M.S thesis, University of New Orleans. Biological and Chemical Oceanography Branch, Oceanography Division, Naval Ocean Research and Development Activity. Harker, J. H., Backhurst, J. R., Coulson, J. M., and Richardson, J. F. 'Coulson & Richardson Chemical Engineering' 5th Edition, Volume 2, pp. 827 - 844 Harkness, J. S., Dwyer, G. S., Warner, N. R., Parker, K. M., Mitch, W. A. and Vengosh, A. (2015) ‘Iodide, Bromide, and Ammonium in Hydraulic Fracturing and Oil and Gas Wastewaters: Environmental Implications’, Environmental Science & Technology. American Chemical Society, 49(3), pp. 1955–1963. doi: 10.1021/es504654n. Hartog, F. A. Jonkers, G., Schmidt, A.P., Schuiling, R.D. (2002) ‘Lead Deposits in Dutch Natural Gas Systems’, SPE Production & Facilities. Society of Petroleum Engineers, 17(02), pp. 122–128. doi: 10.2118/78147-PA. Harvey, D. (2000) ‘Modern Analytical Chemistry’. 1st ed. London; McGraw-Hill Heaton, B and Lambley, J. (1995) ‘TENORM in the oil, gas and mineral mining industry’, Applied Radiation and Isotopes, 46(6), pp. 577–581. doi: https://doi.org/10.1016/0969-8043(95)00094-1. Heaton, B. and Lambley, J. (1995) ‘TENORM in the oil, gas and mineral mining industry’, Applied Radiation and Isotopes, 46(6–7), pp. 577–581. doi: 10.1016/0969- 8043(95)00094-1. Hedström, H., Persson, I., Skarnemark, G. and Ekberg, C. (2013) ‘Characterization of Radium Sulphate’, Journal of Nuclear Chemistry, 2013, pp. 1–4. doi: 10.1155/2013/940701. Hedvall, R. and Erlandsson, B. (1996) ‘Radioactivity concentrations in non-nuclear industries’, Journal of Environmental Radioactivity, 32(1), pp. 19–31. doi: https://doi.org/10.1016/0265-931X(95)00077-N. Her Majesty’s Stationery Office. London (2011). The Environmental Permitting (England and Wales ) Regulations. Hladik, M. L., Focazio, M. J. and Engle, M. (2014) ‘Discharges of produced waters from oil and gas extraction via wastewater treatment plants are sources of disinfection by-products to receiving streams’, Science of The Total Environment, 466–467, pp. 1085–1093. doi: https://doi.org/10.1016/j.scitotenv.2013.08.008. Holdway, D. A. (2002) ‘The acute and chronic effects of wastes associated with offshore oil and gas production on temperate and tropical marine ecological processes’, Marine Pollution Bulletin, 44(3), pp. 185–203. doi: https://doi.org/10.1016/S0025-326X(01)00197-7. Horrocks, D. (1974) ‘Applications of liquid scintillation counting’, Information Storage and Retrieval, p. 357.

Hosseini, A., Beresford, N. A., Brown, J. E., Jones, D. G., Phaneauf, M., Thorring, H. and Yankovich, T. (2010) ‘Background dose-rates to reference animals and plants

269 arising from exposure to naturally occurring radionuclides in aquatic environments’, Journal of Radiological Protection, 30(2), pp. 235–264. doi: 10.1088/0952- 4746/30/2/S03. Hosseini, A., Brown, J. E., Gwynn, J. P. and Dowdall, M. (2012) ‘Review of research on impacts to biota of discharges of naturally occurring radionuclides in produced water to the marine environment’, Science of the Total Environment. Elsevier B.V., 438, pp. 325–333. doi: 10.1016/j.scitotenv.2012.08.047. Housecroft, C. E. (2012) Inorganic chemistry . 4th ed. Edited by A. G. Sharpe. Harlow, England: Pearson. Houston, S. J. (2007) ‘Formation waters in petroleum reservoirs : Their controls and applications.’. Hu, Y. (2016) ‘Analysis of produced water data to model and identify reactions occuring in the reservoir', Thesis, Institute of Petroleum Engineering, School of Energy, Geoscience, Infrastructure and society, Herriot Watt University, Ediburgh, UK. Huang, C.K. and Kerr, P.F. (1960) 'Infrared study of the carbonate minerals. Am. Mineral. 45, 311 324.

Huh, C. A. and Bacon, M. P. (1985) ‘Determination of Thorium Concentration in Seawater by Neutron Activation Analysis’, Analytical Chemistry, 57(11), pp. 2138– 2142. doi: 10.1021/ac00288a030. Hunt, J. M. (1979) 'Petroleum Geochemistry and Geology - A series of books in geology'. San Francisco: W. H. Freeman and Company. pp. 617. ISBN 0716710056 Hylland, K. (2003) ‘Bioaccumulation in marine organisms—effect of contaminants from oil well produced water: Jerry M. Neff; Elsevier Science, Amsterdam; ISBN (hardcover) 0-080-43716-8. Eu 160, US$160; 452 pp.’, Journal of Experimental Marine Biology and Ecology, 287(1), pp. 135–137. doi: https://doi.org/10.1016/S0022-0981(02)00598-1. IAEA (2004) ‘Radiation protection and the management of radioactive waste in the oil and gas industry’, Environmental health perspectives, 34(34), pp. 130. IAEA (2013) Management of NORM Residues - IAEA-TECDOC-1712. International Atomic Energy Agency, 2014. Radiation Protection and Safety of Radiation Sources: International Basic Safety Standards (GSR Part 3). Int. At. Energy Agency Vienna 3, 471. https://doi.org/STI/PUB/1578

ICRP (2007) ‘ICRP Publication 103 The 2007 Recommendations of the International Commission on Radiological Protection’. Idelchik, I. E. (1998) ‘Handbook of Hydraulic Resistance, 3rd Edn. By I. E. Idelchik. Begell House, 1994. 790 pp. ISBN 0-8493-9908-4. 138.95.’, Journal of Fluid Mechanics, pp. 376–378. Igunnu, E. T. and Chen, G. Z. (2014) ‘Produced water treatment technologies’, (July 2012), pp. 157–177. doi: 10.1093/ijlct/cts049.

270

International Atomic Energy Agency (2014) ‘Radiation Protection and Safety of Radiation Sources: International Basic Safety Standards (GSR Part 3)’, International Atomic Energy Agency Vienna, 3, p. 471. doi: STI/PUB/1578. IOGP (2016) ‘Managing Naturally Occurring Radioactive Material ( NORM ) in the Oil and Gas Industry’, International Association of Oil & Gas Producers, 412(March), p. 68. Ismail, B., Teng, I. L. and Muhammad Samudi, Y. (2011) ‘Relative radiological risks derived from different tenorm wastes in Malaysia’, Radiation Protection Dosimetry, 147(4), pp. 600–607. doi: 10.1093/rpd/ncq577. Jacobs, R. P. W. M., Grant, R. O. H., Kwant, J., Marquenie, J. M. and Mentzer, E. (1992) ‘The Composition of Produced Water from Shell Operated Oil and Gas Production in the North Sea - Produced Water: Technological/Environmental Issues and Solutions’, in Ray, J. P. and Engelhardt, F. R. (eds). Boston, MA: Springer US, pp. 13–21. doi: 10.1007/978-1-4615-2902-6_2. Jenkins, R. (1999) X-ray fluorescence spectrometry . 2nd ed. New York: Wiley (Chemical analysis ; v. 152).

Jerez Vegueria, S. F., Godoy, J. M. and Miekeley, N. (2002) ‘Environmental impact studies of barium and radium discharges by produced waters from the “Bacia de Campos” oil-field offshore platforms, Brazil’, Journal of Environmental Radioactivity, 62(1), pp. 29–38. doi: https://doi.org/10.1016/S0265-931X(01)00148-5. Jiménez, S., Mico, M. M., Arnaldos, M., Medina, F. and Contreras, S. (2018) ‘State of the art of produced water treatment’, Chemosphere, 192, pp. 186–208. doi: https://doi.org/10.1016/j.chemosphere.2017.10.139. Johnston, R. F., Pickett, S. C. and Barker, D. L. (1990) ‘Autoradiography using storage phosphor technology’, Electrophoresis, 11(5), pp. 355–360. doi: 10.1002/elps.1150110503.

Jones, M. J., Butchins, L. J., Charnock, J. M., Pattrick, R. A. D., Small, J. S., Vaughan, D. J., Wincott, P. L. and Livens, F. R. (2011) ‘Reactions of radium and barium with the surfaces of carbonate minerals’, Applied Geochemistry, 26(7), pp. 1231–1238. doi:https://doi.org/10.1016/j.apgeochem.2011.04.012. Jonkers, G., Harog, F. A., Knaepen, W. A,. J. and Lancee, P. F. J. (1997) ‘Characterization of NORM in the oil and gas production (E&P) industry’, Proceedings of the International Symposium on Radiological Problems with Natural Radioactivity in the Non-Nuclear Industry, pp. 23–47. Keating, G. E., McCartney, M. and Davidson, C. M. (1996) ‘Investigation of the technological enhancement of natural decay series radionuclides by the manufacture of phosphates on the Cumbrian coast’, Journal of Environmental Radioactivity, 32(1), pp. 53–66. doi: https://doi.org/10.1016/0265-931X(95)00079- P. Keith-Roach, M. J. (2002) 'Interactions of microorganisms with radionuclides' in M. J. Keith-Roach and F. R. Livens. (eds). Amsterdam, Elsevier (Radioactivity in the

271 environment), pp. 61-100., v. 2. ISBN 9780080534909 Kelland, M. A. (2014) Production Chemicals for the Oil and Gas Industry, Second Edition. 2nd edn. Baton Rouge: CRC Press. Khan, T. S., Alshehhi, M., Stephen, S. and Khezzar, L. (2015) ‘Characterization and preliminary root cause identification of black powder content in a gas transmission network – A case study’, Journal of Natural Gas Science and Engineering, 27, pp. 769–775. doi: https://doi.org/10.1016/j.jngse.2015.09.022. Kondash, A. J., Warner, N. R., Lahav, O. and Vengosh, A. (2014) ‘Radium and Barium Removal through Blending Hydraulic Fracturing Fluids with Acid Mine Drainage’, Environmental Science and Technology, 48(2), pp - 1334-1342. doi: 10.1021/es403852h. Konhauser, K. O., Mortimer, R. J. G., Morris, K. and Dunn, V. (2002) 'The role of microorganisms during sediment diagensis: Implications for radionuclide mobility, Interactions of Microorganims with Radionuclides'. Kowacz, M. and Putnis, A. (2008) ‘The effect of specific background electrolytes on water structure and solute hydration: Consequences for crystal dissolution and growth’, Geochimica et Cosmochimica Acta, 72(18), pp. 4476–4487. doi: https://doi.org/10.1016/j.gca.2008.07.005. Kowacz, M., Putnis, C. V. and Putnis, A. (2009) ‘The control of solution composition on ligand-promoted dissolution: DTPA-Barite interactions’, Crystal Growth and Design, 9(12), pp. 5266–5272. doi: 10.1021/cg9007894. Kozich, J. J., Westcott, S. and baxter, N. (2013) ‘Development of a Dual-Index Sequencing Strategy and Curation Pipeline for Analyzing Amplicon Sequence Data on the MiSeq Illumina Sequencing Platform’, Applied and Environmental Microbiology, 79(17), pp. 5112 LP – 5120. doi: 10.1128/AEM.01043-13.

Krauskopf, K. B. (Konrad B. (1995) Introduction to geochemistry. 3rd ed / K. Edited by D. K. Bird. New York ; McGraw-Hill (Earth science and geology series). Laing, N., Graham, G. M. and Dyer, S. J. (2003) ‘Barium Sulphate Inhibition in Subsea Systems - The Impact of Cold Seabed Temperatures on the Performance of Generically Different Scale Inhibitor Species’, International Symposium on Oilfield Chemistry. Houston, Texas: Society of Petroleum Engineers, p. 12. doi: 10.2118/80229-MS. Landa, E. R., Phillips, E. J. P. and Lovley, D. R. (1991) ‘Release of 226Ra from uranium mill tailings by microbial Fe(III) reduction’, Applied Geochemistry, 6(6), pp. 647–652. doi: https://doi.org/10.1016/0883-2927(91)90075-Z. Landa, E. R. and Reid, D. F. (1983) ‘Sorption of radium-226 from oil-production brine by sediments and soils’, Environmental Geology, 5(1), pp. 1–8. doi: 10.1007/BF02381293. Langmuir, D. and Melchior, D. (1985) ‘The geochemistry of Ca, Sr, Ba and Ra sulfates in some deep brines from the Palo Duro Basin, Texas’, Geochimica et Cosmochimica Acta, 49(11), pp. 2423–2432. doi: https://doi.org/10.1016/0016-7037(85)90242-X.

272

Langmuir, D. and Riese, A. C. (1985) ‘The thermodynamic properties of radium’, Geochimica et Cosmochimica Acta, 49(7), pp. 1593–1601. doi: https://doi.org/10.1016/0016-7037(85)90264-9. Lauer, N. E., Warner, N. R. and Vengosh, A. (2018) ‘Sources of Radium Accumulation in Stream Sediments near Disposal Sites in Pennsylvania: Implications for Disposal of Conventional Oil and Gas Wastewater’, Environmental Science & Technology. American Chemical Society, 52(3), pp. 955–962. doi: 10.1021/acs.est.7b04952. Laul, J. C. (1994) 'U-234/U-238 ratio: Qualitative estimate of groundwater flow in Rocky Flats monitoring wells'. United States: American Nuclear Society, Inc. Leblans, P., Vandenbroucke, D. and Willems, P. (2011) ‘Storage Phosphors for Medical Imaging’, Materials (Basel, Switzerland). MDPI, 4(6), pp. 1034–1086. doi: 10.3390/ma4061034.

Li, H., Carlson, K.H., Sharvelle, S. and Stednick, J. (2013) ‘Produced Water Quality Characterization and Prediction for Wattenburg Field'. Thesis. Department of Civil and Environmental Engineering. Colorado State University. Li, J., Tang, M., Ye, Z., Chen, L. and Zhou, Y. (2017) ‘Scale formation and control in oil and gas fields: A review’, Journal of Dispersion Science and Technology. Taylor & Francis, 38(5), pp. 661–670. doi: 10.1080/01932691.2016.1185953. Lippincott, E.R., Vanvalkenburg, A., Weir, C.E. and Bunting, E.N. (2012) 'Infrared studies on polymorphs of silicon dioxide and germanium dioxide'. J. Res. Natl. Bur. Stand.(1934). 61, 61. https://doi.org/10.6028/jres.061.009

Liu, Y. and Wang, A. J.J.F. (2009) 'Raman, MIR, and NIR spectroscopic study calcium sulphates: gypsum, bassanite, and anhydrite. 40th Lunar Planet. Sci. Conf. 3, 89 100.

Lloyd, J. R., Newsome, L., Shaw, S. and Morris, K. (2019) ‘Bioremediation of strontium and technetium contaminated groundwater using glycerol phosphate’, Chemical Geology, 509(October 2018), pp. 213–222. doi: 10.1016/j.chemgeo.2019.02.004. López-Ruiz, B. (2000) ‘Advances in the determination of inorganic anions by ion chromatography’, Journal of Chromatography A, 881(1–2), pp. 607–627. doi: 10.1016/S0021-9673(00)00244-2.

Lovley, D. R. (1987) ‘Organic matter mineralization with the reduction of ferric iron: A review’, Geomicrobiology Journal, 5(3–4), pp. 375–399. doi: 10.1080/01490458709385975. Lovley, D. R. and Chapelle, F. H. (1995) ‘Deep subsurface microbial processes’, Reviews of Geophysics, 33(3), pp. 365–381. Lovley, D. R. and Phillips, E. J. (1987) ‘Rapid assay for microbially reducible ferric iron in aquatic sediments.’, Applied and environmental microbiology, 53(7), pp. 1536–40. Luptáková, A., Kotulicova, I., Balintova, M. and Demcak, S. (2015) ‘Bacterial Reduction Of Barium Sulphate By Sulphate-Reducing Bacteria’, Nova

273

Biotechnologica et Chimica, 14(2), pp. 135–140. doi: https://doi.org/10.1515/nbec- 2015-0022. Mackay, E. J. (2005) ‘Scale Inhibitor Application in Injection Wells to Protect Against Damage to Production Wells: When does it Work?’, SPE European Formation Damage Conference. Sheveningen, The Netherlands: Society of Petroleum Engineers, p. 9. doi: 10.2118/95022-MS. Masella, A. P. et al. (2012) ‘PANDAseq: paired-end assembler for illumina sequences’, BMC Bioinformatics, 13(1), p. 31. doi: 10.1186/1471-2105-13-31.

Matthews, J. A. (2014) ‘Inductively coupled plasma atomic emission spectrometry (ICP-AES)’, Encyclopedia of Environmental Change, 2(1), pp. 1–19. doi: 10.4135/9781446247501.n2036.

Mazumder, Q. H. (2012) ‘CFD Analysis of the Effect of Elbow Radius on Pressure Drop in Multiphase Flow’, 2012. doi: 10.1155/2012/125405. Mccartney, R. A. (2008) ‘Conditions under which anhydrite precipitation can occur in oil reservoirs as a result of seawater injection’. McCready, R. G. L., Bland, C. J. and Gonzales, D. E. (1980) ‘Preliminary studies on the chemical, physical, and biological stability of Ba/RaSO4 precipitates’, Hydrometallurgy, 5(2), pp. 109–116. doi: https://doi.org/10.1016/0304- 386X(80)90032-8. McDevitt, B., McLaughlin, M., Cravotta, C. A., Ajemigbitse, M. A., Van Sice, K. J., Blotevogel, J., Borch, T. and Warner, N. R. (2019) ‘Emerging investigator series: radium accumulation in carbonate river sediments at oil and gas produced water discharges: implications for beneficial use as disposal management’, Environmental Science: Processes & Impacts. The Royal Society of Chemistry. doi: 10.1039/C8EM00 336J. McKinsey Global Institute (2019) ‘Global Energy Perspective 2019 : Reference Case’, Energy Insights, (January). Meissner, H. P. and Tester, J. W. (1972) ‘Activity Coefficients of Strong Electrolytes in Aqueous Solutions’, Industrial & Engineering Chemistry Process Design and Development. American Chemical Society, 11(1), pp. 128–133. doi: 10.1021/i2600 41a025. Menzie, C. A., Southworth, B., Stephenson, G. and Feisthauer, N. (2008) ‘The Importance of Understanding the Chemical Form of a Metal in the Environment: The Case of Barium Sulfate (Barite)’, Human and Ecological Risk Assessment: An International Journal. Taylor & Francis, 14(5), pp. 974–991. doi: 10.1080/108070308 02387622. Meyers, K. O., Skillman, H. L. and Herring, G. D. (1985) ‘Control of Formation Damage at Prudhoe Bay, Alaska, by Inhibitor Squeeze Treatment’, Journal of Petroleum Technology. Society of Petroleum Engineers, 37(06), pp. 1019–1034. doi: 10.2118/12472-PA. Miezitis, Y. and Mernagh, T. P. (2007) ‘A Review of the Geochemical Processes

274

Controlling the Distribution of Thorium in the Earth’s Crust and Australia’s Thorium Resources’. Geoscience Australia. Mitchell, R. W., Grist, D. M. and Boyle, M. J. (1980) ‘Chemical Treatments Associated With North Sea Projects’, Journal of Petroleum Technology, 32(05), pp. 904–912. doi: 10.2118/7880-PA. Miyake, M., Minato, I., Morikawa, H. and Iwai, S. (1978) ‘Crystal structures and sulphate force constants of barite, celestite and anglesite’, American Mineralogist; 63(5-6), pp. 506–510. Moghadasi, J., Jamialahmadi, M., Muller-Steinhagen, H. and Shariff, A. (2003) ‘Scale formation in Iranian oil reservoir and production equipment during water injection’, Society of Petroleum Engineers. doi: 10.2118/80406-MS Moghadasi, J., Jamialahmadi, M., Muller-Steinhagen, H. and Shariff, A. (2007) ‘Scale Formation in Oil Reservoir and Production Equipment during Water Injection (Kinetics of CaSO4 and CaCO3 Crystal Growth and Effect on Formation Damage)’, Society of Petroleum Engineers. doi: 10.2118/82233-MS. Monnin, C. and Galinier, C. (1988) 'The solubility of celestite and barite in electrolyte solutions and natural waters at 25°C: A thermodynamic study'. Chem. Geol. 71, pp.283-296. https://doi.org/10.1016/0009-2541(88)90055-1

Moore, W. S. (1981) ‘The thorium isotope content of ocean water’, Earth and Planetary Science Letters, 53(3), pp. 419–426. doi: https://doi.org/10.1016/0012- 821X(81)90046-7. Neame, K. D. (1974) 'Introduction to liquid scintillation counting'. Edited by C. A. Homewood. London ((Borough Green, Sevenoaks, Kent)): Butterworth.

Neff, J. M. (2002) ‘Barium in the Ocean’, Bioaccumulation in Marine Organisms, (January), pp. 79–87. doi: 10.1016/B978-008043716-3/50005-1. Newsome, L., Morris, K. and Trivedi, D. (2014) ‘Microbial reduction of uranium(VI) in sediments of different lithologies collected from Sellafield’, Applied Geochemistry. Elsevier Ltd, 51, pp. 55–64. doi: 10.1016/j.apgeochem.2014.09.008.

Newsome, L., Cleary, A., Morris, K. and Lloyd, J. R. (2017) ‘Long-Term Immobilization of Technetium via Bioremediation with Slow-Release Substrates’, Environmental Science and Technology, 51(3), pp. 1595–1604. doi: 10.1021/acs.est.6b04876. Newsome, L., Morris, K. and Lloyd, J. R. (2014) ‘The biogeochemistry and bioremediation of uranium and other priority radionuclides’, Chemical Geology. The Authors, 363, pp. 164–184. doi: 10.1016/j.chemgeo.2013.10.034.

Newville, M. (2004) ‘Fundamentals of X-ray Absorption Fine Structure’, http://xafs.org/Tutorials?action=AttachFile&do=get&target=Newville_Intro.pdf abgerufen März 2016.

Newville, M. (2014) ‘Fundamentals of XAFS’, Reviews in Mineralogy and Geochemistry, 78(1), pp. 33–74. doi: 10.2138/rmg.2014.78.2.

275

Nichols, M. L. and Smith, E. C. (1941) ‘Coprecipitation with Barium Sulfate.’, The Journal of Physical Chemistry. American Chemical Society, 45(3), pp. 411–421. doi: 10.1021/j150408a006. Nirdosh, I., Trembley, W. B. and Johnson, C. R. (1990) ‘Adsorption-desorption studies on the 226Ra-hydrated metal oxide systems’, Hydrometallurgy, 24(2), pp. 237–248. doi: https://doi.org/10.1016/0304-386X(90)90089-K. Nixon, A., Keller, D., Fritze, K., Pidruczny, A. and Corsini, A. (1983) ‘Radium removal from Elliot Lake uranium-mill solids by EDTA leaching’, Hydrometallurgy, 10(2), pp. 173–186. doi: 10.1016/0304-386X(83)90004-X. NRC (1999) Health Effects of Exposure to Radon: BEIR VI. Washington, DC: The National Academies Press. doi: 10.17226/5499. Nurk, S., Bankevich, A. and Antipov, D. (2013) ‘Assembling Genomes and Mini- metagenomes from Highly Chimeric Reads BT - Research in Computational Molecular Biology’, in Deng, M. et al. (eds). Berlin, Heidelberg: Springer Berlin Heidelberg, pp. 158–170.

O’Brien, R. S. and Cooper, M. B. (1998) ‘Technologically enhanced naturally occurring radioactive material (NORM): Pathway analysis and radiological impact’, Applied Radiation and Isotopes, 49(3), pp. 227–239. doi: https://doi.org/10.1016/S0969-8043(97)00244-3. Ohta, K., Morikawa, H. and Tanaka, K. (2000) ‘Ion-chromatographic behavior of alkali metal cations and ammonium ion on zirconium-adsorbing silica gel’, Journal of Chromatography A, 884(1–2), pp. 123–130. doi: 10.1016/S0021-9673(00)00214- 4.

Oil&GasUK (2018) ‘Environment Report 2018’. Ojovan, M. I. (2019) An introduction to nuclear waste immobilisation. Third edit. Edited by W. E. Lee and S. N. Kalmykov. Amsterdam: Elsevier. Okonkwo, P.C. and Adel, M.A.M., (2014) 'Erosion-Corrosion in Oil and Gas Industry: A Review'. Int. J. Metall. Mater. Sci. Eng. 4, 7–28.

Ollivier Bernard, M. M. (2005) ‘Symptoms of Souring’, Petroleum Microbiology. American Society for Microbiology (ASM). Olsvik, P. A., Berntssen, M. H. G., Hylland, K., Eriksen, D. O. and Holen, E. (2012) ‘Low impact of exposure to environmentally relevant doses of 226Ra in Atlantic cod (Gadus morhua) embryonic cells’, Journal of Environmental Radioactivity, 109, pp. 84–93. doi: https://doi.org/10.1016/j.jenvrad.2012.02.003. OSPAR (2001) ‘Overview assessment of the implementation of OSPAR Recommendation 2001 / 1 for the management of produced water from offshore installations Offshore Industry Series’. OSPAR (2018) ‘Annual report on discharges of radioactive substances from the non- nuclear sector in 201 6 Radioactive Substances Series Draft Report on Discharges of

276

Radionuclides from the Non - nuclear Sectors in 2016’. OSPAR Comission 2014 (2013) 'OSPAR report on discharges, spills and emissions from offshore oil and gas installations in 2013', OSPAR Commission. Ouyang, B., Akob, D. M., Dunlap, D. and Renock, D. (2017) ‘Microbially mediated barite dissolution in anoxic brines’, Applied Geochemistry, 76, pp. 51–59. doi: 10.1016/j.apgeochem.2016.11.008. Palanisamy, K., Raj, K. S., Bhuvaneswari, S. and Subramanian, V. K. (2017) ‘A novel phenomenon of effect of metal on calcium carbonate scale, morphology, polymorphism and its deposition’, Materials Research Innovations. Taylor & Francis, 21(5), pp. 294–303. doi: 10.1080/14328917.2016.1214230. Pardue, J. H. and Guo, T. Z. (1998) ‘Biogeochemistry of 226Ra in contaminated bottom sediments and oilfield waste pits’, Journal of Environmental Radioactivity, 39(3), pp. 239–253. doi: https://doi.org/10.1016/S0265-931X(97)00047-7. Parkhurst, D. L. and Appelo, C. a. J. (2013) ‘Description of Input and Examples for PHREEQC Version 3 — A Computer Program for Speciation , Batch-Reaction , One- Dimensional Transport , and Inverse Geochemical Calculations.’, U.S. Geological Survey Techniques and Methods, book 6, chapter A43, p. 678.

Parsons, J. G., Aldrich, M. V. and Gardea-Torresdey, J. L. (2002) ‘Environmental and biological applications of extended X-ray absorption fine structure (EXAFS) and X- ray absorption near edge structure (XANES) spectroscopies’, Applied Spectroscopy Reviews, 37(2), pp. 187–222. doi: 10.1081/ASR-120006044.

Passo, C.J. and Cook, G.T. (1994) ‘Handbook of Environmental Liquid Scintillation Spectrometry - A Complilation of Theory and Methods’, Scottish Univeristies Research and Reactor Centre.

Paschoa, A. S. (1997) ‘Naturally occurring radioactive materials (NORM) and petroleum origin’, Applied Radiation and Isotopes, 48(10), pp. 1391–1396. doi: https://doi.org/10.1016/S0969-8043(97)00134-6. Paschoa, A. S. (1998) ‘Potential environmental and regulatory implications of naturally occurring radioactive materials (NORM)’, Applied Radiation and Isotopes, 49(3), pp. 189–196. doi: 10.1016/S0969-8043(97)00239-X. Pauwels, S. J. (2002) ‘Transplantation and in situ Exposures. Testing for Toxicity and Bioaccumulation in Freshwater Sediments’, Handbook on Sediment Quality. WEF. Payne, G. E. (1987) ‘A History Of Downhole Scale Inhibition By Squeeze Treatments On The Murchison Platform’, Offshore Europe. Aberdeen, United Kingdom: Society of Petroleum Engineers, p. 12. doi: 10.2118/16539-MS. Penner-Hahn, J. E. (2001) ‘X-ray Absorption Spectroscopy’, Encyclopedia of life sciences. New York : John Wiley & Sons Incorporated, p. 414. doi: 10.1038/npg.els.0002984.

Periasamy, A., Muruganand, S. and Palaniswamy, M. (2009) 'Vibrational studies of Na2SO4, K2SO4, NaHSO4 and KHSO4 crystals'. Rasayan J. Chem. 2(4), 981–989.

277 https://doi.org/10.1016/00222860(76)82004-2

Perkampus, H.-H. (1992) 'UV-VIS spectroscopy and its applications'. Berlin ; Springer-Verlag (Springer laboratory).

Petrucci, R. H. (2007) 'General Chemistry: Principle and Modern Applications'. 9th Edition, Upper Saddle River, NJ, Prentice Hall. ISBN 10: 0131493647/ISBN 13: 9780131493643

Phillips, E. J. P., Landa, E. R., Kraemer, T. and Zielinski, R. (2001) ‘Sulfate-reducing bacteria release barium and radium from naturally occurring radioactive material in oil-field barite’, Geomicrobiology Journal, 18(2), pp. 167–182. doi: 10.1080/01490450120549. Pontedeiro, E. M., van Genuchten, M.T., Cotta, R.M., Simunek, J. (2010) ‘The effects of preferential flow and soil texture on risk assessments of a NORM waste disposal site’, Journal of Hazardous Materials, 174(1), pp. 648–655. doi: https://doi.org/10.1016/j.jhazmat.2009.09.100. Pontedeiro, E. M., Heilbron, P. F. L. and Cotta, R. M. (2007) ‘Assessment of the mineral industry NORM/TENORM disposal in hazardous landfills’, Journal of Hazardous Materials, 139(3), pp. 563–568. doi: https://doi.org/10.1016/j.jhazmat.2006.02.063. Postma, D. (1983) ‘Pyrite and siderite oxidation in swamp sediments’, Journal of soil science, Volume 34, Issue 1. https://doi.org/10.1111/j.1365-2389.1983.tb00821.x Proske, U., Wood, R., Fallon, S. and Stevenson, J. (2015) ‘Use of heavy liquid density separation to remove pyrite from sediment samples for radiocarbon dating’, Quaternary Geochronology, 25, pp. 66–71. doi: https://doi.org/10.1016/j.quageo.2014.10.002. Punjungsari, T. N. (2017) ‘Characterisation of sulfate reducing bacteria from biofilm sulfate reduction and CU precipitation bioreactor (batch culture)’, 2(2), pp. 1–9. Puntervold, T. and Austad, T. (2008) ‘Injection of seawater and mixtures with produced water into North Sea chalk formation: Impact of fluid-rock interactions on wettability and scale formation’, Journal of Petroleum Science and Engineering. Elsevier B.V., 63(1–4), pp. 23–33. doi: 10.1016/j.petrol.2008.07.010. Putnis, C. V, Kowacz, M. and Putnis, A. (2008) ‘The mechanism and kinetics of DTPA- promoted dissolution of barite’, Applied Geochemistry, 23(9), pp. 2778–2788. doi: https://doi.org/10.1016/j.apgeochem.2008.07.006. Raahauge, P.J., Olsson, J., Dideriksen, K., Stipp, S.L.S., Oelkers, E.H. and Zhen-Wu, B.Y.(2016). 'Experimental determination of barite dissolution and precipitation rates as a function of temperature and aqueous fluid composition'. Geochim. Cosmochim. Acta 194, 193–210. https://doi.org/10.1016/j.gca.2016.08.041

Ramaswamy, V., Vimalathithan, R. M. and Ponnusamy, V. (2010) ‘Synthesis and characterization of BaSO4 nano particles using micro emulsion technique’, Advances in Applied Science Research, 1(3), pp. 197–204.

278

Rauret, G., Lopez-Sanchez, J.F., Sahuquillo, A., rubio, R., Davidson, C., Ure, A. and Quevauviller, Ph. (1999) ‘Improvement of the BCR three step sequential extraction procedure prior to the certification of new sediment and soil reference materials’, Journal of Environmental Monitoring, 1(1), pp. 57–61. doi: 10.1039/a807854h. Ravel, B. and Newville, M. (2005) ‘ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT’, Journal of Synchrotron Radiation, 12(4), pp. 537–541. doi: 10.1107/S0909049505012719. Rawahi, Y. M. Al (2017) ‘Studies on Scale Deposition in Oil Industries & Their Control’, 3(12), pp. 152–167. Read, D., Read, G.D. and Thorne, M.C. (2013) 'Background in the context of land contaminated with naturally occurring radioactive material'. J. Radiol. Prot. 33, 367–380.https://doi.org/10.1088/0952-4746/33/2/367

Reed, S. J. B. (2005) 'Electron microprobe analysis and scanning electron microscopy in geology'. Second edi, Electron Microprobe Analysis & Scanning Electron Microscopy in Geology. Second edi. Cambridge: Cambridge University Press.

Risthaus, P., Bosbach, D., Becker, U. and Putnis, A. (2001) ‘Barite scale formation and dissolution at high ionic strength studied with atomic force microscopy’, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 191(3), pp. 201– 214. doi: https://doi.org/10.1016/S0927-7757(00)00843-8. Robert, H. (1981) ‘Isotopic and geologic studies to identify the sources of sulfate in groundwater containing high barium concentrations’. Roberts, C. J., Britt Quinby, J., Duggan, W. P. and Yuan, Y. (1998) ‘Disposal options and case-study pathway analyses’, Applied radiation and isotopes. Oxford ; Pergamon Press, pp. 241–258. doi: 10.1016/S0969-8043(97)00245-5. Robertson, J. O. and Chilingar, G. V. (2017) 'Environmental Aspects of Oil and Gas Production - 5.2. Studies Supporting Hydraulic Fracturing. John Wiley & Sons, pp. 211–267. Røe Utvik, T. I. (1999) ‘Chemical characterisation of produced water from four offshore oil production platforms in the North Sea’, Chemosphere, 39(15), pp. 2593–2606. doi: https://doi.org/10.1016/S0045-6535(99)00171-X. Rosenberg, Y. O., Metz, V., Oren, Y., Volkman, Y. and Ganor, J. (2011(a)) ‘Co- precipitation of radium in high ionic strength systems: 2. Kinetic and ionic strength effects’, Geochimica et Cosmochimica Acta, 75(19), pp. 5403–5422. doi: https://doi.org/10.1016/j.gca.2011.07.013. Rosenberg, Y. O., Sadeh, Y., Metz, V., Pina, C. M. and Ganor, J. (2014) ‘Nucleation and growth kinetics of RaxBa1−xSO4 solid solution in NaCl aqueous solutions’, Geochimica et Cosmochimica Acta, 125, pp. 290–307. doi: 10.1016/j.gca.2013. 09.041. Rosenberg, Y. O., Metz, V. and Ganor, J. (2011(b)) ‘Co-precipitation of radium in high ionic strength systems: 1. Thermodynamic properties of the Na–Ra–Cl–SO4–

279

H2O system – Estimating Pitzer parameters for RaCl2’, Geochimica et Cosmochimica Acta, 75(19), pp. 5389–5402. doi: https://doi.org/10.1016/j.gca.2011.06.042. Rowan, E. L. (2011) ‘Radium Content of Oil- and Gas-Field Produced Waters in the Northern Appalachian Basin (USA)’: Roychoudhury, A. N., Kostka, J. E. and Van Cappellen, P. (2003) ‘Pyritization: A palaeoenvironmental and redox proxy reevaluated’, Estuarine, Coastal and Shelf Science, 57(5–6), pp. 1183–1193. doi: 10.1016/S0272-7714(03)00058-1. Rumble, J. R. (1978) ‘CRC handbook of chemistry and physics’, CRC handbook of chemistry and physics (Online). Electronic. Boca Raton, Fla: CRC Press. Rutt, H.N. and Nicola, J.H. (1974) 'Raman spectra of carbonates of calcite structure'. J. Phys.C Solid State Phys. 7, 4522–4528. https://doi.org/10.1088/0022 3719/7/24/015

Ruus, A., Schaanning, M., Oxnevad, S. and Hylland, K. (2005) ‘Experimental results on bioaccumulation of metals and organic contaminants from marine sediments’, Aquatic Toxicology, 72(3), pp. 273–292. doi: https://doi.org/10.1016/j.aquatox. 2005.01.004. Sajih, M., Bryan., N. D., Livens, F. R., Vaughan, D. J., Descostes, M., Phrommavanh, V., Nos, J. and Morris, K. (2014) ‘Adsorption of radium and barium on goethite and ferrihydrite: A kinetic and surface complexation modelling study’, Geochimica et Cosmochimica Acta, 146, pp. 150–163. doi: https://doi.org/10.1016/j.gca.2014. 10.008. Shakkthivel, P. and Vasudevan, T. (2006) ‘Acrylic acid-diphenylamine sulphonic acid copolymer threshold inhibitor for sulphate and carbonate scales in cooling water systems’, Desalination, 197(1), pp. 179–189. doi: https://doi.org/10.1016/j.desal. 2005.12.023. Shannon, R. D. (1976) ‘Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides’, Acta Crystallographica Section A, 32(5), pp. 751–767. doi: 10.1107/S0567739476001551. Shi, W., Xia, M., Lei, W. and Wang, F. (2013) ‘Molecular dynamics study of polyether polyamino methylene phosphonates as an inhibitor of anhydrite crystal’, Desalination, 322, pp. 137–143. doi: https://doi.org/10.1016/j.desal.2013.05.013. Siddeeg, S. M., Bryan, N. D. and Livens, F. R. (2015) ‘Behaviour and mobility of U and Ra in sediments near an abandoned uranium mine, Cornwall, UK’, Environmental Science: Processes & Impacts. The Royal Society of Chemistry, 17(1), pp. 235–245. doi: 10.1039/C4EM00230J. Singh, R. (2015a) ‘Chapter 2 - Water and Membrane Treatment’, in Singh, R. B. T.- M. T. and E. for W. P. (Second E. (ed.) Membrane Technology and Engineering for Water Purification: Application, Systems Design and Operation. Oxford: Butterworth-Heinemann, pp. 81–178. doi: https://doi.org/10.1016/B978-0-444- 63362-0.00002-1. Singh, R. (2015b) ‘Chapter 3 - Hybrid Membrane Systems – Applications and Case

280

Studies’. (Second E. (ed.) Membrane Technology and Engineering for Water Purification: Application, Systems Design and Operation. Oxford: Butterworth- Heinemann, pp. 179–281. doi: https://doi.org/10.1016/B978-0-444-63362-0.00003- 3. Skoog, D. A. (2007) 'Principles of instrumental analysis'. 6th ed. /. Edited by F. J. Holler and S. R. Crouch. Belmont, Calif. ; Thomson Brooks/Cole.

Smith, A. L. (1985) ‘Radioactive Scale Formation’, Offshore Technology Conference. Houston, Texas: Offshore Technology Conference, p. 10. doi: 10.4043/5081-MS. Smith, B. C. (1996) 'Fundamentals of Fourier transform infrared spectroscopy'. Boca Raton: CRC Press.

Smith, G.D., Firth, S., Clark, R.J.H. and Cardona, M. (2002) 'First- and second-order Raman spectra of galena (PbS)'. J. Appl. Phys. 92, 4375–4380. https://doi.org/10.1063/1.1505670

Smith, K. P. (1992) ‘An Overview of Naturally Occurring Radioactive Materials (NORM ) in the Petroleum Industry’. SNIFFER (2003) ‘Executive summary’, Robotics and Computer-Integrated Manufacturing, 7(3–4), pp. 193–195. doi: 10.1016/0736-5845(90)90003-q. Somayajulu, B. L. K. and Goldberg, E. D. (1966) ‘Thorium and uranium isotopes in seawater and sediments’, Earth and Planetary Science Letters, 1(3), pp. 102–106. doi: https://doi.org/10.1016/0012-821X(66)90084-7. Steffan, E. M. (2013) 'NORM scale forming processes'.Hanover 2012/2014. Version 1.9 Stevenson, A.G., Tait, B.A.R., Richardson, A. E., Smith, T., Nicolson, R. A. and Stewart, H. R. (1995) ‘The geochemistry of seabed sediments of the United Kingdom Continental Shelf; the North Sea, Hebrides and West Shetland Shelves, and the Malin-Hebrides sea area’, British Geological Survey Technical Report, pp. 484. (WB/95/028) (unpublished). Stookey, L. L. (1970) ‘Ferrozine a new spectrophotometric reagent for iron’, 42(7), pp. 779–781.

Štrok, M. and Smodiš, B. (2013) ‘Soil-to-plant transfer factors for natural radionuclides in grass in the vicinity of a former uranium mine’, Nuclear Engineering and Design, 261, pp. 279–284. doi: https://doi.org/10.1016/j.nucengdes.2013. 03.036. Summerlin, J. and Prichard, H. M. (1985) ‘Radiological Health Implications of Lead- 210 and Polonium-210 Accumulations in LPG Refineries’, American Industrial Hygiene Association Journal. Taylor & Francis, 46(4), pp. 202–205. doi: 10.1080/15298668591394662. Sutherland, R. (1998) ‘Loss-on-ignition estimates of organic matter and relationships to organic carbon in fluvial bed sediments’, Hydrobiologia. Edited by R. Sutherland, 389(1–3), pp. 153–167.

281

Tanner, R. N. and Wittingham, K. P. (1986) ‘Scale Control During Waterflooding Operations: A Field Appraisal of Inhibitor Requirements and Performance’, International Meeting on Petroleum Engineering. Beijing, China: Society of Petroleum Engineers, p. 12. doi: 10.2118/14127-MS. Taskayev, A.I.; Ovchenkov, V.Y.; Aleksakhin, R. M. (1978) ‘Forms of 226Ra in the horizons of soils with a high concentration of this isotope’, Soviet Soil Science, 10(1), pp. 45–51. Telmadarreie, A., Shadizadeh, S. R. and Alizadeh, B. (2012) ‘Investigation of Hydrogen Sulfide Oil Pollution Source : Asmari Oil Reservoir of Marun Oil Field in the Southwest of Iran’, Iranian Journal of Chemical Engineering, 9(3), pp. 63–74. Teo, B. K. (1986) EXAFS : basic principles and data analysis . Berlin ; Springer-Verlag (Inorganic chemistry concepts ; v. 9).

Tessier, A., Campbell, P. G. C. and Bisson, M. (1979) ‘Sequential extraction techniques for the speciation of particulate trace metals’, Analytical Chemistry, 51(7), pp. 844–851. Thiry, Y. and Van Hees, M. (2008) ‘Evolution of pH, organic matter and 226radium/calcium partitioning in U-mining debris following revegetation with pine trees’, Science of The Total Environment, 393(1), pp. 111–117. doi: https://doi.org/10.1016/j.scitotenv.2007.12.020. Thorpe, C. (2012) ‘Strontium sorption and precipitation behaviour during bioreduction in nitrate impacted sediments’, Chemical Geology. Elsevier B.V., 306– 307, pp. 114–122. doi: 10.1016/j.chemgeo.2012.03.001.

Todd, A. C. and Yuan, M. (1990) ‘Barium and Strontium Sulfate Solid Solution Formation in Relation to North Sea Scaling Problems’, SPE Production Engineering. Society of Petroleum Engineers, 5(03), pp. 279–285. doi: 10.2118/18200-PA. Todd, A. C. and Yuan, M. D. (1992) ‘Barium and Strontium Sulfate Solid-Solution Scale Formation at Elevated Temperatures’, SPE Production Engineering. Society of Petroleum Engineers, 7(01), pp. 85–92. doi: 10.2118/19762-PA. Tokunaga, K., Kozai, N. and Takahashi, Y. (2018) ‘A new technique for removing strontium from seawater by coprecipitation with barite’, Journal of Hazardous Materials, 359, pp. 307–315. doi: https://doi.org/10.1016/j.jhazmat.2018.07.044. Trifilieff, O. and Wines, T. H. (2009) ‘Scientific & amp; Technical Report Black Powder Removal from Transmission Pipelines: Diagnostics and Solutions’, Pipeline Rehablitation & Maintenance Conference, pp. 1–12. UK NORM Waste Strategy (2014) ‘Strategy for the management of Naturally Occurring Radioactive Material (NORM) waste in the United Kingdom: A Review of UK Legislation and Policy Relating to NORM Waste Management – Non Technical Summary January 2014’, (January). UNSCEAR (2000) Sources and effects of ionizing radiation : United Nations Scientific Committee on the Effects of Atomic Radiation : UNSCEAR 2000 report to the General

282

Assembly, with scientific annexes. Edited by UNSCEAR. New York: United Nations. USGS (1999) ‘Naturally Occurring Radioactive Materials (NORM) in Produced Water and Oil-Field Equipment — An Issue for the Energy Industry’, (September). van Beek, P., Reyss, J. and Paterne, M. (2002) ‘226Ra in barite: Absolute dating of Holocene Southern Ocean sediments and reconstruction of sea-surface reservoir ages’, Geology, 30(8), pp. 731–734. doi: 10.1130/0091- 7613(2002)030<0731:RIBADO>2.0.CO;2.

VanLoon, G. W. (2000) 'Environmental chemistry : a global perspective'. in S. J. Duffy (eds). New York, Oxford: Oxford University Press, pp. 560. ISBN 0198564406 Van Sice, K., Cravotta, C. A., McDevitt, B., Tasker, T. L., Landis, J. D., Puhr, J. and Warner, N. R. (2018) ‘Radium attenuation and mobilization in stream sediments following oil and gas wastewater disposal in western Pennsylvania’, Applied Geochemistry, 98, pp. 393–403. doi: https://doi.org/10.1016/j.apgeochem.2018.10.011. Van Weers, A. W., Pace, I. and Strand, T. (1997)' Current practices of dealing with natural radioactivity from oil and gas production in EU member states, Rep. EU 17621, Office for Official Publications of the European Communities, Luxembourg Vazirian, M., Charpentier, T.V.J., de Oliveira Penna, M. and Neville, A., (2016) ‘Surface Inorganic Scale Formation in Oil and Gas Industry : as Adhesion and Deposition Processes’. J. Pet. Sci. Eng. 137, 2232. https://doi.org/10.1016/j.petrol.2015.11.005

Vearrier, D., Curtis, John A. and Greenberg, M. I. (2009) ‘Technologically enhanced naturally occurring radioactive materials’, Clinical Toxicology, 47(5), pp. 393–406. doi: 10.1080/15563650902997849. Vearrier, D., Curtis, J. A. and Greenberg, M. I. (2009) ‘Technologically enhanced naturally occurring radioactive materials’, Clinical Toxicology. Taylor & Francis, 47(5), pp. 393–406. doi: 10.1080/15563650902997849. Vetter, O. J. (1972) ‘An Evaluation of Scale Inhibitors’, Journal of Petroleum Technology. Society of Petroleum Engineers, 24(08), pp. 997–1006. doi: 10.2118/3188-PA. Vetter, O. J. (1976) ‘Oilfield Scale---Can We Handle It?’, Journal of Petroleum Technology. Society of Petroleum Engineers, 28(12), pp. 1402–1408. doi: 10.2118/5879-PA. Vetter, O. J., Kandarpa, V. and Harouaka, A. (1982) ‘Prediction of Scale Problems Due to Injection of Incompatible Waters’, Journal of Petroleum Technology. Society of Petroleum Engineers, 34(02), pp. 273–284. doi: 10.2118/7794-PA. Vidic, R. D. (2013) ‘Impact of Shale Gas Development on Regional Water Quality’. doi: 10.1126/science.1235009. Vinograd, V.L., Kulik, D.A., Brandt, F., Klinkenberg, M., Weber, J., Winkler, B., Bosbach, D., (2018) 'Thermodynamics of the solid solution - Aqueous solution

283 system (Ba,Sr,Ra)SO4 + H2O: I. The effect of strontium content on radium uptake by barite'. Appl. Geochemistry 89,59–74. https://doi.org/https://doi.org/10.1016/j.apgeochem.2017.11.009

Vinograd, V.L., Winkler, B., Bosbach, D., Brandt, F., Rozov, K., Klinkenberg, M., Refson, K., (2013) 'Solid–aqueous equilibrium in the BaSO4–RaSO4–H2O system: First-principles calculations and a thermodynamic assessment'. Geochim. Cosmochim. Acta 122, 398–417. https://doi.org/10.1016/j.gca.2013.08.028

Viollier, E., Inglett, P. and Hunter, K. (2000) ‘The Ferrozine method revisted’, Applied Geochemistry, 15, pp. 785–790. doi: 10.1016/S0883-2927(99)00097-9.

Waldichuk, M. (1971) 'Radioactivity in the marine environment'. Washington. Div. of Biology and Medicine, U.S. Atomic Energy Commission. Wang, K.-S., Resch, R., Dunn, K., Shuler, P., Tang, Y., Koel, B. E. and Yen, T. F. (1999) ‘Dissolution of the barite (001) surface by the chelating agent DTPA as studied with non-contact atomic force microscopy’, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 160(3), pp. 217–227. doi: https://doi.org/10.1016/S0927- 7757(99)00183-1. Wang, K.-S., Resch, R., Dunn, K., Shuler, P., Tang, Y., Koel, B. E. and Yen, T. F. (2000) ‘Scanning Force Microscopy Study of Etch Pits Formed during Dissolution of a Barite (001) Surface in CDTA and EDTA Solutions’, Langmuir. American Chemical Society, 16(2), pp. 649–655. doi: 10.1021/la990285x. Wang, K.-S., Resch, R., Dunn, K., Shuler, P., Tang, Y., Koel, B. E. and Yen, T. F. (2002) ‘Effects of Scale Dissolvers on Barium Sulfate Deposits: A Macroscopic and Microscopic Study’, Corrosion 2002, (02309), pp. 1–11. Wangersky, P. J. (2003) 'Methods of seawater analysis', Marine Chemistry. doi: 10.1016/0304-4203(78)90045-2.

Wang, Q. (2008) ‘Generation mechanism and control measures for H2S in oil wells, Liaohe Oilfield’, Petroleum Exploration and Development. Research Institute of Petroleum Exploration & Development, PetroChina, 35(3), pp. 349–354. doi: 10.1016/S1876-3804(08)60082-8. Wang, Q., Garrity, G. and Tiedje, J. (2007) ‘Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy’, Applied and environmental microbiology. 2007/06/22. American Society for Microbiology, 73(16), pp. 5261–5267. doi: 10.1128/AEM.00062-07.

Warner, N. R. et al. (2013) ‘Impacts of Shale Gas Wastewater Disposal on Water Quality in Western Pennsylvania’, Environmental Science & Technology. American Chemical Society, 47(20), pp. 11849–11857. doi: 10.1021/es402165b. Warren, E. A. and Smalley, P. C. (1993) ‘The chemical composition of North Sea formation waters: a review of their heterogeneity and potential applications’, Geological Society, London, Petroleum Geology Conference series. Geological Society of London, 4(1), pp. 1347–1352. doi: 10.1144/0041347.

284

Wat, R. M. S. et al. (1992) ‘Kinetics of BaSO4 Crystal Growth and Effect in Formation Damage’, SPE Formation Damage Control Symposium. Lafayette, Louisiana: Society of Petroleum Engineers, p. 9. doi: 10.2118/23814-MS. WEC (2019) ‘Global Energy Scenarios Comparison Review About the World Energy Council’, pp. 35. Wilkins, M. J., Vaughan, D. J., Beadle, I. and Livens, F. R. (2007) ‘The influence of microbial redox cycling on radionuclide mobility in the subsurface at a low-level radioactive waste storage site’, Geobiology. John Wiley & Sons, Ltd (10.1111), 5(3), pp. 293–301. doi: 10.1111/j.1472-4669.2007.00101.x. Wilson, A. J. and Scott, L. M. (1992) ‘Characterization of Radioactive Petroleum Piping Scale With an Evaluation of Subsequent Land Contamination’, Health Physics, 63(6). Willmott, P. (2011) ‘An Introduction to Synchrotron Radiation’. doi: 10.1002/97811 19970958.

Wolicka, D. and Borkowski, A. (2012) ‘Microorganisms and Crude Oil’, Introduction to Enhanced Oil Recovery (EOR) Processes and Bioremediation of Oil-Contaminated Sites. doi: 10.5772/48016. Woods, C. and Rowland, A. P. (1997) ‘Applications of anion chromatography in terrestrial environmental research’, Journal of Chromatography A, 789(1–2), pp. 287–299. doi: 10.1016/S0021-9673(97)00923-0.

Worden, R. H., Manning, D. A. C. and Lythgoe, P. R. (2000) ‘The origin and production geochemistry of radioactive lead (210Pb) in NORM-contaminated formation waters’, Journal of Geochemical Exploration, 69–70, pp. 695–699. doi: https://doi.org/10.1016/S0375-6742(00)00073-X. Yoshida, Y., Yoshikawa, H. and Nakanishi, T. (2008) ‘Partition coefficients of Ra and Ba in calcite’, Geochemical Journal. Japan, 42(3), pp. 295–304. Yuan, M., Todd, A. C. and Sorbie, K. S. (1994) ‘Sulphate scale precipitation arising from seawater injection: a prediction study’, Marine and Petroleum Geology, 11(1), pp. 24–30. doi: 10.1016/0264-8172(94)90006-X. Zardin, B., Cillo, G., Borghi, M., D'Adamo, A. and Fontanesi, S. (2017) ‘Pressure Losses in Multiple-Elbow Paths and in V-Bends of Hydraulic Manifolds’. doi: 10.3390/en10060788. Zeissler, C. J., Lindstrom, R. M. and McKinley, J. P. (2001) ‘Radioactive particle analysis by digital autoradiography’, Journal of Radioanalytical and Nuclear Chemistry, 248(2), pp. 407–412. doi: 10.1023/A:1010640411441. Zhang, T., Gregory, K., Hammack, R. W. and Vidic, R. D. (2014) ‘Co-precipitation of Radium with Barium and Strontium Sulfate and Its Impact on the Fate of Radium during Treatment of Produced Water from Unconventional Gas Extraction’, Environmental Science & Technology. American Chemical Society, 48(8), pp. 4596– 4603. doi: 10.1021/es405168b.

285

Zhang, Z., Liu, Y., Dai, Z., Bhandari, N., Zhang, F., Yan, F., Ruan, G., Alsaiari, H., Lu, Y., Deng, G., Kan, A. and Tomson, M. (2016) ‘Impact of FeIII/FeII on Scale Inhibition’, SPE International Oilfield Scale Conference and Exhibition. Aberdeen, Scotland, UK: Society of Petroleum Engineers, p. 17. doi: 10.2118/179905-MS. Zhu, C. (2004) 'Coprecipitation in the barite isostructural family: 1. Binary mixing properties'. Geochim. Cosmochim. Acta 68, 3327–3337. https://doi.org/10.1016/j.gca.2003.10.014.

Zhu, G., Liu, H., Su, J., Zhu, Y., Wang, Y, and Sun, C. (2017) ‘Genesis and distribution of hydrogen sulfide in deep heavy oil of the Halahatang area in the Tarim Basin, China’, Journal of Natural Gas Geoscience. Elsevier Ltd, 2(1), pp. 57–71. doi: 10.1016/j.jnggs.2017.03.004.

286

Appendix 1

Conferences and Courses

•Produced Water Society Seminar; 5 – 7th February 2019, Sugar Land, Texas, USA

•BP-ICAM (International Centre for Advanced Materials) Annual Conference; November 2015-2018; Manchester, UK

•NERC Centre for Doctoral Training in Oil and Gas Annual Conference; November 2015-2018, Edinburgh, UK

•The School of Earth and Environmental Science (SEES) Postgraduate Annual Conference; December 2015-2018, Manchester, UK

•Produced Water Middle East; 11 – 12th November 2018, Abu Dhabi, UAE

•Abu Dhabi International Petroleum Exhibition & Conference (ADIPEC); 11- 15th November 2018; Abu Dhabi, UAE

•Water, Wastewater & Environmental Monitoring (WWEM) Conference; November 2018, Telford, UK

•13th International Symposium on Nuclear and Environmental Radiochemical Analysis; 17 – 20th September 2018, Cambridge, UK

• envEXPO; 27th Feb – 1st March 2017, Norwich Research Park, UK

287

Field- and classroom-based courses completed through the NERC CDT in Oil and Gas training development programme:

288