Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Uranous sulfate precipitation as a novel route to uranium purification in extractive metallurgy Burns, Alexander D. 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_september_burns_alexander.pdf [ 9.15MB ]
Metadata
JSON: 24-1.0166486.json
JSON-LD: 24-1.0166486-ld.json
RDF/XML (Pretty): 24-1.0166486-rdf.xml
RDF/JSON: 24-1.0166486-rdf.json
Turtle: 24-1.0166486-turtle.txt
N-Triples: 24-1.0166486-rdf-ntriples.txt
Original Record: 24-1.0166486-source.json
Full Text
24-1.0166486-fulltext.txt
Citation
24-1.0166486.ris

Full Text

Uranous sulfate precipitation as a novel route to uraniumpurification in extractive metallurgybyAlexander D. BurnsB.Sc. Mining Engineering, Queen’s University, B.A. Computer Science, Queen’s University,            Doctor of Philosophyin      (Materials Engineering)e University of British Columbia(Vancouver)July © Alexander D. Burns, AbstractUranous sulfate can be crystallized from uranium(IV)-containing solutions by rais-ing the temperature and adding sulfuric acid. Several important aspects of the pro-cess have never been investigated, however, making its successful application as areal-world extractive metallurgy technology far from certain. is dissertation ad-dresses several fundamental questions surrounding the crystallization of uranoussulfate from acidic process solutions. e effects of various parameters on the sol-ubility of uranous sulfate and the kinetics of its precipitation are demonstrated, in-cluding temperature, acid concentration, and agitation, based on the results from aseries of bench-scale experiments. e effects of various impurities on the selectivityand efficiency of the crystallization process are also determined. Two new uranoussulfate x-hydrate polymorphs, the hexahydrate and the octahydrate, are character-ized using single-crystal x-ray diffraction, vibrational spectroscopy, and chemicalassay data, and an understanding of the conditions under which they form is de-veloped. e thermal stability and decomposition characteristics of uranous sulfatetetrahydrate, hexahydrate, and octahydrate are demonstrated through fundamentalthermodynamic calculations and through the examination of thermal analysis data.e fundamental kinetics of uranium(IV) oxidation in acidic solutions are quanti-fied through the interpretation of experimental data under various conditions ofacidity, temperature, and oxygen partial pressure. Finally, a hydrometallurgy flowsheet incorporating uranous sulfate precipitation is presented, and the viability ofthe complete process is demonstrated experimentally, including electrolytic reduc-tion, precipitation, filtration, drying, and calcining. is work demonstrates thaturanous sulfate precipitation is viable as a hydrometallurgical process technology,and that further work is justified.iiPrefacee original concept for this project was developed by Cameco Corporation, andwas the subject of several prior investigations at their research centre in Port Hope,Ontario, from  to . eir work generally focused on the electrolytic re-duction phase of the proposed flow sheet, along with some general studies on thesolubility of uranous sulfate in the context of plant design. e work presented inthis dissertation is more fundamental in nature and focuses on the characteristics ofthe precipitate itself, and the kinetics of several related phenomena. It is my originalwork and does not replicate or otherwise make use of Cameco’s previous work.Chapters  and . Portions of the introductory text and background informationwere originally published in my PhD proposal titled e electrolytic reduction andprecipitation of uranous sulfate ().Chapter . A version of this material has been published in Burns, Patrick, Lam,and Dreisinger []. Dr. Mati Raudsepp in the Department of Earth, Ocean & At-mospheric Sciences was involved in the early collection of powder diffraction pat-terns of unknown precipitates that ultimately led to the initiation of this study. Datacollection and refinement, as well as the preparation of the Crystallographic Infor-mation Framework (CIF) files, were conducted by Dr. Brian Patrick and Anita Lamin the Department of Chemistry. e theory of a possible uranous sulfate hexa-hydrate supercell structure was formulated by Dr. Patrick. e rest of the work,including the synthesis and preparation of the crystals, vibrational spectroscopy,powder x-ray diffraction, analysis, and discussion were conducted by me.Chapter . Preliminary thermogravimetric data were collected using instru-mentation at Simon Fraser University, Department of Chemistry, with the assis-tance of Dr. Dev Sharma. ese data were subsequently made redundant by theiiimore detailed studies conducted at UBC and so were not used in this dissertation.Chapter . e portion of the work concerning the oxidation of uranium(IV)in perchloric acid was presented as a conference paper at Hydro  in Victoria,BC, Canada []. All of the work was conduced by me, with supervision from Dr.David Dreisinger.Chapter . e flow sheet was developed as a part of a study for Cameco’s re-search centre titled Electrolytic reduction and precipitation of uranous sulfate: Flowsheet development, mass balance and operating cost analysis (), with supervisionfrom Dr. David Dreisinger and input from Dr. Michael Murchie and Dr. AngeloFernando at Cameco. e flow sheet portion of that report is reproduced with mi-nor modifications in this dissertation.Analysis. Most of the analytical methods were developed and conducted in-house with input from Dr. Bé Wassink. e total uranium and uranium(IV) titra-tion methods were based on ASTM standard C- []. e titration portion ofthe sulfate determination method was based on an Application Bulletin publishedby Metrohm []. e full method is available in Appendix C. Free acid determi-nation by standard addition followed Dr. Wassink’s method [], which is repro-duced with permission in Appendix D. Many of the titrations were conducted bymy undergraduate research assistants, Nicole Kosloski, Jason Midgley, and Kam-ran Rostam Sadeghi. Atomic absorption (AA) analysis was conducted by ParisaAbbasi, a research engineer in our laboratory, with supervision and help from Dr.Wassink. FTIR spectroscopy was conducted using instrumentation in the Depart-ment of Mining Engineering, with training and support from Sally Finora. Ramanspectroscopy was conduced using instrumentation in Dr. Guangrui Xia’s labora-tory, with training and support from her Master’s student Ye Zhu.ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiiList of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xviiiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxii Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview of chapters . . . . . . . . . . . . . . . . . . . . . . . . . .  Background information . . . . . . . . . . . . . . . . . . . . . . . . . . . A brief history of uranium . . . . . . . . . . . . . . . . . . . . . . . . Relevant thermodynamic quantities . . . . . . . . . . . . . . . . . . . e aqueous chemistry of uranium . . . . . . . . . . . . . . . . . . . Electrolytic reduction of uranium(VI) . . . . . . . . . . . . . . . . . Industrial processes . . . . . . . . . . . . . . . . . . . . . . . . . . v. Analytical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Total uranium . . . . . . . . . . . . . . . . . . . . . . . . . .. Uranium(IV) . . . . . . . . . . . . . . . . . . . . . . . . . .. Total sulfate . . . . . . . . . . . . . . . . . . . . . . . . . . .. Free acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Safe handling of uranium . . . . . . . . . . . . . . . . . . . . . . .  Crystallization of uranous sulfate: solubility, speed, selectivity, and form . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background information . . . . . . . . . . . . . . . . . . . . . . . . . Experimental setup and data treatment . . . . . . . . . . . . . . . . .. Series A: Slow equilibration . . . . . . . . . . . . . . . . . . .. Series B: Fast precipitation with impurities . . . . . . . . . . .. Miscellaneous tests . . . . . . . . . . . . . . . . . . . . . . .. Determining waters of hydration . . . . . . . . . . . . . . . .. Minimizing sampling error due to uranium(IV) oxidationand evaporation . . . . . . . . . . . . . . . . . . . . . . . . . Results and analysis: solubility and kinetics . . . . . . . . . . . . . . .. e effect of sulfate and temperature on uranium(IV) sol-ubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. e effect of seeding on the kinetics of precipitation . . . . . .. e effect of impurities . . . . . . . . . . . . . . . . . . . . . Results and analysis: precipitate characterization . . . . . . . . . . . .. eoretical chemical composition and x-ray patterns . . . . .. Solid phase stability under various conditions . . . . . . . . .. Precipitate quality in the presence of impurities . . . . . . . . Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . .  e crystal structures of uranous sulfate hexahydrate and octahydrateand a comparison to the other known hydrates . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Synthesis and crystallization . . . . . . . . . . . . . . . . . vi.. Data collection and refinement . . . . . . . . . . . . . . . . .. Vibrational spectroscopy . . . . . . . . . . . . . . . . . . . .. Chemical analysis . . . . . . . . . . . . . . . . . . . . . . . .. Soware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Crystal structures . . . . . . . . . . . . . . . . . . . . . . . .. Vibrational spectroscopy . . . . . . . . . . . . . . . . . . . . Discussion and comparison with other known uranium(IV) sulfatehydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Note on the observed superstructure of uranous sulfate hexahydrate . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ermal stability of uranous sulfate I: ermodynamics and theory . . . . Water loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . e SO2/SO3 equilibrium . . . . . . . . . . . . . . . . . . . . . . . . Anhydrous uranous sulfate decomposition . . . . . . . . . . . . . . . Uranous sulfate decomposition phase diagram . . . . . . . . . . . .  ermal stability of uranous sulfate II: Experimental examination . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background information . . . . . . . . . . . . . . . . . . . . . . . . . Experimental procedures and data treatment . . . . . . . . . . . . . .. Bulk drying and calcining for x-ray analysis . . . . . . . . . .. ermal analysis instrumentation and calibration . . . . . . .. TGA data treatment . . . . . . . . . . . . . . . . . . . . . . .. DSC data treatment and baseline correction . . . . . . . . . . Validation of thermal analysis method . . . . . . . . . . . . . . . . .. Selection of representative samples . . . . . . . . . . . . . . .. Choice of scan rate . . . . . . . . . . . . . . . . . . . . . . .. e effect of particle size . . . . . . . . . . . . . . . . . . . .. Reproducibility . . . . . . . . . . . . . . . . . . . . . . . . . Results: x-ray analysis of bulk sample decomposition . . . . . . . . . Results: ermal analysis . . . . . . . . . . . . . . . . . . . . . . . vii.. Decomposition in nitrogen . . . . . . . . . . . . . . . . . . .. Decomposition in air . . . . . . . . . . . . . . . . . . . . . .. Decomposition under hydrogen and ammonia . . . . . . . .. e use of isothermal holds to identify intermediary products.. Further study on the phase change in the hexahydrate . . . . Interpretation of DTG curves . . . . . . . . . . . . . . . . . . . . . .. Peak deconvolution methodology . . . . . . . . . . . . . . .. Peak assignment and interpretation . . . . . . . . . . . . . . Interpretation of DSC curves . . . . . . . . . . . . . . . . . . . . . .. Heats of transformation . . . . . . . . . . . . . . . . . . . . . Reaction kinetics during thermal decomposition . . . . . . . . . . . .. eoretical kinetics under ideal behaviour . . . . . . . . . . .. Inferring reaction kinetics from peak shape . . . . . . . . . . Analysis and mechanism proposal . . . . . . . . . . . . . . . . . . . .. An argument in support of the occurrence of a uranous sul-fate recrystallization phase transformation . . . . . . . . . . .. Proposed decomposition mechanism in nitrogen . . . . . . .. e influence of oxygen . . . . . . . . . . . . . . . . . . . . .. Estimation of reaction rates and gas-phase composition . . . .. A thermodynamic interpretation of uranous sulfate decom-position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary and conclusion . . . . . . . . . . . . . . . . . . . . . . .  e kinetics of uranium(IV) oxidation with molecular oxygen . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background information . . . . . . . . . . . . . . . . . . . . . . . . .. Oxidation with molecular oxygen in perchloric acid . . . . . .. Oxidation with molecular oxygen in sulfuric acid . . . . . . .. Tracer studies . . . . . . . . . . . . . . . . . . . . . . . . . .. Underlying reaction mechanism . . . . . . . . . . . . . . . .. Other related studies . . . . . . . . . . . . . . . . . . . . . . Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Solution preparation . . . . . . . . . . . . . . . . . . . . . . viii.. Continuous monitoring of uranium(IV) concentration byUV-Vis spectroscopy . . . . . . . . . . . . . . . . . . . . . .. Gas injection . . . . . . . . . . . . . . . . . . . . . . . . . . .. Temperature monitoring and control . . . . . . . . . . . . . . Validation of experimental method . . . . . . . . . . . . . . . . . . .. UV-Vis spectroscopy . . . . . . . . . . . . . . . . . . . . . .. Gas flow rate and stirring speed . . . . . . . . . . . . . . . . .. Evaporative losses . . . . . . . . . . . . . . . . . . . . . . . .. Reproducibility . . . . . . . . . . . . . . . . . . . . . . . . . Rate equation methodology . . . . . . . . . . . . . . . . . . . . . . . Results and discussion: oxidation in perchloric acid . . . . . . . . . .. Reaction order in uranium(VI) . . . . . . . . . . . . . . . . .. Reaction order in uranium(IV) . . . . . . . . . . . . . . . . .. Reaction order in H+ and oxygen . . . . . . . . . . . . . . .. e effect of temperature . . . . . . . . . . . . . . . . . . . .. Proposed apparent overall rate equation . . . . . . . . . . . . Results and discussion: the effect of sulfate . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Demonstration plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Solution preparation . . . . . . . . . . . . . . . . . . . . . . .. Equipment and procedure . . . . . . . . . . . . . . . . . . . Results and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . .. Electrolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Crystallization . . . . . . . . . . . . . . . . . . . . . . . . . .. Solids analysis . . . . . . . . . . . . . . . . . . . . . . . . . . Implications for plant design . . . . . . . . . . . . . . . . . . . . . .  Flow sheet development . . . . . . . . . . . . . . . . . . . . . . . . . . . Description of unit operations . . . . . . . . . . . . . . . . . . . . . .. Continuous electrolysis . . . . . . . . . . . . . . . . . . . . .. Crystallization . . . . . . . . . . . . . . . . . . . . . . . . . ix.. Drying and calcining . . . . . . . . . . . . . . . . . . . . . .. Residual uranium recovery and impurity removal . . . . . .  Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . Review of objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributions to the art . . . . . . . . . . . . . . . . . . . . . . . . . Further work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Production of uranium(IV) solutions by electrolytic reduction . . . . . B Raman, FTIR, and XRD patterns for the uranous sulfate x-hydrates . . . B. Uranous sulfate tetrahydrate . . . . . . . . . . . . . . . . . . . . . . B. Uranous sulfate hexahydrate . . . . . . . . . . . . . . . . . . . . . . B. Uranous sulfate octahydrate . . . . . . . . . . . . . . . . . . . . . . B. Parisaite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C Total sulfate determination . . . . . . . . . . . . . . . . . . . . . . . . . D Free acid determination . . . . . . . . . . . . . . . . . . . . . . . . . . E Oxidation kinetics worksheet . . . . . . . . . . . . . . . . . . . . . . . F Radioactive uranium safe handling procedures . . . . . . . . . . . . . . xList of TablesTable . Standard state thermodynamic quantities ( °C) relevant to thedecomposition of uranous sulfate. . . . . . . . . . . . . . . . . . Table . Uranium standard reduction potentials . . . . . . . . . . . . . . Table . Formation constants of aqueous uranium(IV) and uranium(VI)complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table . Hydrolysis reactions for uranium(VI) and uranium(IV) . . . . . Table . Summary of industrial processes employing electrolytic reduction Table . Experimental conditions for the precipitation of uranous sulfate . Table . Aqueous-phase impurity assays before and aer each test . . . . Table . XRD identity, chemical assays, andwaters of hydration for solidsprecipitated from pure solutions . . . . . . . . . . . . . . . . . . Table . XRD identity, chemical assays, andwaters of hydration for solidsprecipitated from solutions containing impurities . . . . . . . . . Table . eoretical mass fractions uranium and sulfate for various ura-nous sulfate hydrates . . . . . . . . . . . . . . . . . . . . . . . . Table . Single crystal x-ray diffraction experimental details . . . . . . . . Table . Assay results for uranous sulfate hexahydrate and octahydrate . . Table . Selected bond lengths and hydrogen bond lengths for uranoussulfate hexahydrate and octahydrate . . . . . . . . . . . . . . . . Table . Comparison of the normalized cell volumes, intercell connec-tivity, and sulfate binding modes of the known uranous sulfatehydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiTable . Sulfate tetrahedra angles for uranous sulfate hexahydrate andoctahydrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table . Comparison of crystal parameters for uranous sulfate hexahyd-rate sub- and super-cells . . . . . . . . . . . . . . . . . . . . . . Table . ermal treatment of the uranous sulfate hydrates at °C, °C,°C, and °C . . . . . . . . . . . . . . . . . . . . . . . . . . Table . Integrated areas under the deconvolutedDTGpeaks for the tetra-hydrate, hexahydrate, and octahydrate . . . . . . . . . . . . . . . Table . eoretical change in equivalent molecular weight correspond-ing to the losses of various molecules from a structure. . . . . . . Table . Heats of reaction for thermal events observed byDSCduring thedehydration of uranous sulfate x-hydrate . . . . . . . . . . . . . Table . Test conditions for the oxidation studies . . . . . . . . . . . . . . Table . Comparison of uranium(IV) assays by titration and continuousUV-Vis spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . Table . Demonstration plant aqueous assays . . . . . . . . . . . . . . . . Table . Demonstration plant solids assay . . . . . . . . . . . . . . . . . Table . List and description of flow sheet streams . . . . . . . . . . . . . Table D. Solution compositions for testing HSO analysis by pH electrode Table D. Analytical results for analysis of HSO–metal sulfate solutions . Table D. Calibration data for analytical results in Table D. . . . . . . . . Table D. Analytical results for analysis of HSO–metal sulfate solutionsusing a meter with  mV resolution. . . . . . . . . . . . . . . . . Table D. Calibration data for analytical results in Table D. . . . . . . . . Table F. Isotopic abundance, half-life, and emission types for natural ura-nium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiList of FiguresFigure . Pourbaix diagram of uranium in a non-complexing medium . . Figure . Schematic of the experimental setup for the crystallization ofuranous sulfate. . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . e uranium(IV) concentration over time during slow crystal-lization of uranous sulfate in a shaken vessel at different initialsulfuric acid concentrations at  °C . . . . . . . . . . . . . . . Figure . Equilibrium uranium(IV) and sulfate concentrations achievedaer shaking for – days at  °C,  °C, and  °C . . . . . Figure . e effect of seeding on crystallization kinetics at  °C . . . . . Figure . e effect of temperature and of seeding on crystallization ki-netics at  °C and  °C . . . . . . . . . . . . . . . . . . . . . . Figure . e effect of Cu, Ni, Fe, and Al on uranium recovery duringuranous sulfate precipitation . . . . . . . . . . . . . . . . . . . Figure . Uranous sulfate crystallization kinetics in the presence of copper Figure . Powder x-ray diffraction reference patterns for the known ura-nous sulfate x-hydrates . . . . . . . . . . . . . . . . . . . . . . Figure . Gravimetric analysis of uranous sulfate tetrahydrate, hexahyd-rate, octahydrate, and parisaite . . . . . . . . . . . . . . . . . . Figure . Experimentally-determined powder XRD pattern of parisaite . . Figure . Gravimetric analysis of various over-hydrated samples of ura-nous sulfate tetrahydrate . . . . . . . . . . . . . . . . . . . . . Figure . Map of the different polymorphs of uranous sulfate with respectto temperature, free acid, and test duration . . . . . . . . . . . . xiiiFigure . Schematics of the structures of uranous sulfate hexahydrate andoctahydrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . Displacement ellipsoid model depicting the extended structureof uranous sulfate hexahydrate . . . . . . . . . . . . . . . . . . Figure . Polyhedral model of uranous sulfate hexahydrate . . . . . . . . Figure . Displacement ellipsoidmodel depicting the connectivity of ura-nous sulfate octahydrate . . . . . . . . . . . . . . . . . . . . . . Figure . Polyhedral model of uranous sulfate octahydrate . . . . . . . . Figure . FTIR and Raman spectra of uranous sulfate hexahydrate andoctahydrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . Schematic of the different sulfate binding modes observed inthe known uranous sulfate hydrate complexes. . . . . . . . . . . Figure . Pseudo-precession image for uranous sulfate hexahydrate . . . . Figure . Equilibrium SO/SO ratio as a function of temperature andoxygen partial pressure . . . . . . . . . . . . . . . . . . . . . . Figure . Equilibrium SO/SO ratio as a function of temperature andpSO3 + pSO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . eoretical thermodynamic equilibrium of the decompositionof U(SO) from – °C in an inert atmosphere . . . . . . Figure . eoretical thermodynamic equilibrium of the decompositionof U(SO) from – °C in an atmosphere fixed at pO2 ≈0.209 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . Phase diagram of the U-S-O system for homogenous decom-position of U(SO) . . . . . . . . . . . . . . . . . . . . . . . . Figure . ermal water loss pathways for uranyl sulfate x-hydrate as de-termined by various authors . . . . . . . . . . . . . . . . . . . . Figure . Validation of thermal analyzer temperature and heat flow cali-bration using indium and silver . . . . . . . . . . . . . . . . . . Figure . Correction of the DSC signal . . . . . . . . . . . . . . . . . . . Figure . TGA and DSC curves of U(SO) ·HO at different scan rates xivFigure . eeffect of particle size on theTGAcurves forU(SO) ·HOand U(SO) ·HO . . . . . . . . . . . . . . . . . . . . . . . . Figure . Reproducibility of the TGA and DTA curves . . . . . . . . . . . Figure . TGA, DTG, and DSC curves for U(SO) ·HO in nitrogen . . Figure . TGA, DTG, and DSC curves for U(SO) ·HO in nitrogen . . Figure . TGA, DTG, and DSC curves for U(SO) ·HO in nitrogen . . Figure . TGA, DTG, and DSC curves for U(SO) ·HO in air . . . . . Figure . TGA, DTG, and DSC curves for U(SO) ·HO in air . . . . . Figure . TGA, DTG, and DSC curves for U(SO) ·HO in air . . . . . Figure . TGA scans of U(SO) · HO under nitrogen, air, hydrogen,and ammonia. . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . Isothermal holds at  °C and  °C under nitrogen showingwater loss for U(SO) ·HO, U(SO) ·HO, andU(SO) ·HO. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . Isothermal holds at  °C,  °C,  °C,  °C, and  °Cfor U(SO) ·HO under air and nitrogen. . . . . . . . . . . . Figure . °C isothermal holds of U(SO) ·HO in air and a nitrogen Figure . DSC signal during the sequential heating and cooling ofU(SO) ·HO across the P phase change . . . . . . . . . . . . . . . . Figure . A comparison of the raw and deconvoluted DTG signals of thethree uranous sulfate hydrates from – °C . . . . . . . . . Figure . Non-dimensionalized solutions for three possible reaction ratecontrol mechanisms as temperature is increased at a constantrate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . e thermal decompositionpathways forU(SO) ·HO,U(SO) ·HO, and U(SO) ·HO in nitrogen . . . . . . . . . . . . . Figure . Schematic of oxidation study experimental setup . . . . . . . . Figure . UV-Vis spectra for uranium(IV) and uranium(VI) in perchlo-ric acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . UV-Vis spectra for uranium(IV) and uranium(VI) in sulfuricacid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvFigure . e change in extinction coefficient observed with increasingperchloric acid concentration . . . . . . . . . . . . . . . . . . . Figure . e stability of uranium(IV) over time during the bubbling ofwater-saturated nitrogen . . . . . . . . . . . . . . . . . . . . . Figure . Oxidation rate vs. concentration plots . . . . . . . . . . . . . . Figure . First- and second-order rate plots of two tests in perchloric acid Figure . ln-ln plots of oxidation rate vs. U(IV) concentration . . . . . . Figure . e effect of H+ on the apparent rate constant in perchloric acid Figure . e effect of oxygen partial pressure on the apparent rate constant.Figure . Arrhenius plot showing temperature dependence of oxidationkinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure . Comparison of modelled and experimental data . . . . . . . . . Figure . Results from four identical oxidation tests in . N sulfuric acid Figure . e effect of adding sodium sulfate on the oxidation kinetics in. N perchloric acid . . . . . . . . . . . . . . . . . . . . . . . . Figure . e effect of adding sodium sulfate on the oxidation kinetics in. N perchloric acid . . . . . . . . . . . . . . . . . . . . . . . . Figure . Process flow diagram of the demonstration plant . . . . . . . . Figure . Electrolytic reduction of the synthetic leach solution . . . . . . Figure . Concentrations of U, Al, Fe, and Ni over the course of uranoussulfate precipitation . . . . . . . . . . . . . . . . . . . . . . . . Figure . XRD pattern for demonstration plant solids . . . . . . . . . . . Figure . ermal analysis of the demonstration plant solids . . . . . . . Figure . Proposed flow sheet for the electrolytic reduction and precipi-tation of uranous sulfate. . . . . . . . . . . . . . . . . . . . . . Figure A. Electrolyzer with submersible anode chamber. . . . . . . . . . . Figure A. Cell potential vs. time for a typical electrolysis experiment . . . Figure B. Raman spectrum of uranous sulfate tetrahydrate . . . . . . . . Figure B. FTIR spectrum of uranous sulfate tetrahydrate . . . . . . . . . . Figure B. Powder XRD spectrum of uranous sulfate tetrahydrate . . . . . xviFigure B. Raman spectrum of uranous sulfate hexahydrate . . . . . . . . Figure B. FTIR spectrum of uranous sulfate hexahydrate . . . . . . . . . Figure B. Powder XRD spectrum of uranous sulfate hexahydrate . . . . . Figure B. Raman spectrum of uranous sulfate octahydrate . . . . . . . . . Figure B. FTIR spectrum of uranous sulfate octahydrate . . . . . . . . . . Figure B. Powder XRD spectrum of uranous sulfate octahydrate . . . . . Figure B. Raman spectrum of uranous sulfate octahydrate . . . . . . . . . Figure B. FTIR spectrum of parisaite . . . . . . . . . . . . . . . . . . . . Figure B. Powder XRD spectrum of uranous sulfate octahydrate . . . . . Figure D. Calibration plot for HSO standards in  M MgSO; . mVresolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure D. Calibration plot for HSO standards in MMgSO;  mV res-olution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviiList of TermsAA atomic absorption, an analytical method for determining theconcentration of metals in solution.CIF Crystallographic Information Framework, a standard for informationinterchange in crystallography maintained by the International Union ofCrystallography.CSTR continuous stirred tank reactor, a type of reactor commonly used inchemical engineering, where the reactor is a tank with equal inflow andoutflow rates and aggressive stirring. At steady state, and in the ideal case,the composition is assumed to be uniform throughout the tank. ree orfour tanks are usually used in series.DSA dimensionally stable anode, a titanium anode coated with a proprietarymixed metal oxide consisting of metals such as iridium, ruthenium,platinum, rhodium, and tantalum. Typically used for oxygen evolution.DSC differential scanning calorimetry, a thermoanalytical technique in whichthe amount of energy required to raise the temperature of a sample ismeasured as a function of temperature. It allows for the precisedetermination of heat capacity and the heat of reaction. Oen runsimultaneously with TGA.DTA differential thermal analysis, a thermoanalytical technique similar todifferential scanning calorimetry in which the amount of energy requiredto raise the temperature of a sample is measured as a function ofxviiitemperature. It allows for the precise determination of heat capacity andthe heat of reaction. Oen run simultaneously with TGA.DTG derivative thermogravimetry, the first derivative of a TGA curve, givingthe rate of weight change. Can be used to identify simultaneous chemicalreactions. Oen gives similar data to DSC and DTA.ISE ion-selective electrode, an electrode with a membrane at the junction thatonly allows the specified ion to cross. It can be used to detect the potentialof the specified ion while reducing the effect of other ions in solution.FTIR Fourier transform infrared spectroscopy, a measurement technique forcollecting the infrared spectrum of a sample.m’ equivalent molecular weight, a quantity used in thermal analysis todenote the instantaneous average molecular weight of the material in thecrucible at any given time.parisaite A compound related to uranous sulfate that sometimes forms as ametastable intermediary during crystallization of uranous sulfate fromacid solutions. Possibly the acid salt. Named in honour of Parisa Abassi,the researcher who first produced it.TGA thermogravimetric analysis, a thermoanalytical technique in which theweight of a sample is measured as a function of temperature. Weight lossevents at a particular temperature indicate the occurrence of a chemicalreaction or the loss of a volatile component. Oen run simultaneouslywith DSC.XRD x-ray diffraction, an analytical method for investigating the structure of acrystalline solid. Powder x-ray diffraction can be used to determine theidentity of one or more crystalline phases in a powdered solid based on adatabase of known structures. Single-crystal x-ray diffraction can be usedto elucidate the molecular structure of a crystalline solid.xixAcknowledgmentsis project never would have happened without help from my very good friend,Fortuitous Circumstance. To Tai Yen, Sadan Kelebek, and Chris Pickles, who sentme to the CMP conference, where I was seated next to Chuck Edwards at dinner,who years later put me in touch with Mike Murchie, who agreed to have lunch withme, aer which he agreed to fund this project despite me ordering the most expen-sive dessert on the menu. To Jan De Bakker, who introduced me to Boyd Davis,who gave me Sam Marcusen’s email address, who years later told me, while I wasmicrowaving a vending machine lava cake, that the most important thing aboutstarting a PhD is to choose a project that is interesting even if my life studying itisn’t.Many researchers volunteered their time and equipment to this project free ofcharge. I thank Anita Lam and Brian Patrick in the Department of Chemistry,who taught me everything I know about crystal structures; Dev Sharma in the De-partment of Chemistry at Simon Fraser University, who introduced me to ther-mogravimetry and let me use his TGA; Sally Finora in the Department of MiningEngineering, who let me use her FTIR even aer blowing her pellet press shield tosmithereens; Dr. Mati Raudsepp in theDepartment of Earth, Ocean&AtmosphericSciences, who gave me full use of his laboratory and powder diffractometers, not tomention the time of his staff (particularly Jenny Lai), and never failed to show anhonest interest in my endless supply of purple-green powders; Dr. Maggie Xia andher student Ye Zhu for givingme time on the Raman spectrometer; and Bailey Kelly,Laurie Johnson, and Mike Broczkowski at Cameco’s research centre in Port Hope,Ontario, who assisted with some analysis. I also greatly appreciate the hard workof my three undergraduate summer students, Nicole Kosloski, Jason Midgley, andxxKamran Rostam Sadeghi. I hope I managed to teach them as much as they taughtme.I thank my parents, who encouraged me to pursue further education, and myaunt Joyce, who made certain that there were no barriers to doing so. I also thankAngelo Fernando and Mike Murchie at Cameco, who were just as concerned aboutmy personal wellbeing as they were about their investment. I will ever be gratefulto my supervisor, David Dreisinger, who continues to give me more responsibilityand freedom than I deserve. And most of all, I am grateful for Dr. Bé Wassink, whotaught me humanity.Funding through the National Sciences and Engineering Research Council andfrom Cameco Corporation is gratefully acknowledged.About the typefaceM P,  : Minion is an Adobe Originals typeface designed by RobertSlimbach. It was inspired by classical, old style typefaces of the late Renaissance,a period of elegant, beautiful, and highly readable type designs. Its aesthetic andfunctional qualities, with a full complement of glyphs from all Western alphabets atmultiple widths and weights, serves as a reminder that no matter how agonizinglynarrow my focus becomes, there exists a breed of professional whose obsessivenesseven I cannot approach: the professional typographer.Powered by liquid propellent!I am indebted to the forces of evolution, which gave us Coffea arabica, and to thesocial forces that built establishments to serve it to me roasted and soaked in a cup.YVR: AGRO Cafe, Beyond Coffee, East Van Roasters, Euro Bagel, Finch’s Market,Forty-ninth Parallel, Heartwood Community Cafe, Lost + Found Cafe, MatchstickCoffee Roasters, Melrichie’s, Momento Coffee House, Nelson the Seagull, Prop-house, Uprising Breads, e Wilder Snail Cafe; NYC: B Cafe, Irving Farm CoffeeRoasters, Lost Weekend, Tavern on the Green; YKA: e Art We Are, CommonGrounds, Zack’s; YYZ: Sammy’s, Bicerin; YGK: e Sleepless Goat.xxiDedicationFor LynnxxiiChapter IntroductionHe has been eight years upon a project for extracting sunbeams out ofcucumbers, which were to be put in phials hermetically sealed, and letout to warm the air in raw inclement summers.— Jonathan Swi, Gulliver’s Travels ()It takes many steps to convert uranium from its ore into the most energy-densefuel currently available to humankind. Almost every technique in the extractivemetallurgist’s toolbox has been applied to uranium extraction at one point, includ-ing acidic nitrate, chloride, and sulfate leaching, carbonate leaching, pressure leach-ing, ion exchange, solvent extraction, selective precipitation, direct fluorination,electrolysis, and calcining under oxidizing and reducing atmospheres. It is thereforerather surprising to learn that there exists one known technique that has barely beenstudied, let alone applied in practice: the electrolytic reduction and precipitation ofuranous sulfate. is is the topic of this dissertation.eDepartment ofMaterials Engineering at UBC has a distinguished history ofuranium research. In the s, Dr. Frank Forward co-developed the Beaverlodgecarbonate leach process for Eldorado, Canada’s uranium company at the time [],assisted in part by his master’s student Ernie Peters, who wrote his thesis on thesubject []. Forward also conducted research on acid pressure leaching and thehydrometallurgical production of UO for nuclear power plants [, ], and laterbecame the Director of Research at the Canadian Uranium Research Foundation.e University’s interest in uranium waned aer Forward’s departure, however, asgovernments the world over gradually assumed responsibility for research in thefield.Aer the twenty-five year slumber that followed thereeMile Island andCher-nobyl crises, nuclear energy is now enjoying a renaissance. is is partly due to thegeopolitical and economic uncertainties associated with oil, but it is also due to therise of responsible environmental stewardship: nuclear power produces virtually nowaste compared to conventional sources. Even recent setbacks associated with theFukushimameltdown seem unlikely to stop the long-term growth of nuclear power.As described by the International Atomic Energy Agency in  []:Nuclear power currently generates  of the world’s electricity. It pro-duces virtually no sulfur dioxide, particulates, nitrogen oxides, volatileorganic compounds (VOCs) or greenhouse gases (GHGs). e com-plete nuclear power chain, from resource extraction to waste disposalincluding reactor and facility construction, emits only – grams ofcarbon equivalent per kilowatt-hour (gCeq/kW.h). is is about thesame as wind and solar power including construction and componentmanufacturing. All three are two orders of magnitude below coal, oil,and natural gas (– gCeq/kW.h).Nuclear power plants require a steady supply of fuel, and mining companies areactively preparing to develop new orebodies to meet rising demand. is disserta-tion presents research associated with a novel method of uranium purification thatcould be used at a new or existing uranium mill. Specifically, it addresses variousfundamental and engineering aspects of the selective precipitation of uranous sul-fate from acidic leach solutions. If proven feasible, this technology could be usedas an alternative to solvent extraction, which is relatively costly and hazardous, andcould possibly allow for the recovery of acid in a closed-loop system, thus reducingthe requirement for acid neutralization.. Objectiveseoverarching objective of this dissertation is to advance the knowledge and prac-tice of the selective precipitation of uranous sulfate as a new uranium hydrometal-lurgical processing technology. is will be approached by focusing on six ques-tions, each addressing a specific gap in the literature.. What are the best operating conditions for the precipitation of uranoussulfate?Uranous sulfate is known to precipitate from acidic uranium(IV) solutions with theaddition of sulfuric acid and the application of heat. e relationship between ura-nous sulfate solubility, temperature, and sulfuric acid concentration is fairly wellunderstood, but nothing has been reported on optimizing the process to achievehigh recovery and fast kinetics. iswill be addressed by combining existing knowl-edge with new experimental results in order to recommend the best conditions foroperating a uranous sulfate precipitation process.. How do impurities affect the precipitation process?A successful hydrometallurgical process employing uranous sulfate precipitationmust perform well on impure solutions. No work has been reported in the litera-ture, however, on the effect of impurities on the selectivity and recovery of uranoussulfate precipitation, or on the purity of the resulting solids. is gap in the litera-ture will be addressed by discussing a series of laboratory experiments on the effectsof Al, Cu, Ni, and Fe.. What are the different uranous sulfate polymorphs, and how do they differfrom one another?Several different uranous sulfate hydrate polymorphs have been identified in the lit-erature, with each forming under a specific set of conditions. A deep understandingof the number of different polymorphs, their structures, and the conditions underwhich they form is lacking, however. New experimental data will be combined withresults from the literature to formulate a comprehensive understanding of the gen-esis and form of the uranous sulfate x-hydrates.. How does uranous sulfate respond to drying and calcining?Uranous sulfate x-hydrate will dehydrate and decompose when heated, but a pre-cise understanding of the temperatures and transitions involved is lacking. e de-composition of uranous sulfate will be discussed from both a thermodynamic andexperimental perspective.. Is aqueous uranium(IV) stable against oxidation by oxygen gas?e uranium(IV) solution required for the production of uranous sulfate may beexposed to oxygen gas at many points in a potential process, including from oxy-gen evolved at the anode during electrolytic reduction, or from air while being heldin large tanks during crystallization or during a process upset or shutdown. It istherefore important to know what measures are effective to prevent the undesiredoxidation of uranium(IV).eoxidation kinetics of uranium(IV) in perchloric acid,and in the presence of sulfate, will be discussed.. Can uranous sulfate precipitation be developed into a viable extractivemetallurgical technology?Most aspects of uranous sulfate precipitationhave been studied to some extent (if thework presented in this dissertation is included), but an entire industrial process hasnever been proposed or tested. Aproposal for a complete plant flow sheet employinguranous sulfate precipitation will be discussed, and evidence for its viability will bepresented.. Overview of chaptersis dissertation is divided into nine chapters (with this introduction being thefirst). Each chapter employs a combination of theory, new experimental data, andexisting knowledge to address the objectives given above.In Chapter , background information on the project is given, including histor-ical context, thermodynamics, aqueous chemistry, and a review of related indus-trial processes. It provides context for understanding the purpose and scope of theproject. Literature and background information relevant only to a single area of theproject is not given here, but is instead presented in the relevant chapter.In Chapter , various aspects of the precipitation process are explored, focus-ing on the effect of temperature, free acid concentration, and crystallization time.Five unique uranous sulfate x-hydrate polymorphs are identified, and each sample ismapped according to the conditions of its genesis to broadly define the stable regionsof each polymorph. e effects of various other parameters on the crystallizationprocess, such as seeding and agitation, is also explored.In Chapter , the crystal structures of two of the polymorphs identified in Chap-ter , uranous sulfate hexahydrate and octahydrate, are presented from single crystalx-ray diffraction data. Vibrational spectroscopy on the two polymorphs is also dis-cussed. e structures of these polymorphs is compared to the suite of previously-known uranous sulfate polymorphs, including U(SO) · HO, to draw connec-tions between the number of crystalline and solvent waters in the hydrated salt toits intra- and intermolecular bonding.In Chapters  and , the thermal decomposition of uranous sulfate tetrahyd-rate, hexahydrate, and octahydrate is discussed. e thermal decomposition processis first treated from a theoretical perspective, focusing on the thermodynamics ofthe system. ese predictions are then tested experimentally by thermogravimetricanalysis (TGA), differential scanning calorimetry (DSC), and derivative thermo-gravimetry (DTG). Unique decomposition paths are identified for the three hy-drates, and the existence of multiple forms of U(SO) is inferred.In Chapter , the kinetics of aqueous uranium(IV) oxidation are investigatedthrough a series of experiments in perchloric acid. e irreproducibility of the ki-netics in the presence of sulfate is also discussed.In Chapter , the viability of uranous sulfate precipitation as a processing tech-nology is experimentally demonstrated from beginning to end, including electroly-tic reduction, crystallization, washing and filtration, and calcining. It shows that anintegrated process that uses uranous sulfate precipitation is viable in principle.In Chapter , a complete flow sheet showcasing uranous sulfate precipitationis proposed. Elements of the design are informed by existing knowledge from theliterature and new knowledge presented in this dissertation. e flow sheet couldbe used as a guide to direct further research and development.Finally, Chapter  gives a summary of thework done, answers the five questionsstated in the objectives, and gives suggestions for future work.Chapter Background information. A brief history of uraniume story of uranium is young and brash, complete with politics, pride, celebrity,secrecy, espionage, war, cooperation, and coercion. No other metal has attractedsuch focus from politicians, generals, scientists, activists, and economists alike. Ageneral understanding of the story is important for anyone working in the field ofuranium metallurgy.Uranium was identified as an element in  by Martin Heinrich Klaproth, aGerman apothecary and early analytical chemist. Klaproth had in fact producedonly the oxide, not the pure element, and in  the french chemist Eugène Péligotisolated the metal by reducing uranium tetrachloride with potassium metal. Ura-nium remained a curiosity with no significant commercial use until the late thcentury, when the physicist Henri Becquerel discovered that uranium salts emittedinvisible rays, now known to be electromagnetic radiation. A flurry of scientificactivity followed, quickly leading Marie and Pierre Curie to the discovery of theelement radium.e discovery of radium, and particularly its use in cancer treatment, sparked ademand for the uranium-bearing ore from which it was extracted. e only sourceof such material was initially the tailings dump of the defunct Jáchymov uraniummine in the Czech Republic, but the rich Shinkolobwe deposit in the Belgian Congoand the Great Bear Lake deposit in northern Canada were soon developed to meetthe explosive demand. e cost of radium soared to overUS, per gram (dollars), justifying the enormous cost of extracting the minuscule amount of ra-dium found in the ore. e tailings, still rich in uranium, were discarded as waste.Eldorado, Canada’s radium company, disposed of uranium-containing waste rockwherever it could find space, including in old silos, in the Port Hope harbour, andeven as fill for nearby construction sites.Demand for uranium itself grew in  when physicists announced that nu-clear fission was theoretically possible, and that it could be used to produce a pow-erful weapon. e governments of the United States, Britain and Germany beganbuying uranium from radium producers to fuel their nuclear weapons programs.Aer WWII, when everyone in the world learned of its energy and wartime poten-tial, uranium quickly eclipsed radium in importance, and amarket for uraniumwasfinally established.e study of uranium metallurgy only truly began in the s as part of theManhattan project, but a vast research budget and the commitment and influenceof the U.S. military ensured its rapid development. By the mid-s, many aspectsof uranium metallurgy had been investigated, designed, piloted, and built into op-erating plants. Most of the processes used today were developed and piloted in the-year period following the war.e uranium industry today is a fully-developed supply chain for the many nu-clear power plants around the world. Corporations from many countries, includingCanada’s Cameco (né Eldorado), France’s Areva, Kazakhstan’s Kazatomprom, andAnglo-Australian BHP Billiton, mine and refine uranium from orebodies scatteredacross the world. From its humble position at the beginning as valueless gangue,uranium has become a critical commodity for the world energy market.Sources: [, , , , ]. Relevant thermodynamic quantitiese rapid development of nuclear technologies following the second world war cre-ated an intense need for fundamental knowledge of uranium chemistry. ere istherefore no lack of fundamental thermodynamic data related to uranium process-ing. e thermodynamic values used in this dissertation were curated by the Or-ganisation for Economic Co-operation and Development (OECD) Nuclear EnergyAgency, which maintains an internally-consistent database of the thermodynamicsof uranium, neptunium, plutonium, americium, and technetium derived from overone thousand peer-reviewed and government publications. An exhaustive review ofthe uraniumportion of the data set was compiled byGrenthe et al. [], and updatedmore recently by Guillaumont and Mompean []. A subset of the thermodynamicquantities relevant to the present work is reproduced in Table ..Table .: Standard state thermodynamic quantities relevant to the decompo-sition of uranous sulfate ( °C). Reference: Guillaumont and Mompean[].Compound State ∆G°, kJ/mol ∆H°, kJ/mol S°, J/molU(SO) ·HO c -. -. U(SO) ·HO c -. -. U(SO) c -. -. UOSO c -. -. .UO c -. -. .UO† c -. -. .SO g -. -. .SO g -. -. .HO g -. -. .O g   .U+ aq -. -. -.U+ aq -. -. -.USO+ aq -. -. -.U(SO) aq -. -. -.UO+ aq -. -. -UO+ aq -. - -.UOSO aq -. -. .UO(SO)– aq -. -. .UO(SO)– aq -.† For γ-UO. e aqueous chemistry of uraniumUranium in solution can exist in the (III), (IV), (V) and (VI) oxidation states. Ta-ble . shows the standard reduction potentials for the transitions between oxi-dation states. Uranium(VI) and uranium(IV) are both stable in water, makingTable .: Uranium standard reduction potentials. Calculated from the ther-modynamics of Guillaumont and Mompean [].Ox. State Change Half-reaction E◦, VVI−−→ V UO++ e– −−→ UO+ .VI−−→ IV UO++H++e– −−→ U++HO .V−−→ IV UO++H++ e– −−→ U++HO .IV−−→ III U++ e– −−→ U+ -.IV−−→  U++e– −−→ U -.III−−→  U++e– −−→ U -.them relevant in hydrometallurgical processes. Uranium(III) is formed at a po-tential below that of hydrogen evolution, so it is generally not found in aqueousprocesses. Uranium(V), if formed, rapidly disproportionates into uranium(IV) anduranium(VI), and so is rarely found in measurable quantities. Since uranium(III)and uranium(V) are only observed under laboratory conditions, their chemistrieswill not be discussed further. e chemistries of uranium(VI) and uranium(IV),however, are vital to any discussion about the aqueous processing of uranium.Uranium(VI) is highly soluble, is easily leached, and forms complexes with a va-riety of ligands. Uranium(IV), in contrast, is much less soluble and generally cannotbe leached without first being oxidized. Figure . shows a Pourbaix diagram of theuranium–water system, constructed usingHSC ., for a .molal uranium solu-tion, omitting complexing ligands and hydrolyzed compounds. UO+ is the stableaqueous species under oxidizing, neutral, and acidic conditions, while uranium(IV)is soluble only at very low pH. Uranium(VI) is predicted to precipitate by hydroly-sis around pH . In reality, however, uranium forms complexes readily with manysubstances, making the actual pH of precipitation higher.Like the other actinides and lanthanides, uranium complexes readily with sul-fate, which distinguishes it from most base metals. It also complexes readily withchloride, fluoride, and nitrate. Table . shows the formation constants of uraniumcomplexes with these ligands.Majima et al. [] calculated the theoretical speciation of a mixed uranyl/ura-nous sulfate system by solving the simultaneous equilibrium relationships for a so-lution containing  g L− uranium and / g L− sulfur. ey showed thatFigure .: Pourbaix diagram of uranium in a non-complexingmedium, ◦C,. molal U. Generated by HSC ..Table .: Formation constants of aqueous uranium(IV) and uranium(VI)complexes ( °C, I = 0). Reference: Guillaumont and Mompean [].log10 β ◦1 log10 β◦2 log10 β◦3 log10 β◦4U+F– . . . .Cl– .SO– . .NO– . .UO+F– . . . .Cl– . -.SO– . . .NO– .the negatively-charged uranium(VI) disulfate complex UO(SO)– is dominantover the neutral monosulfate complex UO(SO)aq at high sulfate levels, but onlysomewhat dominant at lower sulfate levels. e effect is more pronounced for ura-nium(IV), where the neutral disulfate complex U(SO) aq is dominant over themonosulfate U(SO)+ even at low sulfate levels. e authors did not include thetrisulfate uranium(VI) complex UO(SO)– in their analysis.e hydrolysis of uranium (i.e., complexation with OH–) is quite complex. Asubset of the known hydrolysis equilibria are given in Table .. Uranium(VI) formsa plethora of hydrolyzed species, but generally remains as UO+ at low pH. Ura-nium(IV), however, will hydrolyze even at pH  to form UOH+.Table .: Hydrolysis reactions for uranium(VI) and uranium(IV) ( °C, I =0). Reference: Guillaumont and Mompean [].Hydrolysis reaction log10K◦U+U+ +HO(l) ←−→ UOH+ +H+ -.U+ +OH– ←−→ U(OH)(aq) -.UO+UO+ +HO(l) ←−→ UOOH+ +H+ -.UO+ +HO(l) ←−→ UO(OH)(aq) +H+ -.UO+ +HO(l) ←−→ UO(OH)– +H+ -.UO+ +HO(l) ←−→ UO(OH)– +H+ -.UO+ +HO(l) ←−→ (UO)OH+ +H+ -.UO+ +HO(l) ←−→ (UO)(OH)+ +H+ -.UO+ +HO(l) ←−→ (UO)(OH)+ +H+ -.UO+ +HO(l) ←−→ (UO)(OH)+ +H+ -.etc.. Electrolytic reduction of uranium(VI)e electrolytic reduction of uranium(VI) to form uranium(IV) is a critical pre-cursor to the precipitation of uranous sulfate. While the electrolysis process is notdiscussed in depth in this dissertation, it is important to understand the process inorder to be aware of design constraints.e electrolytic reduction of uranium(VI) can be described by the followinghalf-cell reactions:Cathodic half-cell: UO+ +H+ +e– −−→ U+ +HO E◦ = 0.27V (.)Anodic half-cell: HO−−→ O +H+ +e– E◦ =−1.23V (.)Overall: UO+ +H+ −−→ U+ + O E◦cell =−0.96V (.)Since E◦cell is negative, the reaction does not proceed spontaneously, and so willonly occur if a voltage is applied using an external power supply.Equation (.) shows the reduction of uranium(VI) taking place as a simultane-ous two-electron transfer. e true reaction path, however, is more complex. Casa-dio and Lorenzini [] investigated the reduction of uranium(VI) at the millimolarlevel by cyclic voltammetry, and showed that the supporting electrolyte composi-tion and scan speed can cause different reaction mechanisms to dominate. eyshowed that uranium(VI) −−→ uranium(V) always proceeds by a single electrontransfer, but that uranium(V) −−→ uranium(IV) can proceed either by chemicaldisproportionation or by a second electron transfer. Under high acid conditions,they found that disproportionation dominated. ey also found that the additionof sulfate enhanced the disproportionation kinetics, likely due to the strong com-plexing power of sulfate towards uranium(IV). Kern and Orlemann [] found thatthe disproportionation of uranium(V) is extremely rapid, except at millimolar lev-els in a low-acid, non-complexing matrix, and even then it only survives for tensof seconds. Under the high-acid, high-sulfate conditions being considered in thisdissertation, the disproportionation mechanism would clearly dominate. is hasfew practical implications, however, since the ultimate result – two electrons andtwo protons consumed per uranium – is the same. It can therefore be assumed fordesign purposes that the reduction process involves a direct two-electron transfer.Gurinov and Frolov [] showed that uranium(VI) reduction is diffusion-limit-ed under typical operating conditions. In electrochemical terms, this means that anelectrolyzer designed to produce uranium(IV) will operate at the reaction’s limitingcurrent density, where an increase in cell potential does not produce an increase inreduction rate. Under these conditions, the rate-controlling step is the rate of masstransfer of uranium(VI) from the bulk electrolyte to the cathode surface. As theelectrolytic reduction of uranium(VI) proceeds, the concentration gradient betweenthe bulk solution and that in contact with the cathode surface becomes smaller,causing the rate of mass transfer to decline. Eventually, when the uranium cannottransfer from the bulk solution fast enough to consume all of the supplied current(assuming a constant-current cell is being used), hydrogen evolution will occur tomake up the difference. Hydrogen evolves according to the following half-cell re-actions:Cathodic half-cell: H+ +e– −−→H E◦ = 0V (.)Anodic half-cell: HO−−→ O +H+ +e– E◦ =−1.23VOverall: HO−−→ O +H E◦cell =−1.23V (.)Hydrogen bubbles can “mask” part of the electrode surface, reducing the avail-able surface area and consequently reducing the reaction rate. Conversely, the for-mation and release of tiny bubbles can actually increase the reaction rate by dis-turbing the solution next to the cathode and inducing convective flow to a degreenot possible by bulk turbulence alone. In theory, turbulence or agitation shoulddecrease the thickness of the mass transfer boundary layer, and thus increase thereaction rate. Gurinov and Frolov [] found that the rate of reduction can be madetens of times faster by inducing forced convection of the electrolyte. In fact, at veryhigh flow rates it became impossible to distinguish the point at which hydrogenevolution began. Awakura et al. [] showed that the limiting current increases withtemperature, as does overall current efficiency, which is consistent with a diffusion-limited process, since the diffusion coefficient increases with temperature.e diffusion coefficient (D) and boundary layer thickness are clearly essentialparameters for the design of an electrolyzer. Both of these parameters have been de-termined by several groups under various conditions. Awakura et al. [] determinedthe diffusion coefficient of uranium(VI) at several temperatures by comparing thelimiting current density to that of copper reduction, for which D is well known.ey measured an apparent diffusion coefficient of .–. × - cm s−, with ahigher value observed at higher uranium concentrations. is compares well to thediffusion coefficient determined by Casadio and Lorenzini [] at  °C of (. ±.) × - cm s− in  N KSO (pH = ). ey also concluded that the thick-ness of the mass transfer boundary layer is controlled mainly by the agitation speedunder conditions of forced convection.For the purposes of this dissertation, this brief review of the electrolytic reduc-tion of uranium(VI) can be summarized as follows:• e electrolytic reduction of uranium(VI) is mass transfer-limited.• e evolution of hydrogen bubbles can be beneficial because it induces con-vection and disturbs the mass-transfer boundary layer surrounding the cath-ode.• e mass transfer boundary layer can be made smaller, and the reaction rateincreased proportionally, by operating the electrolytic cell under conditionsof forced convection.• e actual reduction mechanism involves a single-electron transfer, followedby the disproportionation of uranium(V). However, the process can be con-sidered a two-electron transfer for design purposes.. Industrial processese precipitation of uranous sulfate has never been practiced on an industrial or pi-lot scale. However, the electrolytic reduction of uranium sulfate solutions has beentested in a number of other contexts, mainly with the goal of producing high-purityUF, and sometimes uranous sulfate precipitated as scale or silt as an unintendedconsequence. It is therefore useful to examine previous attempts to electrolyze ura-nium solutions on a large scale. eUnited States, Japan, U.K., Spain, and France alloperated pilot plants on various scales. e processes are summarized in Table ..Table .: Summary of industrial processes employing electrolytic reductionProcess Nation Year Concept Ref.UK  UFwater−−−→ UO+e−−−→ UF []Excer USA  Ore HCl/H2SO4−−−−−−−→ UO+IX/DHF−−−−−→ UOFe−/DHF−−−−−→ UF []SIMO France  UOHNO3−−−−→ UO+H2SO4−−−−→ UO+e−−−→ U+ DHF−−−→ UF []SAEC Spain  UOH2SO4−−−−→ UO+e−−−→ U+ DHF−−−→ UF []PNC Japan  Ore H2SO4−−−−→ UO+SX/HCl−−−−−→ UO+e−−−→ UClDHF−−−→ UF []DHF: dilute hydrofluoric acidSX: solvent extractionOre: uranium ore concentrateefirst attempt at an industrial electrolytic reduction process was developed bythe Imperial Chemical Industries Company of Great Britain. A patent filed in describes a process by which a uranyl fluoride solution is electrolysed to producesolid UF []. e patent specifies that the starting material must be a pure uranylfluoride solution, such as would be obtained by dissolving UF in water. e authormentions that the process could be applied to sulfate or chloride systems, althoughthat was not the objective of the invention.e Excer process was developed by American researchers at the Oak RidgeNa-tional Laboratory in  as a cost-effective way to produce high-purity UF [].e process involved ion exchange of a uranium-containing solution, stripping withHCl to create a high-purity uranyl chloride solution, electrolytic reduction and pre-cipitation in the presence of HF, filtration, and finally dehydration. e feed solu-tion could be sulfuric acid leach liquor, sulfate, or chloride concentrate, or nitrateconcentrate from solvent extraction. e electrolytic cell consisted of a mercurycathode, lead anode, and an Ionics CR- cation exchange membrane. e elec-trolysis had to be conducted at – °C because operating at a lower temperaturecaused the gelatinous UF· HO to precipitate rather than the more convenientUF· HO.e SIMO process was developed by the French organization Société UgineKuhlmann for use in the Eurochemic reprocessing plant in Mol, Belgium []. eprocess involved the dissolution of a uranium feed in nitric acid, followed by con-tact with sulfuric acid, then distillation of the nitric acid to make a uranyl sulfatesolution. e authors emphasized that it was essential to transition from the nitratesystem to the sulfate system because of its suitability to the downstream fluorina-tion process. A variety of cathode materials were tested, including platinum andtitanium, but a horizontal mercury cathode was chosen because of its resistanceto HF and its ability to absorb contaminant cations. Platinum and iridium wereused as anodes, and a polypropylene porous membrane separated the anodic andcathodic compartments of the cell. e pilot plant used three reduction cells inseries to achieve  reduction. While the process seemed to be successful, a con-sistent problem was fouling of the mercury cathode by precipitated uranous sulfate.To prevent precipitation, the feed solution had to be diluted, resulting in a lowerthroughput.e Spanish Atomic Energy Commission (SAEC) also developed an electroly-tic sulfate-based process for the production of UF []. In addition to electrolysis,their study included details on fluorination, precipitation, filtering, and drying. Avariety of electrode and cell body materials were tested for their ability to resist thecorrosive electrolyte. Monel, titanium, Hastelloy B, Hastelloy C, graphite, and leadwere tested. Platinum, palladium, zirconium, and other expensive metals were nottested because of their excessive cost if used on an industrial scale. Batch electrolyticreduction tests were conducted using a synthetic solution at  g L− HSO and g L− uranium(VI), a PVC cell, lead cathode, graphite anode, and a PVC porousdiaphragm. e electrolysis was run in three stages, with a fluorination/precipi-tation step between each reduction. e three reduction phases together achieveda . conversion to U(IV), but taking into account reoxidation and loss of en-trained mother liquor, an overall uranium recovery of . was achieved. A sec-ond study was conducted to investigate the influence of sodium and chloride. Asolution containing  g L− HCl,  g L− HSO and  g L− UO was reducedin a batch electrolyzer using lead electrodes. An unidentified white precipitate, as-sumed to be lead sulfate, precipitated in the cell. e authors speculated that chlo-ride reacted with the lead cathode to form a soluble lead chloride, which in turnreacted with sulfate to form an insoluble sulfate, thereby freeing the chloride ion tocontinue to oxidize the lead cathode. e authors concluded that the catholytemustbe totally free of Cl–. Sodium was not found to interfere with electrolysis, but theauthors acknowledged that it might interfere with UF precipitation. A large-scalepilot plant was operated for twomonths to test the process. e plant was apparentlyquite successful and few problems were reported, but it was necessary to clean thecells periodically to remove a build-up of sludge, assumed to be composed of UF,lead oxide, and other metallic oxides, but perhaps also containing uranous sulfate.e authors suggested that a -week cell cleaning rotation would suffice to keep theplant operating continuously.Japan’s PNC operated a chloride-based batch electrolytic reduction pilot plantfor the production of UF in Ningyo-toge as part of their uranium enrichment re-search program in the s []. No publicly-available information on the projecthas been published. However, the researchers did publish a series of papers in thescientific literature regarding electrolytic reduction and precipitation in the sulfatesystem (reviewed in Chapter ), apparently as a precursor to converting the plantfrom the chloride to the sulfate system. It is unlikely that the conversion ever tookplace, however, since work at the Ningyo-toge site was discontinued in .. Analytical methodsMany different analytical methods were used to gather and interpret experimen-tal data for this dissertation. X-ray diffraction, Raman spectroscopy, infrared spec-troscopy, and atomic absorption spectroscopywere all used, but they arewell knowntechniques and do not require further explanation here. e determinations of to-tal uranium, uranium(IV), total sulfate, and free acid used non-standard or lesser-known techniques, and so are described below. ermogravimetric analysis (TGA)and differential scanning calorimetry (DSC) are described in detail in Chapter ... Total uraniumIf the sample was a solid, it was first dissolved in nitric acid. Aqueous sampleswere used directly. Total uranium was determined using a variation of the well-established Davies and Gray method, as described by ASTM standard C-[, ]. e potassium dichromate titrant was periodically validated against a ura-nium standard solution (AccuTrace AAN-,  µgmL−).In the modified Davies and Gray method, an aliquot containing - mg ofuranium is quantitatively transferred to a beaker, where it is combined with phos-phoric acid and an excess of ferrous sulfate. e ferrous sulfate reduces all of theuranium to uranium(IV). e residual ferrous sulfate is selectively oxidized usingnitric acid, catalyzed by molybdenum. Sulfamic acid shields the uranium(IV) fromoxidation by the nitric acid. e solution is rapidly titrated with potassium dichro-mate using potentiometric endpoint detection on a platinum electrode against astandard calomel reference electrode.For a detailed description of the reagents, equipment, and procedure, refer toASTM standard C- []... Uranium(IV)Uranium(IV) was determined by direct oxidative potentiometric titration with po-tassium dichromate in phosphoric acid. e procedure was similar to the modifiedDavies and Gray method described above, except that the ferrous reduction stepand the associated reagent additions were not done. A small amount of ferric sul-fate was added just before each titration in order to match the conditions of thewell-tested Davies and Gray method (which includes residual ferric). Failure to addferric made it difficult to detect the endpoint, suggesting that the presence of theferric-ferrous couple is required for detecting the uranium(IV)-uranium(VI) cou-ple on a platinum electrode... Total sulfateTotal sulfate was determined by potentiometric titration with lead perchlorate in isopropanol using a Pb++ ion-selective electrode. Lead sulfate is insoluble inisopropanol, so the titration endpoints were indicated by a sharp electrode responsecorresponding to the appearance of unprecipitated Pb++ ions in solution. Uraniumwas found to interfere with the electrode response, so it was first removed by hydro-gen peroxide precipitation at ∼pH . Titrations were conducted in  isopropanolto reduce the solubility of lead sulfate to negligible levels.is method is based on a method published by Metrohm []. A full descrip-tion of the method can be found in Appendix C... Free acidFree acid could not be determined by neutralization because some metals in so-lution, most notably iron, would hydrolize and precipitate before the equivalencepoint was reached. Instead, free acid was determined by the method of standardaddition, using a method developed internally by Dr. Bé Wassink []. A copy ofDr. Wassink’s method is provided in Appendix D, with permission.In the method of standard addition, a pH probe is used to measure the solutionpotential before and aer the addition of a known quantity of sulfuric acid. esevalues are used to solve two simultaneous Nernst equations, which gives the initialacid concentration. e electrode slopemust be known precisely, and is determinedimmediately before the analysis using standard solutions. Both the sample and thecalibration standards are prepared in a matrix of . M magnesium sulfate, whichprovides a very strong and relatively constant background ionic strength betweensamples. e potential is allowed to fully stabilize before readings are taken. etitration is performed using a pH meter with a precision of .mV. e additionsof standard acid can be performed using an automatic or manual pipette.. Safe handling of uraniumUranium is an alpha-emitting radioactive substance, and thus requires special han-dling procedures beyond what is typical in a metallurgical laboratory. Special pre-cautions undertaken for this project include radiation safety training for everyoneworking in the laboratory, shielded sample storage, rigorous housekeeping stan-dards, regular monitoring of radiation levels, and special waste disposal arrange-ments. Details on the safe handling of uranium for this project can be found inAppendix F.Chapter Crystallization of uranous sulfate:solubility, speed, selectivity, and form. IntroductionUranous sulfate hydrate, U(SO) · xHO, is sparingly soluble under acidic condi-tions, although its solubility varies widely with temperature and sulfate concentra-tion. It can be crystallized out of solution by adding sulfate (typically as sulfuricacid) or by increasing the temperature, although a high recovery is by no meansguaranteed. It also does not necessarily result in a quickly-forming, easily-handledprecipitate. If le undisturbed, uranous sulfate crystals tend to grow from super-saturated solutions slowly – very slowly – onto preexisting nucleation sites. eresulting large, purple-green crystals make for an excellent show-and-tell piece, buta rather poor hydrometallurgical process. To be useful in the plant, uranous sulfatemust crystallize quickly and selectively, with high uranium recovery, into a precip-itate with a known composition and good handling characteristics.In the present work, four aspects of the uranous sulfate x-hydrate crystalliza-tion process were investigated: solubility under a wide range of temperatures andsulfuric acid concentrations; the kinetics of crystallization at different temperatures( °C,  °C, and  °C) with stirring and seeding; the purity of precipitates formedfrom solutions containing Cu, Ni, Fe(II), and Al; and the crystalline form of the pre-cipitate (i.e., the value of x) under different conditions.. Background informatione fully oxidized form of uranium, uranium(VI), is very soluble in sulfuric acid.e reduced form, uranium(IV), is much less soluble, and crystallizes from sulfuricacid solutions as the hydrated sulfate salt, U(SO) · xHO. Electrolytic reductioncan be used to convert a highly-concentrated (and fully soluble) uranium(VI) sulfatesolution into a supersaturated uranium(IV) solution without the addition of anyreagents [, , ].A cornucopia of U(SO) · xHO polymorphs have been described in the liter-ature, ranging from x = 0 to x = 9 [, , , , ]. Of these, the most stable isthought to be Kierkegaard’s U(SO) · HO, which has four waters of hydration.e octahydrate, U(SO) · HO, has classically been considered the other stablepolymorph of uranous sulfate, forming at lower temperatures (< °C) []. Vari-ous authors [, , ] have shown that the octahydrate displays normal solubilitycharacteristics, becoming more soluble with increasing temperature, but the tetra-hydrate displays inverse solubility characteristics, becoming less soluble upon heat-ing. is leads to the unusual, and potentially useful, situation where uranium(IV)can be precipitated from solution by increasing the temperature and adding sulfate.is is particularly felicitous in the context of selectivity, since many of the contam-inants expected in leach solutions become more soluble at higher temperature andacid concentration.e most thorough study of the solubility of uranous sulfate under plant-typeconditions was conducted by Suzuki et al. [] of the Japanese Power Corporation.ey studied the solubility of uranous sulfate by heating acidic uranous sulfate solu-tions for several weeks while periodically assaying the supernatant for uranium(IV)and sulfate. e group also approached equilibrium from the other direction bydissolving an excess of solid uranous sulfate in sulfuric acid solutions, with ex-cellent agreement with the precipitation studies. e authors acknowledged thatsolids with a form different than U(SO) · HO precipitated at lower tempera-tures, but they did not investigate this in depth. ree important conclusions can bedrawn from their work: that solubility decreases with temperature and free sulfate;that precipitation can be very slow, oen taking a week or more to reach equilib-rium, with faster kinetics at higher temperatures; and that precipitates other thanU(SO) ·HO can form at lower temperatures.Virtually no data have been published on the effect of impurities on the solu-bility of uranous sulfate or the selectivity of its precipitation. e effect of fluoridehas been studied in the context of direct uranous tetrafluoride precipitation from asulfate medium [], where it was found that uranium(IV) solubility was enhancedby the presence of fluoride up to a F:U molar ratio of :, beyond which solubilitydeclined. e effect of common leach solution impurities, such as Cu, Ni, Fe(II),and Al, have not been reported.. Experimental setup and data treatmentAll of the experiments described in this section proceeded in a similar manner. Aquantity of previously-prepared uranium(IV) stock solution was measured into asealed vessel, typically a mL Pyrex bottle, along with sulfuric acid and metalsulfate salts, if required, to make up the required test solution. e solutions weremade on a mass basis, not volumetrically, in order to minimize oxidation of ura-nium(IV) by contact with air, so the starting uranium concentrations varied slightlyfrom test to test. Deionized water was used to make all solutions using ACS-gradereagents. e exception was uranium, which was supplied from Cameco Corpora-tion’s Blind River Refinery as granulated nuclear fuel-grade UO. All test solutionswere purged with nitrogen or argon to minimize the risk of uranium(IV) oxidation.If periodic sampling was required, an Omnifit Q-series cap with sampling ports wasused, and the vessel was purged with inert gas aer each sample was taken. All testswere thermostatted in a ermo Scientific SWB shaking water bath, with an at-tached Isotemp  circulating chiller for cooling. For stirred tests, the solutionwas agitated with an immersible magnetic stir plate and a PTFE stir bar. For stag-nant tests, no agitation was used. A schematic of the general experimental setup isshown in Fig. ..e procedure for the production of the uranium(IV) stock solutions by elec-trolytic reduction can be found in Appendix A.Aer each test, the contents of the vessels were filtered through an OsmonicsTstagnant magnetically stirred shakenFigure .: Schematic of the experimental setup for the crystallization of ura-nous sulfate.nylon . µm filter. e filtrate was stored in a polypropylene bottle, and analyzedfor uranium, sulfate, and impurity metals, as required. e solids were washed witha sulfuric acid solution at the same temperature and concentration used for the test,and then rinsed with ethanol or isopropanol. e solids were then air-dried in afume hood overnight, or until the odour of alcohol was no longer evident, thenstored in a plastic jar. All precipitates were identified by powder x-ray diffraction(XRD), and then digested in . nitric acid and analyzed for total uranium andsulfate. In tests containing impurities (Al, Cu, Fe(II), or Ni), these elements wereassayed by atomic absorption (AA).e test solutions were always slightly higher in sulfate than in HSOfree, withtwo extra moles of sulfate per mole of uranium. is was because of the acid con-sumption inherent in the the dissolution of UO and the electrolytic reduction pro-cess. Likewise, if the solution was created by dissolving previously-prepared ura-nous sulfate, it also resulted in the dissolution of two additional moles of sulfate permole of uranium. To avoid confusion, the concentrations of both sulfate and freesulfuric acid are tabulated.e experimental conditions for all tests are shown in Table .. e experi-mentswere divided into several different series, eachwith different goals and slightlydifferent procedures, as described below.Table .: Experimental conditions for the precipitation of uranous sulfateInitial test conditions, mol L−1Test T, °C HSOfree SO–T UT U(IV) Impurity Duration, h Agitation Seeded?-A*  . . . . -  shake no-A*  . . . . -  shake no-A  . . . . -  shake no-A  . . . . -  shake no-A  . . . . -  shake no-A  . . . . -  shake no-A*  . . . . -  shake no-A*  . . . . -  shake no-A  . . . . -  shake no-A*  . . . . -  shake no-A  . . . . -  shake no-A  . . . . -  shake yes-A  . . . . -  shake no-A  . . . . -  shake noContinued on next pageTable . – continued from previous pageInitial test conditions, mol L−1Test T, °C HSOfree SO–T UT U(IV) Impurity Duration, h Agitation Seeded?-A*  . . . . -  shake no-A  . . . . -  shake no-A  . . . . -  shake no-A  . . . . -  shake yes-A  . . . . -  shake no-A  . . . . -  shake yes-A  . . . . -  shake yes-A  . . . . -  shake no-A  . . . . -  shake yes-A  . . . . -  shake no-A  . . . . -  shake no-A  . . . . -  shake no-A  . . . . -  shake noContinued on next pageTable . – continued from previous pageInitial test conditions, mol L−1Test T, °C HSOfree SO–T UT U(IV) Impurity Duration, h Agitation Seeded?-B  . . . . -  stir no-B  . . . . -  stir no-B  . . . . -  stir no-B*  . . . . -  stir no-B  . . . . -  stir no-B  . . . . -  stir yes-B  . . . . -  stir no-B  . . . . -  stir yes-B  . . . . -  stir yes-B  . . . . Cu: .E-  stir no-B  . . . . Cu: .E-  stir no-B  . . . . Cu: .E-  stir no-B*  . . . . Cu: .E-  stir no-B  . . . . Cu: .E-  stir noContinued on next pageTable . – continued from previous pageInitial test conditions, mol L−1Test T, °C HSOfree SO–T UT U(IV) Impurity Duration, h Agitation Seeded?-B*  . . . . Ni: .E-  stir no-B  . . . . Ni: .E-  stir no-B  . . . . Ni: .E-  stir no-B†  . . . . Ni: .E-  stir no-B  . . . . Ni: .E-  stir no-B  . . . . Fe: .E-  stir no-B  . . . . Fe: .E-  stir no-B†  . . . . Fe: .E-  stir no-B*  . . . . Fe: .E-  stir no-B  . . . . Fe: .E-  stir no-B  . . . . Al: .E-  stir no-B  . . . . Al: .E-  stir noContinued on next pageTable . – continued from previous pageInitial test conditions, mol L−1Test T, °C HSOfree SO–T UT U(IV) Impurity Duration, h Agitation Seeded?-B  . . . .Al: .E- stir yesCu: <.E-Fe: .E-Ni: .E--S  . . . . -  stir no-S  . . . . -  stir yes-S  . . . . -  stir no-S  . . . . -  stagnant yes-S  . . . . -  stagnant yes-S  . . . . -  stagnant no-S  . . . . -  stagnant yes-S  . . . . -  stir no-S  . . . . -  stir yes* no precipitation was observed† precipitation was observed, but there was too little sample for analysis.. Series A: Slow equilibratione goal of Series A was to explore how temperature and acid concentration influ-ence the crystalline form of the precipitate, as well as to validate their effect on solu-bility and precipitation kinetics. e solutions were allowed to come to equilibriumover a long period of time (weeks or months), which generated large, well-definedcrystals. e tests were generally terminated when the uranium(IV) concentrationdid not decline significantly between samples. Shaking at  rpm provided gentleagitation, but was not strong enough to suspend the precipitates, and so the crystalsgrew on the bottom and sides of the vessel. e water bath was set to either  °C, °C, or  °C. e stock uranium(IV) solution was created by dissolving UOin sulfuric acid, followed by electrolytic reduction in a divided cell using a coppercathode, dimensionally stable anode (DSA), and a Nafionmembrane. Nitrogen wasused as the inert atmosphere.Series A tests are indicated with the suffix ‘-A’... Series B: Fast precipitation with impuritiese goal of Series B was to study the effect on precipitate purity and the crystalliza-tion process of four impurities known to be common in Saskatchewan’s uraniumore: Al, Cu, Fe(II), and Ni. Crystallization was encouraged to proceed as quicklyas possible by using vigorous magnetic stirring and high temperatures ( °C and °C). e impurities were all added as their sulfate salts. Each was dissolved atroom temperature in a minimal amount of water and sulfuric acid to avoid unnec-essary dilution, and then mixed with uranium(IV) stock solution to create a su-persaturated uranous sulfate solution when brought to the test temperature. In thecase of iron, Fe(II) was used to simulate the likely form aer undergoing electrolyticreduction. In some cases, sodium sulfate was also added to hold the sulfate concen-tration approximately constant across tests, while in others the sulfate concentrationwas allowed to vary.e stock uranium(IV) solution was created by dissolving previously-preparedU(SO) · HO in sulfuric acid, and was stored in a polypropylene bottle underargon to prevent oxidation. e uranium(IV) content of the stock solution was pe-riodically verified by titration. e impurities M were added such that the molarM:U ratio varied between . and .. e tests were generally operated for  h,although if a test did not show signs of precipitation, it was allowed to continuelonger, and some tests were terminated with no sign of precipitation.Samples of the supernatant were taken daily, and were analyzed for uranium bytitration and impurities by AA. e tests were flushed with argon during sampling.Series B tests are indicated with the suffix ‘-B’... Miscellaneous testsSome additional experiments were conducted in order to produce bulk amountsof uranous sulfate, typically for use in generating uranium(IV) stock solutions forother purposes. e tests were carried out inmuch the sameway as described above,but usually in larger quantities. Additionally, small amounts of uranous sulfate wereproduced for the work presented in Chapter , primarily with the goal of growingcrystals for characterization by single crystal x-ray diffraction. All of the precipi-tates from these experiments were collected and analyzed in the same way as thosefrom Series A and B, and the results have been included in the discussion whereappropriate.ese miscellaneous tests are indicated with the suffix ‘-S’... Determining waters of hydrationWaters of hydration (i.e., x in U(SO) · xHO were determined from TGA data.A complete description of thermal analysis and data processing methods are givenin Chapter . In summary, the number of waters of hydration was calculated bynormalizing the data to weight of its final decomposition product (UO at  °C),and back-calculating the value of x in U(SO) · xHO to account for the observedinitial weight.x =m′0−MWU(SO4)2MWH2O(.)where m′ 0 is the ‘equivalent molecular weight’ of the dry room temperaturesample, based on the normalization procedure. In some cases, partial weight losswas observed at uncharacteristically low temperatures of < °C, probably indicat-ing the existence of absorbed, not crystalline, water. ese occurrences are indi-cated.Only a subset of the solid samples were analyzed by TGA owing to the longduration of each determination (approximately  hours each)... Minimizing sampling error due to uranium(IV) oxidation andevaporationWhile all analytical methods suffer from error to some degree, the determination ofuranium(IV) presents a particular challenge due to the ease with which it reoxidizesin air. Tominimize exposure to air, sampleswerewithdrawnusing a syringe througha downcomer tube permanently installed in the lid of the reaction vessel, while alow-pressure stream of nitrogen was attached a second port in the lid. is hadthe effect of replacing the volume lost to sampling with inert nitrogen. Sampleswere diluted and analyzed as soon as possible following withdrawal to minimizeair exposure time. A certain amount of waiting time was nevertheless necessaryin order to allow the samples to cool to avoid volume-measurement errors relatedto thermal expansion. In later tests, the sampling syringe was placed into an ice-cooled holder in order to more quickly lower the sample’s temperature, speeding upthis process.Evaporative losses were minimized by using sealed reaction vessels, with thepressure equalized by temporarily releasing the sealing stopper several times duringthe initial heat-up period.. Results and analysis: solubility and kinetics.. e effect of sulfate and temperature on uranium(IV) solubilityIt has already been established by Suzuki et al. [] that the solubility of uranium(IV)decreases with increased sulfate concentration and at higher temperatures. eseresults were validated through two sets of experiments, where equilibrium was ap-proached slowly from shaken supersaturated solutions (Series A). As expected, thesolubility of uranous sulfate at  °C declined with higher sulfate concentration, asdid the time required to reach equilibrium, as shown in Fig. .. e initial ura-Figure .: e uranium(IV) concentration over time during slow crystalliza-tion of uranous sulfate in a shaken vessel at different initial sulfuric acidconcentrations at  °C (initial sulfate concentrations shown in parenthe-ses). Tests, from top to bottom: -A, -A, -A, -A, -A.nium(IV) concentration in all tests was .×− mol L−, and the free acid (andsulfate) concentrationwas adjustedwith concentratedHSO. At the highest sulfateconcentration tested, with an initial concentration of .mol L− initial SO–T(.mol l− HSOfree), the solution reached equilibrium at .×− mol L−uranium(IV). At the lowest acid concentration tested, .mol L− initial SO–T(.mol l− HSOfree), the uranium concentration declined much less to only.×− mol L− in the same length of time, although it was not clear whetherequilibrium had yet been established.Similar tests were conducted at  °C and  °C in order to confirm the rela-tionship between uranium(IV) solubility, sulfate concentration, and temperature.Figure . shows the equilibrium uranium(IV) and SO–T concentrations reachedaer – days for tests at  °C,  °C, and  °C. e data align reasonably wellwith the solubility curves reported by Suzuki et al, and extend their work to lowertemperatures and higher acid concentrations.Some of the tests did not show any signs of precipitation, even though referenceFigure .: Equilibriumuranium(IV) and sulfate concentrations achieved aershaking for – days at  °C,  °C, and  °C. e solubility curvesfrom Suzuki et al. [] are shown for reference. Tests: -A,-A, -A,-A, -A, -A, -A, -A, -A, -A, -A, -A, -A.to the solubility curves by Suzuki et al suggests that precipitation should have oc-curred. is may have been a (very long) example of an induction period related toa barrier to homogeneous nucleation... e effect of seeding on the kinetics of precipitatione kinetics of precipitation are driven by two phenomena: nucleation, which isthe formation of a new crystal, and growth, which is the addition of additionalatoms to an existing crystal. Many of the precipitation tests had an induction period,where initially no precipitation occurred for a period extending from several hourstomany days. e induction period was particularly protracted under conditions ofhigher solubility, such as at lower temperatures or lower HSOfree concentration –in other words, tests with less of a driving force to precipitate. is suggests that theinduction period was reflective of a barrier to homogeneous nucleation, and that itslength could be influenced by the availability of nucleation sites for heterogeneousnucleation, or other uranous sulfate particles for crystal growth.It would be preferable from a process engineering standpoint if precipitationwere to occur immediately and quickly. In theory, adding seed uranous sulfateparticles would provide sites for heterogeneous nucleation and crystal growth, thusavoiding the induction period. To test this theory, . g of previously-prepared finepowdered U(SO) · HO precipitate was added to one of two identical tests at °C using rapid magnetic stirring to create a suspension. Figure . shows thatthe induction period all but vanished in the presence of seed, reducing the timeneeded to reach equilibrium in a stirred vessel by several hours in an -hour test.e ultimate concentration reached was the same whether seeded or not.e addition of seed was found not only to shorten or eliminate the inductionperiod, but also to enhance the kinetics of precipitation even aer the inductionperiod had been overcome. Figure . shows how seeded tests at  °C and  °Cproceededmore quickly than identical unseeded tests, even though the shaking agi-tation was not aggressive enough to create a suspended slurry. In this case, the layerof fine seedmaterial settled on the bottom of the vessel, providing a layer of materialfor nucleation and growth. It is likely that the particle size of the seed would havea significant effect on the precipitation kinetics, with a smaller particle size offeringmore surfaces for growth, and therefore faster kinetics... e effect of impuritiesWithout exception, the concentration of impurity was virtually unchanged over theduration of a precipitation test (see Table .), suggesting that no impurities precipi-tated alongside the uranium sulfate. However, the presence of impurities did appearto increase the solubility of uranium(IV). Figure . shows how various M:U molarratios influenced the amount of uranium remaining in solution aer  hours. Inall cases, the uranium recovery declined as the amount of impurity was increased1.In other words, the presence of impurities had no deleterious effect on selectivity,but did worsen recovery.It is worth noting that the four impurities tested (Al, Cu, Fe, and Ni) all havesubstantially smaller ionic radii than uranium, as well as very different chemistries,and so would not be expected to substitute for uranium in the crystal structure of1Note that the starting uranium concentrations were slightly different in each test because thesolutions were prepared by mass, not volumetrically, to minimize exposure to oxygen during mixing.Figure .: e effect of seeding with . g of U(SO) · HO on crystalliza-tion kinetics.  °C, magnetic stirring. Tests: -B, -B.Figure .: e effect of temperature and of seeding with . g of U(SO) ·HO on crystallization kinetics. Initial composition: . M [SO–]T,. M [HSO]free. Flasks agitated on a shaking table at  shakes perminute. Final sulfate concentrations indicated as boxed number. Tests:-A, -A, -A, -A.uranous sulfate. Other actinides and lanthanides, however, do have similar atomicradii and chemistries to uranium, and thus might be more problematic. For exam-ple, both thorium and cerium are known to form M(SO).HO salts.e tests with the highest impurity concentrations did not precipitate at all. It isunclear whether the apparent increase in solubility was due to an abnormally longinduction period, or whether the solubility of uranium(IV) was enhanced by thepresence of the impurity. ese cases are marked with *. All other tests showedevidence that equilibrium, or something close to it, had been reached.Table .: Aqueous-phase impurity assays before and aer each test. Sulfatewas adjusted to approximately .mol L− with sodium sulfate (actual con-centration varied from .–.mol L−).Test Impurity M [M]t=0, mol L− [M]t=72h, mol L−-B Cu .E- .E--B Cu .E- .E--B Cu .E- .E--B Cu .E- .E--B Cu .E- .E--B Ni .E- .E--B Ni .E- .E--B Ni .E- .E--B Ni .E- .E--B Ni .E- .E--B Fe .E- .E--B Fe .E- .E--B Fe .E- .E--B Fe .E- .E--B Fe .E- .E--B Al .E- .E--B Al .E- .E--BAl .E- .E-Cu <.E- .E-Fe .E- .E-Ni .E- .E-is apparent increase in the solubility of uranium(IV) with the addition of im-purities does not have an obvious explanation. One culprit could be the increasedionic strength of the solutions, since that could have made the solutions more hos-pitable to highly-charged ions such as U+. However, sodium sulfate was added tothe tests to equalize their sulfate concentrations, so the ionic strengths were nearlythe same. Another explanation could be that sulfate was ‘tied up’ in metal-sulfatecomplexes, lowering the amount of sulfate available to uranium(IV).is also seemsunlikely, however, since the sulfate concentration was at least an order of magnitudegreater than both the uranium and impurity concentrations.It is possible that the impurities somehow decreased the kinetics of the crystal-lization process, and that the observed decline in recovery was in fact only a failureto reach equilibrium in the  h period allowed for each test. It has already beenshown that a substantial induction period can precede precipitation, so this is cer-tainly possible. is theory was tested by taking samples of the supernatant dailyduring the copper tests to follow the uranium concentration over time. Figure .shows that there was no induction period at low copper, a -hour induction periodat medium copper, and no precipitation at all at high copper. It seems, then, thatthe length of the induction period did indeed increase in the presence of copper,but that equilibrium was nevertheless reached aer  h. Similar results for iron didnot serve to clarify the situation further.It is apparent that Cu, Ni, Fe, and Al all act to suppress crystallization of uranoussulfate in some way, though it is unclear whether it is a kinetic or thermodynamiclimitation. Equation (.) shows the general relationship between the observed sol-ubility with the activity coefficientsα and the ostensibly-constant solubility productksp:αU4+ [U4+]α2SO2+4 [SO2+4 ]2 = ksp[U4+][SO2+4 ]2 =kspαU4+α2SO2+4= Ksp(.)If the presence of impuritymetals increases the activity coefficients, the apparentsolubility product Ksp would show a corresponding increase. It is unclear, however,what might cause such a change in the activity coefficients. e sulfate concen-tration of every test was adjusted to approximately mol L− with sodium sulfate,Figure .: e effect of impurities on uranium recovery during uranous sul-fate precipitation. From top to bottom, Cu, Ni, Fe, and Al. Bars show theaqueous uranium concentration at  and  hours. Cu, Ni and Fe tests: °C, . M SO, . N HSOfree. Al tests:  °C, . M SO, . NHSOfree. Tests: -B, -B through -B.Figure .: Uranous sulfate crystallization kinetics in the presence of copper. °C, ∼.mol L− free sulfuric acid, agitation by stirring. Tests: -B,-B, -B, -B.so it is unlikely that the effect can be explained on the grounds of differing ionicstrength. One possibility is that some impurities form a bisulfate complex, thuslowering the free acid concentration and as a consequence increasing the solubilityof uranium(IV). In any case, it seems that these metals make a passive contributionon uranium(IV) solubility, since their concentrations did not change as uraniumdropped from solution. e effects of impurities on the solids themselves are dis-cussed in the next section.. Results and analysis: precipitate characterizationediscussion thus far has focused on aqueous-phase phenomena, relating the con-centration of uranium in solution to various parameters. It is equally important,however, to look at the solid precipitates themselves.All of precipitates generated over the course of work were collected and ana-lyzed, providing there was enough material to allow it. In the shaken or stagnanttests, large crystals tended to grow slowly on the bottom of the vessel, whereas inthe stirred tests a finer powder-like precipitate formed. e fine precipitates settledrapidly, usually within a few seconds, when stirring was stopped. In all,  differ-ent solid samples were characterized. Chemical and XRD analysis results for thesolids derived from pure solutions are shown in Table ., and those from solutionscontaining impurities in Table .... eoretical chemical composition and x-ray patternse theoretical chemical composition of the different uranous sulfate hydrates caneasily be calculated from stoichiometry, as shown in Table .. A simple way ofdetermining a precipitate’s most likely crystalline form is to compare its chemicalassays to these theoretical values.Powder XRD offers a more direct method of identifying the crystal structureof an unknown precipitate. Figure . shows the theoretical powder x-ray patternsfor the various known uranous sulfate hydrates generated from published Crystall-ographic Information Framework (CIF) files (disclaimer: one of these publicationswas authored using data presented in Chapter  of this dissertation). e XRD pat-terns are different enough from one another (and from the pattern of any otherknown substance, for that matter) to easily distinguish one structure from anotherin experimental x-ray data, with one exception. e pentahydrate, U(SO) ·HO,is virtually indistinguishable from the tetrahydrate. In the pentahydrate identifiedby Schnaars and Wilson [], the additional interstitial water molecule rests in va-cancies within the planes of cross-linked U(SO) ·HO monomers, which appar-ently has no effect on its XRD pattern. is means that supplementary analysis isrequired to be certain of the amount of water present in solids nominally showingthe pattern of U(SO) ·HO.Table .: XRD identity, chemical assays, and waters of hydration for solidsprecipitated from pure solutions. Waters of hydration determined by TGA.Note that not all tests listed in Table . yielded enough solid sample foranalysis.XRDIdentitySolids assay, mass  Waters ofhydrationTest  U SO-A U(SO) ·HO . .-A U(SO) ·HO . . .-A U(SO) ·HO . . .-A U(SO) ·HO . .-A U(SO) ·HO . .-A mix* . . .-A mix* . .-A mix* . .-A mix* . . .-A U(SO) ·HO . .-A U(SO) ·HO . .-A U(SO) ·HO . .-A U(SO) ·HO . . .-A U(SO) ·HO . . .-A U(SO) ·HO . .-A U(SO) ·HO . . .-A U(SO) ·HO . . .-A U(SO) ·HO . .-A U(SO) ·HO . . .-A unknown . .-A U(SO) ·HO . . .-B U(SO) ·HO . . .-B parisaite . .Continued on next pageTable . – continued from previous pageXRDIdentitySolids assay, mass  Waters ofhydrationTest  U SO-B parisaite . .-B parisaite . .-B U(SO) ·HO . .-B U(SO) ·HO . .-B U(SO) ·HO . . .-B U(SO) ·HO . .-S U(SO) ·HO . . .-S mix† . . .-S U(SO) ·HO . . .-S U(SO) ·HO NES NES .-S U(SO) ·HO . NES .-S U(SO) ·HO . .-S U(SO) ·HO . . .* mixture of U(SO) ·HO and U(SO) ·HO† mixture of U(SO) ·HO and U(SO) ·HOTable .: XRD identity, chemical assays, and waters of hydration for solids precipitated from solutions containing im-purities. Waters of hydration determined by TGA. Note that not all tests listed in Table . yielded enough solidsample for analysis.XRDIdentitySolids assay, mass  Waters ofhydrationTest  U SO Al Cu Fe Ni-B parisaite . . - <. - --B parisaite . . - . - --B U(SO) ·HO . . - . - - .*-B parisaite . . - - - <.-B U(SO) ·HO . . - - - <. .*-B parisaite . . - - <. --B U(SO) ·HO . . - - . --B NES . NES <. - - --B U(SO) ·HO . . <. - - - .*-B U(SO) ·HO . . <. <. <. <. .* e TGA signatures and chemical assays of these samples suggest a greater amount of water in the structure thantheir XRD identities of U(SO) ·HO suggest. Some weight loss was observed immediately upon commencing TGAanalysis, suggesting that some of this water was loosely bound.Table .: eoretical mass fractions uranium and sulfate for various uranoussulfate hydrates.theoretical mass U SO HOU(SO) . . .U(SO) ·HO . . .U(SO) ·HO . . .U(SO) ·HO . . .U(SO) ·HO . . ... Solid phase stability under various conditionsMost of the solid samples matched one of the theoretical XRD reference patternsunambiguously, and chemically assayed close to the theoretical values, and thuswere easy to classify. e sulfate to uranium molar ratio in these samples was closeto :, matching the expected stoichiometry. ey also had well-defined TGA sig-natures with weight-loss events corresponding to the expected number of watersof hydration, as shown in Fig. . (these decomposition reactions are discussed inmore detail in Chapter ). In general, the octahydrate formed at cold temperatures(≤ °C), the tetrahydrate at high temperatures (≥ °C), and the hexahydrate attemperatures in between, although only from long-duration tests. Some of the sam-ples, however, defied easy classification and required further analysis.One of these was a precipitate with the characteristic, but unknown, XRD pat-tern shown in Fig. .. is solid has been christened parisaite in this disserta-tion, aer the research assistant who first produced it. It was frequently found asa product during the Series B tests. Parisaite was characterized by a relatively lowuranium content of –, a high sulfate content of –, and a penchant forreactingwith, and expensively destroying, the alumina crucibles used for TGA anal-ysis. e sulfate to uraniummolar ratio varied between . and .. e single TGAscan that was collected showed a series of small weight loss events, culminating ina slow apparent weight loss probably linked to the destruction of the crucible. Itsexperimentally-determined XRD pattern is shown in Fig. .. e extra sulfate,and possibly even its tendency to chemically attack the alumina crucible at highFigure .: Powder x-ray diffraction reference patterns for the known uranoussulfate x-hydrates, from 2θ =–°, generated byMercury .. [] frompublished CIF files [, , , ].Figure .: Gravimetric analysis of uranous sulfate tetrahydrate, hexahydrate,octahydrate, and parisaite. Samples: -S, -A, -S, -B.temperatures, may reflect the presence of crystalline sulfuric acid, HSO, in thestructure. Formulations such as U(SO) · HSO or U(HSO) might be possi-ble, but it is impossible to reach a conclusion in the absence of specific evidence.A second type of difficult-to-classify precipitate, referred to here as ‘overhy-drated tetrahydrate’, or +hydrate, displayed an XRD pattern matching that of ura-nous sulfate tetrahydrate, and had a sulfate to uranium ratio of :. However, theamount of water varied quite substantially, from slightly over , in the case of -B,to as high as ., in the case of -B. e TGA signatures varied, oen showing aseries of small water loss events, as shown in Fig. .. is extra water was notsimply a symptom of careless sample drying, as oven-drying overnight at  °C hadlittle effect. A possible explanation is that the extra water was strongly absorbedto U(SO) · HO in such a way as not to interfere with the x-ray pattern, and togive it some measure of stability against drying, in a similar manner to Schnaars’spentahydrate. Interestingly, +hydrate only precipitated from solutions containingimpurities.It could be that both +hydrate and parisaite represent intermediary compoundsalong the pathway to U(SO) ·HO formation. Parisaite was only found in  °CFigure .: Powder XRD pattern of parisaite. Sample: -B.Figure .: Gravimetric analysis of several over-hydrated samples of uranoussulfate tetrahydrate, +hydrate, aer oven-drying overnight at  °C.e pattern for U(SO) · HO is shown for reference. Samples: -S, -B, -B, -B.tests less than  h in duration, and could apparently be avoided by increasing thetemperature or test duration, lowering the amount of acid relative to sulfate, or byadding seed. ere was no obvious cause or cure for the appearance of +hydrate.e conditions under which each of the five identified phases – uranous sul-fate tetrahydrate, hexahydrate, and octahydrate, +hydrate, and parisaite– form areeasier to visualize as a phase map. Figure . shows three projections of the re-lationship between crystalline phase and temperature, acid concentration, and testduration, incorporating every solid sample produced over the course of the presentwork. U(SO) ·HO formed at temperatures≥ °C under a wide variety of con-ditions. U(SO) · HO formed at temperatures ≤ °C. For long-duration slowcrystallization tests of more than  days, U(SO) ·HO formed at temperaturesas high as  °C. Parisaite and +hydrate aremore difficult to classify with respect totemperature because they were only ever observed at  °C, the most common tem-perature tested. Generally, though, it seems that both were intermediary products,with parisaite forming first, followed by +hydrate, then finally settling on tetra-hydrate as a final product. None of the solids showed a clear preference for free acidconcentration except hexahydrate, although this may simply have been because thelongest-duration low-temperature tests tended to be higher in acid. Note that thepossible effects of the different agitation types (i.e., stagnant, shaken, or stirred)on fluid circulation and mass transfer, which might influence the same underlyingmechanisms as test duration, are not considered here... Precipitate quality in the presence of impuritiesPerhaps the most important question for the viability of uranous sulfate precipita-tion as a process technology – and one that has never been addressed in the litera-ture – is whether the process can be done selectively in the presence of impurities.It has already been shown that the aqueous-phase assays for Cu, Ni, Fe(II), and Alremained unchanged during the crystallization of uranous sulfate, even when thesemetals were present at very high concentrations. Nevertheless, direct analysis of thesolids themselves was necessary to confirm this apparent selectivity.In all cases, the uranous sulfate precipitate remained very pure despite the pres-ence of impurities in the mother liquor, as shown in Table .. In all cases, the im-Figure .: Mapof the different polymorphs of uranous sulfate with respect tothe temperature, free acid, andduration of the tests fromwhich theywereharvested. Phases determined by XRD and TGA. Shaded areas indicatethe apparent regions of stability for each polymorph.purity was near or below the detection limit of the AA analytical instrument, evenwhen the sample was prepared with the lowest possible dilution. For the purpose ofinterpreting these results, it is useful to consider the worst case observed, CRYST-, which was precipitated from a solution containing a : molar ratio of Fe:U.e solid assayed at . Fe and . uranium, which is a -fold difference bymass, or a -fold difference by atom-fraction – rather good, considering the highamount of Fe in the mother liquor – but nevertheless not a negligible amount. eFe in the solid could not have derived from entrained mother liquor, since .mLwould have been required – an unlikely prospect in a . g sample of thrice-washedsolids. It is possible that the Fe derived from residual undissolved Fe(SO) ·HO,particularly since the experimental procedure required the preparation of an ex-tremely concentrated Fe solution. Only .mg of undissolved ferrous sulfate wouldbe necessary to produce an iron assay of ., so this is entirely possible. Moresimply, the value may represent an error in the analysis, caused by imperfect matrixmatching or dri in the calibration. In any case, such fastidiousness sidelines themore important conclusion of these experiments: that uranous sulfate precipitationis highly selective in the presence of large amounts of Al, Cu, Fe(II), or Ni.e question arises whether uranous sulfate precipitation is so selective that itmight be possible to convert the product directly into nuclear fuel. Given the ex-tremely rigorous specifications applied to nuclear fuel, this seems quite unlikely,and indeed, ASTM standard C lists its impurity specification for uranyl nitratefeed to a UO conversion facility in micrograms per gram of uranium []. Whenthese specifications are applied to uranous sulfate tetrahydrate (. uranium byweight), each of Al, Cu, Fe, and Ni must be <., and the sum of all impuritiesmust be <.. In most cases, the AA instrument used for solids analysis wasnot sensitive enough to even detect such small quantities, and when it could detectthe presence of these metals, they exceeded the specification (see Table .). It isalmost certain, then, that uranous sulfate precipitated from impure solutions wouldnot be able to bypass the refining currently applied to mill concentrates.. Summary and conclusionsUranous sulfate crystallization is different from a typical hydrometallurgical pre-cipitation process because it takes place at very high acid concentration, proceedsrelatively slowly, and is greatly affected by temperature. To make a viable process,it must produce a crystalline powder in the shortest possible time, while maximiz-ing uranium recovery. e optimal operating conditions in the lab were achievedat  °C and >.mol L− sulfuric acid, with aggressive stirring and the addition ofseed material. e kinetics of crystal growth appeared be a significant limit on thespeed of the process.Five different polymorphs of uranous sulfate were identified over the courseof work, each of which formed under different conditions. Each demonstratedunique x-ray and thermogravimetric signatures. U(SO) · HO formed at hightemperatures, U(SO) · HO at low temperatures, and U(SO) · HO at lowtemperatures and long duration. e so-called over-hydrated tetrahydrate, +hy-drate, formed from tests containing impurities, and were identical to U(SO) ·HO except for the presence of additional absorbed water. Parisaite formed asan intermediary crystallization product at  °C. It had an unknown structure, al-though the unusually high amount of sulfate relative to uranium suggests the incor-poration of sulfuric acid or bisulfate in the structure.e selectivity of the process was unaffected by the presence of Cu, Ni, Fe(II),or Al, with no evidence of precipitation in either the aqueous or solid assays. eseimpurities did negatively affect the solubility or kinetics of the process, however,with a substantial suppression of uranium recovery associated with higher levels ofimpurities.Chapter e crystal structures of uranoussulfate hexahydrate and octahydrateand a comparison to the other knownhydratesNote: is work has been previously published [] in collaboration with Dr. BrianPatrick andAnita Lam in theDepartment of Chemistry andDr. DavidDreisinger intheDepartment ofMaterials Engineering. ework presented in this dissertation islargely taken from that publication, but with additional analysis to provide context.. IntroductionIn Chapter , four different crystallographically-distinct uranous sulfate polymor-phs were identified in the solids precipitated from uranium(IV)-sulfuric acid solu-tions. ese have thus far been described as uranous sulfate ‘tetrahydrate’, ‘hexahyd-rate’, ‘octahydrate’, and parisaite. Each of these phases was found to give a uniquepowder XRD pattern, which provided a convenient ‘fingerprint’ for quickly iden-tifying which phase (or combination of phases) was present in a given precipitate.When their four x-ray patterns were searched for in two popular crystallographicdatabases, the Powder Diffraction File (PDF) and the Open Crystallographic Dat-abase (OCD), surprisingly only the tetrahydrate was found. A further search ofthe literature confirmed that the octahydrate had never before been crystallograph-ically characterized (although it had been identified chemically), and hexahydrateand parisaite had never been observed at all. Given this gap in the literature, it wasdecided to structurally characterize the undescribed uranous sulfate hydrates.While the pursuit of detailed crystallographic data may seem peripheral to theneeds of a process engineer, knowledge of a structure actually has many uses be-yond the arcane practice of tabulating bond distances and torsion angles. One tan-gible benefit is the ability to calculate a substance’s theoretical powder XRD pattern,which grants a concretemethod of identifying its presence in powder samples. It canalso provide explanations for macroscopic behaviour. For example, knowledge ofthe positioning and connectivity of coordinated and solvent water molecules (col-lectively described here as ‘waters of hydration’) could explain why the differenturanous sulfate polymorphs lose their water at such different temperatures whenheated (this phenomenon is explored in Chapter ).In the presentwork, two newuranium(IV) sulfate salt hydrateswere structurallycharacterized by single crystal x-ray diffraction and analyzed by vibrational spec-troscopy: {[U(SO)(HO)] ·HO}n (complex ) and [U(SO)(HO)] · HO(complex ). e editors ofActa Crystallographica helpfully advised that the propernames for these compounds are catena-poly[[pentaaquauranium(IV)]-di-µ-sulf-ato-κ4O:O′] monohydrate] and hexaaquabis(sulfato-κ2O,O′)uranium(IV)-dihydr-ate, respectively, but to mitigate the risk of causing a nomenclature-induced aneur-ysm in the reader, they will simply be referred to as hexahydrate and octahydrate.Other compounds with the general formula [U(SO)(HO)a] · bHO, wherea is the number of coordinated waters and b is the number of solvent waters, havebeen characterized in the past by other authors [, , ], all of which consist ofa central uranium(IV) atom coordinated to at least two sulfates and between zeroand seven water molecules. Some of the complexes have additional solvent watermolecules that contribute to hydrogen bonding. e binding mode of the sulfate tothe uranium and the degree of bridging between units is different in each case.Parisaite was not characterized because no single crystals of sufficient size wereisolated.. Experimental.. Synthesis and crystallizationFor complex , a solution of  g L− uranium and  g L− sulfate was madeby dissolving . g of UO in . g of concentrated sulfuric acid diluted toml with deionized water. e solution was electrolytically reduced at a currentof . A for five hours to . conversion in an electrolytic cell consisting of acopper cathode, a TIR-® dimensionally-stable anode (titanium substrate witha proprietary coating of iridium and tantalum oxidides), and a Nafion membrane.In order to create a supersaturated solution, . g of concentrated sulfuric acidwas slowly mixed with  mL of the reduced solution in a nitrogen-purged glassbottle sitting in a chilled water bath, being mindful to not allow the temperature torise above  °C. e bottle was then sealed and placed in a  °C shaking waterbath set at  shakes per minute. e solution was allowed to crystallize for days, aer which the crystals were separated from the supernatant using a .mmnylon filter. e solids were washed three times with a  sulfuric acid solution,three times with deionized water, and twice with ethanol. e crystals were driedin air for  hours, then placed in a desiccator with silica gel desiccant overnight.e yield was . g of small purple-green rod-like crystals. Powder XRD analysisconfirmed that the the crystals consisted entirely of an unidentified phase, with notraces of any known uranium(IV) or uranium(VI) compounds.To synthesize complex , . g of previously-prepared crystalline uranoussulfate (identified by powder XRD as the tetrahydrate) was dissolved in ml of.mol L− sulfuric acid at room temperature under nitrogen. ml of the result-ing uranium(IV) solution was added to a glass vial. e vial was placed without acap inside a larger bottle containing ml of absolute ethanol, such that the ethanolwas able to diffuse into the mother liquor, thus lowering the uranium(IV) solubilityover time and encouraging slow crystallization. e outer bottle was purged withnitrogen, sealed, and placed inside a  °C water bath. Aer  days, the contentswere filtered, and the resulting crystals were washed with  °C .mol L− sulfuricacid, followed by  °C ethanol... Data collection and refinementCrystal data, data collection methods, and structure refinement details are summa-rized in Table ..For complex , a green rod crystal having approximate dimensions of . ×. × . mm was mounted on a glass fibre with epoxy. Measurements weremade on a Bruker X APEX II diffractometer with graphite monochromated Mo-Kα radiation. e data were collected in a series of φ and ω scans in .° oscilla-tions using . s exposures at a temperature of . °C, to a maximum 2θ value of.°. e crystal-to-detector distance was .mm.For complex , a green crystal having approximate dimensions of . × .× . mm was mounted on a glass fibre under a drop of oil. Measurements weremade on a Bruker APEX DUO diffractometer with graphite monochromated Mo-Kα radiation. e data were collected in a series of φ and ω scans in .° oscilla-tions using . s exposures at a temperature of −. °C, to a maximum 2θ valueof .° . e crystal-to-detector distance was .mm.e data were collected and integrated using the Bruker SAINT soware pack-age []. For complex , face-indexed absorption correction was performed on thedata, with minimum and maximum transmission coefficients of . and .,respectively. All hydrogen atoms were refined isotropically. For complex , thedata were corrected for absorption effects using the multi-scan technique (SAD-ABS) []. All hydrogen atoms were refined isotropically and located on the watermolecules by difference maps, except hydrogen atoms A and B, which wereplaced in calculated positions and were refined as riding on O. e data for bothwere corrected for Lorentz and polarization effects... Vibrational spectroscopyIR samples were diluted with dry potassium bromide to .–. wt. and pressedinto a pellet. Spectra were collected on a PerkinElmer Spectrum  Fourier trans-form infrared spectroscopy (FTIR) Spectrometer using  scans over the range – cm−1 with a resolution of  cm−1. e background, determined by scanning apure potassiumbromide pellet, was subtracted from each data set. Raman data werecollected on single crystals using a HORIBA Jobin Yvon LabRAM HR Raman mi-Table .: Single crystal x-ray diffraction experimental detailsComplex  Complex Crystal datachemical formula [U(SO)(HO)] ·HO [U(SO)(HO)] ·HOcrystal habit, color rod, purple-green prism, purple-greenformula weight (gmol−) . .crystal system, space group monoclinic, I2/a monoclinic, P21/ncollection temperature (K)  a, (Å) .() .()b, (Å) .() .()c, (Å) .() .()α , β , γ , (°) , .(),  , .(), volume (Å3) .() .Z  abs. coefficient µ (mm−1) . .crystal size (mm) .× .× . .× .× .Data collectiondiffractometer Bruker X APEXII area-detector diffractometerBruker DUO APEXII CCDarea-detector diffractometerradiation type Mo Kα Mo KαAbsorption correctionTmin, Tmax ., . ., .total no. reflections  unique reflections  Rint . .(sinΘ/λ )max(Å−1) . .RefinementR[F2 > 2σ(F2)], wR(F2), S ., ., . ., ., .no. of reflections  no. of parameters  H-atom treatment all H-atom parameters refined H atoms treated by a mix-ture of independent and con-strained refinement∆ρmax,∆ρmin (e Å−3) ., -. ., -.Table .: Assay results for uranous sulfate hexahydrate and octahydrate (com-plexes  and ). Values in parentheses are the theoretical stoichiometricvalues.Complex  Complex [U(SO)(HO)] ·HO [U(SO)(HO)] ·HOUranium (wt ) . (.) . (.)Sulfate (wt ) . (.) . (.)Water (wt ) . (.) . (.)Implied waters of hydration . (.) . (.)Trace metals below detection limitcroscope with a linearly polarized excitation line of  nm (complex ) and  nm(complexes  and ). e second laser was used for complex  in order to confirmthat certain broad, strong peaks were due to fluorescence and not Raman scattering... Chemical analysisA sample of each crystal was digested in . nitric acid and diluted to a knownmass and volume. Uranium and sulfate were determined by titration, trace metalsby ICP-MS, and waters of hydration by thermal analysis. e assay results, shownin Table ., were within  of the theoretical stoichiometric values... SowareAll refinements were performed using SHELXL- [] using the OLEX []interface for complex  and the WinGX [] interface for complex . Both struc-tures were solved by direct methods using the SIR program []. Images of thestructures were generated using CrystalMaker® [].. ResultsSchematic representations of the two complexes are shown in Fig. .. Select bondlengths are shown in Table .. e complete set of crystallographic data can bedownloaded in CIF format from the Acta Crystallographica C website.S       O       O       O       O       U       S       O       O       O       O       O       O       O       O       O       O       ·2H       2       O    U       S       O       O       O       O       S       O       O       O       O       O       O     O       O       O       U       S       O       O       O       O       S       O       O       O       O       n       ·H       2       O     complex 1 complex 2nFigure .: Schematic of the structures of uranous sulfate hexahydrate andoctahydrate (complexes  and ).Table .: Selected bond (–) and hydrogen bond (donor ... acceptor) lengths(Å) for [U(SO)(HO)] ·HO () and [U(SO)(HO)] ·HO ().Complex  Complex [U(SO)(HO)] ·HO [U(SO)(HO)] ·HOU−O .() O···O .() U−O .() O···O .()U−O .() O···O .() U−O .() O···O .()U−O .() O···O .() U−O .() O···O .()U−O .() O···O .() U−O .() O···O .()U−O .() O···O .() U−O .() O···O .()S−O .() O···O .() U−O .() O···O .()S−O .() O···O .() U−O .() O···O .()S−O .() U−O .() O···O .()S−O .() U−O .() O···O .()U−O .() O···O .()S−O .() O···O .()S−O .() O···O .()S−O .() O···O .()S−O .() O···O .()S−O .() O···O .()S−O .() O···O .()S−O .() O···O .()S−O .().. Crystal structurese neutral U(IV) sulfate {[U(SO)(HO)] ·HO}n, (), crystallizes as purple-green rods in the body-centred monoclinic space group I2/a. e molecular com-plex consists of a uranium atom coordinated to four bidentate bridging sulfate lig-ands and five water molecules, resulting in a nine-coordinate complex which formsherringbone chains in the direction of the c-axis (Figs. . and .). Four of the fivecoordinated water molecules are approximately the same distance from the ura-nium centre (.() Å and .() Å for U–O and U–O, respectively),but the fih bound water, O, is further, at .() Å. e solvent water atom Oresides in a vacant space between neighbouring chains. ere is extensive hydrogenbonding both within each unit and extending between chains.e neutral U(IV) sulfate [U(SO)(HO)] ·HO, (), crystallizes as purple-green prisms in the primitive monoclinic space group P21/n. e molecular com-plex consists of a uranium atom bound by two terminal bidentate chelating sulfateligands and coordinated to six water molecules, resulting in a -coordinate com-plex with no bridging (Figs. . and .). Each of the free waters, O and O,is held by hydrogen bonding to two bound waters and two bound sulfate oxygens,connecting a total of three neighbouring molecules in the crystal structure. Hydro-gen bonds are also observed between each of the unbound sulfate oxygens and thebound water of a neighbouring molecule in the crystal structure.Figure .: Displacement ellipsoid model ( probability level) depicting theextended structure of {[U(SO)(HO)] ·HO}n (). Hydrogen atomshave been omitted for clarity. Symmetry codes: (i) x,−y+ 12 ,z−12 ; (ii)−x+ 12 ,−y+12 ,−z+32 ; (iii)−x+12 ,y,−z+1.Figure .: Polyhedral model of [U(SO)(HO)] ·HO (). (a) A view downthe a axis, showing the herringbone chain connectivity in the directionof the c axis. (b) A view down the c axis, showing the chains head–onand demonstrating that individual chains are connected only through hy-drogen bonding (dashed lines). Colour key: uranium in green, sulfate inyellow, coordinated water in red, and free water in blue.Figure .: Displacement ellipsoid model depicting the connectivity of[U(SO)(HO)] ·HO (). Hydrogens are omitted for clarity.Figure .: Polyhedral model of [U(SO)(HO)] ·HO () showing hydro-gen bonding (uranium in green, sulfate in yellow, coordinated water inred, and free water in blue)... Vibrational spectroscopye vibrational spectra of both complexes are dominated by signals correspondingto the sulfate anion and the total water. Free sulfate, which has tetrahedral geometry(Td), has four Raman-active and two IR-active vibrational modes []. As the sym-metry of the sulfate is reduced through coordination, additional vibrational modesarise from the liing of degeneracies, resulting in a change in the number and po-sition of the peaks seen in the Raman and IR spectra. e different sulfate bindingmodes – bidendate bridging for complex  and chelating terminal for complex – can therefore be distinguished by their Raman and IR spectra. e most activeregion of the Raman and IR spectra from – cm− are given in Fig. .. eextended spectra are available in Appendix B.For complex , the intense signal at  cm− in the Raman spectrum has beenassigned to the ν1 symmetric stretching mode, with the corresponding signal in theIR spectrum at  cm−. Also in the Raman spectrum, there are two ν2 ( and [sh]1 cm−), two ν3 ( [sh] and  [sh] cm−) and three ν4 signals ([w],  [sh] and  cm−) for a total of nine ligand vibrational modes. In the IRspectrum, signals are observed at ν1 (one band), ν2 (two bands), ν3 (three bands)and ν4 (three bands). e IR spectrum also exhibits a medium band with two localpeaks at  and  cm−, corresponding to the total water.For complex , all four vibrational modes are also active. e strongest Ramansignal is again ν1 at  cm−, with the corresponding band in the IR spectrum at cm−. Two ν2 ( [sh] and  [sh] cm−] and three ν3 signals (, [sh] and  cm−) are also clear in the Raman spectrum. Two weak ν4 bendingbands are visible in the Raman spectrum ( [w] and  [w] cm−), with muchstronger corresponding peaks in the IR spectrum ( and  cm−). e IR spec-trum shows up to six closely-overlapping signals in the ν3 region, compared withthree in the Raman spectrum. Similar to complex , a multicomponent band cor-responding to the total water is observed at – cm−. Intense fluorescencewas observed from – cm− when using the  nm laser, but these peaksdisappeared when using the  nm laser.1sh = sharp, w = weakFigure .: FTIR (blue) and Raman (red) spectra of uranous sulfate hexahyd-rate and octahydrate (complexes  and ).. Discussion and comparison with other knownuranium(IV) sulfate hydratesWith the addition of the two structures presented in this work, five distinct struc-tures of the neutral U(IV) sulfate x-hydrate complex are now known: anhydrousU(SO) [], tetrahydrate U(SO)(HO) [], hexahydrate [U(SO)(HO)] ·HO, octahydrate [U(SO)(HO)]·HO, and-hydrate [U(SO)(HO)]·HO[]. A pentahydrate structure, [U(SO)(HO)] ·HO [], is also known [],but it is structurally identical to the -hydrate except for an additional loosely co-ordinated water molecule situated in vacancies within the cross-linked planes. euranium atoms in these compounds are bound by between two and eight sulfate lig-ands, zero to seven water molecules in the primary coordination sphere, and zeroto two free (solvent) water molecules.e coordination of an increasing number of water molecules significantly al-ters the structures of these compounds, affecting the binding mode of the sulfateligands, bond lengths and angles, and the degree of bridging (Table .). Generally,the addition ofmore coordinatedwaters is correlatedwith an increase in the unit cellvolume and a reduction in the degree of bridging between individualmonomers. Asadditional waters are incorporated into the structure, they appear to displace coor-dinated sulfate bonds, thus reducing the opportunity for sulfate bridging. Four dif-ferent sulfate binding modes, shown schematically in Fig. ., are observed in thesecomplexes: tetradentate bridging, bidentate bridging, chelating terminal, and mon-odentate terminal. Bidentate bridging results in either sheet-like connectivity in thecase of the tetrahydrate, or chain-like connectivity in the case of the hexahydrate.Differences in the sulfate binding mode for each complex can be observed intheir Raman and IR spectra. In all of the hydrated species, all four vibrationalmodes for sulfate are active; in particular, the ν1 mode is very strong. e positionof the ν1 peak in the Raman spectrum is shied depending on the binding mode: cm− for both the bidentate-bridging complexes, the tetrahydrate [] and thehexahydrate;  cm− for the monodentate terminal nonahydrate; and  cm−for the bidentate chelating octahydrate. is seems to reflect a lowering of symme-try of the sulfate group, compared to the Td tetrahedral geometry of free sulfate,with lower symmetry associated with a shi of the ν1 band towards lower energy.Table . shows the O−S−O angles within the sulfate tetrahedra for complexes and . Compared to the ideal sulfate tetrahedron angles of .°, the octahydrate() is less symmetric than the hexahydrate (), supporting this hypothesis. ere isno spectrographic data available for the anhydrous compound, but it is reasonableto postulate that the near-ideal tetrahedral geometry of the sulfate group shouldproduce spectra similar to free sulfate, with a ν1 band at higher energy than the cm− seen for cases of bidentate coordination.Table .: Comparison of the normalized cell volumes, intercell connectivity, and sulfate binding modes of the knownuranous sulfate hydrates.Complex NormalizedCell Volume∗IntercellConnectivityCoordinatedSulfatesSulfate BindingModeRef.U(SO) . D  Tetradentatebridging[][U(SO)(HO)] . D sheets  Bidentatebridging[, ][U(SO)(HO)] ·HO . D sheets  Bidentatebridging[][U(SO)(HO)] ·HO . D chains  Bidentatebridging[][U(SO)(HO)] ·HO . None † Chelatingterminal[][U(SO)(HO)] ·HO . None  Monodentateterminal[]∗ Å2. Cell volume divided by Z, the number of molecules per cell†Each of the two sulfates is coordinated twice in a chelating fashion, for a total of four bondsU       O       S       O       O       O       U       O       O       S       O       O       O       S       O       O       O       U       U       U       U       S       O       O       O       O       U       U       S       O       O       O       O       S       O       O       O       O       U       S       O       O       O       O       U       U       Tetradentate Bridging       Bidentate Bridging (chains)       Bidentate Chelating       Monodentate Terminal       Bidentate Bridging (sheets)       Figure .: Schematic of the different sulfate binding modes observed in theknown uranous sulfate hydrate complexes.. Note on the observed superstructure of uranous sulfatehexahydrateDuring the data collection and refinement on complex , it was noticed that veryfaint additional reflections were present that could not be accounted for by the basicstructure presented here. Figure . shows a pseudo-precession image of the h0lzone, with the extra faint reflections visible only to a resolution of approximately. Å. ese extra reflections could simply be interference from the inclusion of aminor amount of a second phase in the analyzed crystal. More intriguing, however,is the possibility that they identify the existence of a larger super-cell within thestructure, with slightly offset identical sub-cells.If refined as a supercell, the solution for complex  gives a super-cell eight timesthe volume of the sub-cell described previously. While the super-cell structuresolves and refines well in space group I2 as a racemic twin, the hydrogen atomscould be located but not refined, and bond length and angle standard uncertaintiesTable .: Sulfate tetrahedra angles (°) for uranous sulfate hexahydrate andoctahydrate (complexes  and ).Complex  Complex [U(SO)(HO)] ·HO [U(SO)(HO)] ·HOO−S−O .() O−S−O .()O−S−O .() O−S−O .()O−S−O .() O−S−O .()O−S−O .() O−S−O .()O−S−O .() O−S−O .()O−S−O .() O−S−O .()O−S−O .()O−S−O .()O−S−O .()O−S−O .()O−S−O .()O−S−O .()Figure .: Pseudo-precession image of h0l zone for uranous sulfate hexahyd-rate (complex ), and a magnified view of the boxed region showing thefaint spots resulting from the super-cell.were considerably worse than the sub-cell. In the reported sub-cell structure (solvedand refined in space group I2/a) all hydrogen atoms could be refined isotropically,and standard uncertainties of bond lengths and angles were considerably improved.Table . gives a comparison of the sub- and super-cell structures.e connectivity in the superstructure is identical to the substructure, and thebond lengths and angles are very similar. e cell dimensions and angles remainidentical, except for the doubling in size in each direction. e main differenceappears to be a slight variation in the placement and orientation of the free waterTable .: Comparison of crystal parameters for uranous sulfate hexahydrate(complex ) sub- and super-cells.Sub-cell Super-cella (Å) .() .()b (Å) .() .()c (Å) .() .()α (°)  β (°) .() .()γ (°)  volume (Å3) .() .()Unique reflections  Reflections (I/sigI)  space group I2/a I2 (racemic twin)S−O bond precision (Å) . .molecules, labeled O in the substructure.. ConclusionTwo new uranium(IV) sulfate hydrate species have been synthesized and character-ized, and their characteristics compared to four other known structures. ese twonew compounds add to the current body of structurally characterized uranium(IV)sulfates and advance the understanding of how waters of hydration affect their con-nectivity. ey also offer insight into the wide variation in sulfate binding modepresent in uranium(IV) compounds with otherwise-similar chemical formulae.Chapter ermal stability of uranous sulfate I:ermodynamics and theorye stability of solid uranous sulfate hydrate under various conditions, as well asthe decomposition products and gases, can be predicted by thermodynamics. efollowing topics are treated here from a theoretical perspective: water loss; decom-position of U(SO); the SO/SO equilibrium; and the construction of a phase di-agram.Chemical equations in this section are written in the following format:AT ◦eq−−→ B+Cwhere T ◦eq is the temperature that satisfies ∆G◦ (T ) = 0. e T ◦eq values werecalculated using the thermodynamic values given by Guillaumont and Mompean[], which were introduced in Chapter , according to Eq. (.):Teq =∆H◦∆S◦−R lnQ(.)whereQ is the reaction quotient, which is the ratio of the activities of the productspecies divided by the activities of the reactant species, each raised to the power ofits stoichiometric coefficient.Data for the construction of the phase diagram was obtained from the sowarepackage HSC . [], which takes into account the slight temperature dependenceof ∆H◦ and S◦, giving slightly more accurate results than if the values at  °C wereused.. Water losse thermodynamic database published by Guillaumont and Mompean [] con-tains thermodynamic quantities for two uranous sulfate hydrates: the tetrahydrateand the octahydrate. Reactions for their complete dehydration can be written asfollows:U(SO4)2 ·4H2O85◦C−−−→ U(SO4)2 +4H2O(g) ∆G◦ = 206.30−0.5763T (.)U(SO4)2 ·8H2O90◦C−−−→ U(SO4)2 +8H2O(g) ∆G◦ = 418.38−1.1527T (.)e uncertainties in the published S◦ values for the two hydrates is around±, however, making the calculated T ◦eq values only accurate to ± °C. eenthalpies of reaction for the tetrahydrate and the octahydrate are . kJmol−and . kJmol−, respectively, which amounts to approximately  kJmol− perwater molecule. If stable intermediary hydrates exist, as seems likely, the true tran-sition temperatures and heats of reaction would be different than calculated here.. e SO2/SO3 equilibriumWhen uranous sulfate decomposes, the sulfate is released as either SO or SO gas.us an important aspect of this system, from a theoretical perspective, is the ther-modynamic equilibrium between SO and SO, and the kinetics of this reaction.e gas-phase equilibrium between SO, SO, and O can be written as shownin Eq. (.):SO3779◦C←−−→ SO2 +12O2 ∆G◦ = 98.89−0.0940T (.)As temperature increases, there is a shi in gas-phase equilibrium from SOtowards SO. e ratio of SO to SO at equilibrium can be calculated using therelationship shown in Eq. (.), and is a function of temperature (via the equilibriumconstant K) and oxygen partial pressure.logpSO2pSO3= logK−0.5log pO2 (.)Aplot of this relationship is shown in Fig. .. As pO2 decreases, the equilibriumshis to the le. In air (pO2 = . atm), SO and SO are equimolar at  °C.If pO2 is not fixed, and instead is allowed to float according to the stoichiometryof Eq. (.), the calculations become more complex. To understand this scenario, itis useful to imagine a sample of uranous sulfate decomposing at a fixed rate undera stream of nitrogen in an unpressurized vessel (i.e., PT = 1atm). e released SOwould decompose to SO and O to a certain extent, but the combined flow of SOand SO would remain constant. If the flow of nitrogen increases,[pSO2 + pSO3]must decrease accordingly.is system can be solved by applying stoichiometric restraints, and by fixing[pSO2 + pSO3], as shown in Eq. (.). is leads to Eq. (.), which can be solvednumerically for any[pSO2 + pSO3], and then plotted to show the relationship be-tween gas composition and temperature, as shown in Fig. ..pO2 = 0.5pSO2pSO2 + pSO3 = constant(.)logK = 0.5log(0.5)+1.5log pSO2− log([pSO2 + pSO3]− pSO2)logpSO2pSO3= log pSO2− log([pSO2 + pSO3]− pSO2)(.)As[pSO2 + pSO3]decreases, the equilibrium shis to the le, with SO be-coming more favourable at lower temperatures. Put another way, SO is favoured ifthe decomposition gases are dilute, while SO is favoured if they are concentrated.Quite by coincidence, equimolar SO/SO is again reached at  °C for PT = 1atm(pSO3 = pSO2 = 0.4 atm, pO2 = 0.2 atm).It would be foolish to accept these thermodynamic predictions without castinga critical eye towards kinetics. e kinetics of the SO/SO equilibrium are wellFigure .: Equilibrium SO/SO ratio as a function of temperature and oxy-gen partial pressure. e dashed line indicates the equilibrium if pO2 isnot fixed, but rather builds up according to the stoichiometry of the reac-tion with pSO3 + pSO2 = 1 atm.Figure .: Equilibrium SO/SO ratio as a function of temperature and pSO3+ pSO2, with pO2 set according to reaction stoichiometry. e dashedline indicates the equilibrium in air, with pO2 = 0.209 atm.known to extractive metallurgists due to its importance in SO-capture technolo-gies. According to Louie [], the reverse reaction (i.e., the homogenous oxidationof SO with O), is severely kinetically limited, occurring primarily above  °C,where in any case it is not thermodynamically favoured. In the presence of a cata-lyst such as VO or FeO, however, the reaction can take place at an appreciablerate below  °C, though still not particularly fast, with the highest conversion ob-tained between – °C. Below  °C the kinetics are too slow for appreciableconversion to take place. e forward reaction scarcely fares better, with Yilmazet al. [] determining that the homogenous decomposition of SO in nitrogen isslow below  °C.Given the dubious kinetics of the SO/SO equilibration reaction, it seems un-likely that it would play a significant role during the thermal decomposition of ura-nous sulfate. Nevertheless, it must still be understood in order to properly interpretthe results from thermodynamic simulations, and also could be useful when con-sidering the downstream affect of the off-gases.. Anhydrous uranous sulfate decompositionTo begin understanding the thermodynamics of the decomposition of anhydrousuranous sulfate, U(SO), it is useful to start with a simulation. HSC . [], whichcontains both an extensive thermodynamic database and tools for solving equilib-rium equations, was used for this purpose. Only a subset of the possible speciesin the database were found to have regions of stability under the conditions tested:U(SO), UOSO, UO, and UO in the solid phase, and O, N, SO, and SOin the gaseous phase. All other species were removed to simplify the calculations.e gaseous uranium species were also removed to prevent HSC from erroneouslypredicting their formation.In the first simulation, shown in Fig. ., mol of U(SO) was allowed to cometo equilibrium in an isobaric batch reactor initially containing an excess of nitro-gen as the temperature was raised from – °C, and allowing the decomposi-tion gases to stay in equilibrium with the solids. e first stage of decomposition,the homogenous oxidation of U(SO) to UOSO, was accompanied by the releaseof SO, with sulfur acting as oxidant to uranium. e second stage of decomposi-Figure .: eoretical thermodynamic equilibrium of the decomposition ofU(SO) from – °C in an inert atmosphere, initially containingmol N per mol uranium as solid. (g) gas phase; (c) condensed phase.Data generated by HSC . []tion, to UO, saw the release of the remaining sulfate as SO. A small amount ofoxygen gas was also generated in the second step as the uranium reduced to UO.e SO released in the first decomposition step remained as such because of a lackof oxygen to react with. Above  °C, however, the SO/SO equilibrium preferredSO.Figure . shows a simulation of the same system, except carried out in air in-stead of nitrogen. e presence of oxygen has a significant effect on the thermo-dynamic stability of the various species. Of particular note is a broadening of theUOSO stability region to both lower and higher temperatures. Also of signifi-cance is the increased stability of UO.In air, UOSO was predicted to be stable at both higher and lower temper-atures. e low-temperature stability was due to the heterogeneous oxidation ofU(SO) with O, which has a lower equilibrium temperature than for homoge-neous oxidation. At higher temperatures, further decomposition was delayed dueto the leward pressure on the reaction equilibrium from the presence of oxygenFigure .: eoretical thermodynamic equilibrium of the decomposition ofU(SO) from – °C in an atmosphere fixed at pO2 ≈ 0.209, ini-tially containing mol N and mol O per mol uranium. (g) gasphase; (c) condensed phase. Data generated by HSC . []gas. Under air, the predicted stability of UO was enhanced.ese two simulations suggest that the following five solid decomposition reac-tions (as well as the SO/SO equilibrium) could play a part in the thermal decom-position of uranous sulfate:U(SO4)2451◦C−−−→ UO2SO4 +SO2 ∆G◦ = 167.65−0.2314T (.)U(SO4)2 +12O2227◦C−−−→ UO2SO4 +SO3 ∆G◦ = 68.76−0.1374T (.)UO2SO4889◦C−−−→ 13U3O8 +SO3 +16O2 ∆G◦ = 257.84−0.2219T (.)UO2SO4916◦C−−−→ UO3 +SO3 ∆G◦ = 225.64−0.1897T (.)UO3725◦C−−−→ 13U3O8 +16 O2 ∆G◦ = 32.20−0.0323T (.)In order to evaluate the importance of UO as an intermediary product, it isuseful to examine the conditions under which Eqs. (.) and (.) would be ex-pected to take place. Given that the T ◦eq value for the first reaction is higher than forthe second, UO should not form under standard state conditions. Only at the rightcombination of high pO2 and low pSO3 could UO be formed, and even then onlyover a relatively narrow temperature range. It therefore seems unlikely that UOwould play a significant role in the thermal decomposition of uranous sulfate.e equilibrium temperature of the final decomposition toUO, Eq. (.), canbe calculated by substituting the appropriate thermodynamic and stoichiometricvalues into Eq. (.):Tfinal =257840Jmol−1(221.942Jmol−1K)−(8.3145Jmol−1K)ln(pSO3 p1/6O2) (.). Uranous sulfate decomposition phase diagrame simulation results shown in Figs. . and . are misleading because they as-sume that the product gases remain in contact with the solids. A phase diagramcan give a better sense of the stability of the various species at different gas compo-sitions. Figure . shows a phase diagram of the system between – °C. ey-axis represents the partial pressure of either SO or SO, depending on which gasis evolved during that stage of decomposition. e effect of oxygen partial pres-sure on the second decomposition step, Eq. (.), is shown with three pO2 isobars.Also shown is the stability region of UO at two different pO2 isobars. UO is thedominant phase at any oxygen partial pressure above ∼ °C, demonstrating thatit is always possible to convert U(SO) into UO as long as it is calcined at a highenough temperature.Figure .: Phase diagram of the U-S-O system for homogenous decomposi-tion of U(SO). pO2 isobars are shown, along with their effect on thestability of UO.Chapter ermal stability of uranous sulfateII: Experimental examination. Introductionree different uranous sulfate x-hydrate polymorphs were identified in Chapter :U(SO) · HO, U(SO) · HO, and U(SO) · HO. Each was found to becrystallographically unique, with different sulfate bindingmodes anddifferent num-bers of crystalline waters. In this chapter, the thermal stability and decompositionpathways of the three hydrates are described, and decomposition mechanisms areproposed.To begin understanding the system, samples of the three hydrates were heated inan oven or furnace for a period of time, allowed to cool, and then analyzed by pow-der x-ray diffraction in an attempt to directly identify the resulting material. eseexperiments were only partly successful, withmany of the intermediary compoundsfound to be x-ray amorphous, thus making them unidentifiable. However, weightloss measurements provided indirect evidence of the processes involved. Further-more, these experiments brought attention to the slow solid-phase recrystallizationkinetics of U(SO).e system was studied more closely using the thermoanalytical techniques ofdifferential thermal analysis (DTA) and differential scanning calorimetry (DSC).From these data, the onset temperatures and enthalpies of transformation of thevarious weight loss and thermal events from – °C were identified, and theintermediate decomposition products were identified. e effect of operating underdifferent atmospheres, primarily nitrogen vs. air, was also explored.is chapter is concluded by combining the experimental results with thermo-dynamic and kinetic theory to propose chemical decomposition pathways for thethree uranous sulfate hydrates. e different behaviour of the three hydrates areexplained by a transition from amorphous to crystalline anhydrous uranous sul-fate. is fills a gap in the literature and brings the understanding of the thermaldecomposition of the uranous sulfate hydrates to the same level as the uranyl sulfatehydrates.. Background informationIt has long been recognized that there are several different uranous sulfate x-hydrates[], and that they have distinct thermal decomposition fingerprints []. Leroy andTridot [] reported that U(SO) · HO dehydrates in two steps, with U(SO) ·HO as an intermediary, followed by oxidation to UOSO, and finally decompo-sition to UO, according to the following reaction pathway in air (temperaturesapproximated from published curves):U(SO4)2 ·4H2O∼ °CU(SO4)2 ·H2O∼ °CU(SO4)2∼ °CUO2SO4∼ °C U3O8Gil et al. [] reported on the thermal decomposition of ten different uranous-M sulfate x-hydrate double salts (with M = Cd, La, Li, Mg, Mn, Ni, V, and Zn), butnot on pure uranous sulfate. Suzuki et al. [] used TGA to identify a tetrahydrate,a trihydrate, and a n-hydrate (n≈ 1.7−2.1), all of which decomposed to U(SO)by  °C, but they did not study the process in depth. No data on the thermaldecomposition of the hexahydrate or octahydrate have been reported.Given the lack of published information on the compounds of interest, it is use-ful to review a closely-related compound, uranyl sulfate x-hydrate. e decomposi-tion of UOSO ·xHOhas been studied by a number of authors [, , , , ],who have proposed the various water loss mechanisms shown in Fig. .. In eachstudy, the authors used TGA and DTA to suggest a decomposition pathway for wa-ter loss. e sheer quantity of identified phases, as well as the lack of agreementbetween authors, only serves to confirm Walter McCrone’s droll observation on theusefulness of studying polymorphism1. Notz and Jaffe [] observed three distinctendotherms during the decomposition of UOSO ·HO, which they attributed tothe stepwise loss of single water molecules. Leroy et al. [] found that two differentdecomposition pathways exist, starting with either the tetrahydrate or the hemihep-tahydrate, each distinguished by different decomposition temperatures and inter-mediaryXRDpatterns. Cordfunke [, ] subsequently confirmed the twodistinctpathways, but identified the .-hydrate as the true form of the compound previ-ously assumed to be -hydrate, and also identified several different phases of themonohydrate. Sato et al. [] found a stepwise pathway similar to Notz’s, and thex-ray diffraction patterns for their - and .x-hydrates were the same, supportingCordfunke’s theory that the -hydrate is simply a more-hydrated version of the .-hydrate. In all, at least twelve crystallographically-unique phases of uranyl sulfatehydrate and anhydrate have been identified.e high-temperature thermal decomposition of both uranous and uranyl sul-fate is more straightforward. Tridot [] reported that U(SO) first oxidizes toUOSO, accompanied by the release of SO. Notz and Jaffe [] and Tridot []both observed a sharp endothermic event centred at  °C which they attributedto a phase transition from αUOSO to βUOSO. Upon heating in an oxidiz-ing atmosphere to high temperature, all uranium oxides, and many other uraniumcompounds, ultimately decompose or convert to UO []. Uranyl sulfate is noexception, itself decomposing to UO at high temperature.e operating atmosphere during thermal decomposition has been shown tohave an effect on the decomposition pathway and the stability of the intermedi-ary products. Tridot [] observed that the onset temperature for the oxidation ofU(SO) was approximately  °C in a 10−2 mmHg vacuum,  °C in dry oxygen,1Walter McCrone stated that ‘every compound has different polymorphic forms, and that, in gen-eral, the number of forms known for a given compound is proportional to the time and money spentin research on that compound.’ Physics and Chemistry of the Organic Solid State,, vol. , pp.-.Notz and Jaffe [] ()UO2SO4 ·3H2O∼ °CUO2SO4 ·2H2O∼ °CUO2SO4 ·H2O∼ °CUO2SO4Leroy et al. [] ()Series A: UO2SO4 ·4H2O∼ °CUO2SO4 ·3H2O∼ °CUO2SO4 ·H2O∼ °CUO2SO4Series B: UO2SO4 ·3.5H2O RT UO2SO4 ·3H2O∼ °C UO2SO4 ·H2O∼ °C UO2SO4Cordfunke [] ()Series A: UO2SO4 ·2.5H2O∼ °C x-ray amorphous + °C UO2SO4Series B: UO2SO4 ·3.5H2O RT UO2SO4 ·3H2O∼ °CβUO2SO4 ·H2O∼ °CUO2SO4 ·12 H2O∼ °CUO2SO4Sato et al. [] ()UO2SO4 ·3H2O∼ °C UO2SO4 · xH2O(2.5≤ x≤ 3)∼ °CUO2SO4 ·2H2O∼ °C UO2SO4 ·H2O∼ °CUO2SO4Figure .: ermal water loss pathways for uranyl sulfate x-hydrate, - °C, as determined by various authors. Quoted temperatures are theonset temperatures estimated graphically from the published DTA curves(no DTA curve was presented by Cordfunke, so the TGA curve was usedinstead). ‘RT’ indicates room temperature, with immediate water loss ob-served at the commencement of the test.and  °C in SO. Under vacuum, the decomposition proceeded directly to UO,with only an inflection point in the TGA curve (as opposed to a plateau) to indicatethe transitory presence of UOSO. Notz and Jaffé [] found that the atmospherealso affected the onset temperature for the final decomposition to UO:  °C inhelium,  °C in air, and  °C in a mixture of SO and O.. Experimental procedures and data treatmentermal decomposition experiments were conducted on samples of solid uranoussulfate tetrahydrate, hexahydrate, and octahydrate. All of the solidswere crystallizedfrom aqueous uranium(IV)–sulfuric acid solutions in the course of other researchactivities. A description of sample genesis can be found in Chapter ... Bulk drying and calcining for x-ray analysisBulk drying of solids was conducted in a standard laboratory oven set to  °C or °C for  h. e weights of the solids were recorded before and aer each exper-iment, and XRD was used to identify the crystalline phases.Calcining at  °C was conducted on approximately mg of sample in analumina crucible in the STA- furnace in an air atmosphere. High-temperaturecalcining was conducted at  °C in an electric furnace inside a ceramic crucible... ermal analysis instrumentation and calibrationermogravimetric analysis (TGA) and differential scanning calorimetry (DSC)were used to investigate the stability and decomposition of the three uranous sulfatex-hydrate phases from – °C. Most tests were carried out at UBC on approx-imately mg of ground sample in an alumina crucible under nitrogen or air usinga PerkinElmer STA- TGA/DSC. e tests under ammonia and hydrogen at-mospheres were carried out at Cameco Corporation’s laboratory in Port Hope, On-tario using a TA Instruments SDT Q TGA/DSC. Unless otherwise mentioned,all scans used the following program:1. Switch gas to nitrogen at 20.0mlmin−12. Hold for 1.0min at 30°C3. Heat from 30°C to 995°C at 10.0°Cmin−14. Hold for 5min at 995°CUnless otherwise stated, quoted temperatures for thermal events refer to theDSC extrapolated peak onset temperature, Te, which is the “temperature where theinflectional tangent at the ascending peak slope intersects the linearly extrapolatedinitial baseline” []. is method proved to be reasonably objective in its applica-tion, corresponded well with DTG data, and showed less dependence on scan ratethan the equally-arbitrary ‘peak’ temperature. Te should not be confused with theinitial peak temperature, Ti, which is the temperature at which the DSC curve be-gins to deviate from the extrapolated baseline2. For events above  °C, for whichthe DSC data was not clear enough to define an onset temperature, an approximatetemperature or range is given based on visual inspection.e PerkinElmer STA- was calibrated for mass against a certified-weightsteel bead. TGA data were baseline-corrected using data collected on an empty cru-cible. is succeeded in correcting for systematic error (i.e., the buoyancy effect),but could not correct for the irregular dri in the microbalance seen over longertests. During one  h test, the weight error, as measured by the indicated weight ofthe empty crucible, reached ±.mg. is represents an error of± in a typicalmg sample, or the weight of one third of a water molecule, which explains whyresults from long tests did not match the theoretical values as crisply as those fromshort tests.Temperature was calibrated against the known transition temperatures of in-dium (. °C) and silver (. °C) using the onset temperature determined attwo scan rates,  °Cmin− and  °Cmin−. Heat flow was calibrated against thetransition heat of indium (. J g−). e calibration was checked by runningthese samples again, as shown in Fig. .... TGA data treatmentAll weight measurements were normalized to the weight of the final decomposi-tion product, UO, in order to eliminate uncertainty in the amount of water in theoriginal samples, and to provide a common point to compare the different hydrates.2A more thorough explanation of the terminology applied to the graphical interpretation of DSCcurves can be found in Gmelin and Sarge []Figure .: Validation of thermal analyzer temperature and heat flow calibra-tion using indium and silver.  °Cmin−, nitrogen atmosphere.is was possible because uranous sulfate x-hydrate decomposes stoichiometricallyto UO at high temperatures (this was shown to be true thermodynamically andexperimentally). us the number of moles of uranium in the crucible, NU, couldbe calculated by dividing the final decomposition weight at  °C by the molecularweight of UO., regardless of the starting material, as in Eq. (.) (note: UO =UO.). e TGA data were then scaled to ‘equivalent molecular weight’, m′, bydividing the raw weight readings by NU.NU =m(850 ◦C)MWUO2.666[mol] (.)m′ (T ) = m(T )NU[gmol−] (.)First derivatives of the TGA curves (i.e., derivative thermogravimetry (DTG))were calculated by finite difference approximation using  s intervals ( °C at a scanrate of  °Cmin−), and were expressed as a time-based rate by multiplying thedata by the constant scan rate. Peak deconvolution and integration of the DTG datawere performed using Fityk .. [], a soware package for generic peak fittingand signal deconvolution... DSC data treatment and baseline correctionIn classical DSC analysis, the difference in heat flow is measured between a sampleand an inert reference. e STA- does not include a reference sample holder,instead relying on an integrated ‘reference ring’. ismethod is reasonably effective,insomuch as thermal events are clearly visible as peaks (endothermic) or troughs(exothermic) in the DSC curve, with the area under the curve representing ∆Hrxn.However, the lack of a true reference results in a significant background signal, par-ticularly at high temperatures, that can obscure small or broad features and makeit difficult to determine the onset temperature of an event, and impossible to accu-rately integrate the peak area.e DSC curves were relatively flat below  °C, but suffered from a continu-ously-increasing background signal above  °C. A similarly-shaped backgroundsignal was observed on a scan of a mg sample of previously-calcined uranous sul-fate (i.e., UO), but without the peaks associated with chemical reactions or phasechanges. With this in mind, all DSC data were corrected for baseline effects by sub-tracting theweight-correctedUO scan, as shown in Fig. .. eweight-corrected‘transformed baseline’ was calculated from the raw baseline by multiplying it by theweight fraction remaining of the sample, which accounted for the declining sampleweight as decomposition progressed, and thus the sample’s lower heat capacity, rel-ative to the unchanging weight of the reference.  °C, the same temperature usedfor weight normalization, was selected as the ‘zero’ point when scaling the base-line. is yielded relatively flat baseline-corrected curves with identifiable thermalevents, but the substantial uncertainty at higher temperaturesmade it unsuitable forquantitative use above  °C.. Validation of thermal analysis method.. Selection of representative samplesRepresentative samples of uranous sulfate tetrahydrate (-S), hexahydrate (-A),and octahydrate (-S) were chosen from the forty-six solids that were producedover the course of work described in Chapter . e choice was largely arbitrary,since nearly all of the solids were found to be of high quality and purity. e sam-Figure .: Correction of the DSC signal using a baseline collected on a pre-calcined sample of UO to approximate the heat capacity effect.ples were therefore chosen based on the available sample volume, each being largeenough to accommodate dozens of thermoanalytical experiments if necessary.e identity of each was confirmed by XRD and chemical analysis. Samples -S and -S were both precipitated as powders, and so could be used directly, but-A had to be ground with a mortar and pestle... Choice of scan rateeeffect of scan ratewas evaluated onU(SO) ·HOat  °Cmin−,  °Cmin−and  °Cmin−, with the results shown in Fig. .. In theory, a slow scan rate is bettersuited to pinpointing the start of a thermal or weight loss event, but it also makesthe differential heat flow less pronounced, resulting in a weaker DSC signal.e TGA weight loss curve was sharpest at the slowest scan rate, but the samefinal weights were ultimately achieved regardless. e difference between the onsetFigure .: TGA and DSC curves for the same sample of U(SO) · HO atdifferent scan rates (a,  °Cmin−; b,  °Cmin−; c,  °Cmin−). Sam-ple: -S.temperatures for water loss at  and  °C was negligible (unlike the peak tem-peratures), at  °C and  °C, respectively. However, at  °Cmin−, the onsettemperature registered substantially lower, at  °C. A similar difference was seenfor higher-temperature events.e temperature shi observed at different scan rates makes it clear that theapparent onset temperature should not be regarded as absolute. It also highlightsthe importance of only comparing data collected at identical scan rates. With thisin mind, a default scan rate of  °C/min was chosen for all tests in order to balanceaccuracy with expediency... e effect of particle sizeParticle size could theoretically influence both heat and mass transfer during TGAanalysis, with both becoming less efficient with increasing size. is would seemparticularly important if the kinetics of a process were solid-diffusion limited, sincethe larger diffusion distance inherent in a larger particle would slow the reactionrate. Ideally, the samples could be ground to the same particle size in a micronizingmill to minimize this effect, but efforts to do so were unsuccessful because the sam-ples were found to either dissolve (in water) or dehydrate (in ethanol). e sampleswere instead ground by hand with a mortar and pestle to a fine powder.e effect of particle size was investigated experimentally by collecting TGAscans on ground and unground examples of the same material. For tetrahydrate,two different samples were tested: -S, a powder sample, and -A, a sample con-sisting of large, chunky crystals. e crystals of -A were crushed slightly to allowthem to fit into the crucible, but were not finely ground, while -S was used as-is. For hexahydrate, the same sample of -A was tested twice, the first time usingthe original  µm×mmneedle-like crystals, and the second time grinding themwith a mortar and pestle.e difference in particle size had no obvious effect for tetrahydrate, while forhexahydrate the decomposition reactions took place slightlymore quickly, as shownin Fig. .. It was concluded that no further sample preparation beyond hand-grinding was necessary to generate comparable results between the three hydrates... ReproducibilitySeveral thermal analysis tests were repeated to ensure reproducibility, as shown inFig. .. e TGA curves overlapped so closely that they were nearly indistinguish-able. e DSC signals were also reproducible, with peak temperatures within  °C.Two samples of U(SO) ·HO produced under very different conditions alsohad nearly the same TGA signature, as shown in Fig. .. -A was grown slowlyas large crystals at  °C in a shaken flask over  days, while -S was precipitatedquickly as a powder at  °C in a stirred vessel over  hours. ese results confirmedthat the thermoanalytical behaviour of uranous sulfate x-hydrate was independentof preparation method.Figure .: e effect of particle size on the TGA curves for U(SO) · HOand U(SO) · HO, by analyzing whole crystals vs. powders. Samples:-S, -A, -A.. Results: x-ray analysis of bulk sample decompositione weight loss and powder XRD results for the three uranous sulfate hydrates aersustained heating in an air atmosphere are shown in Table .. e results providebasic information on their stability at  °C,  °C,  °C, and  °C.None of the solids lost a significant amount of weight or changed in crystallo-graphic identity at  °C, showing that all three hydrates are stable under regular am-bient conditions. Aer  h at  °C, the tetrahydrate again remained unchanged,but the hexahydrate and octahydrate both lost weight equivalent to . and . crys-talline waters, respectively, not including the small amount of adsorbed water lostat  °C. ese samples were found to be be x-ray amorphous, suggesting that thesamples had not recrystallized as a coherent lower hydrate.At  °C, uranous sulfate tetrahydrate quickly lost weight corresponding tocomplete dehydration aer one hour, but was found to be x-ray amorphous. Asample of hexahydrate held for  h at  °C did appear to undergo partial solid-phase recrystallization, with its characteristic x-ray peaks visible above the broadamorphous background.Uranous sulfate tetrahydrate completely decomposed to UO at  °C. AerFigure .: Reproducibility of the TGA andDTA curves (a, U(SO) ·HO; b,U(SO) ·HO) in a nitrogen atmosphere, scan rate  °C/min. Samples:-A, -S. h at  °C, the furnace was turned off, and the sample was allowed to return toroom temperature slowly inside the oven over a further  h. X-ray analysis con-firmed that the resulting black powder wasUO. eweight loss was notmeasuredbecause of damage to the crucible.e bulk dehydration and calcining tests provided some useful information onthe decomposition of the three uranous sulfate hydrates, in particular confirmingthat the final decomposition product at high temperatures is UO. However, thedifficulty in obtaining coherent x-ray diffraction patterns of the intermediary prod-ucts in a reasonable length of time limited the use of this method, and it was notpursued further, instead favouring the more detailed data provided by TGA.Table .: ermal treatment of the uranous sulfate hydrates at °C, °C,°C, and °C. Identity determined by XRD.-S -A -SU(SO) ·HO U(SO) ·HO U(SO) ·HO°CDuration  h  h  hIdentity U(SO) ·HO U(SO) ·HO U(SO) ·HOWt loss -. -. -.°CDuration  h  h  hIdentity U(SO) ·HO x-ray amorphous x-ray amorphousWt loss -. -. -.°CDuration  h  h-Identity x-ray amorphous U(SO)†Wt loss -. -.°C Duration  h - -Identity UO† Sample wasmostly x-ray amorphous, but showed very faint peaks correspond-ing to U(SO). Sample was entirely amorphous aer a  h hold.. Results: ermal analysisTGA, DTG, and DSC curves for uranous sulfate tetrahydrate, hexahydrate, andoctahydrate in nitrogen and air are shown in Figs. . to .. ese figures will bereferred to throughout the discussion. e onset temperatures for thermal eventsbelow  °C are marked on the DSC curves. e deconvoluted DTG peaks areshown overlaid on the raw data (these will be discussed in a later section). Variouscharacteristic peaks or regions are marked, corresponding to water loss (W), phasechange (P), and sulfur loss (S) events.Figure .: TGA, DTG, and DSC curves for U(SO) ·HO in a nitrogen at-mosphere.Figure .: TGA, DTG, and DSC curves for U(SO) ·HO in a nitrogen at-mosphere.Figure .: TGA, DTG, and DSC curves for U(SO) ·HO in a nitrogen at-mosphereFigure .: TGA, DTG, and DSC curves for U(SO) ·HO in an air atmo-sphere.Figure .: TGA, DTG, and DSC curves for U(SO) ·HO in an air atmo-sphere.Figure .: TGA, DTG, and DSC curves for U(SO) ·HO in an air atmo-sphere... Decomposition in nitrogenFor all three hydrates, there were four stages of decomposition: initial rapid wa-ter loss to form an intermediary lesser hydrate; slow loss of the remaining water toform anhydrous uranous sulfate; oxidation to UOSO; and finally decompositionto UO. In addition, the hexahydrate showed a small but prominent thermal eventat  °C which was absent from the other two hydrates. e initial and final mate-rials were weight-stable, but the intermediary compounds oen declined in weightsomewhat over time.e number of waters of hydration implied by the TGA data agreed broadlywith the theoretical values, giving ., ., and . waters of hydration for thetetrahydrate, hexahydrate, and octahydrate, respectively. e slight excess over thetheoretical value in each case corresponded with the adsorbed water that was re-moved at  °C in the oven-drying experiments described earlier.e onset temperatures for initial water loss were different for the three species,at  °C,  °C, and  °C for the tetrahydrate, hexahydrate, and octahydrate, re-spectively. An inflection point in the TGA curves aer initial water loss points tothe existence of a transient lower hydrate. e remaining water in the lower hydratewas then lost slowly and continuously over a several hundred degree range. Inter-estingly, neither the hexahydrate nor the octahydrate showed a tendency to formU(SO) ·HO as an intermediarye three (former) hydrates showed different decomposition behaviour evenaer converting to U(SO). e hexahydrate had the greatest stability as U(SO),maintaining a steady weight to a higher temperature than the other two speciesand decomposing to UOSO at  °C. e octahydrate did not maintain a steadyweight as U(SO), but rather lost weight slowly starting around  °C, followedby more rapid decomposition to UOSO around  °C. e tetrahydrate also lostweight slowly starting around  °C, with an acceleration around  °C, but didnot show the characteristic ‘s-shaped’ curve of the other two hydrates, instead losingweight at a more-or-less constant rate.Aer reaching UOSO, all three hydrates behaved the same, with decomposi-tion to UO starting in the range – °C... Decomposition in airermal analysis in an air atmosphere was performed in the same manner as thenitrogen tests. Operating in air had no effect on water loss (W) or the exother-mic event in the hexahydrate (P). e sulfur loss decomposition steps (S), fromU(SO) to UO, however, were affected by the presence of oxygen, with the reac-tions generally shiing to a higher temperature. efirst stage, U(SO) toUOSO,was delayed to higher temperature in the tetrahydate and the hexahydrate, but wasunaffected in the octahydrate. e second stage of decomposition, from UOSOto UO, proceeded more or less the same for each polymorph, but approximately °C hotter than under nitrogen... Decomposition under hydrogen and ammoniaA sample of U(SO) ·HO (-S), was sent to Cameco Corporation’s technologycentre in Port Hope, Ontario for further TGA analysis under four different atmo-spheres: nitrogen, air, hydrogen, and ammonia. Hydrogen and ammonia are bothreducing gasses, and thus could be used to produce UO, which is the fuel used innuclear reactors, directly from uranous sulfate. e following program was used:1. Switch gas to nitrogen2. Hold for 20min at 30°C3. Switch gas to chosen atmosphere (N2, air, H2, or NH3)4. Heat from 5°C to 1000°C at 5°Cmin−15. Hold for 20min at 1000°C6. Switch gas to air7. Hold for 20min at 1000°Ce data were provided as a function of time, rather than temperature, so thetemperature at any given point was estimated by dividing the elapsed time by thescan rate. e results were normalized as before, with the final weight at step (°C, air) used as the UO basis weight. Results for all four gases are shown inFig. ..e sample behaved identically between – °C under nitrogen, air, and hy-drogen, losing water at the same point and showing the characteristic slow loss ofthe last water. Under nitrogen and air, the sulfate was driven off between – °C in the same manner as was observed at UBC, with air delaying the onset ofFigure .: Normalized TGA scans of U(SO) · HO under nitrogen, air,hydrogen, and ammonia. e atmosphere was switched to air at the endof every test (*). Scan rate  °C /min. Temperature axis approximatedfrom the recorded time using the scan rate.decomposition, and leaving UO as the residual product. Under hydrogen, the in-termediary UOSO was skipped altogether, instead decomposing directly to UOaround  °C. When the atmosphere was switched to air, the UO immediatelyoxidized to UO.Under ammonia, decomposition proceeded quite differently. As soon as theatmosphere was switched to ammonia, the sample increased in weight equivalentto ∆m′ = 27.47, which corresponds to the absorption of . NH molecules. iswas followed immediately by a slow and consistent weight loss until  °C, at whichpoint the mass dropped to the equivalent of UO. e UO again oxidized to UOimmediately upon switching the atmosphere to air... e use of isothermal holds to identify intermediary productse thermoanalytical data presented so far have pointed to the existence of severalintermediary compounds that form during the thermal decomposition of uranoussulfate x-hydrate. In some cases, their existence is obvious: anhydrous U(SO), forexample, has a clear region of stability around  °C. For others, only an inflec-tion point in the TGA data identifies a possible intermediary. UOSO is one suchcase, occurring around  °C. e existence of at least one lower hydrate is alsosuggested by an inflection point between – °C, depending on the startingmaterial.ese intermediaries were investigated further by running theTGAwith a seriesof isothermal hold steps at  °C,  °C,  °C,  °C and  °C, in both nitrogenand air, using to the following program:1. Switch gas to nitrogen or air at 20.0mlmin−12. Hold for 1.0min at 30°C3. Heat from 30°C to 90°C at 50.0°Cmin−14. Hold for 180min at 90°C5. Heat from 90°C to 160°C at 50.0°Cmin−16. Hold for 180min at 160°C7. Heat from 160°C to 300°C at 50.0°Cmin−18. Hold for 60min at 300°C9. Heat from 300°C to 640°C at 50.0°Cmin−110. Hold for 60min at 640°C11. Heat from 640°C to 750°C at 50.0°Cmin−112. Hold for 60min at 750°C13. Heat from 750°C to 995°C at 50.0°Cmin−114. Hold for 1min at 995°Ce TGA curves from the isothermal holds at  °C and  °C, showing waterloss, are given in Fig. .. e hexahydrate and octahydrate both decomposed toa lower hydrate at  °C, reaching a weight equivalent to U(SO) ·1.25HO, whilethe tetrahydrate remained unchanged. At  °C, all three solids decomposed to alower hydrate with between . and . waters.Figure . shows the full isothermal hold program for uranous sulfate tetra-hydrate under both air and nitrogen. During the  °C hold, the solid dehydratedcompletely to U(SO), although, interestingly, it did so at a much slower rate un-der nitrogen than under air. e behaviour diverged further at  °C, with stableUOSO appearing to form under air, but under nitrogen showing only an inflec-tion point. During the  °C hold, the sample decomposed rapidly to UO underboth air and nitrogen.Figure .: Isothermal holds at  °C and  °C under nitrogen showing wa-ter loss for U(SO) ·HO, U(SO) ·HO, and U(SO) ·HO.Figure .: Isothermal holds at  °C,  °C,  °C,  °C, and  °C forU(SO) ·HO under air and nitrogen.Figure .:  °C isothermal holds of U(SO) ·HO in air and nitrogen.e UOSO intermediary was investigated more closely by using a minhold at  °C in both nitrogen and air according to the following program:1. Switch gas to nitrogen or air at 20.0mlmin−12. Hold for 1.0min at 30°C3. Heat from 30°C to 600°C at 40.0°Cmin−14. Hold for 400min at 600°C5. Heat from 600°C to 995°C at 40.0°Cmin−1e weight loss profiles shown in Fig. . confirm that stable UOSO formedat  °C under air, but not under nitrogen... Further study on the phase change in the hexahydrateUranous sulfate hexahydrate was unique in producing an exothermic event (P) inits DSC signature. is event corresponded to a small peak in the DTG curve thatmarked the completion of water loss. is was unusual because, in all other cases,water loss was associated with an endothermic event. It is therefore likely that theenergy released from the P event was actually due to a phase change.To see if the P event was reversible, a sample of hexahydrate was temperature-cycled according to the following program:Figure .: DSC signal during the sequential heating and cooling ofU(SO) ·HO, showing that the P peak is not reversible. Heating/cooling rate °Cmin−, nitrogen atmosphere.1. Switch gas to nitrogen at 20.0mlmin−12. Hold for 1.0min at 30°C3. Heat from 30°C to 360°C at 10.0°Cmin−14. Hold for 5min at 360°C5. Cool from 360°C to 250°C at 10.0°Cmin−1As shown in Fig. ., there was no corresponding endothermic event when thesample was cooled, meaning that the phase change was irreversible.. Interpretation of DTG curveseDTGcurves included in Figs. . to . are the numerically-determinedderiva-tives of the TGAdata, dm′dt , not independent data sets. e TGA andDTG curves aresimply different representations of the same reaction rate data, with the former inintegral form and the latter in differential form. Nevertheless, the DTG curves areuseful for identifying simultaneously-occurring reactions, being composed of a setof overlaid peaks, each representing an individual weight-losing chemical reaction.In addition, since DTG peaks oen mirror DSC peaks, they offered an opportunityto examine events obscured in the DSC signal by background interference above °C... Peak deconvolution methodologye magnitude of the DTG signal at any point can be considered a proxy for theinstantaneous rate of the underlying weight-losing chemical reaction, with the in-tegrated area under the curve representing the sum change in equivalent molecularweight over the period of the integration. If two or more weight-losing events takeplace simultaneously, the observed DTG curve is the sum of the underlying reac-tion rates when scaled to represent the stoichiometric weight change associatedwitheach reaction. For instance, the decomposition of U(SO) to UO was composedof at least two overlapping events, which registered as an inflection point in the TGAsignal, but resolved into several overlapping peaks in the DTG signal.Peak deconvolution techniques were used to estimate the size and shape of dis-tinct, but overlapping, DTG peaks. e peak-fitting soware Fityk .. [] wasused to minimize the error between the experimentally-determined DTG curvesand the sum of any number of adjustable log-normal peaks. e log-normal modelwas chosen because of its ability to simulate asymmetric peaks superficially similarto those seen in the data, thereby granting a good approximation of peak area andposition, even without having a physical significance to this system. As such, infor-mation derived from the deconvoluted peaks, particularly when several events over-lap closely, should not be considered conclusive, but rather as supporting evidenceto be combined with thermodynamic and kinetic theory or other experimental re-sults. No attempt was made to fit a model to the slow weight-loss regions without aclear peak.Integrating under a deconvoluted peak yielded the total change in equivalentmolecular weight associated with a single event n, ∆m′n.∫[dm′ndt][dTdt]−1dT = dm′n = ∆m′n [gmol−] (.)is represents the change in the equivalent molecular weight of the material inthe crucible as a result of a single deconvoluted peak n (e.g., W, S, etc.). eintegrated DTG peak areas for the three uranous sulfate hydrates in nitrogen andTable .: Integrated areas under the deconvoluted DTG (thermogravimetry)peaks for the tetrahydrate, hexahydrate, and octahydrate, representing∆m′ for each weight loss event.Tetrahydrate Hexahydrate OctahydratePeak  nitrogen air nitrogen air nitrogen airW . . . . . .P - - . . - -S . . - - . .Sa . . - - . .Sb . . . . - -S . . . . . .air are tabulated in Table ..e expected ∆m′n for a given event can be calculated from the stoichiometry ofits associated chemical reaction, recognizing that only solid (and liquid) species reg-ister on the TGA. For example, the expected change in equivalent molecular weightexpected for the oxidation of U(SO) to UOSO can be calculated as follows:∆m′{U(SO) oxidation}= MWU(SO4)2−MWUO2SO4= 430.15−366.09 (.)= 64.07If the reaction is for a homogenous decomposition of a single solid reactant (whichis the case for all reactions in this system), ∆m′n is simply the stoichiometrically-weighted sum of the molecular weights of the evolved product gases (SO in thecase of Eq. (.)). For the decomposition of uranous sulfate x-hydrate, the possibleproduct gases are HO, SO, SO, and O, with the equivalent molecular weightlosses shown inTable .. ese can bematched to the areas under the deconvolutedpeaks to gain insight into the underlying reactions... Peak assignment and interpretatione weight loss events fell into three categories: water loss (W), sulfur loss (S), anda small weight change associated with the hexahydrate phase change (P). Each peakTable .: eoretical change in equivalent molecular weight, ∆m′, corre-sponding to the losses of various molecules from a structure.Stoichiometric loss ∆m′·.HO .·.HO .·HO .·.HO .·HO .·HO .·HO .·HO .·HO .SO .SO .SO+  O .was numbered in the same way for each dataset according to its behaviour and tem-perature of occurrence.Water loss always proceeded as two concurrent events: initial rapid loss of mostof the water (W), and slow loss of the remaining water (W). W occurred at dif-ferent temperatures for the three hydrates, but W showed less variation, occurringover a similar temperature range for all three hydrates. ere was no significant dif-ference between air or nitrogen. e areas under theW curves for the tetra-, hexa-,and octahydrate were equivalent to approximately ., , and  waters, respectively.is is fairly consistentwith the isothermal hold results presented earlier in Fig. ..e second type of event, the exothermic phase change (P), appeared defini-tively only in the hexahydrate, and was the same under both nitrogen and air. Itseemed to mark an acceleration of water loss aer a longer period of slow diffusion-controlled dehydration. A possible faint indication of the same event was also ob-served for the octahydrate, although it was not accompanied by a correspondingexothermic peak in the DSC curve. ere was no P peak for the tetrahydrate. earea of this peak corresponded roughly to . HO.e final category of weight-loss event, sulfur loss, was markedly different foreach of the three hydrates. Close examination of the deconvoluted peaks, however,reveals commonalities. e highest-temperature event, the S peak, occurred atthe same temperature for all of the hydrates, peaking at – °C under nitrogenand – °C under air. ermodynamic calculations indicate that the S eventrepresents the decomposition of UOSO to UO, and indeed, the area of the Speaks corresponded quite well to the loss of SO+  O, with a ∆m′ of – undernitrogen. Under an air atmosphere, the S peak was slightly larger, representing a∆m′ of –. e reasons for this are unclear.e remaining peaks, S, Sa, and Sb, must have therefore collectively repre-sented the oxidation of U(SO) to UOSO. Although each of the three hydratesgave a seemingly unique response in this zone, close examination of the deconvo-luted peaks reveals telling similarities. Figure . gives a closer look at the DTGsignals of the three uranous sulfate hydrates in the sulfur loss region between – °C. e S zonewas active in the tetrahydrate and octahydrate (albeit at a highertemperature in the latter), and absent entirely in the hexahydrate. is behaviourwas repeatable in air and nitrogen, across several runs of the same sample, and intwo different samples of tetrahydrate. e bulk of the weight loss occurred aerthis, either as the Sa peak for the octahydrate (– °C in nitrogen and air), theSb peak for the hexahydrate ( °C in nitrogen, – °C in air), or a combina-tion of both for the tetrahydrate. e sum of the areas of the S, Sa, and Sb peakscorresponded approximately to the loss of SO.. Interpretation of DSC curves.. Heats of transformationDifferential scanning calorimetry (DSC) was conducted alongside TGA. In mostcases, each peak in the DTG signal was accompanied by a corresponding change inthe DSC signal. In theory, the integrated area under a peak in a DSC curve repre-sents the heat of reaction, ∆Hrxn, associated with that event.As has already been discussed, the quality of the DSC data above  °C waspoor, requiring aggressive background correction. In addition, the simultaneousoccurrence of the S, Sa, Sb, and S events made it difficult to distinguish theheat flow associated with a single event. is limited the quantitative use of theFigure .: A comparison of the raw and deconvoluted DTG signals of thethree uranous sulfate hydrates from – °C. Nitrogen atmosphere,scan rate  °Cmin−.DSC data to below  °C: in other words, the W, W, and P events.e initial water loss (W) and phase change (P) events both gave clear peaksin the DSC signal. e calculated peak areas are shown in Table .. For the Wpeak, the energy per water molecule is also shown, assuming the number of waterscalculated earlier from theDTGdata (. for the tetrahydrate,  for the hexahydrate,and  for the octahydrate).Table .: Heats of reaction for thermal events observed by DSC (differentialscanning calorimetry) during the dehydration of uranous sulfate x-hydrate,– °C. A positive value indicates an endothermic event.Tetrahydrate Hexahydrate OctahydratePeak  nitrogen air nitrogen air nitrogen airW (kJmol−) . . . . . .W (kJmol− of HO) . . . . . .P (kJmol−) - - -. -. - -In Chapter  it was shown that the theoretical ∆H◦ for complete dehydrationof both the tetrahydrate and octahydrate is approximately  kJmol− of HO. isis consistent with the experimental results for the octahydrate and the hexahydrate,which both gave  kJmol− of HO (under nitrogen), only slightly higher thanpredicted by the thermodynamic calculations. For the tetrahydrate, however, theobserved heat of reactionwas higher, at  kJmol− ofHO(in nitrogen). If the totalenergy is divided by  waters instead of ., however, the value becomes  kJmol−of HO, bringing it in line with the other two hydrates. It is likely, then, that thewater loss peak in the DSC signal incorporates the energies for both the W andW events, as might be expected considering the considerable overlap of the twoevents.e P event, which occurred only in the hexahydrate, registered as an exother-mic peak with an integrated area of −. kJmol−. e significance of this will bediscussed later.. Reaction kinetics during thermal decompositione thermoanalytical experiments described in this chapter were conceived as a wayto study the stability and decomposition pathways of the uranous sulfate x-hydrates,not their decomposition kinetics. Still, a deeper look at the DTG data allows for aqualitative assessment of the kinetics of the system... eoretical kinetics under ideal behaviourBroadly speaking, two different kinetic regimes could control the kinetics of decom-position – diffusion control, and activation energy-control (i.e., Arrhenius). To an-alyze these in the context of the current system, it is useful to introduce the unitlessparameter conversion, X , which represents the mass (or mole) fraction of a reactantthat has been consumed by the reaction. Since the temperature increases at a fixedrate in TGA analysis, there is a linear relationship between temperature and time:[dTdt]= constant [°C s−] (.)e governing rate laws during TGA analysis can therefore be written as a func-tion of temperature instead of time.Under diffusion control, mass transfer of the product gases across a boundarylayer controls the reaction rate. is can bemodelled in its simplest sense accordingto Fick’s Law:[dXdt]= D(pG− pbulkG)[s−] (.)where pG is the partial pressure of the product gas ‘G’ in direct contact with thesolids, and pbulkG is its partial pressure in the bulk gas phase above the sample. Ifboth pG and pbulkG are assumed to be constant in the short term (the former beingset by thermodynamic equilibrium, and the latter a function of reaction rate andpurge gas flow), this results in a constant reaction rate.Under activation energy control, the reaction rate follows the Arrhenius ratelaw:[dXdt]= kexp(−EaRT)[s−] (.)Both diffusion-controlled and activation energy-controlled kinetics could alsobe influenced by conversion X , with the rate decreasing as the amount of remainingreactant declines. For example, this might occur if a growing product layer acts asa mass transfer barrier that increases in length as the reaction progresses. A roughestimation of this effect can be included in the kinetic models by including a (1−X)term:[dXdt]= (1−X)D(pG− pbulkG)[s−] (.)[dXdt]= (1−X)kexp(−EaRT)[s−] (.)e solutions to these four scenarios are plotted in both differential and integralform in Fig. .. Each has a signature shape. us, by comparing the experimenta-lly-determined DTG peaks to these theoretical differential rate curves, it is possibleto gain insight into the underlying kinetic controlmechanismat a given point duringthermal analysis.(a) Diffusion control (b) Activation energy control(c) Diffusion control with (1− X) depen-dency(d) Activation energy control with (1−X)dependencyFigure .: Non-dimensionalized solutions for three possible reaction ratecontrol mechanisms as temperature is increased at a constant rate... Inferring reaction kinetics from peak shapeIt is possible to make a qualitative assessment of the kinetics behind each decompo-sition reaction by comparing the shape of the DTG curves to the theoretical shapesgiven in Fig. .. In general, flat, non-zero regions in the DTG represent diffusion-like kinetics, while peaks represent activation energy-controlled kinetics.Initial water loss (W) clearly had a strong element of activation energy-control.e rate declined somewhat as the reaction progressed, however, suggesting that therate was dependent somewhat on the amount of water remaining in the structure.e remaining water (W) departed at a slow, fairly constant rate akin to diffusion,again with the rate slowing as the amount of water approached zero. e phasechange peak P, seen only in hexahydrate, was too small to interpret easily, but itcould be seen as the sudden vaporization of released bound water as a result of astructural phase change, since free water would rapidly evaporate at > °C.e S, Sa, and Sb peaks all appeared to represent the same chemical reaction,the oxidation of U(SO), but under different kinetic regimes. e first event, S,initially seemed to follow diffusion-like kinetics, but was followed by an accelerationand drop-off similar in nature to the P peak. is was immediately followed bythe activation energy-controlled kinetics of the Sa and/or Sb events. It seems,then, that the oxidation of U(SO) proceeded slowly by diffusion-like kinetics atlow temperature, then switched to activation energy-controlled kinetics at highertemperature, with different variations depending on the original structure of thesolid.e final event, S, corresponded to the final decomposition of UOSO toUO, and consistently showed an activation energy-controlled shape for all sam-ples.. Analysis and mechanism proposal.. An argument in support of the occurrence of a uranous sulfaterecrystallization phase transformationeoxidation of U(SO) to UOSO proceeded alongmarkedly different paths de-pending on the amount of crystalline water in the original hydrate. What could ex-plain the different high-temperature decomposition behaviours of three chemically-similar uranous sulfate hydrate polymorphs that should ostensibly converge to thesame anhydrous form at a thermodynamically-predicted temperature of less than °C? It has already been shown that these differences could not have been causedby differences in particle size, and the data have proven reproducible, so experimen-tal error cannot be blamed. A clue lies in the fundamentally different structures ofthe three hydrates, which was discussed in Chapter . Betke and Wickleder []showed that crystalline U(SO) has a highly cross-linked structure, with each ura-niumatom coordinated in amonodentate fashion to eight sulfate groups, a structurenot shared by any of the uranous sulfate hydrates. is suggests that the conversionto U(SO) involves not only the ejection of water molecules, but also the reconfig-uration of the uranium-sulfur bonds within the crystal lattice.It can be supposed, then, that aer water ejection, the three hydrates must tran-sition through an amorphous phase, am-U(SO), on their way to recrystallizingas orthorhombic uranous sulfate, cry-U(SO). It is possible that this phenomenonwas observed during the bulk decomposition of U(SO) · HO, where a samplecalcined in air at  °C for  h was found to be x-ray amorphous, but a sample ofU(SO) ·HO calcined for  h showed signs of the characteristic x-ray pattern ofcrystalline U(SO). It might be, then, that the exothermic P event represents thecross-linking of the hexahydrate’s chains into the structure of crystalline anhydrousuranous sulfate. Following the same logic, S could represent the same event, butoccurring at a higher temperature, reflecting a higher energy barrier for reconfig-uring the sulfate bonds in the tetrahydrate and the octahydrate.Continuing with the theory that the P peak represents the conversion of amor-phous am-U(SO) to crystalline cry-U(SO), the difference in internal energy be-tween the two phases should manifest as an irreversible thermal event in the DSCcurve. is was indeed observed in the P event in the hexahydrate, as an exother-mic peakwith an integrated area of −. kJmol−. is implies that the crystallinephase has a lower internal energy than the amorphous phase, and therefore shoulddecompose to UOSO at a higher temperature. is, too, was observed, with theSa peak representing the decomposition of am-U(SO), and the Sb peak repre-senting the decomposition of cry-U(SO). If S represents the same phase changeas P, an exothermic thermal event should be associated with it too. Unfortunately,the low quality of the DSC data above  °C made it impossible to unequivocallydistinguish a small exothermic phase change event from the much larger endother-mic events that occurred simultaneously. Nevertheless, slope changes in the DSCcurve around  °C suggest that something did occur... Proposed decomposition mechanism in nitrogene following thermal decomposition mechanism for the uranous sulfate hydratesis proposed based on the experimental and thermodynamic evidence.Water loss for all three uranous sulfate hydrates proceeds in two stages: initialrapid water loss of most of the water (W), followed by the slow diffusion-like lossof the remaining water (W). Each has a transiently-stable intermediary hydrate,containing approximately ., ., and . waters, respectively, with the two mono-hydrates having different structures, denoted α and β , respectively. e remainderof the water departs more slowly under diffusion-like kinetics.Aer all of the water is ejected from the crystal structure, the resulting anhy-drous uranous sulfatemust undergo structural reorganization to reach its crystallineform. Two forms of U(SO) therefore exist at different points during thermal de-composition: first, the amorphous form am-U(SO), then the crystalline formcry-U(SO). e temperature at which this happens depends on the structure ofthe original solid: ∼ °C for the chain-bridged hexahydrate, ∼ °C for the sheet-bridged tetrahydrate, and possibly not at all for the solitary U(SO) ·HO.e oxidation of am-U(SO) to UOSO is initially characterized by slow diff-usion-like kinetics (S) starting around  °C, followed by a transition to rapidactivation energy-controlled kinetics (Sa) starting around  °C. cry-U(SO), incontrast, is weight-stable, showing no low-temperature diffusion-like decomposi-tion, until activation energy-controlled oxidation (Sb) starts around  °C.e final decomposition of UOSO to UO (S) starts around  °C for eachof the three hydrates, overlapping significantly with the oxidation of cry-U(SO).e resulting final product, UO, is stable to at least  °C, the highest temperaturetested.Each of the weight loss and heat flow peaks can be matched to specific chem-ical reactions, as shown in Eqs. (.) to (.). A graphical representation of theproposed decomposition pathways for the three uranous sulfate hydrates is given inFig. ..W: U(SO4)2 ·xH2O−−→ U(SO4)2 · yH2O+(x-y)H2O (.)W: U(SO4)2 · yH2O−−→ U(SO4)2 + yH2O (.)Sa: am-U(SO) −−→ UO2SO4 +SO2 (.)Sb: cry-U(SO) −−→ UO2SO4 +SO2 (.)S: UO2SO4 −−→ 13U3O8 +SO3 +16O2 (.)dehydrationU(SO) ·HO.HOU(SO) ·  HO HOam-U(SO)U(SO) ·HOHOα-U(SO) ·HOHOam-U(SO)U(SO) ·HOHOβ -U(SO) ·HOHOam-U(SO)decompositionam-U(SO)?cry-U(SO)SOUOSOSO + OUOSO§Figure .: e thermal decomposition pathways for U(SO) · HO,U(SO) ·HO, and U(SO) ·HO in nitrogen. Solids in blue, gasesin red. ?erecrystallization temperature of amorphous uranous sulfatedepends on the structure of the original hydrate. § e decompositionof am-U(SO) takes place at a lower temperature than cry-U(SO)... e influence of oxygenOperating in an air atmosphere had only a small effect on the thermal decomposi-tion of the uranous sulfate hydrates. e water loss steps, W and W, were unaf-fected by the presence of oxygen gas. e amorphous-to-crystalline phase changesteps, P and S, were likewise unaffected. is is unsurprising, since oxygen doesnot feature in any of these reaction.e Sa and S peaks, representing the decompositions of am-U(SO) andUOSO, respectively, also behaved as expected. e Sa peak did not shi in thepresence of oxygen, again because oxygen gas is not featured in that reaction. edecomposition of UOSO shied by approximately  °C towards higher tempera-ture in an air atmosphere, reflecting the leward pressure on the equilibrium by theincreased presence of oxygen gas. is shi explains why UOSO was stable dur-ing a long isothermal hold at  °C in air, but not under nitrogen, as was shown inFig. .. It also matches the observations of Tridot [], who found that UOSOwas stable under air and SO, but not under vacuum.e Sb peak, representing the decomposition of cry-U(SO), showed a  °Cshi to higher temperature, which at first seems surprising given that the reaction isnot envisioned to be any different than the decomposition of am-U(SO) in termsof reactants and products. One explanation could be be that, given the significantoverlap of the Sb and S events, their product gases would effectively dilute oneanother, lowering their respective partial pressures and shiing both equilibriumtemperatures lower than they would be at standard state (i.e.,  atm). If S takesplace at a higher temperature, the rate of SO evolution shiswith it, diluting the Sbreaction less, and consequently shiing its equilibrium temperature as well. is didnot affect the Sa reaction simply because it takes place at too low of a temperatureto be influenced by the product gases of S.e behaviour of the system in the presence of oxygen also serves to clarify therole of (and ultimately show that they don’t take place) two reactions that are pre-dicted by thermodynamics, but have thus far been ignored: the heterogeneous oxi-dation of U(SO) with oxygen gas, and the equilibrium-shi reaction between SOand SO. e heterogeneous oxidation of U(SO) is predicted to take place nearly °C lower than homogeneous oxidation, and yet the temperature at which oxi-dation takes place actually increases in air, suggesting that heterogeneous oxidationsimply does not take place due to kinetic or mechanistic limitations. Likewise, ifthe SO/SO equilibrium were active, there would be a gas-phase equilibrium shiaway from SO in the presence of air between – °C, where the reaction isknown to be both thermodynamically and kinetically feasible. e resulting lower-ing of the SO partial pressure could cause a corresponding shi in the oxidationof U(SO) to lower temperature, which did not occur. is suggests that even ifthe SO/SO equilibrium-shi reaction does take place, it does not happen quicklyenough to be relevant during TGA analysis... Estimation of reaction rates and gas-phase compositione molar reaction rates, the associated gas evolution rates, and from them the gas-phase composition, can be estimated from the peaks in theDTGcurves (i.e.,W, S,etc.). First, each of the peaks must be normalized by dividing the dm′dt values by thepeak area (i.e., ∆m′n) to give the reaction rate in terms of conversion X . e molarreaction rate associated with each peak,−rn, can then be calculated by multiplyingthe normalized rate by the total number of moles of uranium present, NU, whichremains unchanged throughout the decomposition process3.− rn =(dNdt)n=dm′ndt∆m′nNU [mol s−] (.)It has already been shown that each peak is associated with an individual chem-ical reaction (Eqs. (.) to (.)). It is therefore straightforward to estimate thegas production rates from stoichiometry:rH2O = (x− y)rW+(x− y)rWrSO2 = rSrSO3 = rSa+ rSbrO2 =rSa+ rSb6rtotal = rH2O + rSO2 + rSO3 + rO2[mol s−] (.)e bulk gas composition above the solids (and in the exhaust) is a function ofthe total rate of gas evolution and the flow of purge gas, rpurge. Neglecting the verysmall contribution to total gas volume from the evolved gases (the purge gas flowrate in the present analysis was three orders of magnitude greater than the maxi-mum gas evolution rate), the partial pressure of a gas in the well-mixed zone abovethe sample is simply its evolution rate divided by the purge gas flow rate. is rep-resents the minimum possible gas partial pressure in thermodynamic contact withthe solids. e maximum possible gas partial pressure, such as would occur deepwithin the sample away from the turbulent flow of the purge gas, is the ratio of that3In the general case, each reaction rate must also be divided by an appropriate stoichiometric coef-ficient. For the homogenous decomposition of uranous sulfate, however, all of the proposed reactionsinvolve exactly one mole of uranium, so no adjustment is necessary.gas’s evolution rate and the total gas evolution rate. e range of possible gas partialpressures in contact with the solid is therefore represented by Eq. (.).rGrpurge< pG <rGrtotal(.)is range of possible partial pressures can be used to compare the observedtemperatures during thermal analysis with the values predicted by thermodynam-ics... A thermodynamic interpretation of uranous sulfatedecompositione thermal decomposition of uranous sulfate x-hydrate broadly followed the ther-modynamic predictions given in Chapter . For some of the weight loss events, theinitial peak temperature (Ti) was below the thermodynamic standard state equi-librium temperature (T ◦eq). is was unambiguously the case for water loss in theoctahydrate (W) and for the final decomposition of UOSO to UO (S). elong diffusion-like period before S, representing the slow oxidation of U(SO),also seemed to begin below its standard state equilibrium temperature, althoughthe signal was too faint to be certain. e reason for this phenomenon can be ex-plained by the way T ◦eq is calculated.T ◦eq is the temperature at which∆Grxn = 0when all reactants and products are atstandard state – meaning, in the ideal system, an activity of  for solids and a partialpressure of  atm for gases. is is always true for the le-hand side of the heteroge-nous decompositions reactions under consideration, which consist of a single solidreactant. However, as has already been shown, the partial pressures of the prod-uct gases in thermodynamic contact with the solid must be < . is lowers thereaction quotient below standard state, which pushes the actual equilibrium tem-perature down.As an example, the thermodynamic feasibility of the final decomposition ofUOSO to UO (S) can be calculated using Eq. (.):∆GS3(T ) = ∆H◦−T∆S◦+RT ln(pSO3 p1/6O2)(.)e minimum and maximum gas partial pressures can be calculated from theexperimental DTG curves using Eqs. (.) to (.). Choosing decomposition ofthe octahydrate in nitrogen at  °C as an example, the range of possible gas partialpressures were .< pSO3 < . and .< pO2 < . atm. is yieldsvalues of −. kJmol− at the minimum gas activity and . kJmol− at themaximumgas activity. erefore it is quite feasible thermodynamically for UOSOto decompose at  °C or lower – but only if the gas phase is substantially diluted.e amount of gas-phase dilution required to make a reaction just thermody-namically feasible (i.e., ∆G = 0) can be calculated iteratively based on thermody-namics and the stoichiometry of the reaction. For the decomposition of UOSOat  °C, a -times dilution is required, with greater dilution required at lowertemperatures. Eventually, the required dilution becomes unachievably high and sothe reaction cannot occur – this is Ti.Two events did not behave according to the gas dilution theory discussed above:the main water loss event for U(SO) ·HO (W), and the oxidation of U(SO)(Sa/Sb). e thermodynamics predict a standard state temperature of  °C fortetrahydrate water loss, and yet the observed Ti was approximately  °C – higherthan the standard state equilibrium temperature. One reason for this could simplybe the uncertainty in the thermodynamic data: the S◦ values for U(SO) · HOlisted in Guillaumont and Mompean [] has a variance of±, which could ac-count for a ± °C difference in the predicted T ◦eq. It could also be related to thethermodynamic stability of the intermediary hemihydrate phase. In the absence ofan easily-distinguishable event in the DSC data to confirm this, however, no expla-nation from the available data is forthcoming.e oxidation of U(SO) (Sa/Sb) also did not conform to thermodynamictheory. While slow oxidation of U(SO) represented by the S zone did commencearound the standard state equilibrium temperature of  °C, rapid decompositiondid not begin until at least  °C higher. is suggests that the reaction is diffusion-controlled to a temperature well above its thermodynamic equilibrium. No obviousexplanation for this behaviour is forthcoming.. Summary and conclusione thermal stability and decomposition of three uranous hydrates was investi-gated: U(SO) · HO, U(SO) · HO, and U(SO) · HO. All were shownto be stable at room temperature, and to follow a different dehydration series, withdifferent intermediary hydrates. e onset temperatures for initial water loss (W)were approximately  °C,  °C, and  °C, respectively, decomposing into lowerhydrates with the estimated composition of hemihydrate, α-monohydrate, and β -monohydrate, respectively. is was followed by a period of slow diffusion-drivenloss of the remaining water.eproduct aerwater losswas amorphous uranous sulfate, am-U(SO), whichtransitioned to its crystalline form, cry-U(SO), aer a period of time. In thehexahydrate, this transformation took place around  °C (P) at a scan rate of °Cmin− and was associated with an exothermic release of energy equivalent to. kJmol−. e same phase transformation appeared to take place in the tetra-hydrate (S), but at ∼ °C, although it was partly masked by the simultaneousoccurrence of other decomposition events. e considerable differences betweenthe hydrates in this respect might reflect the different sulfate bonding modes andcross-linking of the crystal structures of the original hydrates.Oxidation to UOSO behaved differently for am-U(SO) than cry-U(SO).e amorphous form began decomposing rapidly at ∼ °C (Sa). is was pre-ceded by slow decomposition with diffusion-like kinetic characteristics (S), possi-bly reflecting the lack of stabilizing cross-linking in the amorphous structure. ecrystalline form began decomposing rapidly at ∼ °C (Sb), which was not pre-ceded by slow diffusion-like decomposition.Final decomposition ofUOSO toUO began at∼ °C for all three hydratesin nitrogen, and  °C higher under air. is shi reflected the leward pressure onthe thermodynamic equilibrium by the increased presence of O(g). is had anindirect effect on the decomposition of cry-U(SO), pushing it to higher tempera-tures as well, possibly due to the shi in the relative quantities of SO, SO, and Othat influence the equilibrium reactions.e results presented in this chapter could be used in the design of a drying orcalcining circuit for uranous sulfate x-hydrate. If U(SO) · HO were to be themain feed material to such a process, as seems likely, a minimum temperature of °C would be required to dehydrate the material fully. e resulting materialcould be either amorphous or crystalline, depending on the operating conditions,which could have downstream implications. Further decomposition toUO wouldrequire a temperature of > °C, with a higher requirement in the presence of anair atmosphere. is would see the release of equal amounts of SO and SO, as wellas a small amount of O, which would have implications for equipment corrosionand acid plant design.Chapter ekinetics of uranium(IV)oxidation withmolecular oxygenNote: eportion of this work concerning the oxidation of uranium(IV) in perchlo-ric acid was presented as a conference paper at Hydrometallurgy  in Victoria,BC, Canada [].. IntroductionUnder aqueous processing conditions, uranium can exist in its oxidized form (ura-nium(VI) as UO+) or in its reduced form (uranium(IV) as U+). Uranium(IV)will oxidize in the presence of oxygen gas according to the overall reaction given inEq. (.):U4+ +0.5O2 +H2O−−→ UO22+ +2H+ ∆G◦ =−184.5kJmol−1 (.)e unintended oxidation of uranium(IV) would be detrimental to a uranoussulfate precipitation process. It is therefore important to understand how differ-ent parameters affect the kinetics of oxidation. In the present work, the kinetics ofuranium(IV) oxidation by molecular oxygen are examined in the non-complexingperchlorate medium under varying conditions of acid concentration (.–. N),oxygen partial pressure (. and . atm), uranium(VI) concentration, and tem-perature (– °C). An apparent overall rate equation is proposed as a function ofuranium(IV), H+, and oxygen partial pressure. e effect of sulfate, both by oper-ating in sulfuric acid and by adding sodium sulfate to a perchloric acid solution, isalso examined, although the results are inconclusive.. Background information.. Oxidation with molecular oxygen in perchloric acide first rigorous study of the oxidation of aqueous uranium(IV) by molecular oxy-gen in perchloric acid was conduced by Halpern and Smith []. ey proposed asimplified overall rate law that was first order in uranium(IV), first order in oxygen,and inverse-first order in H+. eir rate law was based on the initial reaction rates,not data from the entire duration of the tests, and significant deviation was notedas oxidation progressed. is was particularly true when acid was >. N, whichthey attributed to a slow secondary reaction or the influence of an unknown im-purity or by-product. Halpern and Smith did not state how they calculated oxygenconcentration, making it difficult to validate the results.Halpern and Smith also investigated the effect of various other ions on the ox-idation kinetics. ey found that Fe+, Cl–, and Ag+ inhibit the reaction, Hg+ andCu+ enhance it, and Co+ and Mg+ have no effect. ey attributed the enhancingeffect of Cu+ to the cycling of Cu+ and Cu+, while the effect of the inhibiting ionswas explained as an interruption of the chain mechanism through the destructionof the intermediary products HO or UO+.Sobkowski [] investigated the same system a decade later, observing that thereaction kinetics fell somewhere between first-order and second-order in urani-um(IV). He noted that the underlying reactionmechanism appeared to change withfree acid and temperature, with the most notable difference occurring at low acid.He observed a five-fold increase in kinetics when using oxygen gas instead of air, inline with Halpern and Smith’s proposed first-order relationship in oxygen.Most recently, Shilov et al. [] studied the oxidation of uranium(IV) with oxy-gen gas in weakly acidic and neutral solutions, finding that the reaction proceedsaccording to pseudo-first order kinetics aer an induction period. ey found thatthe kinetics of oxidation are enhanced by uranium(VI), essentially demonstratingan autocatalytic effect since uranium(VI) is a reaction product. ey also observedthat rapid oxidation was usually preceded by an induction period. It is unlikelythat their results apply to the strongly acidic system described in the present work,however... Oxidation with molecular oxygen in sulfuric acidMcCoy andBunzel []were the first to study the oxidation of aqueous uranium(IV)in sulfuric acid, nearly fiy years before Halpern and Smith studied the simpler per-chloric acid system. e authors noted that the apparent rate constant varied capri-ciously between tests, which they blamed on inconsistency in the free acid resultingfrom their method of producing uranium(IV) by zinc metal reduction, althoughthey claimed to obtain more consistent results when using electrolytic reduction.e authors concluded that the oxidation of uranium(IV) in sulfuric acid is firstorder in uranium(IV) and inverse second-order in free acid.Russian researchers Sudarikov et al. [] studied the same system fiy yearslater. Similar to McCoy and Bunzel, they found an inverse second-order relation-ship in H+. ey noted that their assumed first-order rate constant appeared toincrease over the duration of the test. e addition of more sulfate as sodium sul-fate slowed the kinetics of the reaction, which the authors took as evidence for theformation of inactive polynuclear U(IV)-SO– complexes. Sobkowski [] investi-gated the same system much more thoroughly a few years later (and helpfully pub-lished their results in English). He also found that sulfate has an inhibitory effecton the kinetics of uranium(IV) oxidation. e kinetics were complex, followingneither a first-order or a second-order relationship in uranium(IV), and displayedautocatalytic behaviour, becoming faster as the reaction progressed. Južnič and Fe-dina [] confirmed that the reaction rates in sulfuric acid were much slower thanin perchloric acid. e authors proposed a first-order rate law, but noted that itsvalidity was restricted to certain concentrations, and that there appeared to be asignificant induction period. ey found the rate to be inversely proportional toHSO activity... Tracer studiese underlying mechanism for the oxidation of uranium(IV) was studied by Gor-don and Taube [] in a series of clever tracer studies using oxygen gas labeled withthe isotope O. eir results showed that the final oxidation product, UO+, con-tained one atom of O derived from the isotopically-labeled oxygen gas, and oneatom of O presumably derived from hydrolysis with the water of the solution.Given that the oxidation of uranium(IV) to uranium(VI) involves the donation oftwo electrons, and each oxygen atom in O can accept two electrons, this representsthe maximum theoretically-possible amount of oxygen that can be supplied by Oas the oxidizing agent.  of labeled oxygen reported to the UO+, with nonefound in the water or elsewhere.e authors did not report on the effect of sulfate, so it is impossible to saywhether their results also apply in the presence of sulfate... Underlying reaction mechanismIn all of the studies described above, the authors conclude that the hydrolyzed ionUOH+ is the actively-oxidized species [, , ]. is is supported by the inverserelationship between oxidation speed and free acid, since any increase in acid wouldshi the hydrolysis reaction away from UOH+.Many of the studies concluded that the oxidation of uranium(IV) is a first-orderprocess, despite evidence from their own data to the contrary. More cautious au-thors only noted that the rate behaved somewhere between first- and second-orderin uranium(IV), with substantial variation depending on conditions. is was par-ticularly true in the sulfate system.Sobkowski suggested that the apparent change in kinetic behaviour at low acidis due to the formation of the higher hydrolysis product U(OH)+ with differentkinetics than UOH+, although he claimed that the kinetics of the latter are so slowthat they can be neglected. Lotnik et al. [] point out that, in any case, U(OH)+only exists in appreciable amounts above pH ., which is well outside of the rangeunder consideration here.Halpern and Smith proposed a chain reaction mechanism in perchloric acid in-volving a series of single-electron transfers, with the intermediary products UO+and HO. However, this mechanism was disproved by Gordon and Taube, whopointed out that such amechanismwould result in .–. oxygens in UO+ deriv-ing from the O gas, rather than the . observed from their isotopic tracer studies.Fallab [] proposed a direct two-electron process that is compatible with Gor-don andTaube’s tracer studies involving the formation of a uranium–oxygen adduct:UOH3+ +O2 −−→ U(O2)OH3+U(O2)OH3+ +UOH3+ −−→ 2UO22+ +2H+(.)Shilov et al. [] proposed a multi-step mechanism involving uranium(IV) hy-drolysis, the reaction of both uranium(IV) and uranium(V) with oxygen, and thedissociation of uranium(V).eir experimentswere conducted in the pH range .–., however, and so likely do not apply to the present work.Explanations for the effect of sulfate on the oxidation kinetics are less robustand fraught with speculation. Sobkowski [] proposed that the presence of sulfateslows the oxidation kinetics because uranium(IV) forms strong sulfate complexeswhich compete with the hydrolysis reaction, thus reducing the amount of UOH+available to be oxidized. In their study on photoelectrical properties of uranium(V)[], the authors also advanced a theory that the presence of sulfate promotes thedisproportionation of uranium(V), thus interrupting the oxidation reaction chainand consequently slowing the oxidation reaction.Elliot et al. [] found that introducing -propanol as a radical scavenger madethe reaction rate faster in sulfuric acid, implying that free radicals, such as the hy-droxyl radical, are involved in the mechanism. ey proposed a complex mecha-nism involving a free-radical chain sequence that applies to the sulfate system at pH.... Other related studiesSobkowski [] showed that exposure to light from a tungsten lamp causes the redoxpotential of uranium(IV)/uranium(VI) to fall, corresponding to the photoreductionof uranium(VI) to uranium(V). ey also observed that UV light has a significanteffect on the kinetics of oxidation by oxygen gas [].Lotnik, Khamidullina, Kazakov, published a series of papers on the chemilumi-nescence produced during the oxidation of uranium(IV) with molecular oxygen inperchloric acid [, , , ]. e authors were only able to detect chemillumines-cence in <. N perchloric acid, and not at all in sulfuric acid. ey proposed thatthe oxidation of UO+ to UO+ is the luminescent step, with the electron-excited(UO+)* being the chemiluminescent emitter. e authors followed the kineticsof the reaction using light intensity as a proxy for concentration, apparently with-out ever measuring uranium(IV) concentration directly, so it is difficult to knowwhether their conclusions on the reaction kinetics are valid. ey found that theoxidation kinetics depend heavily on the surface area of glass packing in contactwith the solution, which they explained as the heterogeneous decay of radicals atthe solid-liquid interface. ey also found that UO+ catalyzes the reaction, possi-bly due to a shi in the uranium(V) dissociation reaction. ey noted that it takessome time to reach themaximum light emission, suggesting that themaximum rateof UO+ generation does not occur at the time of maximum U+ removal (the be-ginning of the reaction), which is indicative of consecutive reactions.. ExperimentalTwenty oxidation tests were conducted in perchloric acid under varying conditionsof acid concentration, oxygen partial pressure, temperature, total uranium concen-tration, and sulfate concentration. Four additional tests were run in sulfuric acid.All tests were conducted in a jacketed  mL glass reactor with a five-port glass lid.A titanium gas-shearing impeller rotating at  rpmwas used to bothmix the solu-tion and disperse the gas. Gas was injected through a fritted glass lance positioneddirectly below the impeller. A schematic of the setup is shown in Fig. .. e testparameters are given in Table .. A blank copy of the operational worksheet andchecklist that was completed for each test is given in Appendix E... Solution preparatione uranium used to create all test solutions came from unenriched reactor-gradeUO from Cameco Corporation’s Blind River Refinery in Blind River, Ontario,Canada. e oven-dried UO assayed . uranium (theoretical value .),and ICP analysis confirmed that no trace metals were present (<. µg g− Ag, Al,Figure .: Schematic of oxidation study experimental setupAs, Ba, Be, Cd, Co, Cr, Cs, Cu, Fe, Ga, Mn, Mo, Ni, Pb, Rb, Se, Sr, Tl, V, Zn).e stock uranium(IV) solutions were prepared differently depending on whe-ther the oxidation studies were to be conducted in perchloric acid (tests –) orsulfuric acid (–). For tests –, to ensure the solutions were free from all im-purities, the UO was dissolved in dilute perchloric acid, precipitated as uranyl per-oxide by hydrogen peroxide addition, and then filtered and washed thoroughly withdeionized water. e precipitate was then redissolved in hot .mol L− perchlo-ric acid to create a pure uranium(VI) stock solution. For tests –, the uraniumwas purified by dissolving UO in sulfuric acid, precipitating it as uranous sulfateby applying heat, and then redissolving the precipitate in sulfuric acid. e purifiedsolutions were then reduced in several batches by electrolytic reduction in a two-electrode membrane cell according to the procedure outlined in Appendix A. Eachbatchwas assayed for uranium(IV), total uranium, and free acid, andwas stored un-der a nitrogen atmosphere in a Sigma Aldrich Sure/Stor air-free storage flask untilneeded.e test solutions for each experiment were made as needed by diluting the re-Table .: Test conditions for the oxidation studies. All studies were conductedin perchloric acid, except tests –, which were conducted in sulfuricacid.Test  [UT], mol L− [H+], N Temp., °C Gas [SO–]T, mol L− . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Oxygen - . . . Air - . . . Air - . . . Oxygen . . . . Oxygen . . . . Oxygen . . . . Oxygen . . . . Oxygen .–† . . . Oxygen .† sulfuric acidquired reagents in a mL volumetric flask while purging the headspace of theflask with inert gas to minimize oxidation during mixing. For tests –, aciditywas adjusted with  perchloric acid, which was first diluted to .mol L− andstandardized by titration with standard sodium hydroxide. For tests –, acid-ity was adjusted with concentrated sulfuric acid similarly diluted and standardized.e compositions of all test solutions were validated by chemical analysis, and in allcases agreed well with the theoretical dilution values... Continuous monitoring of uranium(IV) concentration by UV-Visspectroscopyeuranium(IV) concentration wasmonitored continuously in situ using a Perkin-Elmer Lambda  UV-Vis spectrometer connected to a flow-through cuvette andcontinuous sampling pump. Using this method provided immediate feedback thr-oughout the experiment, and also eliminated some of the problems associated withthe more traditional technique of periodic sampling, such as sample decay by con-tinued oxidation. So many data points were collected that the concentration vs.time curves could be considered continuous functions, which allowed for numeri-cal differentiation techniques1 to be applied to calculate the instantaneous reactionrate at any point during a test.Solution was continuously drawn from a sampling port at the bottom of the re-actor using a peristaltic pump, passed through a debubbler to remove any entrainedgas bubbles, and then pumped through a  cm quartz flow-through cuvette insidethe spectrometer. A soware interface to the spectrometer recorded the absorbanceof the solution at . nm once per second (less frequently for long tests). esampling system had a total holdup volume of mL and a flow rate of mLmin−,which was deemed adequate for providing a representative sample of the contentsof the reactor at any given time without skewing the kinetics by withholding a largeamount of solution from exposure to oxygen.Before each experiment, a calibration curve was generated relating absorbanceat . nm to uranium(IV) concentration using six standard solutions, rangingfrom .–. mol L− uranium(IV). e standards were matrix-matched to thetest solution, and were prepared fresh for each experiment. e composition of theuranium(IV) stock solution used to prepare the standards was validated by titrationevery day. e standards were temperature-controlled in a water bath set to thesame temperature of the test to account for the effect of temperature on absorbance... Gas injection. oxygen gas or breathing-grade air was supplied from a compressed gas cylin-der through a fritted glass lance positioned just below a gas-shearing impeller. Be-1Numerical differentiation was performed using the Python package NUMDIFFTOOLS [].fore entering the reactor, the dry gas was saturated with water vapour by bubbling itthrough amagnesium sulfate solution at the same temperature and ionic strength asthe test solution in order to prevent evaporative losses. Flowwas controlled using anAalborg flow tube meter set to  std mLmin− (oxygen) or  std mLmin−(air). Before the commencement of each test, the reactor was continuously purgedwith water-saturated nitrogen in order to prevent premature oxidation from infil-trate air... Temperature monitoring and controle temperature of the solutionwas controlled using a heated circulating water bathconnected to the jacketed reactor. e temperature was continuously monitoredand recorded using a PTFE-coated RTD probe and an Omega logging temperaturecontroller connected to a computer. e temperature probe was calibrated againstdeionized ice water ( °C) and boiling water ( °C).During preliminary tests, it was found that the solution temperature oen in-creased by  °C or more over the course of an experiment. is may have been dueto the exothermic nature of the reaction, or the heat generated by the aggressive ag-itation. In any case, such variation in the temperature was deemed unacceptable forthe accurate study of kinetics. In order to gain finer control over the temperature,a glass “cooling finger” was immersed in the solution. e temperature controllerwas set to open and close a solenoid valve as needed, pulsing  °C water throughthe finger, on an as-needed basis. Temperature control of ±. °C was achieved byusing both the cooling finger and the water jacket together.. Validation of experimental method.. UV-Vis spectroscopyScans from – nm showed that uranium(IV) has several strong absorbancepeaks in the – nm spectral range in both perchloric and sulfuric acids, withthe two strong peaks centered around . nm and . nm both free from in-terference from uranium(VI) (Figs. . and .). A position near the top of thestrongest peak, at . nm, was chosen for continuousmonitoring in both systems.Figure .: UV-Vis spectra in . N perchloric acid, °C, quartz cell,  cmpath length, deionized water as background reference. (a) UV-Vis spec-tra for . mol/L uranium(IV) or uranium(VI); (b) absorbance vs. ura-nium(IV) concentration at . nm.Figure .: UV-Vis spectra in . N sulfuric acid, °C, quartz cell,  cm pathlength, deionized water as background reference. (a) UV-Vis spectrafor . mol/L uranium(IV) or uranium(VI); (b) absorbance vs. ura-nium(IV) concentration at . nm.e relationship between absorbance and concentration at . nm was found tobe linear in the range .–. mol L− in both perchloric and sulfuric acids, re-gardless of acid concentration, which made it possible to calculate the uranium(IV)concentration of a solution directly from absorbance measurements using Beer’sLaw:A = e`c (.)where A is the measured absorbance, ` is the path length, e is the extinction coeffi-cient, and c is the concentration. e value of ewas calculated from the straight-linefit of the calibration curve constructed from measurements on standard solutions.e oxidation of onemole of uranium(IV) by oxygen gas produces twomoles ofH+ (see Eq. (.)), which caused the acidity of the test solutions to increase slightlyover the course of an experiment. Absorbance spectra are oen affected by pH,so even this minor increase in acid could invalidate a calibration curve to convertabsorbance to concentration. Figure . shows that there is a positive relationshipbetween the extinction coefficient, e, and perchloric acid concentration, with a par-ticularly strong effect at low acid. is introduced a systematic error into everyexperiment, with the magnitude of the error growing as H+ was generated. Basedon the relationship shown in Fig. . for a test containing .mol L− uraniumand ≥. N acid, the corresponding increase in H+ of .mol L− would resultin a measurement error of about  at completion, and proportionally less thanthat earlier in the oxidation process. is error was deemed minor enough to beaccepted within standard experimental error.e method of monitoring uranium(IV) concentration using UV-vis spectro-scopy was validated during test  by periodically withdrawing samples from thereactor and immediately titrating them for uranium(IV), with the results shown inTable .. e error was within  for the first three samples, and showed bias con-sistent with the continued oxidation of the sample between the time of samplingand the time of analysis. e error on the last sample, however, was . issample was taken five hours into the test aer more than  of the uranium(IV)had oxidized. Some of the error can be explained by the inherently lower accu-racy of a titration when titrating for small amounts, and also by the magnifyingeffect of calculating error on small values. Signal dri in the spectrometer’s zero-Figure .: e change in extinction coefficient observed with increasing per-chloric acid concentration, for uranium(IV) at  °C.baseline over the long time span of the test could also have introduced error, as couldthe inherently higher error associated with measuring low absorbances. Whateverthe cause, it was clear that measurements at low absorbance or aer a substantialamount of time had elapsed since calibration were less accurate than those madeat higher absorbance or soon aer calibration. For this reason, only data collectedabove .mol L− were used in the present analysis. is was not detrimentalto the quality of the results, however, since it still allowed for the collection of thou-sands of data points per test spanning a four-fold drop in concentration.Table .: Comparison of uranium(IV) assays by titration and continuous UV-Vis spectroscopy on samples withdrawn from the reactor during test .Elapsed [U(IV)], mol L− Errortime, min Titration UV-Vis mol L− Percent . . +. . . . +. . . . +. . . . -. -... Gas flow rate and stirring speede fastest observed oxidation rate of any test was .×− mol L− s, which oc-curred at the beginning of test . Assuming stoichiometric oxygen consumption anda reactor volumeof . L, this amounted to anO consumptionof .×− mol s−,or ∼.mLmin−. Oxygen was injected at mLmin−, or  times the max-imum consumption rate. No change in the oxidation rate was observed when re-ducing the gas flow rate by , confirming that enough gas was being supplied toensure oxygen saturation.No decrease in reaction rate was observed with a  reduction in stirring rate,from  rpm to  rpm, confirming that the reactor was well-mixed and that theoxygen bubbles were being adequately dispersed... Evaporative lossese dry gas was pre-saturated with water vapour in order to prevent evaporationof the test solution. To test the effectiveness of this method, nitrogen was bubbledthrough the reactor at  std mL s− for . h while monitoring the uranium(IV)concentration byUV-Vis spectroscopy. As can be seen in Fig. ., [U(IV)] remainedstable throughout the test, neither increasing (from evaporation) or decreasing sub-stantially (from oxidation), confirming that the practice of pre-saturating the gaswith water vapour with a bubbler was effective. e very slight decrease over themin duration of the test amounted to approximately . and was deemed tobe acceptable within experimental error... ReproducibilitySeveral tests were repeated in order to confirm the reproducibility of the results. Inperchloric acid, the repeated tests  and ,  and , and  and  all gave near-identicalresults. In sulfuric acid, however, repeated tests – gave widely different results.is will be discussed later in the chapter.. Rate equation methodologye rate of uranium(IV) oxidation could be affected by several parameters, includ-ing uranium(IV) concentration, [H+], oxygen partial pressure, uranium(VI) con-Figure .: e stability of uranium(IV) over time during the bubbling ofwater-saturated nitrogen. Main graph: exaggerated y-axis, showing a veryslight decline in concentration over time; Inset: full range.centration, and temperature. e following generalized overall rate law is thereforeproposed as a basis for analyzing and interpreting the oxidation kinetics2:−d[U(IV)]dt= k[U(IV)]a[H+]b(pO2)c[U(VI)]d k = Aexp(− EaRT)(.)In this model, the reaction rate is an exponential function of the various concentra-tions or partial pressures (a, b, c, and d), and follows an Arrhenius relationship withtemperature. is leads to the concept of a reaction order. For example, if a= 1, thereaction is said to be first order in uranium(IV), and if b = −2, the reaction is saidto be inverse-second-order in H+.2e general rate equation shown in Eq. (.) includes separate terms for uranium(IV) and ura-nium(VI), even through the two are related through a constant representing total uranium, [U(VI)] =[UT]− [U(IV)]. Applying this substitution removes a degree of freedom from the equation, thus tech-nically making it mathematically simpler. Nevertheless, uranium(VI) is a distinct entity in solution,and removing it from the rate equation using a stoichiometric link obscures the fundamental under-standing of the oxidation rate with respect to the two distinct forms of aqueous uranium. It also makeit more difficult to isolate uranium(IV) for the graphical interpretation of kinetic data. Given thatthe uranium(VI) concentration is trivial to calculate, the stoichiometric simplification has not beenapplied here.e general rate equation shown in Eq. (.) can be rearranged such that all ofthe parameters except [U(IV)] are lumped into the apparent rate constant, k′. isgives the following equation:−d[U(IV)]dt= k′[U(IV)]a k′ = k[H+]b(pO2)c[U(VI)]d (.)Each of the parameters shown in Eq. (.) can be resolved by measuring the oxida-tion rate across several tests where one parameter is varied while every other param-eter is held constant. For pO2 this is straightforward, and can be accomplished byvarying the composition of the gas sparged into the reactor. H+ is similarly straight-forward, if it is assumed that the acidity is constant throughout a given test. is isnot strictly true, as examination of Eq. (.) shows that two moles of H+ are gen-erated per mole of uranium(IV) oxidized. However, if the amount of acid initiallyin the test solution is much greater than the amount generated, the effect should besmall. e effects of uranium(IV) and uranium(VI) are more difficult to examineindependently, given that both are in flux throughout a test. It is therefore essentialthat comparisons between different tests are only made at a point where the con-centrations of uranium(IV) and of uranium(VI) match.is technique can be demonstrated using H+ as an example. In this case, ξ isthe lumped parameter incorporating the effects of everything except H+. Providingthat a point of identical uranium(VI) and uranium(IV) is selected for comparingthe rates of different tests, ξ should be a constant when comparing tests at different[H+]. Taking the logarithm of both sides of Eq. (.) and rearranging to separateout the constant ξ yields Eq. (.):ξ = k(pO2)c[U(VI)]d = constantlnk′ = b ln [H+]+ lnξ(.)If the apparent rate constants k′ for several tests at different [H+] are plotted against[H+] on a log-log scale, the resulting data should be linear, where the slope is b,the reaction order in acid. e same procedure can be repeated for a series of testsvarying pO2 to get the reaction order c in oxygen.Figure .: Oxidation rate vs. concentration plots for tests -, illustrating theeffect of (a) [H+], (b) temperature, (c) pO2 , and (d) [U(VI)], inset: scaledaxes. Test number shown in square brackets. Refer to Table . for testconditions.. Results and discussion: oxidation in perchloric acidPlots of the oxidation rate vs. uranium(IV) concentration for tests – are shownin Fig. .. ese plots show clear trends with respect to H+, oxygen partial pres-sure, and temperature, indicating that all three are involved in the rate law. esedata can be used to determine the reaction orders in uranium(IV), H+, pO2 , anduranium(VI)... Reaction order in uranium(VI)To determine the reaction order with respect to uranium(VI), d, several tests wererun at  °C with oxygen gas, but with total uranium concentrations of . (test), . (test ), and . (test ) mol L−. is had the effect of introducing adifferent ratio of uranium(VI) to uranium(IV) at equivalent points of uranium(IV)concentration (at . mol L− uranium(IV), this amounted to a VI to IV ratioof ., ., and ., respectively). e oxidation rates for these tests are shown inFig. .d.On the main graph shown in Fig. .d, the oxidation rates appear to overlapclosely, regardless of uranium(VI) concentration. When the graph is scaled to bettershow the differences (Fig. .d inset), the rates still appear to overlap quite closely,and do not show a particular trend with respect to increasing uranium(VI). It isclear, then, that [U(VI)] does not significantly affect the rate of uranium(IV) ox-idation under the conditions tested, leading to the conclusion that the reaction iszero-order with respect to uranium(VI) (i.e., d = )... Reaction order in uranium(IV)It has been shown that the reaction kinetics are zero-order in uranium(VI). eapparent overall rate constant k′ can therefore be considered a constant over theduration of a single test (neglecting the small change in [H+] caused by acid gener-ation). Since the only kinetically-important parameter that changes throughout atest is the uranium(IV) concentration, the reaction order in uranium(IV), a, can bedetermined directly from the concentration vs. time data.e classic technique for determining the reaction order is to integrate the ap-parent rate equation shown in Eq. (.) for various assumed values of a and plotthe results, with a linear plot indicating a good fit to the assumed order. Figure .shows two such plots, for a =  (ln [U(IV)] vs. time) and a =  ([U(IV)]−1 vs. time),for representative tests at high and low acid. e low-acid test (test , . N) fitsa first-order model well. e high-acid test (test , . N), however, fits neither afirst-order or a second-order model, falling somewhere in between. All tests with[H+]≥ . N gave plots where 1 < a < 2.e actual value of a could be guessed by solving Eq. (.) for various valuesof 1 < a < 2, plotting the results, and trying to find a value for a that causes thedata to fall in a straight line. Passing judgement on a line’s “straightness”, however,seems to be a rather subjective exercise, particularly when a is not limited to par-ticular whole-number values. Fortunately, the availability of instantaneous reactionFigure .: First- and second-order rate plots of two tests in perchloric acid.Straight lines indicate that the model applies to the data.rate data made available an alternate method for determining the reaction order inuranium(IV). Taking the logarithm of both sides of Eq. (.) yields Eq. (.):ln(−d[U(IV)]dt)= lnk′+a ln [U(IV)] (.)By plotting ln(−d[U(IV)]/dt) vs. ln [U(IV)], the reaction order a can be deter-mined directly from the slope of the line. ese are shown in Fig. .. For tests with≥ . N acid (Fig. .a), the straight-line fit to the data consistently has a slope veryclose to .. is suggests an effective . order relationship in uranium(IV). Testsat ∼. N acid, however, appeared to follow first-order kinetics (Fig. .b), agreeingwith the straight-line fit to the first-order rate plot seen in Fig. .b. is points tothe potential existence of two competing reaction pathways, one dominant at highacid and the other at low acid. Notable in the low-acid tests is the slight negative de-viation from the first-order straight line that grows as the reaction progresses. iscould be reflective of the small systematic error introduced by the H+ generated bythe oxidation process, which has the effect of slowing the reaction, and is not nec-essarily a departure from the first-order relationship. e effect of H+ generation ismuch less significant in tests with proportionally more acid, hence the absence ofthe phenomenon in those tests.Figure .: ln-ln plots of oxidation rate vs. [U(IV)], with data from a represen-tative selection of tests in perchloric acid. e slope of the line indicatesthe reaction order... Reaction order in H+ and oxygene apparent rate constant k′ was calculated for each experiment by dividing theinstantaneous oxidation rate at [U(IV)] = .mol L− ( completion for mosttests) by .. (the assumed reaction order a), according to Eq. (.). oughthe choice of .mol L− as the reference point to compare tests is somewhat ar-bitrary, it is suitable because it is well within the range where the UV-Vis gives ac-curate results (.–. mol L−), and it also avoids the “ramp up” period whenoxygen gas is still reaching saturation at the beginning of the test.e reaction order in [H+], b, was determined by varying [H+] between .–. N (initial concentration) in a series of tests at  °C using oxygen gas. eacid concentrations used in the calculations were based on the assays on the fully-oxidized final solutions, which were then back-calculated to the assumed value at.mol L− uranium(IV) to take into account the H+ generated by the oxidationprocess. Figure . shows a plot of lnk′ vs. ln [H+]. Between .–. N, the relation-ship is linearwith a slope of -., suggesting that inverse second-order kinetics withrespect to [H+] are valid in this region. At . N, there is some deviation from theFigure .: e effect of H+ on the apparent rate constant in perchloric acid,showing the linear region when [H+]≥ . N.inverse second-order relationship, although it is not clear whether the deviation issignificant. At . N, the data show a significant departure from the inverse second-order relationship. If the fundamental reaction order in uranium(IV) is different inlow acid, however, as was shown earlier, the apparent rate constants cannot be di-rectly compared and thus cannot be expected to fall on the same line.To determine the reaction order with respect to oxygen, c, the oxygen partialpressure was varied by running tests with oxygen gas (. O) and air (.O). At atmospheric pressure, this corresponds to pO2 = . and . atm, re-spectively. Two sets of tests were run: the first set in . N acid, and the second in. N acid. In both cases, the oxidation rate was found to be very close to first-orderwith respect to oxygen partial pressure, as shown in Fig. ..If these data are plotted on a linear scale instead of a linear scale, a third datapoint can be added at the origin. is is because the reaction rate was observed tobe zero under a nitrogen atmosphere, or pO2 = 0. All three points fall on the sameline passing through the origin for both the high and low acid conditions.Figure .: eeffect of oxygen partial pressure on the apparent rate constant... e effect of temperaturee effect of temperature on the rate constant was assumed to follow an Arrheniusrelationship, k = Aexp(−Ea/RT ). Tests at  °C,  °C, and  °C were run in .N perchloric acid using oxygen gas. e Arrhenius plot of lnk′ vs. 1/T is shown inFig. .. From the plot, the Arrhenius parameters were calculated as Ea = ,Jmol− and A = .× mol.L−.atm−s−, with A calculated from Eq. (.)assuming . order kinetics in uranium(IV), inverse second-order kinetics in H+,and  order kinetics in uranium(VI). is activation energy is consistent with achemical reaction-controlled mechanism... Proposed apparent overall rate equationBased on the analysis above, the apparent overall rate law for the oxidation of ura-nium(IV) by oxygen gas in perchloric acid when [H+]≥ . N is as follows:− d[U(IV)]dt= k[U(IV)]1.5pO2[H+]2(.)k = 1.25×1013 exp(−9.65×104 Jmol−1RT)Figure .: Arrhenius plot showing temperature dependence of oxidation ki-netics at . mol L− uranium(IV), [H+] = . N.In this case, k has the units mol.L−.atm−s−. ere has been no attempt totake into account or otherwise interpret the underlying elementary reactions thatultimately dictate the overall reaction rate. Not enough data were collected at lowacid to draw conclusions on the rate law under those conditions, although the cur-rent evidence suggests it is st order with respect to uranium(IV).Figure . shows a comparison of experimental data to the modelled proposedrate law. For themodel, the inputs consisted of the initial operating parameters fromeach test, as shown in Table ., and took into account the H+ generated over thecourse of the reaction. e experimental data are shown time-shied in order toestablish a uniform starting condition. e model appears to give a good fit to theexperimental data, accurately predicting the effect of acid, oxygen partial pressure,and temperature. Low-acid samples exhibit substantial deviation from themodelledrate law, as expected given that .-order kineticswere foundnot to apply at low acid.Figure .: Comparison of modelled and experimental data. Model fromEq. (.) using initial experimental parameters. Solid lines: experimen-tal data. Dashed lines: simulation results.. Results and discussion: the effect of sulfatee oxidation of uranium(IV) in perchloric acid gives a glimpse of the underlyingchemistry under simple, non-complexing conditions, but does not represent thereality of an industrial processes. Sulfuric acid would almost certainly be the choiceof acid in a real plant design. It would therefore be more useful to understand theoxidation kinetics in sulfuric acid.Figure . shows the results from four identical oxidation studies, at . Nsulfuric acid,  °C, using oxygen gas. e results from a test in . N perchloricacid is shown for reference. e sulfuric acid tests were completely unreproducible,with the initial oxidation rate of the fastest case (test ) approximately  timesgreater than the slowest (test ). e slopes of the ln-rate vs. ln-concentrationplots (Fig. .c) also did not give a consistent reaction order, ranging from a slopeof . in the fastest case (test ), to nearly zero (i.e., constant rate) in the slowestcase (test ).No explanation for the unreproducible behaviour in sulfuric acid is forthcom-ing. e tests used identical equipment, reagent bottles, stock solutions, and pro-cedures, and were even run at the same time of day (starting between :-:).Given that the same reagents were used to prepare each test, there was no obvioussource of an inhibiting or enhancing spectator ion that would vary between tests.ere was no obvious induction period in any of the tests.To gain some insight into why the tests in sulfuric acid were unreproducible,several tests were run in perchloric acid with sulfate addedwith sodium sulfate. Fig-ure . shows the effect of adding sodium sulfate in a :, :, and : SO–:Umo-lar ratio to tests in . N perchloric acid. e presence of sulfate caused a dramaticslowdown in kinetics; a -fold decrease in initial oxidation rate was observed at thehighest sulfate:uranium ratio. More revealing, however, is the change in the appar-ent reaction order in uranium(IV). Figure .c shows that the apparent reactionorder in uranium(IV) (as indicated by the slope of the ln-rate vs. ln-concentrationplots) drops from . in the case of no sulfate, to . at medium sulfate, to essen-tially zero at high sulfate. is means that the reaction kinetics become less depen-dent on uranium(IV) in the presence of sulfate, suggesting a change in the underly-ing reaction mechanism to a rate-limiting step not directly involving uranium(IV).Figure .: Results from four identical oxidation tests in . N sulfuric acid, °C, oxygen gas, showing non-repeatability. (a): concentration vs.time; (b): rate vs. concentration; (c): ln-ln plot of rate vs. concentra-tion. Dashed line: . N perchloric acid, for comparison.Similar tests were conducted at . N acid, where sodium sulfate was added invarious sulfate:uranium ratios, as shown in Fig. .. e addition of sulfate in a.: and : molar ratio with uranium again caused a slowdown in the kinetics, al-though the reaction order appeared to remain .-order in uranium(IV). At a :ratio, however, the reaction sped up, and appeared to change to a different underly-ing mechanism (nominally .-order in uranium(IV) according to the slope of theln-rate vs. ln-concentration plot).It would seem, then, that small amounts of sulfate, up to a : ratio with ura-nium, cause the kinetics of the oxidation reaction to slow down. is is likely dueFigure .: e effect of adding sodium sulfate on the oxidation kinetics in. N perchloric acid. (a): concentration vs. time; (b): rate vs. concen-tration; (c): ln-ln plot of rate vs. concentration.the formation of the inactive U(SO)+ andU(SO) aq species, which effectively re-duces the concentration of active UOH+ in solution available to be oxidized. How-ever, as the sulfate:uranium ratio increases to : or above, the fundamental reac-tion mechanism appears to change, with a trend towards a constant (i.e., -order)reaction rate.No obvious explanation for the strange kinetic behaviour in the presence of sul-fate is forthcoming. Normally, a constant reaction rate could indicate a mass trans-fer limitation, as might occur if oxygen cannot dissolve and diffuse quickly enoughto match the underlying fundamental kinetics. is cannot be the case here, how-ever, because no oxygen mass-transfer limitation was observed in the perchloricFigure .: e effect of adding sodium sulfate on the oxidation kinetics in. N perchloric acid. (a): concentration vs. time; (b): rate vs. concen-tration; (c): ln-ln plot of rate vs. concentration.acid tests, which were operated under the same conditions. It would seem, then,that the presence of a large amounts of sulfate causes a change in the fundamen-tal underlying reaction mechanism to one independent of uranium(IV) concentra-tion. e unusual non-fractional reaction orders observed in varying amounts ofsulfate may indicate the occurrence of two parallel reaction mechanisms, one a .-order reaction involving UOH+ and the other a constant-rate reaction involvingU(SO)+ or U(SO) aq.Further attempts to elucidate the underlying kinetics in sulfuric acid were un-successful. e non-repeatability in high-sulfate tests highlighted in Fig. . madeit impossible to objectively compare tests at different acid concentrations and tem-peratures, so the dependency of the kinetics on acid and temperature could not beestablished. e non-repeatability in the sulfate system may have indicated advan-tageous catalysis with an unidentified trace impurity, although the source of suchan impurity is unknown given that the same reagents and conditions were used forall of the tests. It could also have been due to an unidentified physical phenomenon,such as differences in the surface conditioning of the titanium impeller, or the for-mation of a nano-scale intermediary precipitate. However, the good repeatabilityobserved in the perchloric acid system makes these explanations unlikely.. ConclusionsIn the present work, an overall equation to approximate the rate of uranium(IV) ox-idation in perchloric acid has been proposed. In the range [H+] = .–. N, it wasfound that the kinetics follow a .-order relationship with respect to uranium(IV),with the apparent rate constant inversely proportional to the square of [H+], pro-portional to oxygen partial pressure, and unaffected by uranium(VI).When [H+]≤. N, the underlying mechanism appears to shi towards a first-order relationshipin uranium(IV). e results follow the same trends described in work published byother authors [, ], but the use of continuous data logging and the applicationof numerical differentiation in the present work permitted a more detailed studyof the kinetics. e fractional exponents in the proposed overall rate law point toa non-elementary, multi-step reaction mechanism possibly involving simultaneousparallel reactions. Further study and analysis is needed, however, beforemeaningfulconclusions on the mechanism can be drawn.e presence of small amounts of sulfate cause the oxidation rate to decrease,likely due to the preferential formation of uranium(IV) sulfate complexes at the ex-pense of the active species UOH+. e presence of large amounts of sulfate appearsto cause a shi in the underlying reaction mechanism. Substantial variability in theobserved kinetics, however, made it impossible to suggest a rate law in the sulfatesystem.It may be possible to design conditions under which uranium(IV) is resistantto oxidation by molecular oxygen: namely high-acid and low-temperature. e un-predictable nature of the oxidation kinetics in sulfuric acid, however, as well as theextremely potent catalytic effect of some cations [], suggests that it would be un-wise to leave a uranium(IV) exposed to air for any length of time in an operatingindustrial process.Chapter Demonstration plante processes of electrolytic reduction and precipitation of uranous sulfate seemswell-suited for consideration as a new hydrometallurgical processing technology.Several authors have demonstrated that the electrolytic reduction of uranyl sulfatesolutions proceeds readily, and is a viable method to produce a uranium(IV) solu-tion [, , ] (not to mention its routine use in the production of uranium(IV)stock solutions for the present work). In Chapter , it was shown that uranous sul-fate has a low solubility at high temperature and high acid, and that it will precip-itate selectively in the presence of Al, Cu, Fe, and Ni. It was also shown that theprecipitate takes the form of uranous sulfate tetrahydrate under those conditions.In Chapter , it was shown that uranous sulfate tetrahydrate can be converted toU(SO), UOSO, or UO by thermal decomposition in air, depending on theoperating temperature.Although each of the above steps have been proven individually, a test of thecomplete process was needed to show its viability as a whole. To achieve this, theprocess was operated in sequence as a bench-scale demonstration of the technology(or, to take journalistic latitude, a . microtonne per hour ‘demonstration plant’,UO basis). Starting with a synthetic leach solution containing impurities, theprocess was taken through electrolytic reduction, acid addition, crystallization ofuranous sulfate, filtration, and drying. ermal decomposition was simulated withTGA analysis on the dried product. e process flow diagram is shown in Fig. ..electrolysis(a) syntheticleach solutionHSOcrystallization filtration(d) filtratedryingU(SO) ·HO(b) (c)Figure .: Process flow diagram of the demonstration plant. Experimental setup.. Solution preparatione synthetic leach solution was designed to partly mimic a highly concentratedleach solution that might derive from Saskatchewan’s high-grade uranium deposits.Four prominent impurities, Al, Cu, Ni, and Fe, were included, with their concen-trations based on an estimate of their prevalence in a high-grade leach solution de-riving from a known orebody, as provided by Cameco Corporation. Other less-prevalent impurities, such as As, were not included. e solution was created bydissolving . g UO, . g Al(SO) · HO, . g CuSO · HO, . gFeSO ·HO, . gNiSO ·HO, and. g concentrated sulfuric acid in deion-ized water and then diluting to mL in a volumetric flask. e resulting solutionwas filtered to remove the small amount of particulate matter that remained aerdissolution of the salts.e anolyte for electrolysis was prepared by diluting . g of concentratedsulfuric acid to mL with deionized water in a volumetric flask, matching thefree acid of the synthetic leach solution.A solution for washing the precipitate was prepared by diluting . g of con-centrated sulfuric acid to mLwith deionized water in a volumetric flask, match-ing the expected free acid concentration of the crystallization mother liquor. ewash solution was heated to  °C before use... Equipment and procedureElectrolytic reduction of the synthetic leach solution was conducted in situ in a mL Corning Pyrex No.  bottle using a stainless steel mesh cathode connectedto a DC power supply. For the anodic half-cell, a self-contained anolyte chamberwith a DSA anode was inserted into the bottle, with a Nafion N membrane toallow flow of H+ between catholyte and anolyte. e superficial single-side cathodeand anode surface areas were . cm. e power supply was set to a fixed currentof . A, giving a current density of Am−. e catholyte was stirred with aPTFE-coatedmagnetic stir bar, and the electrolyzer was immersed in a chilled waterbath set to  °C in order to keep the temperature low enough to avoid prematureprecipitation of uranous sulfate.Precipitation was conducted in the same bottle as electrolysis. Aer removingthe electrolysis apparatus, . g of concentrated sulfuric acid and . g of powderedU(SO) ·HO (-S, as seed) were added to the bottle, a lid with sampling portswas applied, and the bottle was purgedwith argon. e sealed bottle was then heatedto  °C in awater bath placed on top of amagnetic stir plate, and stirred aggressivelyfor  hours. Samples of the supernatant were withdrawn for analysis at , , , and hours. e sampling syringes were quenched in ice water to immediately halt theprecipitation process, and then samples were filtered through a syringe filter.Aer four hours, the hot slurrywas filtered over suction on anOsmonics . µmfilter. e supernatant was collected for analysis, then the precipitate was washedtwice with the previously-prepared sulfuric acid wash solution at  °C, and thentwice with isopropanol, until the filtrate ran clear. e solids were allowed to suc-tion-dry for several minutes, and then were transferred to a  °C oven in a petridish for drying overnight.e dried filter cake was gently broken up into powder with a mortar and pes-tle, and then transferred to a polypropylene tub for storage. e final sample wasanalyzed chemically for uranium and sulfate, thermogravimetrically for water, andstructurally by XRD.. Results and analysise results from each phase of the demonstration are described below. Chemicalassays of the solutions and solids are given in Table . and Table ., respectively... Electrolysise electrolytic reduction of uranium(VI) to uranium(IV) was effective and stra-ightforward, if not particularly efficient. Figure . shows the progress of reduction,as well as the cell voltage, over the duration of the test. e progress of reductionwas determined by redox titration for uranium(IV) with potassium dichromate. Anincrease in voltage was observed aer approximately min (†), occurring aroundthe same time as a marked increase in hydrogen evolution. e completion of re-duction was accompanied by a characteristic increase in cell voltage (*).e theoretical minimum electrolysis time was min at  current effi-ciency and a current of . A, taking into account the reduction of both uranium(VI)and copper(II). In reality, complete reduction was achieved aer  minutes (asindicated by the midpoint of the voltage increase *), giving an overall current effi-ciency of . Judging from the substantial gas generation on the cathode, it wouldseem that hydrogen evolution was almost entirely responsible for the low efficiency.e aqueous copper assays suggested that nearly all of the copper was removedby electrolysis. is was not surprising, given that the standard reduction poten-tial for copper (.V) is higher than uranium (.V), and thus should reduceTable .: Demonstration plant aqueous assays, mol L−UT HSOfree SO–T Al Cu Fe Ni(a) Syn. leach sol. . . . . . . .(c) Start of cryst. . . . . <.E- . .(d) Filtrate . . . . . . .Table .: Demonstration plant solids assaySolids assay, mass  Waters ofhydrationXRDIdentityU SO Al Cu Fe Ni. . <. <. <. <. . U(SO) ·HOFigure .: Electrolytic reduction of the synthetic leach solution, showing re-duction progress (orange) and cell voltage (turquoise). e two notablevoltage increase events corresponded with the onset of significant hydro-gen evolution (†) and the completion of reduction (*).first. At the end of the test, the cathode was found to be plated with copper, with aweight gain of . g, representing approximately  of the copper in the syn-thetic leach solution. Aluminum, iron, and nickel were not plated, reflecting theirlower reduction potentials, which are all below that of hydrogen evolution. Iron wasadded to the synthetic leach solution as Fe(II), and so did not consume any current,but in a reality the conversion of Fe(III) to Fe(II) would also be a source of currentinefficiency. ere was no sign of uranous sulfate precipitate on the cathode or inthe cell.As the hydrogen evolution rate increased, large hydrogen bubbles formed andattached to the cathode surface, masking a portion of the cathode that would haveotherwise been in contact with solution, and thus lowering the effective bulk currentdensity. Experience with other electrolysis systems suggests that this could havebeen solved by improving convective flow in the cell, thereby sweeping the emergenthydrogen bubbles away before they could agglomerate into large bubbles.Figure .: Concentrations of U, Al, Fe, and Ni over the course of uranous sul-fate precipitation. Cu assays were all below detection limit... Crystallizatione electrolysed solution was fed directly to the crystallization phase, where it wasdiluted somewhat by the addition of concentrated sulfuric acid. e resulting ura-nium concentration of .mol L− was about  lower than in the syntheticleach solution. About half of this dilution can be accounted for by sulfuric acidaddition, and a small amount by the generation of HO by the cathodic reaction,while the rest was likely due to the movement of water across the membrane duringelectrolysis (Nafion is permeable to water).e concentrations of uranium, aluminium, iron, and nickel over the course ofthe -hour test are shown in Fig. .. e uranium concentration dropped from.mol L− to .mol L− in about an hour, representing a recovery of ..No change was observed for the remainder of the  h test, suggesting that equilib-rium had been reached. None of the other metals changed in concentration overthe duration of the test. Copper is not shown because it was found to be below thedetection limit (<.×− mol L−) for all samples.e final equilibrium uranium concentration of .mol L− was substantiallyhigher than predicted by the solubility curves given in Chapter  at  °C and aFigure .: XRD pattern for demonstration plant solids, showing a match withU(SO) ·HO.free sulfuric acid of .mol L−. However, it was also shown in Chapter  that thepresence of impurities increases the solubility of uranous sulfate. Given the highlevel of impurities in the synthetic leach solution, the results were not surprising... Solids analysis. g of fine dry solid precipitate was recovered aer filtration, washing, and dry-ing. XRD analysis confirmed that the final product was pure U(SO) · HO, asshown in Fig. .. Solid assay results are shown in Table ., and confirmed that thematerial was free from impurities (subject to the detection limit of AA analysis).ermogravimetric analysis gave TGA and DSC signals characteristic of ura-nous sulfate tetrahydrate (see Chapter ), as shown in Fig. .. e amount of crys-talline water, at .HOper uranium, was slightly higher than the theoretical valueof , which mirrored earlier observations of increased water in solids crystallizedfrom impure solutions. e main water loss peak (W) in the derivative thermo-gravimetry (DTG) curve had an area of ∆m′ = ., corresponding to a loss of .HO, matching the expected hemihydrate intermediary.High temperature decomposition also behaved in a manner characteristic ofuranous sulfate tetrahydrate. ere were three peaks in the DTG signal collectivelycorresponding to the oxidation of U(SO) to UOSO (S, Sa, Sb), with the firststarting around  °C. Conversion to UO began around  °C (S), with thereaction rate increasing rapidly at higher temperatures.Figure .: ermal analysis of the demonstration plant solids, matching thesignature for U(SO) ·HO.. Implications for plant designe basic electrolyzer design used in the demonstration, with oxygen evolution onan anode separated from the catholyte by a proton exchangemembrane, functionedquite well, albeit at an uninspiring overall current efficiency. A continuous flow-through variation of this design is recommended. e anolyte composition shouldbe chosen tominimize water transport into the catholyte across themembrane. Hy-drogen evolution appeared to be the main competing reaction. An important focusof a future cell design, then, should be on the efficient dispersion and removal ofhydrogen bubbles. Optimization of hydrogen evolution could also be a focus of celldesign – enough to disrupt the diffusion boundary layer and enhance the mass dif-fusion rate of uranium(VI), but not so much as to compromise solution-electrodecontact. In the demonstration, copper plated onto the cathode, quickly covering thestainless steel mesh surface. Given that any metal with a higher reduction potentialthan uranium(VI) would quickly plate onto the cathode during continuous opera-tion, the pursuit of an ideal cathode material to maximize current efficiency wouldseem somewhat futile. erefore it would be prudent to design a cell that could ei-ther facilitate the quick swapping of fouled electrodes for cleaning, or allow for aperiodic cleaning cycle, either through the use of an oxidizing cleaning solution orby running the cell in reverse polarity.e crystallization of uranous sulfate proceeded quickly and effectively in thedemonstration, reaching equilibrium in about an hour in a batch reactor and pro-ducing pure U(SO) · HO. A series of three continuous stirred tank reactors(CSTRs) at  °C, with acid and seed addition, is recommended for a potential plantdesign. Unfortunately, the uranium recovery, at ., was too low to justify theuse of this technology alone. Methods to concentrate the solution, such as mem-brane water removal or an evaporator, would help with this, but could be costlyto operate. Even then, the remaining soluble uranium would need to be recoveredeither through a secondary recovery circuit, or via a recycle stream. All of theseefforts would carry implications related to the buildup of impurities.e thermal processing of U(SO) · HO should be straightforward, with itsdesign dictated largely by the desired final product. Drying in air at – °Cwould be sufficient to produce a dry, flowable U(SO) · HO product theoreti-cally containing . uranium, which could be packaged as a final product froma mill. To increase the weight percent uranium, the product could be fully dehy-drated by heating at – °C, giving U(SO) at . uranium. If the productwere calcined at  °C or above to produce UO, the released SO and SO gasescould theoretically be recovered in an acid plant as HSO, which could be used tooffset the acid needed for other parts of the process.Chapter Flow sheet developmentIn Chapter , it was demonstrated that the electrolytic reduction and precipitationof uranous sulfate is a viable process for selectively extracting uranium from a high-grade leach solution containing impurities at a bench scale. Here, a plant flow sheetis proposed, incorporating electrolysis, crystallization, evaporative concentration,washing and filtration, drying, and calcining. A schematic of the flow sheet is givenin Fig. ., and a brief description of each stream is given in Table ..Note that the flow sheet is presented here as an overview, not a complete massand energy balance, in order to respect the confidentiality of the project sponsor.A more complete treatment, including a mass balance and rudimentary economicanalysis, was published in a confidential report [].Off-gasSO2/SO3to acid plantBleedto secondary recoveryFinal ProductU3O8u601CrystallizationTank #1u602CrystallizationTank #2u603CrystallizationTank #3602603601 conc. sulfuric acid604605U604S/L Separationu605Filter Press607610u701Dryer704609rseed recycle501Pregnant leach solutionu501Continuous Electrolyzer502anolyte503606803r evaporator recycle801802steam611wash solution612702steam613evap.608spl606Evaporatorspl611u801Evaporatorinputsoutputs504anolyte out506anode gas505cathode gasu702Calciner701Filtraterecycled to leach circuit703Figure .: Proposed flow sheet for the electrolytic reduction and precipitation of uranous sulfate.Table .: List and description of flow sheet streamsStream Circuit Description Electrolysis Electrolyzer feed - high-grade pregnantleach solution Electrolysis Fresh anolyte to electrolyzer Electrolysis Reduced catholyte Electrolysis Anolyte discharge from electrolyzer Electrolysis Cathode gas (hydrogen) Electrolysis Anode gas (oxygen) Crystallization Sulfuric acid to crystalliation circuit Crystallization Crystallization tank  discharge Crystallization Crystallization tank  discharge Crystallization Crystallization tank  discharge Crystallization S/L sep. underflow Crystallization S/L sep. underflow to filter press Crystallization Filter cake to dryer Crystallization Filtrate recycled to leach circuitr Crystallization S/L sep. underflow seed recycle Crystallization S/L sep. overflow Crystallization Wash solution to filter press Crystallization S/L sep. overflow bleed to secondary recov-ery and impurity removal Crystallization Evaporation from crystallization tanks Drying and calcining Dried uranous sulfate tetrahydrate Drying and calcining Water vapour from dryer Drying and calcining Off-gas from calciner to acid plant Drying and calcining Final product - UO Residual U recovery S/L sep. overflow to evaporator Residual U recovery Water vapour from evaporatorr Residual U recovery Evaporator recycle to crystallization tank. Description of unit operations.. Continuous electrolysisA high-grade acidic uranium(VI)-sulfate solution (s) is continuously fed into amembrane-divided electrolyzer (u). e uranium is electrolytically reduced touranium(IV), yielding a reduced uranium(IV) stream (s), hydrogen gas fromthe cathode (s), and oxygen gas from the anode (s). e temperature is heldat approximately  °C in order to maintain maximum solubility of uranium(IV),thus preventing precipitation within the electrolyzer. e anolyte is a sulfuric acidsolution, and here is shown as a slow-flowing non-recycled stream (s), althoughit could potentially be reused and renewed via a slow bleed. e reduced uraniumsolution is fed onward to the crystallization circuit.It is worth noting that the cost of electricity is unlikely to be as significant for thistechnology as it is for other electrometallurgical systems, such as copper electrowin-ning. e very high molecular weight of uranium (. times greater than copper)means that each kilogram of uranium requires significantly less electricity to pro-cess than other metals. In addition, uranium sells for a much higher price than anyelectrolytically-produced base metal (e.g., Cu, Ni, Zn), even at the depressed pricesprevalent in –, making the cost of electricity easier to justify... Crystallizatione reduced uranium solution (s) is fed into a series of CSTR crystallizationtanks (u, u, and u). e tanks are covered, and preferably sealed, to min-imize reoxidation of uranium(IV) by air. e precipitation tanks are operated at °C, and extra sulfate is introduced in the form of sulfuric acid (s), in order tominimize the solubility of uranous sulfate. Any impurities are expected to stay in so-lution. Also introduced in the first tank is the concentrated uranium(IV)-bearingsolution recycled from the evaporator (sr). Crystal growth is encouraged byrapid agitation and the introduction of seed crystals from the S/L sep. underflow(sr). e output slurry from the last precipitation tank (s) contains precipi-tated uranous sulfate, as well as some residual aqueous uranium(IV).e slurry from the crystallization tanks is fed to a solid-liquid separation (S/Lsep.) step (u), with the clear overflow (s) feeding to the evaporator, and theunderflow (s) split between seed recycle (sr) and the filter press (s). Ableed of the S/L separation overflow (s) is necessary to prevent the uncontrolledbuildup of impurities in the recycle loop. e filter press (u) uses a hot sulfu-ric acid solution as wash solution (s), with the aim of displacing the entrainedimpurity-containing spent solution while minimizing the redissolution of uranoussulfate. e filtrate (s) is a dilute solution containing the impurities and some re-dissolved uranium, and can be recycled to the leaching circuit as a source of acid andto recover the uranium. e filter cake (s), consisting of solid U(SO) · HOand entrained sulfuric acid, is fed to the dryer... Drying and calciningewashed filter cake from the crystallization circuit (s) is fed to a dryer (u),where water is evaporated and released as steam (s). e dried product (s)has the composition of U(SO) · HO, and can be sold or stored indefinitely inthis form. If acid recovery is desired, the solids can be calcined at approximately °C (u), removing the sulfate as SO and SO gas. e off-gas (s) can besent to an acid plant for conversion to HSO. e final product in this case is UO(s). is is the final product of this flow sheet... Residual uranium recovery and impurity removale solution leaving the crystallization circuit still contains a substantial amountof soluble uranium(IV), which represents a major inefficiency of this design. erecovery can be improved by introducing a circuit to remove water from the spentsolution, thus increasing its uranium(IV) concentration and encouraging furtherprecipitation. In this flow sheet, this is modelled as an evaporator (u). Water isremoved in the formof steam (s), and the concentrated solution is recycled to thecrystallization tank (sr). e division of flow between the feed to the evaporator(s) and the bleed (s) would depend on the tolerance of the system to thebuildup of impurities. It might also be possible to use membrane filtration as analternative to evaporation.Ion exchange or solvent extraction could also be used for residual uranium re-covery. Both techniques are well-established and are widely used, making thema reliable choice. However, their performance on uranium(IV) would need to beevaluated, and it might be necessary to reoxidize the uranium to achieve adequateseparation from impurities. e use of solvent extraction or ion exchange wouldalso re-introduce the very technologies targeted for elimination by the present work,albeit at a substantially reduced flow rate.Impurities are not considered directly in this flow sheet. In the present config-uration, the ratio of evaporator recycle to bleed would determine the recirculatingload of impurities, which could rise to very high levels. It would almost certainlybe necessary to include some accommodation for impurity removal, either throughpre-treatment of the electrolyzer feed, or by a removal step aer the bleed stage.A bulk neutralization and precipitation process may be the most economical here,although the uranium precipitated thus would need to be recovered. If only a fewimpurities are found to interfere with the process, ion exchange columns could beused to selectively remove these impurities.Chapter Summary and conclusions. Review of objectivesIn Chapter , it was stated that the overarching objective of this dissertation was toadvance the knowledge and practice of the selective precipitation of uranous sulfateas a new uranium hydrometallurgical processing technology. Six questions wereposed, covering various aspects of uranous sulfate precipitation deemed critical tothe further development of the technology. ese questions are revisited below.. What are the best operating conditions for the precipitation of uranoussulfate?e solubility of uranous sulfate tetrahydrate is inversely proportional to tempera-ture and sulfuric acid concentration. e kinetics of precipitation are fastest at hightemperature, and nucleation and crystal growth can be enhanced by the inclusionof ‘seed’ crystals. Aggressive agitation promotes fast, uniform precipitation of fine-grained crystals, and prevents the tubing and reactor walls from being encrustedwith precipitate. It is therefore recommended that a future process operate at > °Cand >.mol L− sulfuric acid, with fast agitation, a substantial seed recycle, and aresidence time of several hours. is will give fine crystalline U(SO) ·HO pre-cipitate as a product.. How do impurities affect the precipitation process?e precipitates recovered from impure solutions containing Al, Cu, Fe, or Ni arevery pure, containing little or no trace of the impurities. Uranium recovery suffersin the presence of impurities, however, as these impurities all appear to increase thesolubility of uranous sulfate, even in solutions with otherwise-identical sulfate andfree acid concentrations. It seems, then, that uranous sulfate precipitation is highlyselective towards uranium in the presence of impurities, but that recovery can suffer.. What are the different uranous sulfate polymorphs, and how do they differfrom one another?Uranous sulfate can formmany different hydrated salts of the formU(SO) ·xHO,with x ranging between  and . ey differ not only in the number of water mole-cules, but also in the structural arrangement and cross-linking of the sulfates. etetrahydrate always forms at temperatures > °C, while the octahydrate and hexa-hydrate form at lower temperatures. Lower hydrates and the anhydrate can be pro-duced by controlled thermal decomposition, which drives off water, although it isstill unknown whether these compounds are stable under ambient conditions.A crystallographically-unique compound, called parisaite in this work, was fou-nd to form as an intermediary during the crystallization of uranous sulfate tetrahyd-rate. Although its exact form was not identified, it was found to contain approxi-mately four sulfates per uranium, perhaps suggesting that sulfuric acid is incorpo-rated into its structure. Parisaite can be eliminated by controlling various parame-ters, including sulfuric acid concentration and test duration.. How does uranous sulfate respond to drying and calcining?Uranous sulfate x-hydrate decomposes in air or nitrogen in three general steps: wa-ter loss toU(SO); oxidation toUOSO; and decomposition toUO. Initial rapidwater loss occurs between – °C, depending on the polymorph, followed by theslow loss of the remaining water. e resulting anhydrous uranous sulfate behavesdifferently depending on the level of hydration of the original solid, appearing toreflect fundamental differences in their molecular structures. is may be relatedto the relative difficulty involved in the reconfiguration of bonds during the recrys-tallization of amorphous anhydrous uranous suflate into crystalline anhydrous ura-nous sulfate. In a reducing atmosphere ofH orNH, uranous sulfate can be directlyreduced toUO. With proper control of temperature and atmosphere, it is thereforepossible to convert any uranous sulfate polymorph into U(SO), UOSO, UO,or UO using a calcining process.. Is aqueous uranium(IV) stable against oxidation by oxygen gas?Uranium(IV) will oxidize readily in the presence of oxygen gas. In perchloric acid,the rate of oxidation is .-order in uranium(IV), first-order in oxygen, and inverse-second-order in acid, and follows an Arrhenius relationship with respect to temper-ature. e presence of sulfate slows oxidation, likely due to the formation of stablesulfate complexes. e presence of sulfate in high amounts, such as when operatingin sulfuric acid, gives unreproducible results, making it difficult to predict or modelits behaviour.. Can uranous sulfate precipitation be developed into a viable extractivemetallurgical technology?ere is ample evidence to suggest that uranous sulfate precipitation can be usedto selectively extract uranium from an impure leach solution. A possible processmight involve electrolytic reduction, precipitation, filtration, drying, and calcining,giving a final product of either U(SO) or UO. e main obstacle to the success-ful implementation of this process is the relatively high solubility of uranous sulfate,which seems to be enhanced by the presence of impurities. An improvement in re-coverymight be possible through the introduction of recycle streams or a secondaryrecovery circuit, but this would require further study.. Contributions to the artMany of the results presented in this dissertation represent new contributions to thebody of scientific knowledge. ese contributions span several areas of inorganicchemistry and engineering.Contributions to the field of inorganic chemistry were mainly through the de-termination of the crystal structures of {[U(SO)(HO)] ·HO}n (uranous sulfatehexahydrate) and [U(SO)(HO)] · HO (uranous sulfate octahydrate). esehave revealed striking differences in the molecular connectivity of the different hy-drates, which furthers the understanding of how so-called ‘waters of hydration’ canimpact the molecular structure of uranous sulfate. One tangible benefit of the studyis that the powder XRD, Raman, and infrared spectra of the uranous sulfate hydratesare now known, making it much easier to identify their presence in unknown pre-cipitates.New contributions to the fundamentals of hydrometallurgy were centred on thegenesis of uranous sulfate x-hydrate from acidic uranium(IV) solutions. e effectsof the impurities Al, Cu, Ni, and Fe on the precipitation process were demonstrated,showing that uranous sulfate can be selectively crystallized from impure solutions.A useful phase map was developed from experimental data showing how acid con-centration, temperature, and crystallization time dictate which polymorphic formof uranous sulfate is stable, giving a deeper understanding of the polymorphism ofuranous sulfate than was known previously. e studies on the oxidation on ura-nium(IV) with molecular oxygen contributed to the understanding of that systemin several ways, namely by identifying the apparent reaction order under variouscircumstances. Perhaps more importantly, a method of continuous data acquisi-tion and differential analysis was applied which could be applied to other systemsto advance the understanding of non-first-order kinetics.Fundamental contributions were also made to other areas of extractive metal-lurgy. e study of the thermal decomposition of the uranous sulfate hydrates hasgranted a solid understanding of how uranous sulfate tetra-, hexa-, and octahydraterespond to calcining, including the temperatures and decomposition pathways in-volved. is level of detail was previously available only for the tetrahydrate. estudies also revealed significant differences in the decomposition pathways of thethree hydrates, the first time such a phenomenon has been identified for uranoussulfate, which could be important when selecting the operating conditions of a cal-cining process.Finally, this dissertation has advanced the understanding of uranous sulfatefrom an engineering perspective. It was demonstrated that the kinetics of the crys-tallization process can be enhanced significantly using simple engineering controls,such as agitation, seeding, and temperature, whereas previously it was unclearwhetherthe process could bemade to proceed quickly enough to be useful. It was also shownthat uranous sulfate can be precipitated selectively from impure solutions, ratherthan just from pure solutions. A complete potential metallurgical process, includ-ing electrolysis, precipitation, filtration, drying, and calcining, was tested beginningto end, and then formulated into a complete conceptual flow sheet. Together, theseefforts represent the first time that uranous sulfate precipitation has been demon-strated as a viable processing option under realistic conditions.. Further worke pursuit of a PhD degree usually raises more questions than it answers, and thisproject was no exception. Some of the further work suggested below could be doneusing equipment presently available in the UBC materials engineering hydrometal-lurgy laboratory, but most would require the procurement of additional equipmentor partnership with another lab.e solubility of uranous sulfate is fairly well understood between – °C,which represents reasonable limits for a simple open-air precipitation tank. If apressurized vessel were used, however, higher temperatures would be possible, andthe associated costs might be justified by the enhanced uranium recovery. A futurefundamental study could focus on the solubility of uranous sulfate at temperatures> °C in an autoclave. Other fundamental work could focus on solutions con-taining impurities, eventually working towards using real process solutions ratherthan the synthetic solutions studied thus far. Of particular importance is the effectimpurities have on uranous sulfate solubility, which is still poorly understood, andextending such study to anions such as Cl–, which could cause substantial interfer-ence to the process due to their ability to complex with uranium. Attempting tocalculate the real ionic strength (in contrast to the formal ionic strength) might behelpful in furthering this understanding. It would also be worthwhile to explorethe relationship between uranium(IV) and HSO– by adjusting [H+] with a non-complexing acid such as perchloric acid, without changing the total sulfate. iscould grant further insight into the intermediate compound parisaite, which, withapproximately four sulfates per uranium,might haveHSO– orHSO incorporatedinto its structure, and thus should be affected by free acid. If parisaite indeed formsas an intermediate compound, the amount of sulfate in the aqueous phase shouldfirst drop drastically as parisaite is formed, and then recover somewhat as it convertsto U(SO) ·xHO. An interesting set of experiments would follow both sulfate anduranium concentrations in solution, paired with solids analysis, to validate (or re-fute) this hypothesis.e fundamental crystallographic characterization work described in Chapter could be extended to the other compounds identified over the course of work, suchas the intermediary hydrates from thermal decomposition. e structure of the so-called parisaite in particular could be very revealing, given that it appears to formas a precursor during the crystallization process. If it proves infeasible to producesingle crystals of these compounds of a sufficient size and purity for single-crystalx-ray diffraction, it might be worthwhile to investigate electron crystallography us-ing a transmission electron microscope (TEM) as an alternative. Also, recent ad-vancements in cryo-electron microscopy (cryo-EM) could be applied to the studyof uncrystallized inorganic compounds, although as of the writing of this disserta-tion (early ) this technique is in its infancy, with near-atomic-level resolutiononly just becoming possible.e thermal decomposition of uranous sulfate x-hydrate described in Chap-ter  could be investigated more rigorously by controlling the purge gas atmosphereduring TGA analysis, either by saturating it with water vapour (for the water lossreactions), or with SO and SO (for the decomposition reactions). is would re-move an unknown from the system (the activity of the gases in thermodynamicequilibriumwith the solids), making validation of the thermodynamics less specula-tive, andmight also stabilize the intermediate compounds over a wider temperaturerange. e use of additional on-line analytical techniques to complement the TGAand DSC data would also be useful. Continuous XRD analysis of a powder sampleon a temperature-controlled stage could be used to identify the crystallographicalform of the various intermediate compounds, and could also be used to study theapparent slow recrystallization of amorphous uranous sulfate. Real-time analysis ofthe off-gases, using a gas chromatograph or similar, could by used to study the reac-tion stoichiometry, and might also prove useful for estimating the reaction rates oftwo simultaneous reactions, a task performed in this work using DTG peak decon-volution. All of these experiments would require equipment currently not availablein the Department.e oxidation of uranium(IV) by molecular oxygen in the presence of largeamounts of sulfate requiresmuchdeeper study in order to explain the irreproducibil-ity of some of the data presented in Chapter . Any future work should be con-ducted with the upmost care, such as by using the purest reagents, controlling lightexposure, using an all-glass system (including the impeller), and scrupulously clean-ing the entire system between tests to ensure uniform surface conditions. A studyspecifically designed to find the onset of irreproducible behaviour would be help-ful – differential reaction order analysis, showing the transition from .-order toa lower order as sulfate is introduced, might be informative. Further developmenton the theoretical side, namely a refinement of the reaction mechanism in the non-perchloric acid system that accounts for the data presented in this dissertation andelsewhere, could also shed light on how sulfate might interfere with the process.From an engineering perspective, a great deal of information could be obtainedby operating the flow sheet proposed in Chapter  on a mini-pilot scale, includ-ing electrolysis, precipitation, filtration, drying, and calcining.  e electrolysis stepmust be sped up and optimized, with particular attention to the effect of impurities.e long-term performance of the electrolyzer is currently unknown, particularlywith respect to the fouling of the proton-exchange membrane and cathode, and itwill almost certainly be necessary to develop techniques to clean the cell using chem-ical or eletrochemicalmeans, possibly on a regular, automated schedule. eprecip-itation of uranous sulfate following electrolysismust be studied from an engineeringperspective as well, such as by optimizing the pulp density (through the establish-ment of a recirculating load) to encourage crystallization. It might also be possibleto improve recovery by boiling the solution during the precipitation step, ratherthan operating at  °C. Solid-liquid separation and filtration using industrial-styleequipment would give a more reasonable idea of losses due to redissolution and theentrainment of mother liquor, rather than the idealized results given by a laboratoryfiltration setup. Of particularly importance would be the development of a suitablewash liquid – not water – for the filtration process to minimize redissolution. Dry-ing and calcining could be run as a separate campaign, focusing on the recovery ofacid from the off-gases and the preparation of the final product in a suitable formfor transport.. Concluding remarkse selective precipitation of uranous sulfate shows great promise as a new primarypurification technique for uranium-containing solutions. is is exciting—new ex-tractive metallurgy technologies surface rarely, and the chance to overturn the sta-tus quo should be embraced. Many obstacles remain, however, and it will require asignificant investment of resources to develop the concepts presented in this disser-tation into a competitive technology. ere is no question that the world’s demandfor energy will increase over the coming decades, and with it the world’s demandfor uranium. e further development of uranous sulfate precipitation technologywould therefore be a wise investment.References[] A. D. Aczel. UraniumWars: e Scientific Rivalry that Created the NuclearAge. Palgrave Macmillan, . → pages [] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giacovazzo,A. Guagliardi, A. G. Moliterni, G. Polidori, and R. Spagna. SIR: a new toolfor crystal structure determination and refinement. Journal of AppliedCrystallography, ():–, . → pages [] ASTM International. Standard test method for uranium by iron (II)reduction in phosphoric acid followed by chromium (VI) titration in thepresence of vanadium. ASTM Standard, . → pages iv, [] ASTM International. ASTM C-, standard specification fornuclear-grade uranyl nitrate solution or crystals. ASTM Standard, . →pages [] Y. Awakura, K. Sato, H. Majima, and S. Hirono. e measurement of thediffusion coefficient of u (vi) in aqueous uranyl sulfate solutions.Metallurgical Transactions B, ():–, . → pages [] Y. Awakura, H. Hiai, H. Majima, and S. Hirono. Fundamental studies on thecontinuous electrolytic reduction of uranyl sulfate. Metallurgical andMaterials Transactions B, :–, . → pages , , [] U. Betke and M. S. Wickleder. Oleum and sulfuric acid as reaction media:e actinide examples UO(SO)-lt (low temperature), UO(SO)-ht (hightemperature), UO(HSO), An(SO) (An= , U), (HSO)(SO) and(HSO)(SO). European Journal of Inorganic Chemistry, ():–, . → pages , , , , , [] R. Bothwell. Eldorado: Canada’s National Uranium Company. University ofToronto Press, . → pages [] A. Brodtkorb. NUMDIFFTOOLS v.. soware package, October , .URL https://code.google.com/p/numdifftools/. → pages [] Bruker AXS Inc. SAINT v.B soware package. Madison, Wisconsin,USA, -. → pages [] Bruker AXS Inc. SADABS soware package. Madison, Wisconsin, USA,. → pages [] A. D. Burns. Electrolytic reduction and precipitation of uranous sulfate: Flowsheet development, mass balance and operating cost analysis. Private report,e University of British Columbia, September . → pages [] A. D. Burns and D. Dreisinger. e apparent overall oxidation kinetics ofuranium(IV) by oxygen gas in aqueous perchloric acid solutions. InE. Asselin, D. Dixon, F. Doyle, D. Dreisinger, M. Jeffery, and M. Moates,editors, Proceedings of the th International Symposium On Hydrometallurgy, volume , pages –. e Metallurgy and Materials Society ofCIM, Canadian Institute of Mining, Metallurgy and Petroleum, June .→ pages iv, [] A. D. Burns, B. O. Patrick, A. E. Lam, and D. Dreisinger. e effect ofcoordinated water on the connectivity of uranium(IV) sulfate x-hydrate:[U(SO)(HO)] ·HO and [U(SO)(HO)] ·HO], and a comparisonwith other known structures. Acta Crystallographica, C(), . → pagesiii, , , , [] S. Casadio and L. Lorenzini. Cyclic voltammetric behavior of uranyl ion insulfuric acid solutions. application to some nuclear materialscharacterization. Analytical Letters, ():–, . → pages , [] E. Cordfunke. e system uranyl sulphate-water—I: Preparation andcharacterization of the phases in the system. Journal of Inorganic and NuclearChemistry, (): – , . → pages , [] E. Cordfunke. e system uranyl sulphate-water—II: Phase relationships andthermochemical properties of the phases in the system UO-SO-HO.Journal of Inorganic and Nuclear Chemistry, (): – , . →pages , , [] CrystalMaker Soware Ltd, Oxford, England (www.crystalmaker.com).Crystalmaker .., . → pages [] W. Davies and W. Gray. A rapid and specific titrimetric method for theprecise determination of uranium using iron (II) sulphate as reductant.Talanta, ():–, . → pages [] O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. Howard, andH. Puschmann. OLEX: a complete structure solution, refinement andanalysis program. Journal of Applied Crystallography, ():–, .→ pages [] A. E. Duque, M. G. Benach, and F. H. Arroyo. Obtencion de UF pourreduccion electrolica. Technical Report -, Spanish Atomic EnergyCommission, . → pages , [] A. J. Elliot, S. Padamshi, and J. Pika. Free-radical redox reactions of uraniumions in sulphuric acid solutions. Canadian journal of chemistry, ():–, . → pages [] S. Fallab. Reactions with molecular oxygen. Angewandte ChemieInternational Edition in English, ():–, . → pages [] L. J. Farrugia. WinGX suite for small-molecule single-crystal crystallography.Journal of Applied Crystallography, ():–, . → pages [] S. Fell. Electrochemical generation of stable uranium(IV) fluoride-sulphatesolutions. Analyst, ():–, . → pages [] F. Forward and J. Halpern. Developments in the carbonate processing ofuranium ores. Journal of Metals, ():–, December . → pages[] F. Forward and J. Halpern. Acid pressure leaching of uranium ores. Journalof Metals, ():–, March . → pages [] W. Freyberger. Electrolytic reduction of uranyl and ferric sulfate solutions.Technical Report AECD- (MITG-A), Massachusetts Institute of Tech,Watertown, Massachusetts, May . → pages , [] J. M. Gil, F. M. Villa, and F. J. M. Gil. ermal decomposition of sulphatecomplexes of uranium(IV) hydrates. Journal of ermal Analysis andCalorimetry, :–, . → pages [] F. Giolitti and G. Bucci. Sui fenomeni di equilibrio fra gli idrati del solfatouranoso. Gazzetta Chimica Italiana, ():, . → pages , [] J. H. Gittus. Metallurgy of the rarer metals. In H. M. Finniston, editor,Uranium, volume . Butterworths London, . → pages [] E. Gmelin and S. M. Sarge. Calibration of differential scanning calorimeters.Pure and applied chemistry, ():–, . → pages [] G. Gordon and H. Taube. Oxygen tracer experiments on the oxidation ofaqueous uranium(IV) with oxygen-containing oxidizing agents. InorganicChemistry, ():–, // . → pages , [] I. Grenthe, H. Wanner, and I. Forest. Chemical thermodynamics of uranium.Technical report, OECD Nuclear Energy Agency, Issy-les-Moulineaux,France, . → pages [] R. Guillaumont and F. J. Mompean. Update on the chemicalthermodynamics of uranium, neptunium, plutonium, americium andtechnetium. Technical report, OECD Nuclear Energy Agency,Issy-les-Moulineaux, France, . → pages , , , , , , [] Y. Gurinov and Y. Frolov. Intensification of the electrochemical process ofreduction of six-valent to four-valent uranium. Zhurnal Prikladnoi Khimii,():–, . → pages , [] F. Habashi. A short history of uranium. In Proceedings of the rdInternational Conference on Uranium, volume , pages –. eMetallurgical Society of the CIM, CIM, . → pages [] J. Halpern and J. G. Smith. Kinetics of the oxidation of uranium(IV) bymolecular oxygen in aqueous perchloric acid solution. Canadian Journal ofChemistry, :–, . → pages , , , , [] I. Higgins, J. Roberts, C. Hancher, and J. Marinsky. e Excer process.preparing uranium tetrafluoride by ion exchange and electrolysis. Industrial& Engineering Chemistry, ():–, . → pages , [] International Atomic Energy Agency. Nuclear power and sustainabledevelopment. Brochure, . → pages [] K. Južnič and Š. Fedina. On the kinetics of oxidation of uranium(IV) insulphuric acid by molecular oxygen. Journal of Inorganic and NuclearChemistry, ():–,  . → pages [] D. Kern and E. Orlemann. e potential of the uranium (V), uranium (VI)couple and the kinetics of uranium (V) disproportionation in perchloratemedia. Journal of the American Chemical Society, ():–, . →pages [] L. Khamidullina and S. Lotnik. Formation of excited uranyl in oxidation oftetravalent uranium with oxygen in HClO aqueous solutions: IV.enhancement of the chemiluminescent reaction in the presence of Fe+. HighEnergy Chemistry, ():–, . → pages [] L. Khamidullina and S. Lotnik. Involvement of ·OH radicals in themechanism of excitation of the uranyl ion in the chemiluminescent oxidationof U (IV) by atmospheric oxygen in aqueous solutions of perchloric acid.Kinetics and Catalysis, ():–, . → pages [] P. Kierkegaard. e crystal structure of U(SO) ·HO. Acta ChemicaScandinavica, ():–, . → pages , , , [] J. M. Leroy and G. Tridot. Sur les hydrates et la décomposition thermique dusulfate uraneux. Comptes rendus hebdomadaires des séances de l’Académie dessciences, :–, . → pages [] J. M. Leroy, J. Tudo, and G. Tridot. Sur les hydrates du sulfate d’uranyle.Comptes rendus hebdomadaires des séances de l’Académie des sciences, :–, . → pages , , [] D. Lide. e CRC Handbook of Chemistry and Physics, th Edition. CRC,Boca Raton, FL, . → pages [] E. Lohouse. Electrolytic production of uranium tetrafluoride. United StatesPatent ,,, September  . → pages , [] S. Lotnik, L. Khamidullina, and V. Kazakov. Uranyl-catalyzedchemiluminescent reaction of U+ oxidation by dioxygen in aqueous HClOsolution. Russian Chemical Bulletin, ():–, . → pages ,[] S. V. Lotnik, L. A. Khamidullina, and V. P. Kazakov. Strong effect of a solidsurface on a liquid-phase radical chain chemiluminescence reaction.oxidation of U(IV) by dioxygen. Russian Chemical Bulletin, ():–, . → pages [] D. K. Louie. Handbook of sulphuric acid manufacturing. DKL Engineering,Inc., . → pages [] C. F. Macrae, P. R. Edgington, P. McCabe, E. Pidcock, G. P. Shields, R. Taylor,M. Towler, and J. van de Streek. Mercury: visualization and analysis of crystalstructures. Journal of Applied Crystallography, ():–, June . →pages , , , [] H. Majima, Y. Awakura, and S. Hirono. Electrolytic reduction of U(VI) toU(IV) in acidic chloride and acidic sulfate solutions. MetallurgicalTransactions B, ():–, . → pages , , , [] H. Majima, Y. Awakura, K. Sato, and S. Hirono. Laboratory reduction rateand current efficiency studies of batch type electrolytic reduction of U(VI) ina sulfate system. Metallurgical and Materials Transactions B, :–, ../BF. → pages [] H. McCoy and H. Bunzel. e speed of oxidation, by air, of uranoussolutions, with a note on the volumetric determination of uranium. Journalof the American Chemical Society, ():–, . → pages , [] Metrohm. Titrimetric determination of sulfate. Application Bulletin. →pages iv, , [] K. Notz and H. Jaffe. Pyrolysis of uranium compounds—I: Uranyl sulphate.Journal of Inorganic and Nuclear Chemistry, ():–, . → pages, , [] K. J. Notz and H. H. Jaffé. Correlation of TGA and DTA temperatures indecomposition reactions. Journal of the American Ceramic Society, ():–, . → pages [] R. A. Nyquist and R. O. Kagel. Handbook of infrared and raman spectra ofinorganic compounds and organic salts: infrared spectra of inorganiccompounds. Academic press, . → pages [] Outotec Oyj. HSC chemistry. Espoo, Finland, -. → pages , ,, [] E. Peters. e leaching of uranium from pitchblende ores by aqueousoxidation techniques. Master’s thesis, University of British Columbia, .→ pages [] J. Plášil, K. Fejfarová, M. Novák, M. Dušek, R. Škoda, J. Hloušek, J. Čejka,J. Majzlan, J. Sejkora, V. Machovič, and D. Talla. Běhounekite,U(SO)(HO), from Jáchymov (St Joachimsthal), Czech Republic: the firstnatural U+ sulphate. Mineralogical Magazine, ():–, . →pages , , , , [] D. Preston. Before the Fallout: From Marie Curie to Hiroshima. WalkerPublishing Company, Inc., . → pages [] J. Renaud. Activités nucléaires de la société ugine kuhlmann. ÉnergieNucléaire, ():–, August . → pages , [] T. Sato, F. Ozawa, and S. Ikoma. ermal decomposition of uranyl sulfatehydrate. Journal of chemical technology and biotechnology, ():–,. → pages , , [] D. D. Schnaars and R. E. Wilson. Uranium(IV) sulfates: Investigatingstructural periodicity in the tetravalent actinides. Inorganic Chemistry, ():–, . → pages , , , , , [] G. M. Sheldrick. A short history of SHELX. Acta Crystallographica A, ():–, . → pages [] V. Shilov, A. Yusov, V. Peretrukhin, C. Delegard, A. Gogolev, A. Fedosseev,and L. Kazansky. Oxidation of U(IV) by atmospheric oxygen in pH .–.aqueous solutions. Journal of Alloys and Compounds, -:–,. → pages , [] J. Sobkowski. e oxidation reduction potential of the U (IV)-U (VI)system—III: Photoelectrochemical properties of the U (V) ion. Journal ofInorganic and Nuclear Chemistry, ():–, . → pages [] J. Sobkowski. Rate of oxidation of tetravalent uranium ions with molecularoxygen. Roczniki Chemii, :–, . → pages , , , , [] A. Stabrovskii. Study of solubility of uranium(IV) and uranium(III) sulfatesin sulfuric acid. Radiokhimiia, (), . → pages [] B. N. Sudarikov, O. I. Zakharov-Nartsissov, and A. V. Ochkin. Oxidation ofquadrivalent uranium in sulfuric acid solutions with atmospheric oxygen.Trudy Instituta - Moskovskii Khimiko-Tekhnologicheskii Institut imeni D. I.Mendeleeva, :–, . → pages [] S. Suzuki, S. Hirono, Y. Awakura, and H. Majima. Solubility of uranoussulfate in aqueous sulfuric acid solution. Metallurgical Transactions B, ():–, . → pages , , , [] S. ein and P. Bereolos. ermal stabilization of 233UO, 233UO and233UO. Technical Report ORNL/TM-/, Oak Ridge NationalLaboratory, . → pages [] G. Tridot. Applications des methodes des d’analyses thermiques a l’evolutiondes sulfates de vanadyle, d’uranyle, d’uranium(IV), de titanyle et de leursderives dans diverses atmospheres controlees. Pure and Applied Chemistry,():–, . → pages , [] I. Warren and F. Forward. Hydrometallurgical production of uraniumdioxide for reactor fuel elements. e Canadian Mining and MetallurgicalBulletin, ():–, October . → pages [] B. Wassink. Analysis of Acidic Metal Sulfate Solutions for Sulfuric Acid UsingpH Electrode and Standard Addition. University of British Columbia,Department of Materials Engineering, February . → pages iv, [] M. Wojdyr. Fityk: a general-purpose peak fitting program. Journal of AppliedCrystallography, ():–, . → pages , [] A. Yilmaz, L. Hindiyarti, A. D. Jensen, P. Glarborg, and P. Marshall. ermaldissociation of SO at - K. e Journal of Physical Chemistry A, ():–, . → pages Appendix AProduction of uranium(IV) solutionsby electrolytic reductionSummaryLaboratory-grade uranium(IV) salts are not commercially available for purchase,so if a uranium(IV) solution is needed for research purposes, it must be producedby reducing uranium(VI). One method is to pass a uranium(VI) solution througha Jones reductor (zinc metal column), but this is cumbersome and could introduceimpurities into the solution. Electrolytic reduction is a proven alternative that doesnot require the uranium solution to come into contact with any reagents, and is alsoeasier to monitor and control.Below is a brief review of the theory behind the electrolytic reduction of ura-nium(VI), an experimental procedure that can be used in the laboratory, and ex-ample results from a typical electrolysis experiment conducted as a part of the workpresented in this dissertation.Background informatione electrolytic reduction of uranium(VI) can be described by the following half-cell reactions:Cathodic half-cell: UO+ +H+ +e– −−→ U+ +HO E◦ = 0.27V (A.)Anodic half-cell: HO−−→ O +H+ +e– E◦ =−1.23V (A.)Overall: UO+ +H+ −−→ U+ + O ∆E =−0.96V (A.)Since ∆E is negative, the reaction does not proceed spontaneously, and so re-quires a voltage to be applied using an external power supply.If the uranium(VI) cannot diffuse to the cathode surface fast enough to con-sume all of the supplied current (assuming a constant-current cell is being used),hydrogen evolution will occur to make up the difference as a competing reaction,lowering the current efficiency. Hydrogen evolves according to the following half-cell reactions:Cathodic half-cell: H+ +e– −−→H E◦ = 0V (A.)Anodic half-cell: HO−−→ O +H+ +e– E◦ =−1.23VOverall: HO−−→ O +H ∆E =−1.23V (A.)e progress of the electrolysis reaction can be described by Faraday’s law ofelectrolysis for a constant-current cell:n = ItzF(A.)where n is the number of moles of uranium reduced, I is the current (A, or C s−),t is the elapsed time (s), z is the number of electrons transferred per mole of ura-nium reduced ( in this case), and F is Faraday’s constant ( Cmol− of e–).Converting Eq. (A.) into concentration units and substituting in the values foruranium(VI) reduction gives the following equation for calculating the minimumpossible time to complete reduction, tmin, assuming  current efficiency:tmin = ( Cmol−)[U(VI)]iVI(A.)where V is the catholyte volume and [U(VI)]i is the initial concentration of ura-nium(VI) in the catholyte.e current efficiency can be approximated by Eq. (A.):η = tmint(A.)where t is the actual electrolysis time.Experimental procedureReagents• Granulated uranium trioxide (UO)• Concentrated sulfuric or perchloric acid• Deionized water• Nitrogen gas (for purging)Equipment• Electrolyzer– Cathode (e.g., stainless steel mesh)– Anode for oxygen evolution (e.g., DeNora DSA)– Nafion N proton exchange membrane– Heat-removal system (e.g., water jacket)• Power supply capable of A current at V• Logging multimeter• Circulating chiller, or other cooling method• Magnetic stir plate or circulating pump, to mix catholyteMethodCatholyte preparationDissolve an appropriate amountUO in sulfuric or perchloric acid to achieve the de-sired catholyte composition, accounting for acid consumption shown in Eq. (A.).Adjust the acidity to the desired concentration, and dilute to a known volume in avolumetric flask.UO3 +2H+ −−→ UO22+ +H2O (A.)Anolyte preparationDilute concentrated sulfuric or perchloric acid to the desired concentration in avolumetric flask.Electrolysis. Turn on the heat-removal system to the electrolyzer (i.e., connect the circu-lating chiller, or immerse in water bath).. Add catholyte and anolyte (as prepared above) to the appropriate chambersof the electrolyzer.. Bubble nitrogen through the catholyte at a slow rate to flush any evolved hy-drogen from the headspace of the electrolyzer. is is to prevent the potenti-ally-dangerous buildup of hydrogen gas.. Start catholyte circulation. If using magnetic stirrer, add stir-bar to the cath-olyte chamber and turn on stir plate. If using a circulating pump, turn onpump.. Connect the logging multimeter to the electrodes and set to DC Volts mode.. Set the power supply to constant current mode and adjust the current untilthe cell voltage is .–.V. e current that can be achieved will depend onthe electrolyzer setup, such as electrode surface area, electrode materials, etc.. Allow electrolysis to proceed while monitoring the cell voltage on the multi-meter. Take samples periodically and assay for uranium(IV), if desired.. When the elapsed time approaches the minimum possible electrolysis timegiven by Eq. (A.), beginmonitoring the voltagemore closely. e cell voltagewill rise by ∼.V in a distinct S-shaped curve when all of the uranium(IV)has been converted to uranium(VI). Once the cell voltage stabilizes at thehigher value, the test is complete. An electrolysis time approximately longer than tmin is typical.. Turn off the power supply. Transfer the catholyte to an air-free container forstorage. e anolyte can be saved for re-use, if desired.. Clean the electrolytic cell thoroughly, and store the membrane wet in tap wa-ter.Example setup and resultsElectrolytic reduction of uranium(VI) was performed many times during the ex-perimental work described in this dissertation, in both sulfuric acid and perchloricacid. Electrolysis was generally performed using the electrolyzer shown in Fig. A..e cathode chamber was a simple glass jacketed reactor, with  °C water circulat-ing through the jacket to maintain uranous sulfate at its maximum solubility. ecathode was a band of stainless steel mesh running around the inner wall of the ves-sel. e anode chamber was a self-contained submersible compartment wrappedin a Nafion N membrane. e anode compartment had a capacity of mL ofanolyte and had a built-in DeNora DSA as the anode. e catholyte was circulatedusing a magnetic stirrer.Stir barNafion membraneCathode chamber withstainless steel mesh cathodeJacketed reactorAnode chamber with DSA anodeV +-Figure A.: Electrolyzer with submersible anode chamber.As an example, the uranium(IV) solution prepared for the production of sampleFigure A.: Cell potential vs. time for a typical electrolysis experiment. In-crease in potential at * indicates the completion of reduction. Constantcurrent = .A.-S described in Chapter  represents a typical example of how electrolytic reduc-tion was applied for the work described in this dissertation. e catholyte wasmadeby dissolving . g of pure UO in . g concentrated sulfuric acid and dilutingto mL, which yielded a  g L− uranium solution with  g L− total sulfateand  g L− free HSO. e anolyte composition was  g L− sulfuric acid.e power supply was set to operate in a constant current regime, with a fixedoperating current of . A. e voltage between the anode and cathode (i.e., the cellpotential) was measured continuously during electrolysis.Within a few minutes of commencing electrolysis, the catholyte turned fromyellow in colour to purple-green colour, and continued to darken as the reductionprogressed. Oxygen gas evolved at the anode for the duration of the test. Hydrogenbegan evolving at the cathode part way through the test. e approach of the endof the test was indicated by a prominant S-shaped rise in cell voltage, as shown inFig. A.. Complete reduction was indicated by the stabilization of the voltage atmin. Titration of the catholyte for uranium(IV) and total uranium confirmedcomplete reduction had taken place.Using Eq. (A.) to calculate the minimum electrolysis time at  currentefficiency, assuming a catholyte volume of mL, gives a value of min. Equa-tion (A.) yields an approximate current efficiency of .Appendix BRaman, FTIR, and XRD patterns forthe uranous sulfate x-hydratesSupporting information forChapter , including uranous sulfate tetrahydrateU(SO) ·HO, uranous sulfate hexahydrate, [U(SO)(HO)] ·HO (complex ), uranoussulfate octahydrate, [U(SO)(HO)] · HO (complex ), and parisaite. Refer toChapter  for details on instrumentation used.B. Uranous sulfate tetrahydrateFigure B.: Raman spectrum for uranous sulfate tetrahydrate,  nm laser.Sample: -S.Figure B.: FTIR spectrum of uranous sulfate tetrahydrate. Sample: -A.Figure B.: Powder XRD spectrum of uranous sulfate tetrahydrate. Simulatedfrom the crystal structure published by Plášil et al. [] usingMercury .[].B. Uranous sulfate hexahydrateFigure B.: Raman spectrum for uranous sulfate hexahydrate, complex , nm laser. Sample: -A.Figure B.: FTIR spectrum of uranous sulfate hexahydrate, complex . Sam-ple: -A.Figure B.: Powder XRD spectrum of uranous sulfate hexahydrate, complex ,simulated from the crystal structure presented in Chapter  using Mer-cury . [] (Cu source).B. Uranous sulfate octahydrateFigure B.: Raman spectrum for uranous sulfate octahydrate, complex , and  nm lasers, showing fluorescence. Sample: -S.Figure B.: FTIR spectrum of uranous sulfate octahydrate, complex . Sample:-S.Figure B.: Powder XRD spectrum of uranous sulfate octahydrate, complex ,simulated from the crystal structure presented in Chapter  using Mer-cury . [] (Cu source).B. ParisaiteFigure B.: Raman spectrum for uranous sulfate octahydrate, complex , nm laser, showing fluorescence. Sample: -B.Figure B.: FTIR spectrum of parisaite. * marks signals originating from thematrix, not the sample. Sample: -B.Figure B.: Powder XRD spectrum of uranous sulfate octahydrate, complex, using Cu source. Sample: -B..Appendix CTotal sulfate determinationSummaryTotal sulfate was determined by titration with lead perchlorate, which causes theprecipitation of lead sulfate. Uranium interferes with the electrode response, so itis first removed by hydrogen peroxide precipitation at ∼pH . e titration is con-ducted in  isopropanol to reduce the solubility of lead sulfate to negligible levels.A Pb++ ion-selective electrode is used to detect the endpoint, which manifests as asharp increase in potential indicating the presence of Pb++ when all of the sulfatehas precipitated.is method is based on a method published by Metrohm titled Titrimetric de-termination of sulfate [].Reagents• Deionized water• Hydrogen peroxide ()• Ammonium hydroxide ()• Isopropanol• Methyl red (. wt.  in  vol.  ethanol)• Lead perchlorate solution (.mol L− in  isopropanol)Equipment• mL volumetric flask• Pb++ ion-selective electrode• Double-junction reference electrode with KNO outer filling solution• Automatic titrator capable of potentiometric titration• . µm syringe filter• mL plastic syringe and narrow tubing• mL beaker• -inch PTFE stir barMethode initial sample for analysis should be an aqueous acidic solution containing noless than  g L− sulfate and any amount of uranium. If the sample for analysis is asolid, it can be first digested first in nitric acid, and all transfers and dilutions shouldbe measured by mass to obtain the best accuracy.Step I: Uranium removal. To the mL volumetric flask add an aliquot of sample containing approxi-mately mg of sulfate.. Add enough  hydrogen peroxide to oxidize any uranium(IV) to uran-ium(VI) and to precipitate all uranium as UO ·HO, leaving approximately. excess peroxide.. Partially neutralize the solutionwith the ammonia solutionusing a drop-per, swirling frequently to mix. When the solution shows a hint of cloudinessfrom uranium precipitation, add – drops of methyl red indicator and con-tinue neutralizing with ammonia until the colour changes from orange/redto canary yellow. Caution: ammonia is toxic. Use proper ventilation.. Dilute to mL with deionized water. Mix well, and then allow the solids tosettle, at least min. e uranium-containing precipitate should be yellow,and the sulfate-containing supernatant should be clear.. Using the syringe and narrow tube, withdraw the supernatant and filter itthough the syringe filter into a labeled sample bottle. Discard the first  mLof filtrate.Step II: Sulfate determination. Pipette a mL aliquot of the filtered uranium-free sample into a  mLbeaker with a -inch stir-bar. Add mL of isopropanol and mL of deion-ized water.. Polish the Pb++ ion-selective electrode, if necessary.. Immerse the electrodes and allow several minutes to reach equilibrium. estarting potential should be between - and  mV.. Titrate with lead perchlorate in .mL doses. A white lead sulfate precipitatewill form with each dose, until all of the sulfate is consumed. e endpointis indicated by a rapid increase in potential associated with the appearance ofunprecipitated Pb++ ions, which should appear as a peak in the differentialdata.Appendix DAnalysis of acidic metal sulfatesolutions for sulfuric acid using a pHelectrode and standard additionBerend Wassinke University of British ColumbiaDepartment of Materials EngineeringFebruary Analysis of aqueous metal-containing solutions for acid is complicated by thepropensity of many metal ions to hydrolyze and produce acid. Standard titrationmethods are generally obviated since hydrolysis in many cases may be extensive atpH values well below . e higher the hydrolyzablemetal ion concentration is rela-tive to the acid concentration, themore serious the problem. Measurement of pH instrongly acidic, high metal concentration solutions is complicated by the high ionicstrength of the solution and possible non-linear response of the pH electrode. Inaddition, in a high sulfate medium, sulfuric acid may not be fully dissociated. Cal-ibration of the electrode response with solutions containing precisely known con-centrations of the various metal salts may be workable, however it is oen the casethat the composition of samples is neither fully known, nor constant. One possibleapproach to the problem is to use the method of standard addition. (See for exam-ple D.C. Harris Quantitative Chemical Analysis W. H. Freeman and Co, rd edn., p. -.)e pH electrode is used like any other ion selective electrode. e response isrecorded in mV rather than in pH units. e electrode is calibrated with standardscontaining known acid concentrations (e.g., HSO) in a high ionic strength solu-tion. Only the “slope” (e.g., mV/g/LHSO) of the electrode is required. e sampleof interest is then diluted into the same high ionic strength medium. e electroderesponse of a known volume of sample is recorded. Next a small volume of knownacid is added. e addition should not significantly change the ionic strength. eelectrode response is recorded again. If desired further additions of standard acidmay be made.e concentration of acid may be calculated based on the Nernst equation asfollows:E0 = E∗+m logC0 (D.)E1 = E∗+m log(C0V0 +CsVsV0 +Vs)(D.)E0 mV reading of the sampleE∗ constantE1 mV reading of sample plus spike of standard acidm electrode slope (from calibration) in mV/concentration unitsC0 unknown acid concentration in diluted sampleV0 sample volume (mL)Cs standard acid concentration (units must be the same as those usedfor calibration)Vs spike volume (mL)Subtracting Eqs. (D.) and (D.):E1−E0 = m log(C0V0 +CsVs(V0 +Vs)C0)(D.)Rearranging this gives:C0 =CsVs[(V0 +Vs)10(E1−E0)/m]−V0(D.)If more than one spike is added, Vs refers to the total volume of standard acidadded. e initial value forE0 is used for calculating the original acid concentrationC0 aer each spike. e results may be averaged. ese equations assume a constantactivity coefficient for the proton under analysis conditions. If the ionic strength isrelatively high and constant, this condition is adequately fulfilled.A high quality electrode is essential. As will be seen later,±mV can easily leadto errors on the order . A pH meter with a resolution of . mV is desirable.Electrodes must be handled with care to avoid scratching or static charging. eglass bulb may be blotted dry, but never rubbed. e requisite procedures for useand maintenance of the electrode, as appropriate, should be consulted. e elec-trode must be clean and free of excess moisture to avoid contamination or dilutionof the sample. Solutions need to be gently stirred to facilitate equilibration. eelectrode response is temperature sensitive. e sample and standards should be atabout the same temperature. Magnetic stirrers can generate enough heat to warmthe sample significantly. Insulation (e.g., a piece of wood) and a thermostatted wa-ter bath are advisable. Enough time is required to acquire a stable reading. ismaytake – minutes. Spike acid amounts should be in the range of - of that inthe original sample.Sample dilution may be necessary to lower the sample ionic strength relative tothe matrix background (i.e., a salt at high concentration). is may be required toensure that the calibration slope is applicable to the samples. However, excessivedilution can lead to errors due to a rise in the pH and metal ion hydrolysis. us adilution should be employed such that the proton concentration remains in excessof . M. Systems capable of buffering (e.g., HSO−SO– and HCl−Cl– and anyweak acids) require special caution.Procedure for HSO–metal sulfate solutionsemethod has been tested with .–. NHSO (– g/L) with varyingmetalion concentrations. Reagents and standards include the following:• MgSO ·HO (certified ACS grade or better)• ., . and . N HSO standards in  M MgSO ·HO ( g/L)• . N HSO in  M MgSO ·HO ( g/L)Prepare the standards. Store in plastic bottles and have a one inch stir bar in each.Dilute your samples so that they contain .–. N HSO (.–. g/L) andMMgSO ·HO. Use a water bath at constant temperature (◦C or less) to ther-mostat the standards. e temperature of samples and standards should be within.◦Cof each other. Measure themV response of a high quality pH electrode to eachstandard starting at the most dilute one. Allow a consistent time (– minutes) forthe reading to stabilize. Constant dri may be indicative of temperature dri ormeter or electrode malfunction. A meter with a resolution of . mV is preferable.A meter with only  mV resolution will give less precise results. e electrode mustbe clean and dry before immersing it in the solution. Gently blot dry the bulb. Donot rub it. It can be easily damaged. Plot the mV readings versus log10 of the acidconcentrations. e slope of the graph is the electrode slope.Pipette . mL of a sample into a clean dry  mL beaker containing a stirbar. Ensure that the sample is at the same temperature as the standards. Measurethe mV reading of the clean and dry pH probe immersed in the solution. Allowa few minutes for equilibration. Add a known volume of the standard acid (.N HSO in  M MgSO · HO. e amount added should be in the range of- of that in your diluted sample. Record the mV reading. Calculate theacid concentration in your sample from Eq. (D.) and the sample dilution factor. Ifdesired a second spikemay be added and the results averaged. e results describedbelow do not suggest a significant advantage to this. emV readings for the dilutedsample and aer additions of spikes must fall within the calibration range.Results and discussione presence of high MgSO · HO in samples, standards and spikes provides ahigh and relatively constant background sulfate concentration. is facilitates thehigh ionic strength required and mitigates the interfering effects of relatively low,but variable sulfate levels. MgSO · HO is a convenient choice since Na+ andK+ salts of sulfate are not soluble enough. e + charge on Mg+ also provides ahigher ionic strength. e samples are diluted so that they are within the calibrationrange. is also lowers the sample sulfate concentration to a value that is smallrelative to the background sulfate. e acid concentration in the spikesmust be highenough to allow addition of a fairly small, but practical volume. However, it cannotbe too high since this would result in a substantial volume of mixing error. It isimportant that the volume of sample plus spikes be accurately known. e accuracyof the method is limited by a number of factors. e two most significant are theintrinsic imprecision associated with pH probes and the fact that standard additionis essentially an extrapolation. An uncertainty of only±. mV in a single readingresults in an uncertainty of ∼± in concentration associated with that reading.At least two readings are required to ascertain the electrode slope. A further twoare needed to establish the concentration. Errors in the readings associated witha sample analysis may be amplified by the extrapolation, i.e., relating back to theinitial sample prior to standard addition. us precision on the order of – maybe reasonably expected.Results for three solutions are presented below. ese are classified as low acid–high metal, medium acid–medium metal and high acid–low metal samples. Notethat the analyst must verify the validity of the method for particular solution composi-tions not within the scope of this work. Solution compositions are shown inTableD..Sulfate due to sulfuric acid ranged from – of total sulfate.Results for the analysis of the above three samples are presented in Table D..Associated calibration data are shown in Table D. and Fig. D.. e equipmentused was a Fisher Accumet pH probe (glass body, ceramic junction and refillable,a Corning pH meter with . mV resolution and a Radiometer ABU autoburretefor the spikes. A thermocouple was used to measure temperatures of the samplesand to control them within±.◦C.Table D.: Solution compositions for testing HSO analysis by pH electrode(g/L).that standard addition is essentially an extrapolation. An uncertainty of only + 0.2 mV in a single reading results in an uncertainty of ~+1% in concentration associated with that reading. At least two readings are required to ascertain the electrode slope. A further two are needed to establish the concentration. Errors in the readings associated with a sample analysis may be amplified by the extrapolation, ie relating back to the initial sample prior to standard addition. Thus precision on the order of 1-5 % may be reasonably expected.   Results for three solutions are presented below. These are classified as low acid-high metal, medium acid-medium metal and high acid-low metal samples. Note that the analyst must verify the validity of the method for particular solution compositions not within the scope of this work. Solution compositions are shown in Table 1. Sulfate due to sulfuric acid ranged from 30-62% of total sulfate.   Table 1. Solution compositions for testing H2SO4 analysis by pH electrode. Speciesa Low Acid- High Metals Medium Acid- Medium Metals High Acid- Low Metals Al+3 (g/L) 2.66 2.0 1.33 Co+2 0.80 0.60 0.40 Fe+3 3.99 3.0 2.0 Mg+2 2.00 1.5 1.0 Mn+2 4.66 3.5 2.33 Ni+2 7.98 6.0 4.0 H2SO4 24.52 (0.500 N) 36.78 (0.750N) 45.43 (0.9264 N) Total SO42- from metals 54.9 41.3 27.5 Total SO42- 78.9 77.3 72.0 a all metals added as sulfate salts.    Results for the analysis of the above three samples are presented in Table 2. Associated calibration data are shown in Table 3 and Figure 1.The equipment used was a Fisher Accumet pH probe (glass body, ceramic junction and refillable), a Corning pH meter with 0.1 mV resolution and a Radiometer ABU80 autoburrete for the spikes. A thermocouple was used to measure temperatures of the samples and to control them within + 0.3oC.    The data suggest that fairly good analyses can be obtained with the types of solutions used. Accuracies of 98-103% were found. There might be a trend toward slight overestimation of low acid concentrations in the presence of relatively high metal sulfate concentrations, and underestimation for high acid-low metal solutions. This would need to be verified. Close inspection of the calibration data suggests a slight curvature. A second order equation fit the data very well. However, calculation of the slopes at points in the middle of the mV range for sample plus spike resulted in poorer accuracy. Linear fitting of the data and use of a constant slope value seems to yield the best results. Possible curvature of the calibration plot may mitigate against extending the useful range    Figure D.: Calibration plot for HSO standards in  M MgSO; . mV res-olutionTable D.: Analytical results for analysis of HSO–metal sulfate solutions us-ing a meter with . mV resolution.Table 2. Analytical results for analysis of H2SO4-metal sulfate solutions using a meter with 0.1 mV resolution. Sample E0 mV Spike Vol. mLa E1 mV Analyzed [H2SO4] N Actual [H2SO4] N Accuracy % Low acid- high metals  (Average) 299.5 299.5 3.014 5.994 322.1 333.1 0.517 0.512  (0.514) (25.2 g/L) 0.500 24.5 g/L (0.025 N in diluted sample) 103.4 102.3  (102.9) Medium acid- medium metals  (Average) 311.0 311.0 3.029 6.491 328.3 338.9 0.747 0.739  (0.743) (36.4 g/L) 0.750 36.8 g/L (0.0375 N in diluted sample) 99.6 98.6  (99.1) High acid- low metals  (Average) 316.2 316.2 2.984 7.013 330.7 341.6 0.919 0.907  (0.913) (44.8 g/L) 0.9264 45.4 g/L (0.04632 N in diluted sample) 99.2 97.9  (98.5) a second spike volume refers to the total added; two spikes added per sample Other conditions: Electrode slope = 60.233 mV Sample volume = 50.00 mL Dilution: 5.00 mL to 100.0 mL in 2 M MgSO4 as MgSO4.7H2O Spike composition: 0.650 N H2SO4 (31.875 g/L) + 2 M MgSO4. Equilibration time: 3 minutes    Table 3. Calibration data for analytical results in Table 2. [H2SO4] N [H2SO4] g/L log10[H2SO4] N mV reading 0.0100 0.4904 -2.000 274.7 0.0750 3.678 -1.1249 326.8 0.100 4.904 -1.000 334.1 0.200 9.808 -0.6990 353.5 Note: standards contain 2 M MgSO4 as MgSO4.7H2O Equilibration time: 3 minutes Slope = 60.233 mV  Table D.: Calibration data for analytical results in Table D..Table 2. Analytical results for analysis of H2SO4-metal sulfate solutions using a meter with 0.1 mV resolution. Sample E0 mV Spike Vol. mLa E1 mV Analyzed [H2SO4] N Actual [H2SO4] N Accuracy % Low acid- high metals  (Average) 299.5 299.5 3.014 5.994 322.1 333.1 0.517 0.512  (0.514) (25.2 g/L) 0.500 4.5 g/L (0.025 N in diluted sample) 103.4 102.3  (102.9) Medium acid- medium metals  (Average) 311.0 311.0 3.029 6.491 328.3 338.9 0.747 0.739  (0.743) (36.4 g/L) 0.750 36.8 g/L (0.0375 N in diluted sample) 99.6 98.6  (99.1) High acid- low metals  (Average) 316.2 316.2 2.984 7.013 330.7 341.6 0.919 0.907  (0.913) (44.8 g/L) 0.9264 45.4 g/L (0.04632 N in diluted sample) 99.2 97.9  (98.5) a second spike volume refers to the total added; two spikes added per sample Other c nditions: Electrode slope = 60.233 mV Sample volume = 50.00 mL Dilution: 5.00 mL to 100.0 mL in 2 M MgSO4 as MgSO4.7H2O Spike composition: 0.650 N H2SO4 (31.875 g/L) + 2 M MgSO4. Equilibration time: 3 minutes    Table 3. Calibration data for analytical results in Table 2. [H2SO4] N [H2SO4] g/L log10[H2SO4] N mV reading 0.0100 0.4904 -2.000 274.7 0.0750 3.678 -1.1249 326.8 0.100 4.904 -1.000 334.1 0.200 9.808 -0.6990 353.5 Note: standards contain 2 M MgSO4 as MgSO4.7H2O Equilibration time: 3 minutes Slope = 60.233 mV  edata suggest that fairly good analyses can be obtained with the types of solu-tions used. Accuracies of –were found. eremight be a trend toward slightoverestimation of low acid concentrations in the presence of relatively high metalsulfate concentrations, and underestimation for high acid–lowmetal solutions. iswould need to be verified. Close inspection of the calibration data suggests a slightcurvature. A second order equation fit the data very well. However, calculation ofthe slopes at points in the middle of the mV range for sample plus spike resulted inpoorer accuracy. Linear fitting of the data and use of a constant slope value seems toyield the best results. Possible curvature of the calibration plot may mitigate againstextending the useful range to higher acid concentrations. Higher MgSO concen-trations might be worth investigating to give a higher background sulfate level andionic strength. Lower concentrationsmay not be practical due tometal ion hydroly-sis. Treating the data for the standards as an ordinary calibration curve yielded quitepoor analytical results. e matrix effects appear to be significant and this is bestovercome by standard addition. It would seem plausible that the higher the pro-portion of acid compared to metal sulfates, the more accurate the analysis shouldbe. e lowest acidity solution contained . times as much sulfate from metal sul-fate salts as that due to the HSO. is still proved amenable to analysis by thismethod. Cases where the acid content relative to the metals is lower would needto be checked first to see if the method is applicable. e higher the proportion ofmetal sulfates, the more serious the problem of metal hydrolysis.Data for the analysis of the same solutions using ameter with mV resolution ispresented in TableD.. Calibration data are presented in TableD. and Fig. D.. eresults show that poorer precision may be anticipated than when a meter with .mV resolution is used. It is up to the analyst to determine if this is acceptable. A±mV uncertainty in the mV reading can produce – uncertainty in the result forthe analyses in Table D.. e extent of the uncertainty depends in the magnitudeof the difference between the initial and spike mV readings.Table D.: Analytical results for analysis of HSO–metal sulfate solutions us-ing a meter with  mV resolution.Table 4. Analytical results for analysis of H2SO4-metal sulfate solutions using a meter with 1 mV resolution. Sample E0 mV Spike Vol. mLa E1 mV Analyzed [H2SO4] N Actual [H2SO4] N Accuracy % Low acid- high metals  (Average) 288 288 2.777 6.785 309 323 0.535 0.534  (0.534) (26.2 g/L) 0.500 24.5 g/L (0.025 N in diluted sample) 107.1 106.7  (106.9) Medium acid- medium metals  (Average) 298 298 3.014 7.012 315 327 0.764 0.747  (0.756) (37.0 g/L) 0.750 36.8 g/L (0.0375 N in diluted sample) 101.9 99.6  (100.7) High acid- low metals  (Average) 303 303 3.003 7.028 318 328 0.891 0.935  (0.913) (44.8 g/L) 0.9264 45.4 g/L (0.04632 N in diluted sample) 96.1 100.9  (98.5) Low acid- high metals (Sample diluted 3 mL to 100 mL) (Average) 273 273 1.400 2.914 294 305 0.471 0.489   (0.480) (23.6L) 0.500 24.5 g/L (0.015 N in diluted sample) 94.3 97.9   (96.1) a second spike volume refers to the total added; two spikes added per sample Other conditions: Electrode slope = 60.464 mV Sample volume = 50.00 mL Dilution: 5.00 mL to 100.0 mL (unless otherwise noted) in 2 M MgSO4 as MgSO4.7H2O Spike composition: 0.650 N H2SO4 (31.875 g/L) + 2 M MgSO4. Equilibration time: 4 minutes   Table 5. Calibration data for analytical results in Table 4. [H2SO4] N [H2SO4] g/L log10[H2SO4] N mV reading 0.0100 0.4904 -2.000 263 0.0750 3.678 -1.1249 315 0.100 4.904 -1.000 323 0.200 9.808 -0.6990 342 Note: standards contain 2 M MgSO4 as MgSO4.7H2O Equilibration time: 4 minutes Slope = 60.464 mV   Table D.: Calibration data for analytical results in Table D..Table 4. Analytical results for analysis of H2SO4-metal sulfate solutions using a meter with 1 mV re olution. Sample E0 mV Spike Vol. La E1 mV Analyzed [H2SO4] N Actual [H2SO4] N Accuracy % Low acid- high metals  (Average) 288 288 2.777 6.785 309 323 0.535 0.534  (0.534) (26.2 g/L) 0.500 24.5 g/L (0.025 N in diluted sample) 107.1 106.7  (106.9) Medium acid- medium metals  (Average) 298 298 3.014 7.012 315 327 0.764 0.747  (0.756) (37.0 g/L) 0.750 36.8 g/L (0.0375 N in diluted sample) 101.9 99.6  (100.7) High acid- low metals  (Average) 303 303 3.003 7.028 318 328 0.891 0.935  (0.913) (44.8 g/L) 0.9264 45.4 g/L (0.04632 N in diluted sample) 96.1 100.9  (98.5) Low acid- high metals (Sample diluted 3 mL to 100 mL) (Average) 273 273 1.400 2.914 294 305 0.471 0.489   (0.480) (23.6L) 0.500 24.5 g/L (0.015 N in diluted sample) 94.3 97.9   (96.1) a second spike volume refers to the total added; two spikes added per sample Other conditions: Electrode slope = 60.464 mV Sample volume = 50.00 mL Dilution: 5.00 mL to 100.0 mL (unless otherwise noted) in 2 M MgSO4 as MgSO4.7H2O Spike composition: 0.650 N H2SO4 (31.875 g/L) + 2 M MgSO4. Equilibration time: 4 minutes   Table 5. Calibration data for analytical results in Table 4. [H2SO4] N [H2SO4] g/L log10[H2SO4] N mV reading 0.0100 0.4904 -2.000 263 0.0750 3.678 -1.12 9 315 0.100 4.90  -1.000 323 0.200 9.808 -0.6990 342 Note: standards contain 2 M MgSO4 as MgSO4.7H2O Equilibration time: 4 minutes Slope = 60.464 mV      Figure D.: Calibration plot for HSO standards in  M MgSO;  mV reso-lutionAppendix EOxidation kinetics worksheete operational worksheet and checklist included on the following pages was usedfor each experiment described in Chapter .Oxidation	  Study	  Checklist	   	   OX-­‐_______	  v.1.1	  May	  26,	  2014,	  A.B.	  Date	   ________________________	  Matrix	   ________________________	  Gas	   ________________________	   T	   ______	   °C	  [UT]	   ______	   mM	  [Acid]	   ______	   M	   ______	   N	  Date/time	  of	  test	  solution	  creation	   ______________________________	  Date/time	  added	  to	  reactor	  with	  N2	   ______________________________	  Notes:	  	  	  	  	  T-­‐Log	  Start	  Time	   _________________	  Reaction	  Start	  Time	   _________________	  Reaction	  End	  Time	   _________________	  STK	  [U(IV)]	   _________________	  Bubbler	  I	  =	   ______	  Full	  scan	  cycle	  time	   _________________	  Initial	  Samples	  AQU-­‐	  FA-­‐	  ICP-­‐	  SUL-­‐	  Final	  Samples	  AQU-­‐	  FA-­‐	  ICP-­‐	  	  Periodic	  Sampling	  Sample	  ID	   Sampling	  Time	   Elapsed	  Time,	  s	   Sample	  Vol.,	  mL	   Frozen?	   	  [U(IV)],	  M	  UV-­‐Vis	   Titration	  	   	   	   	   	   	   	  	   	   	   	   	   	   	  	   	   	   	   	   	   	  	   	   	   	   	   	   	  	   	   	   	   	   	   	  	   	   	   	   	   	   	  	   	   	   	   	   	   	  Calibration	  Curve	  and	  UV-­‐Vis	  	  Wavelength:	  650.9	  nm	  	  Background	  cell:	  	   __________________	  	  	  Std.	  #	   Abs.	  	   	  	   	  	   	  	   	  	   	  	   	  	   	  Oxidation	  Study	  Checklist	   	   OX-­‐_______	  v.1.1	  May	  26,	  2014,	  A.B.	  Part	  1	  –	  Day-­‐Before	  Preparations	  	  	   Get	  the	  worksheet	   Discuss	  with	  Alex	  which	  test(s)	  will	  be	  done	  tomorrow.	  	   Assemble	  the	  reactor	   If	  not	  already	  done,	  assemble	  the	  reactor	  	   Check	  reagents	   Check	  that	  there	  is	  enough	  of	  the	  U(IV),	  U(VI),	  acid,	  and	  MgSO4	  solutions.	  	  If	  not,	  prepare	  them.	  	   Partially	  prepare	  standards	   Set	  out	  and	  label	  volumetric	  flasks	  for	  the	  standards,	  as	  described	  on	  the	  worksheet.	  	  Introduce	  the	  acid	  and	  U(VI)	  solutions	  as	  described,	  and	  insert	  the	  stopper.	  	  Set	  aside	  for	  tomorrow.	  	  Do	  not	  add	  U(IV).	  	   Partially	  prepare	  test	  solution	   Similar	  to	  preparing	  the	  standards,	  but	  for	  the	  test	  solution.	  	  Add	  the	  acid	  and	  U(VI)	  solution	  to	  the	  500	  mL	  flask	  and	  insert	  stopper.	  	  Set	  aside	  for	  tomorrow.	  	   Prepare	  sample	  bottles	   As	  needed,	  prepare	  the	  AQU,	  FA,	  ICP	  and	  SUL	  sample	  bottles.	  	  Fill	  out	  the	  computer	  worksheets	  and	  label	  the	  bottles.	  	  Set	  aside	  for	  tomorrow.	  	   Install	  cell	  holders	   Install	  the	  100mm	  cell	  holders	  into	  the	  UV-­‐Vis.	  	   Set	  out	  bubbler	   Set	  out	  the	  bubbler	  with	  the	  closest	  ionic	  strength	  to	  tomorrow’s	  test	  solution.	  	   Refill	  water	  bath	   Fill	  the	  heating	  water	  bath	  to	  the	  line	  with	  deionized	  water.	  	   Check	  gas	  cylinders	   Check	  that	  there	  is	  enough	  oxygen,	  air,	  and	  nitrogen	  in	  the	  gas	  cylinders.	  	  Make	  sure	  they	  are	  turned	  off	  at	  the	  tank	  at	  the	  end	  of	  the	  day.	  Part	  2	  –	  Morning	  Checks	  and	  Startup	  	  	   Check	  reactor	  tubing	   Check	  that	  the	  reactor	  inlet	  and	  outlet	  tubes	  are	  secure.	  	  Check	  that	  the	  UV-­‐Vis	  circulating	  lines	  are	  all	  connected.	  	  Check	  that	  the	  water	  bath	  return	  line	  is	  pointed	  into	  the	  water	  bath.	  	   Check	  reactor	  body	   Check	  that	  all	  instruments	  are	  in	  place	  through	  the	  ports	  in	  the	  lid.	  	  The	  reactor	  should	  be	  level	  and	  high.	  	  Check	  that	  the	  impeller	  is	  free-­‐spinning.	  	   Check	  chiller	  tubing	   Check	  that	  the	  chiller	  return	  line	  points	  into	  the	  bath.	  	  Check	  that	  the	  chiller	  circulating	  lines	  are	  connected	  to	  the	  lines	  leading	  into	  the	  AA	  room,	  and	  that	  there	  are	  no	  kinks	  in	  the	  line.	  	   Turn	  on	  chiller	   Turn	  on	  the	  chiller.	  	  Set	  the	  temperature	  to	  12°C	  for	  >=20°C	  tests,	  and	  to	  5°C	  for	  colder	  tests.	  	  Verify	  that	  the	  return	  line	  is	  flowing.	  	   Turn	  on	  water	  bath	   Turn	  on	  the	  heating	  water	  bath.	  	  Set	  2°C	  higher	  than	  the	  test	  temperature.	  	   Install	  bubbler	  	   	   Top	  up	  the	  bubbler	  to	  200	  mL	  if	  necessary.	  	  Place	  the	  bubbler	  in	  the	  water	  bath.	  	  Connect	  the	  reactor	  gas	  lance	  to	  the	  bubbler	  outlet.	  	  Connect	  the	  nitrogen	  line	  to	  the	  bubbler	  inlet.	  	  Record	  the	  ionic	  strength	  of	  the	  bubbler	  on	  the	  worksheet.	  	   Turn	  on	  UV-­‐Vis	   Turn	  on	  the	  UV-­‐Vis	  (green	  switch)	  and	  computer.	  	  Once	  it	  has	  warmed	  up,	  start	  the	  software	  “UV	  Winlab”.	  	  Oxidation	  Study	  Checklist	   	   OX-­‐_______	  v.1.1	  May	  26,	  2014,	  A.B.	  Part	  3	  –	  Solution,	  Standard,	  and	  Reactor	  Preparation	  	  	   Finish	  making	  test	  solution	   Mix	  up	  the	  test	  solution	  in	  a	  500	  mL	  volumetric	  flask	  according	  to	  the	  recipe	  on	  the	  worksheet.	  	   Turn	  on	  nitrogen	   Turn	  on	  the	  nitrogen	  to	  100	  cc/min.	  	   Pour	  test	  solution	  into	  reactor	   Using	  the	  funnel,	  pour	  the	  test	  solution	  into	  the	  reactor.	  	  Check	  that	  nitrogen	  is	  bubbling.	  	   Turn	  on	  gentle	  stirring	   Turn	  on	  the	  stirrer	  to	  a	  low	  setting.	  	  It	  should	  be	  enough	  to	  agitate	  the	  solution	  for	  good	  heating/cooling,	  but	  not	  enough	  to	  cause	  churning.	  	   Turn	  on	  T-­‐controller	   Switch	  on	  Omega	  temperature	  controller.	  	  Ensure	  the	  USB	  cable	  labeled	  “solenoid	  controller”	  is	  plugged	  into	  the	  computer.	  	  Open	  the	  program	  “CN7-­‐A”,	  and	  check	  that	  “COM	  Status”	  is	  green.	  	  Check	  that	  it	  is	  reading	  the	  proper	  temperature.	  	   Configure	  T-­‐controller	   In	  the	  “CN7-­‐A”	  software,	  click	  “Read	  Configuration	  from	  File”.	  	  Load	  the	  config	  file	  that	  matches	  the	  temperature	  of	  the	  test.	  	  The	  files	  are	  located	  in	  /Documents/Omega	  Configurations/.	  	  Click	  on	  “Send	  Configuration	  to	  Instrument”.	  	  Listen	  for	  the	  clicking	  sound	  of	  the	  solenoid.	  	   Finish	  making	  standards	   Add	  the	  U(IV)	  stock	  solution	  to	  the	  pre-­‐prepared	  standard	  flasks.	  	  Be	  very	  careful	  to	  avoid	  oxidation,	  using	  a	  nitrogen	  blanket	  when	  possible.	  	   Pre-­‐heat	  standards	   Put	  the	  tightly-­‐capped	  standard	  flasks	  into	  the	  water	  bath.	  	   Fill	  background	  cell	   Fill	  a	  20mm	  quartz	  cell	  with	  the	  specified	  background	  solution,	  and	  place	  in	  the	  reference	  cell	  holder	  of	  the	  UV-­‐Vis.	  	   Connect	  flow	  cell	   Connect	  the	  20mm	  flow-­‐through	  cell,	  with	  the	  white	  fitting	  on	  the	  side	  with	  the	  arrow.	  	   Increase	  N2	  flow	   Increase	  nitrogen	  flow	  to	  300	  cc/min.	  	   Stirrer	  700	  rpm	   Set	  the	  stirrer	  to	  700	  rpm.	  Part	  4	  –	  Calibration	  	  	   Open	  method	   In	  the	  UV	  Winlab	  software,	  open	  the	  method	  /AlexB/”U(IV)	  Oxidation	  Calibration”	  	   Fill	  in	  sample	  info	   Fill	  in	  the	  standard	  names	  (Std2,	  Std3,	  etc.),	  and	  “reactor”	  for	  the	  final	  sample.	  	  Also	  fill	  in	  the	  concentrations	  of	  the	  standards	  listed	  on	  the	  worksheet.	  	   Connect	  sipper	  and	  disconnect	  return	  line	   Disconnect	  the	  pump	  from	  the	  reactor	  and	  attach	  the	  sipper	  straw.	  	  Disconnect	  the	  return	  line	  from	  the	  reactor	  and	  point	  it	  into	  a	  beaker.	  	   Conduct	  calibration	   Press	  “Start”	  and	  follow	  the	  on-­‐screen	  instructions.	  	  Take	  care	  to	  keep	  the	  standards	  in	  the	  water	  bath	  as	  long	  as	  possible.	  	  Flush	  the	  cell	  with	  air	  between	  standards.	  	   Reconnect	  tubes	   Reconnect	  the	  sampling	  and	  return	  lines	  to	  the	  reactor.	  	  Turn	  on	  pump	  at	  full	  speed.	  	   Measure	  initial	  reactor	  absorbance	   Wait	  60	  seconds	  before	  taking	  the	  reading	  for	  initial	  absorbance.	  	  Leave	  the	  circulation	  pump	  running.	  	   Check	  calibration	  curve	   Check	  the	  calibration	  curve	  under	  Beer’s	  Law	  ?	  Calibration	  and	  make	  sure	  it	  is	  linear.	  	   Save	  file	  and	  exit	   File	  ?	  Save	  Results	  As	  ?	  “OX-­‐##	  Calibration”.	  	  Exit	  the	  calibration	  window.	  	  Oxidation	  Study	  Checklist	   	   OX-­‐_______	  v.1.1	  May	  26,	  2014,	  A.B.	  Part	  5	  –	  Startup	  	  	   Load	  and	  configure	  the	  UV-­‐Vis	  Method	   Open	  the	  method	  /AlexB/”Timedrive	  U(IV)	  Oxidation”.	  	  Change	  the	  Sample	  ID	  to	  “OX-­‐##”.	  	  Change	  the	  sampling	  interval	  and	  total	  time	  to	  whatever	  is	  appropriate	  for	  this	  test.	  	   Start	  T	  log	   Start	  the	  temperature	  log	  in	  the	  CN7-­‐A	  software.	  	  Log	  in	  5-­‐second	  intervals,	  10	  hour	  log	  time.	  	  Record	  the	  exact	  time	  when	  the	  log	  was	  started	  on	  the	  worksheet.	  	   Set	  oxygen	  flow	   Turn	  on	  the	  oxygen	  tank	  and	  adjust	  the	  outlet	  pressure	  to	  50	  psi.	  	  Adjust	  the	  flow	  rate	  to	  110	  mm,	  or	  what	  is	  described	  on	  the	  worksheet.	  	   Stop	  oxygen	  flow	   Stop	  the	  oxygen	  flow	  with	  the	  valve	  after	  the	  regulator.	  	  	   Disconnect	  nitrogen	   Disconnect	  the	  nitrogen	  line	  from	  the	  bubbler.	  	   Connect	  oxygen	   Connect	  the	  oxygen	  line	  to	  the	  bubbler.	  	  Since	  the	  valve	  is	  off,	  there	  should	  be	  no	  gas	  flow.	  	   Take	  sample	   FLUSH	  THE	  SAMPLING	  PORT.	  	  Take	  a	  ~30	  mL	  sample	  into	  the	  pre-­‐labeled	  bottle	  “OX-­‐##	  Initial”.	  	   Start	  UV-­‐Vis	  Log	   Press	  “Start”	  in	  UV	  Winlab.	  	   Conduct	  baseline	  correction	   Conduct	  the	  baseline	  correction	  as	  instructed	  on	  the	  screen.	  	   Start	  test	  -­‐ Start	  oxygen	  flow	  -­‐ Start	  UV-­‐Vis	  log	  -­‐ Record	  start	  time	   Start	  the	  oxygen	  flow,	  start	  the	  UV-­‐Vis	  log,	  and	  record	  the	  exact	  start	  time	  as	  close	  to	  simultaneously	  as	  possible.	  	   Check	  oxygen	  flow	   Check	  the	  oxygen	  flow	  and	  adjust	  to	  110	  mm,	  or	  what	  is	  prescribed	  on	  the	  worksheet,	  if	  necessary.	  Part	  6	  –	  During	  Test	  	  	   Take	  periodic	  samples,	  if	  required	   Either	  titrate	  immediately	  for	  U(IV)	  or	  freeze	  in	  liquid	  nitrogen	  for	  future	  analysis.	  	  Record	  samples	  on	  the	  worksheet.	  	   Turn	  off	  nitrogen	   Turn	  off	  the	  nitrogen	  tank,	  both	  at	  the	  tank	  and	  the	  valve.	  	   Prepare	  samples	   Prepare	  Free	  Acid,	  ICP,	  Sulfate	  analysis	  samples	  on	  the	  “OX-­‐##	  Initial”	  sample.	  	  Record	  the	  sample	  names	  (AQU-­‐###,	  etc.)	  on	  the	  worksheet.	  	   Clean	  up	   Wash	  glassware,	  wash	  and	  dry	  pipettes,	  tidy	  lab	  bench.	  	   Monitor	  the	  test	   Periodically	  check	  on	  the	  test.	  	  Ensure	  that	  the	  UV-­‐Vis	  log	  will	  not	  run	  out	  of	  time.	  	  Watch	  for	  blockages	  in	  the	  UV-­‐Vis	  return	  line.	  	  Monitor	  the	  temperature	  to	  ensure	  it	  stays	  ±0.1°C.	  Part	  7	  –	  Shutdown	  	  	   Record	  time	   Record	  the	  exact	  time	  of	  the	  shutdown	  	   Turn	  off	  oxygen	   Turn	  off	  the	  oxygen	  tank	  	   Take	  sample	   FLUSH	  THE	  SAMPLING	  PORT.	  	  Take	  a	  final	  sample	  into	  the	  pre-­‐labeled	  sample	  bottle.	  	   Save	  UV-­‐Vis	  file	   Stop	  the	  test.	  	  File	  ?	  Save	  Results	  ?	  Save	  as	  a	  New	  Task	  ?	  “OX-­‐##”.	  	   Save	  temperature	  file	   Stop	  the	  Omega	  temperature	  control	  log.	  	  Export	  the	  data	  to	  Dropbox	  ?	  Other	  Lab	  Files	  ?	  “OX-­‐##	  Temperature	  Log.csv”.	  	   Turn	  off	  instruments	   Shut	  off	  the	  agitator,	  water	  bath,	  UV-­‐Vis	  pump,	  chiller,	  temperature	  controller,	  and	  computer.	  	  Oxidation	  Study	  Checklist	   	   OX-­‐_______	  v.1.1	  May	  26,	  2014,	  A.B.	  Part	  8	  –	  Post-­‐production	  	  	   Final	  analysis	   Prepare	  Free	  Acid	  and	  ICP	  analysis	  samples	  on	  the	  final	  reactor	  sample.	  	   Turn	  off	  N2	  and	  O2	   Ensure	  the	  oxygen	  and	  nitrogen	  tanks	  are	  closed.	  	   Put	  away	  bubbler	   Attach	  the	  lid	  to	  the	  bubbler	  solution	  and	  put	  away	  in	  the	  cupboard.	  	   Clean	  reactor	   Empty	  and	  dispose	  of	  the	  spent	  solution.	  	  If	  the	  next	  test	  will	  be	  substantially	  different	  than	  this	  one,	  or	  if	  the	  reactor	  will	  not	  be	  used	  soon,	  rinse	  it	  thoroughly	  with	  D.I.	  water.	  	  Suck	  out	  as	  much	  water	  as	  possible	  with	  the	  pump	  and	  dry	  with	  a	  paper	  towel	  as	  much	  as	  possible.	  	   Clean	  reference	  cell	   Empty	  the	  reference	  cell	  and	  flush	  with	  D.I.	  water	  several	  times.	  	  Allow	  to	  try	  upside	  down	  on	  a	  Kimwipe	  with	  a	  beaker	  over	  top	  to	  protect	  it	  from	  dust.	  	   Clean	  flow	  cell	   Run	  D.I.	  water	  through	  the	  flow-­‐through	  cell.	  	  If	  it	  is	  not	  being	  used	  tomorrow,	  flush	  with	  5M	  nitric	  acid,	  and	  then	  again	  with	  D.I.	  water.	  	  Disconnect	  and	  store	  in	  its	  case.	  	   Clean	  U(IV)	  transfer	  beaker	   Empty	  remaining	  liquid	  to	  waste,	  clean,	  hang	  upside	  down	  to	  dry.	  	   Copy	  data	  to	  Dropbox	   Copy	  the	  test	  and	  calibration	  data	  to	  Dropbox	  via	  a	  USB	  key.	  	   Return	  sheets	  to	  Alex	   Ensure	  all	  fields	  are	  filled	  out,	  then	  staple	  and	  return	  the	  worksheet	  and	  checklist	  to	  Alex.	  	  Appendix FRadioactive uranium safe handlingproceduresSamples were prepared from natural uranium in a radioisotope-certified laboratoryby researchers trained in the hazards of radiation. Although the specific activity ofnatural uranium is relatively low, the reader should be aware that additional safetyprecautions are necessary when handling any radioactive substance.Background informationere are threemain types types of ionizing radiation: alpha particles, which are he-lium nuclei; beta particles, which are high-speed electrons; and gamma-type, whichare high-energy x-rays and gamma-rays. Natural uranium contains a mixture ofthree isotopes, all of which are radioactive: U (.), U (.), and U(trace amounts). All have relatively long half-lives exceeding ten thousand years.All three natural isotopes of uranium are alpha-emitters. Alpha radiation travelsonly a short distance in air (approximately  cm), and can be stopped by a piece ofpaper or human skin. Alpha emitters can be harmful to human health if they comeinto close proximity with exposed living tissue, such as in the lungs or gut. ey aregenerally not considered harmful for external exposure.e decay products of uranium are themselves radioactive, and some are alphaand gamma-emitters. ese products build up in natural uranium over time untilpseudo-steady state equilibrium is reached. ese products can be more harmfulto human health than the uranium itself. e decay product radon, which is a gas(and therefore easy to inhale), can be particularly harmful.A table of the natural uranium isotopes, their natural abundances, half-lives,and radiation types are given in F..Table F.: Isotopic abundance, half-life, and emission types for natural ura-nium. Reference: CRC Handbook, Lide [].Isotope Natural abundance, atom  Half-life, years Radiation typeU . .× αU . .× αU . .× αLicensing and traininge laboratorywas licensed toworkwith radioisotopes under the consolidatedUBCradiation licence by the UBC radiation safety officer. e licence was issued pur-suant to section  of the Nuclear Safety and Control Act.All researchers who worked directly with uranium in the laboratory, includ-ing summer students, took the UBC Radionuclide Safety and Methodology course,including the receiving of Class  dangerous goods.All shipments of uranium-containing materials, including laboratory samplesand waste, were packaged and inspected by a person certified in the shipment ofClass  dangerous goods, with training provided by the British Columbia Instituteof Technology (BCIT).Source of uraniumAll uraniumused for test workwas obtained fromCamecoCorporation, in the formof UO and UO. e UO and UO had been through a metallurgical refiningprocess to make it suitable as nuclear fuel, a process which removed most of thenatural decay products such as radium and radon. is process did not affect thenatural ratio of uranium isotopes. is refined natural uranium was therefore safer(i.e., less radioactive) than unrefined natural uranium. It could also be considered anearly pure alpha emitter. A total quantity of  kg of contained uranium was sentfrom Cameco to UBC for test work.Sample storageSolid and liquid samples were stored in sealed glass and plastic bottles in a cup-board under the laboratory bench in Frank Forward . e wooden door to thecupboard was kept closed when not in use. Small working quantities of uranium-containing solids and solutions were stored on the lab bench for short periods oftime during experiments. One bottle of UO containing at most  kg of containeduranium was also stored under the lab bench.As an added precaution, the remainder of the UO and UO was stored in alead-lined box specifically built to house radioactive samples.Routine checks and precautionse laboratory work benchs were covered with either a chemically-resistant epoxycoating or a stick-on PTFE teflon sheet to ease cleaning and prevent the contami-nation of the underlying surfaces. e laboratory work area and all equipment wascleaned regularly with soap and water. A Geiger–Müller counter was used to checkthat the work environment within the acceptable limit for radiation. Wipe tests tocheck for radioactive contamination were conducted periodically on all work areas.ese were analyzed at the UBC radiation safety office in the UBC hospital.Waste managementUranium-containing aqueous and solid wastes were segregated and stored in a  Lplastic pail. When full, the contents were neutralized to pH – with sodium hy-droxide. e neutralized aqueous supernatant was sent to UBC waste managementfor final disposal, while the uranium-containing precipitates was collected for ship-ment back to Cameco Corporation’s Rabbit Lake mine site for disposal by TDG-certified courier. e shipments complied with all relevant packaging and externalradiation regulations.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0166486/manifest

Comment

Related Items