Synthesis and Applications of Monolithic Metal-Organic Frameworks Bethany Miriam Connolly Department of Chemistry University of Cambridge Jesus College This dissertation is submitted for the degree of Doctor of Philosophy October 2019 ii iii I| Declaration This dissertation is the result of my own work and includes nothing which is the outcome of work done in collaboration except as declared in the Preface and specified in the text. It is not substantially the same as any that I have submitted, or, is being concurrently submitted for a degree or diploma or other qualification at the University of Cambridge or any other University or similar institution except as declared in the Preface and specified in the text. I further state that no substantial part of my dissertation has already been submitted, or, is being concurrently submitted for any such degree, diploma or other qualification at the University of Cambridge or any other University or similar institution except as declared in the Preface and specified in the text. It does not exceed the prescribed word limit. iv v II| Acknowledgements The work detailed in this thesis would not have been possible without the much-appreciated help and support of those around me. I would firstly like to thank my project supervisor Dr Andrew Wheatley, whose guidance and advice throughout the entirety of my PhD has been invaluable to its completion. The support of the entire Wheatley group is also greatly appreciated, especially from Joshua Mehta whose previous work on monolithic SnO2@ZIF-8 laid the foundation for this thesis. I would especially like to thank Josh for contributing the synthesis of monolithic NU-1000 to this work. I am also grateful for the hard work of the undergraduate project students that I supervised in the laboratory: Mollie Trueman with whom I worked on SnO2@ZIF-8 and Alexander Elliott for his work on Pd@HKUST-1 synthesis. Also based in The Chemistry Department, I would like to thank Dr Heather Greer for her training and advice regarding electron microscopy analysis and data processing. From the Chemical Engineering Department, I would like to thank Dr David Fairen-Jimenez, whose guidance and discussions were instrumental in directing the project as well as the rest of his research group. In particular, the hard work of Marta Aragones-Anglada under the supervision of Dr Peyman Moghadam, whose novel computational simulations were critical in supporting my experimental results. All computational simulations and theoretical isotherms discussed in this thesis are the result of this collaboration. Additionally, the help provided by Dr Diana Vulpe, whose training and assistance in mastering and maintaining the N2 gas- adsorption equipment, was essential in supporting not only my research but numerous other students. In the department of Materials Science and Metallurgy, I would firstly like to thank Dave Saul for his training on the notoriously time-consuming skill of epoxy resin cold mounting and micron fine polishing. Without this sample preparation guidance, I could not have obtained the mechanical properties data. Additionally, my research collaboration with Dr Giorgio Divitini in the Pd@HKUST-1 project is much valued. In particular his assistance with the collection of high-resolution STEM-EDX maps as well as the X-ray tomography raw data was vital in fully characterising this material. vi Furthermore, a number of international collaborations were undertaken to further support the project. Firstly, I must thank Prof. Joaquin Silvestre-Albero and Jesus Gandara-Loe at the University of Alicante for the Hg porosimetry, CH4 and CO2 isotherms raw data for each of the monolithic Zr-MOF materials. This experimental data was instrumental to the project’s completion and publication. Furthermore, the FLIM data provided and processed by Dr Stefan Wuttke, Dr Dom Lamb and Nader Danaf (University of Munich) was highly informative in underpinning the qualitative observations regarding monolith formation. Their advice on the structure and presentation of the consequent publication was also highly valued. Prof. Mark Allendorf along with Dr Vitalie Stavila and Dr James White (Sandia National Laboratory, USA) collected much of the raw data regarding H2 adsorption testing of monolithHKUST-1 and Pd@monolithHKUST-1. I would especially like to thank them, not only for their time taken to collect this valuable data, but also for their advice and constructive feedback regarding the project’s outcome. This H2 testing data was reinforced by Katie Hurst, Sarah Shulda and Phil Parilla from the U.S. DOE National Renewable Energy Laboratory, Colorado. Preliminary catalytic testing for CO and CH4 oxidation using doped monolithic composites was performed in collaboration with Dr Alexander Orlov at Stony Brook University while preliminary testing of monolithic composites towards hydrogenolysis of allyl alcohols was performed in collaboration with Hiroshi Naka at Nagoya University. I would like to thank each of them for the time taken to test the catalytic capabilities of these materials. I would like to express my gratitude to the Council for the Oppenheimer Studentship for their generous funding over the last 3 years. Additionally, the travel grants provided by Jesus College were much appreciated, enabling me to present this work to an international audience. Furthermore, I am grateful for the generous research prizes awarded by the Cambridge NanoDTC and Cambridge Society for the Application of Research. Finally, I would like to thank my friends and family for all their encouragement, especially James for his continued advice and emotional support. Last, but by no means least, I would like to acknowledge the guidance of my dad, who was so proud when I started this PhD and would have loved to see me finish it. vii III| Abstract Metal-organic frameworks (MOFs) are a diverse family of coordination polymers which result from the crystalline self-assembly of metal ions/oxide clusters with multi-dentate organic linkers. Despite outstanding academic results, these porous materials have not entered the industrial sphere; their powder morphology renders them unsuitable for most potential applications. A new class of monolithic MOF was recently reported - monolithMOF combine the single-crystal density and porosity of the theoretical MOF with the functionality of a pelletised material. Correspondingly, outstanding results were achieved for two MOFs, with monolithHKUST-1 (Cu3(benzene-1,3,5-tricarboxylate)2) setting a benchmark in natural gas storage while monolithZIF-8 (Zn(2-methylimidazolate)2) demonstrated a capacity to host catalytic nanoparticles, NP@monolithMOF, for practical and environmental water purification. With the aim of expanding on these preliminary results, the synthesis of new monolithic MOFs was explored. Zr-MOFs, renowned for their industrially favourable chemical, thermal and mechanical stability, were targeted. A synthetic procedure was developed for the preparation of the archetypal Zr-MOF UiO-66 (Zr6O4(OH)4(1,4-benzenedicarboxylate)6) as a robust and porous monolith, monolithUiO-66. A novel capacity to control bulk density and pore-size distribution in the macroscopic monolith, via the synthetic inclusion of non-crystalline mesoporosity, was also developed. The generality of this synthetic procedure towards Zr-MOFs in general was further explored including functionalised UiO-66 analogues as well as non- isostructural Zr-MOFs. The capacity of monolithMOFs to host nanoparticles was determined through analysis of the effects of nanoparticle surface functionality on doping level of monolithZIF-8. Subsequently, the immobilisation of mono-, bi- and tri-metallic nano-objects in monolithZIF-8 was pursued for a range of possible industrial applications. The more general capability of monolithic MOFs to host nanoparticles was determined via the synthesis of Pd@monolithHKUST-1. Finally, the feasibility of the new materials, monolithUiO-66 and Pd@monolithHKUST-1, towards possible industrial gas storage applications (CH4, CO2 and H2) was recorded. Benchmark volumetric CH4 and H2 storage capacities were collected in each monolith, and the dependence of these capacities on e.g. pore-size distribution and doping level were also elucidated. Overall, the synthesis of new monolithMOF and NP@monolithMOF composites was comprehensively explored, developing novel and practical materials for prospective industrial applications. viii ix IV| Publications Publications directly contributing to the research in this thesis: Tuning Porosity in Macroscopic Monolithic Metal-Organic Frameworks for Exceptional Natural Gas Storage. B. M. Connolly, M. Aragones-Anglada, J. Gandara-Loe, N. A. Danaf, D. C. Lamb, J. P. Mehta, D. Vulpe, S. Wuttke, J. Silvestre-Albero, P. Z. Moghadam, A. E. H. Wheatley and D. Fairen-Jimenez, Nat. Comm., 2019, 10, 2345. From Synthesis to Applications: Metal–Organic Frameworks for an Environmentally Sustainable Future. B. M. Connolly, J. P. Mehta, P. Z. Moghadam, A. E. H. Wheatley and D. Fairen-Jimenez, Curr. Opin. Green Sustain. Chem., 2018, 12, 47. Sol-Gel Synthesis of Robust Metal-Organic Frameworks for Nanoparticle Encapsulation. J. P. Mehta, T. Tian, Z. Zeng, G. Divitini, B. M. Connolly, P. A. Midgley, J. C. Tan, D. Fairen- Jimenez and A. E. H. Wheatley, Adv. Funct. Mater., 2018, 28, 1705588. Composite Metal-Organic Framework Materials, Processes for Their Manufacture and Uses Thereof. Patent Number WO2018/197715 Other publications not directly contributing to the research in this thesis: Formation Mechanisms of ZnO Spherulites and Derivatives. B. M. Connolly, H. F. Greer and W. Zhou, Cryst. Growth Des., 2019, 19, 249. Controlling the Morphology of Au–Pd Heterodimer Nanoparticles by Surface Ligands. M. Kluenker, B. M. Connolly, D. M. Marolf, M. Nawaz Tahir, K. Korschelt, P. Simon, U. Köhler, S. Plana-Ruiz, B. Barton, M. Panthöfer, U. Kolb and W. Tremel, Inorg. Chem., 2018, 57, 13640. Preparation of Bifunctional Au-Pd/TiO2 Catalysts and Research on Methanol Liquid Phase One-Step Oxidation to Methyl Formate. D. Shi, J. Liu, R. Sun, S. Ji, S. M. Rogers, B. M. Connolly, N. Dimitratos, A. E. H. Wheatley, Catalysis Today, 2018, 316, 206. x xi V| Contents Preface I Declaration iii II Acknowledgements v III Abstract vii IV Publications ix V Contents xi VI Abbreviations xv VII List of Figures xix VIII List of Tables xliii IX Methods xlvii X References lv Chapter I| General Introduction 1.0 Context 3 2.0 Adsorbed Gas 6 3.0 Gas Storage in MOFs 8 4.0 Industrially Viable Adsorption 11 5.0 References 16 Chapter II| Monolithic Metal-Organic Frameworks 1.0 UiO-66 25 1.1 Aims and Objectives 28 1.2 Monolith Synthesis 29 1.3 Monolith Characterisation 34 1.3.1 Particle Size and Morphology 34 1.3.2 Elemental Composition and Structure 35 1.3.3 Mechanical and Thermal Stability 38 1.3.4 Porosity and Density 41 1.3.5 Fluorescent Lifetime Imaging Microscopy 48 1.4 Conclusions 50 2.0 UiO-66-NH2 53 xii 2.1 Aims and Objectives 55 2.2 Monolith Synthesis 56 2.3 Monolith Characterisation 59 2.3.1 Morphology and Crystal Structure 59 2.3.2 Thermal and Mechanical Stability 61 2.3.3 Porosity and Density 64 2.4 Conclusions 69 3.0 UiO-66-ndc 71 3.1 Aims and Objectives 73 3.2 Monolith Synthesis 74 3.2.1 Preliminary Synthesis 74 3.2.2 Modulator Study 77 3.2.3 Modulator Influence on Physical Properties 82 3.3 Monolith Characterisation 87 3.3.1 Morphology 87 3.3.2 Porosity and Density 88 3.3.3 Thermal and Mechanical Stability 91 3.4 Conclusions 93 4.0 NU-1000 94 4.1 Aims and Objectives 96 4.2 Monolith Synthesis 97 4.3 Monolith Characterisation 99 4.3.1 Particle Size and Morphology 99 4.3.2 Crystallinity and Porosity 100 4.3.3 Thermal and Mechanical Stability 103 4.4 Conclusions 105 5.0 References 107 Chapter III| Immobilisation of Nanoparticles in Monolithic Metal-Organic Frameworks 1.0 Monolithic SnO2@ZIF-8 117 1.1 Aims and Objectives 121 1.2 Preliminary Composite Characterisation 122 1.3 Tuning NP Surface Functionality 124 xiii 1.4 Photocatalytic Dye Degradation 130 1.5 Mechanism of Catalysis 134 1.6 Conclusions 136 2.0 Monolithic NP@ZIF-8 138 2.1 Aims and Objectives 143 2.2 Au@monolithZIF-8 144 2.2.1 Preliminary Material Synthesis 144 2.2.2 Composite Synthesis Optimisation 146 2.2.3 Characterisation 148 2.3 PdO/TiO2@monolithZIF-8 152 2.3.1 NP Synthesis 152 2.3.2 Monolithic Composite Synthesis and Characterisation 155 2.4 ((Au@PdO)/TiO2)@monolithZIF-8 158 2.4.1 NP Synthesis 158 2.4.2 Monolithic Composite Synthesis and Characterisation 162 2.5 Conclusions 165 3.0 Monolithic Pd@HKUST-1 168 3.1 Aims and Objectives 171 3.2 Monolithic Composite Synthesis 172 3.3 Characterisation 174 3.3.1 Composition 174 3.3.2 Morphology and Crystal Structure 175 3.3.3 XPS 179 3.3.4 Porosity and Density 182 3.3.5 Thermal and Mechanical Stability 185 3.4 Conclusions 188 4.0 References 189 Chapter IV| Practical Applications of Monolithic Materials 1.0 Gas Storage and Separation in monolithUiO-66 199 1.1 Aims and Objectives 201 1.2 CH4 Storage 202 1.2.1 Gravimetric CH4 Storage 202 xiv 1.2.2 CH4 Adsorption Simulations 205 1.2.3 Volumetric CH4 Storage 207 1.2.4 Working Capacity 210 1.3 CO2 Storage 212 1.4 Gas Adsorption Kinetics 216 1.5 Gas Separation 219 1.6 Conclusions 221 2.0 H2 Storage in monolithHKUST-1 and Pd@monolithHKUST-1 223 2.1 Aims and Objectives 224 2.2 Low Pressure H2 Storage 225 2.3 Mechanistic Studies 228 2.3.1 Temperature Dependence of H2 Storage 228 2.3.2 Relative Contribution of Pd and Cu to Total H2 Uptake 230 2.3.3 In situ H2 Adsorption XRD Studies 236 2.4 High Pressure H2 Storage 239 2.5 Conclusions 246 3.0 References 248 Chapter V| Outlook 1.0 Conclusions 255 2.0 Future Work 262 3.0 References 265 Appendix I Experimental iii II Supplementary Figures xiii III Supplementary References xxxv xv VI| Abbreviations 2,6-ndc 2,6-naphthalene-dicarboxylate AA Acetic acid AFM Atomic Force Microscopy ANG Adsorbed Natural Gas bdc 1,4-benzenedicarboxylate BE Binding Energy BET Brunauer–Emmett–Teller BJH Barrett-Joyner-Halenda bp Boiling Point btc Benzene-1,3,5-tricarboxylate CA Citric acid CNG Compressed Natural Gas Co(bdp) Co(1,4-benzenedipyrazolate) CTAB Cetrimonium bromide Dk Kinetic Diameter DMF N,N-Dimethylformamide E Young’s Modulus (GPa) FCV Fuel Cell Vehicle FIB-SEM Focussed Ion Beam - Scanning Electron Microscopy FLIM Fluorescent Lifetime Imaging Microscopy FWHM Full Width at Half Maximum GCMC Grand Canonical Monte Carlo Simulations H Hardness (GPa) HKUST-1 Cu3(benzene-1,3,5-tricarboxylate)2 HR-TEM High Resolution – Transmission Electron Microscopy IAST Ideal Adsorbed Solution Theory ICP-OES Inductively Coupled Plasma – Optical Emission Spectroscopy IPCC Intergovernmental Panel on Climate Change LNG Liquified Natural Gas m-dobdc 4,6-dioxido-1,3-benzenedicarboxylate MB Methylene Blue xvi MIL-121 Al(OH)(1,2,4,5-benzenetetracarboxylate).H2O MIL-53 Al(OH)(bdc) MOF Metal-Organic Framework MOF-177 Zn4O(4,4',4''-benzene-1,3,5-triyl-trisbenzoate) MOF-808 Zr6O4(OH)4(1,3,5-benzene tricarboxylate)2(HCOO)6 NaOl Sodium oleate NbOFFIVE NiNbOF5(pyrazine)2·2H2O ndc 1,4-naphthalene-dicarboxylate NG Natural Gas NLDTF Non-Local Density Functional Theory NP Nanoparticle NU-1000 Zr6O4(μ3-OH)4(–OH)4(–OH2)4(1,3,6,8-tetrakis (p-benzoate)pyrene)2 NU-110 Cu3(1,3,5-tris[((1,3-carboxylic acid-5-(4- (ethynyl)phenyl))ethynyl)phenyl]-benzoate)6(H2O)3]n NU-1103 Zr6O4(OH)4 (4,4',4'',4'''-((pyrene-1,3,6,8-tetrayltetrakis(benzene-4,1- diyl))tetrakis(ethyne-2,1-diyl))tetrabenzoate)6 NU-111 Cu3(1,3,5-tris[(1,3-carboxylic acid-5-(4-(ethynyl)phenyl))butadiynyl]- benzene)(H2O)3]n NU-125 Cu3(C36H21N9O15) OA Oleic acid PCN-14 Cu2(H2O)2(5,5'-(9,10-anthracenediyl)di-isophthalate) PNMA Poly (N‐methylolacrylamide) PSD Pore Size Distribution PVA Polyvinyl alcohol PVP Polyvinylpyrrolidone PXRD Powder X-ray Diffraction pyptp 4,4',4'',4'''-(pyrene-1,3,6,8-tetrayltetrakis(benzene-4,1-diyl)) tetrakis(ethyne-2,1-diyl)tetrabenzoate SBET Brunauer–Emmett–Teller Specific Surface Area SBU Secondary Building Unit SEM Scanning Electron Microscopy Sr2+-SAPO-34 Strontium Substituted Silicoaluminophosphate Zeolite xvii ST-2 Zn4O3(4,4',4''-s-triazine-1,3,5-triyltri-p-aminobenzoate)4(2,6- naphthalene-dicarboxylate) STEM-EDX Scanning Transmission Electron Microscopy – Energy Dispersive X-ray Spectroscopy STP Standard Temperature and Pressure tatb 4,4',4''-s-triazine-1,3,5-triyltri-p-aminobenzoate tbapy 1,3,6,8-tetrakis(p-benzoate)pyrene TEM Transmission Electron Microscopy TFA Trifluoracetic acid TGA Thermogravimetric Analysis U.S. DOE United States Department of Energy UiO-66 Zr6O4(OH)4(1,4-benzenedicarboxylate)6 UiO-66-Br Zr6O4(OH)4(2-bromo-1,4-benzenedicarboxylate)6 UiO-66-F Zr6O4(OH)4(2-fluoro-1,4-benzenedicarboxylate)6 UiO-66-NH2 Zr6O4(OH)4(2-amino-1,4-benzenedicarboxylate)6 UiO-66-NO2 Zr6O4(OH)4(2-nitro-1,4-benzenedicarboxylate)6 UiO-66-SO2 Zr6O4(OH)4(2-sulfo-1,4-benzenedicarboxylate)6 UiO-67 Zr6O4(OH)4(biphenyl-4,4'-dicarboxylate)6 UiO-68 Zr6O4(OH)4(triphenyl-4,4''-dicarboxylate)6 UMCM-152 Cu2(C28H14O8) UTSA-76a Cu2(5, 5'-(pyrimidine-2, 5-diyl)diisophthalic acid)(H2O)2.5DMF.3H2O Vtot Total Pore Volume (cm3 g–1) Wo Micropore Volume (cm3 g–1) wt Weight XPS X-ray Photoelectron Spectroscopy XRD X-ray Diffraction ZIF-8 Zn(2-methylimidazolate)2 Zr-FUM Zr6O4(OH)4(trans-1,2-ethylenedicarboxylate)6 ρb Bulk Density xviii xix VII| List of Figures Chapter I| General Introduction Figure 1 Global warming trends. Change point analysis of global temperature (combined land and ocean) trends (1850 – 2015) for different data sets collected by NOAA (The National Oceanic and Atmospheric Administration), GISTEMP (Goddard’s Global Surface Temperature Analysis collected by The National Aeronautics and Space Administration), Berkeley (Berkeley Earth Surface Temperature), HadCRUT (The Hadley Centre Climate Research Unit) and Cowtan&Way (Revised HadCRUT data). Reproduced from Reference 2 with permission of IOP publishing. 3 Figure 2 Dependence of adsorbate interaction potential on adsorbent pore diameter. a, Model slit-shaped pores of diameters 2.0 nm (sky blue), 1.3 nm (royal blue) and 0.7 nm (navy blue). b, Potential energy profile (𝜙/𝜙o) for a gas molecule (CO2, kinetic diameter, 𝜎 = 0.33 nm) a distance (z) from the centre of model graphite micropores (a). Data digitised from Reference 21. 7 Figure 3 ST-2 linkers and NG storage capacity. a, b, Chemical structures of tatb and 2,6-ndc, linkers in the MOF ST-2. c, Volumetric CH4 adsorption isotherm at 298 K for ST-2 (blue circles, digitised from Reference 34). Threshold value of maximum volumetric CNG capacity (250 bar, blue line) is indicated. 9 xx Chapter II| Monolithic Metal-Organic Frameworks Figure 1 UiO-66 SBUs and crystal structure. a, Zr6O4(OH)4 cluster. b, bdc. c, Structure of UiO-66 which demonstrates how the metal cluster (a) and linker (b) self-assemble to form the crystalline MOF. Colours correspond to C (purple), H (white), O (sky-blue) and Zr (navy blue). 25 Figure 2 Organic linker molecules used to synthesise zirconium MOFs. a, Linker for the synthesis of monofunctionalised UiO-66-X where X = e.g. H, F, Cl, Br, I, CH3, CF3, NO2, NH2, OH, OCH3, COOH, SO3H. b, Tritopic linker (btc) for the synthesis of MOF-808. c, d, Biphenyl-4,4´- dicarboxylic acid and terphenyl‐4,4´´‐dicarboxylic acid; organic linkers used to synthesise the isoreticular UiO-66 structures, UiO-67 and UiO-68 respectively. e, 1,3,6,8-tetrakis (p-benzoate)pyrene, the tetratopic organic linker in Zr-MOF NU-1000. 27 Figure 3 SEM images of UiO-66 primary particles. a, b, Low and high magnification, respectively, SEM images of UiO-66 primary particles (191.0 ± 36.6 nm) produced by room-temperature synthesis. 30 Figure 4 Literature UiO-66 NPs. a, Optical image of viscous UiO-66 gel comprised of 10 nm primary particles. Inversion of the reaction vessel demonstrates the gel’s viscosity. b, Optical micrograph image of a UiO-66 xerogel fragment (scale bar 100 µm) showing optical transparency. c, N2 adsorption isotherm (0 – 1 bar, 77 K) for UiO-66 gel obtained by drying at 200 ºC in a petri-dish (red diamonds, digitised from Reference 24). 31 Figure 5 Optical images of UiO-66 monoliths. a, UiO-66 washed in ethanol, dried at 200 °C (UiO-66_A). b, UiO-66 monolith with truncated cone shape, produced by washing primary particles in ethanol and drying at 30 °C (UiO-66_B). c, UiO-66 washed in DMF, dried at 30 °C (UiO-66_C). d, UiO-66 washed in DMF with extended centrifugation, dried at 30 °C 33 xxi (UiO-66_D). Coloured symbols (bottom-left) in each figure represent the key that will identify each material henceforth. Figure 6 Electron microscope images of UiO-66 gel and monolith. a, b, TEM images of UiO-66 gel. The irregularly shaped MOF primary NPs, ca. 10 nm, adopt a gelatinous network macrostructure. c, d, TEM images of fully dried, densified monolith (UiO-66_D). e, f, Low and high magnification, respectively, SEM images of UiO-66_D. 34 Figure 7 Elemental maps of monolithic UiO-66. Low magnification STEM electron image of UiO-66_D with area selected for EDX analysis indicated (white box) and corresponding EDX elemental maps showing distribution of Zr (pink), O (blue) and C (orange) throughout. 35 Figure 8 UiO-66 XRD patterns. Comparison of simulated XRD pattern for UiO- 66 generated from its ideal crystal structure (black), to PXRD patterns of UiO-66 monoliths: UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green), confirming successful synthesis of the MOF in each case. 37 Figure 9 Mechanical testing of UiO-66 monoliths. a, Load (mN) vs. Penetration into surface of the monolith (h, nm). The inset in a, 3D rendered AFM images showing the 3D topography of a resulting surface indent. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, as a function of Penetration depth (h, nm) into the monolith surface. Mean properties and corresponding errors (inset b and c) were obtained from measurements taken from 16 indents over penetration depths of 250 – 2000 nm. Measurements obtained in the sub-250 nm penetration range were excluded, to eliminate errors due to surface defects/tip artefacts. Data correspond to UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green). 39 xxii Figure 10 Thermal stability of monolithic UiO-66. TGA comparison (50 – 700 °C, under N2 atmosphere) of UiO-66_D (green) with a powdered UiO-66 sample (black) prepared according to a literature procedure (Reference 30). 40 Figure 11 UiO-66 N2 adsorption isotherms. a, N2 isotherms showing gravimetric N2 uptake at 77 K (0 – 1 bar) for monoliths UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles) as well as the theoretical isotherm simulated from the pristine crystal structure (white squares). b, Low pressure semi- logarithmic representation of each N2 isotherm (a) for P/Po < 0.001. c, Crystal structure of UiO-66 where elements correspond to: Zr (navy blue), O (sky blue), C (purple) and H (white). Tetrahedral pore faces are indicated in grey. 42 Figure 12 N2 desorption isotherms for UiO-66 monoliths. Adsorption (filled markers) and desorption (hollow markers) N2 isotherms collected at 77 K. a, UiO-66_A (blue triangles), b, UiO-66_B (red diamonds), c, UiO-66_C (purple squares) and d, UiO-66_D (green circles). 44 Figure 13 NLDFT PSDs of UiO-66 monoliths. Distribution of micro- and mesopore width across monolith samples: UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green), as obtained from Tarazona NLDFT model analysis of N2 isotherm data (Figure 11a). 45 Figure 14 BJH PSDs of UiO-66 monoliths. Distribution of mesopore diameter across monolith samples: UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green), as obtained from BJH model analysis of N2 isotherm data (Figure 12). 46 Figure 15 Hg porosimetry PSDs. PSDs as obtained by Hg porosimetry showing extensive meso- (2 – 50 nm diameter) and macro- (> 50 nm diameter) 47 xxiii porosity where UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green). Figure 16 FLIM studies of UiO-66 monoliths. a, b, FLIM images of UiO-66_A – D showing the aggregates that comprise each monolith. White dashed boxes (a, inset) indicate the areas selected for high magnification imaging (b). Colours correspond to excitation lifetime (see upper colour bar). c, 2D histogram phasor plots generated from FLIM images of each monolith. Colours correspond to the frequency of occurrence (see lower colour bar). An average of four images was used to generate each plot (Appendix, Supplementary Figures 5 – 8). 49 Figure 17 Functionalised precursors to UiO-66-X. a, b, Chemical structures of mono- and di- functionalised terephthalic acid derived precursors, respectively, which have previously been used to synthesise correspondingly functionalised UiO-66-X; X = H, F, Cl, Br, I, CH3, CF3, NO2, NH2, OH, OCH3, COOH, SO3H etc. 54 Figure 18 Optical and TEM images of UiO-66-NH2 gel. a, Optical image of UiO- 66-NH2 gel. b, c, Low and high magnification, respectively, TEM images UiO-66-NH2 gel. 56 Figure 19 Optical images of amine functionalised monoliths. Optical images of UiO-66-NH2 monoliths corresponding to a, UiO-66-NH2_A, b, UiO-66- NH2_B and c, UiO-66-NH2_C. 57 Figure 20 TEM images of monolithic UiO-66-NH2. a, b, Low and high magnification, respectively, TEM images of dried UiO-66-NH2_C. 59 Figure 21 XRD patterns of UiO-66-NH2 monoliths. Comparison of simulated XRD pattern for UiO-66-NH2 generated from its ideal crystal structure (black),1 to PXRD patterns experimentally obtained for the monolithic 60 xxiv MOFs: UiO-66-NH2_A (green), UiO-66-NH2_B (blue) and UiO-66- NH2_C (purple). Figure 22 Thermal stability of UiO-66-NH2. TGA traces comparing the thermal decomposition of UiO-66-NH2_C (purple), powdered UiO-66-NH2 (black, digitised from Reference 59) and UiO-66_D (green) over the temperature range 50 – 650 °C (under N2 atmosphere). 61 Figure 23 Mechanical testing of monolithUiO-66-NH2. Nanoindentation data for UiO-66-NH2_A (green), UiO-66-NH2_B (blue) and UiO-66-NH2_C (purple) showing a, Load (mN) vs. Penetration into the monolith surface (h, nm) for 16 indents. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, plotted as a function of Penetration Depth into the monolith surface (h, nm). Mean properties and corresponding errors (inset) were obtained from measurements taken across 16 indents over penetration depths of 250 – 2000 nm. Measurements obtained in the sub- 250 nm penetration range were excluded with the aim of eliminating errors due to surface defects/tip artefacts. 63 Figure 24 N2 adsorption isotherms for UiO-66-NH2. Linear N2 adsorption isotherms (collected at 77 K) for monoliths UiO-66-NH2_A (green diamonds), UiO-66-NH2_B (blue squares) and UiO-66-NH2_C (purple circles) compared to the theoretical adsorption isotherm for defect-free UiO-66-NH2 (red dots). 64 Figure 25 N2 adsorption-desorption isotherms for monolithUiO-66-NH2. N2 adsorption (filled markers) and desorption (hollow markers) isotherms (recorded at 77 K) for monolithic MOFs a, UiO-66-NH2_A (green diamonds), b, UiO-66-NH2_B (blue squares) and c, UiO-66-NH2_C (purple circles). 66 Figure 26 NLDFT PSDs of UiO-66-NH2 monoliths. Distribution of micro- and mesopore width across monolithic MOFs: UiO-66-NH2_A (green), UiO- 67 xxv 66-NH2_B (blue) and UiO-66-NH2_C (purple), as obtained from Tarazona NLDFT analysis of the N2 isotherm data (Figure 24). Figure 27 BJH PSDs of UiO-66-NH2 monoliths. Distribution of mesopore diameter across monolith samples: UiO-66-NH2_A (green), UiO-66- NH2_B (blue) and UiO-66-NH2_C (purple), as obtained from BJH model analysis of the N2 isotherm data (Figure 25). 68 Figure 28 Di-functionalised terephthalic acid derived structures. Chemical structures of a, ortho, b, meta and c, para di-substituted terephthalic acid molecules used to synthesise functionalised UiO-66 derivative structures. d, Chemical structure of ndcH2. 72 Figure 29 UiO-66-ndc powder. a, Optical and b, TEM images showing the macroscopic and microscopic morphology of UiO-66-ndc powder, respectively. The material was synthesised by equimolar exchange of the terephthalic acid linker applied to UiO-66 gel synthesis with its naphthalene functionalised analogue. 74 Figure 30 MOF crystallisation. a, Metal cluster [Zr6O4(OH)4]12+, an SBU in UiO- 66 formation. b, Coordination of twelve organic linkers around the zirconium cluster (a). c, Cross-linking of two zirconium hexaclusters by a bidentate organic linker. Black dashed line shows the symmetry incurred by crystal growth. Coloured spheres in the chemical structures represent C (purple), H (white), O (sky blue) and Zr (navy blue). 75 Figure 31 Optical images of UiO-66-ndc. Optical images of UiO-66-ndc_1 – 10 synthesised using variable volumes of modulator in the reaction mixture (see Table 7). 77 Figure 32 Optical images of UiO-66-ndc. Optical images comparing UiO-66-ndc monoliths synthesised with variable volumes of HCl modulator for a fixed volume of AA modulator (see Table 8). 81 xxvi Figure 33 TEM images of UiO-66-ndc. TEM images showing the size and morphology of primary MOF particles of UiO-66-ndc_1 – 15 synthesised using a range of different modulator volumes (Table 7 and Table 8). 83 Figure 34 PXRD patterns of UiO-66-ndc. PXRD patterns for UiO-66-ndc_1 – 15 synthesised with variable volumes of modulator (Table 7 and Table 8). 84 Figure 35 Adsorption isotherms of UiO-66-ndc. a, N2 adsorption isotherms (77 K) for UiO-66-ndc materials synthesised with variable volumes of modulator. Marker colours correspond to materials in the key (b). Red line, inset, shows the computationally simulated N2 adsorption isotherm for defect-free UiO-66-ndc, digitised from Reference 70. b, Table showing the colour key for each material as well as the N2 uptake capacity (cm3 (STP) g–1) of each at P/Po = 0.1. 86 Figure 36 SEM images of monolithUiO-66-ndc. SEM images of a monolith fragment with magnification increased from a – d, showing the smooth monolith surface to be comprised of densely packed MOF NPs. 87 Figure 37 N2 isotherms and PSDs of monolithUiO-66-ndc. a, Adsorption (filled markers) and desorption (hollow markers) N2 isotherm of monolithUiO-66- ndc (77 K) compared to the adsorption isotherm simulated from the ideal crystal structure (red dots). b, Distribution of micro- and mesopore diameter in monolithUiO-66-ndc as obtained from Tarazona NLDFT model analysis of N2 isotherm data (a). c, Distribution of mesopore diameter in monolithUiO-66-ndc as obtained from BJH model analysis of the N2 isotherm data (a). 88 Figure 38 PSD of monolithUiO-66-ndc. PSD of monolithUiO-66-ndc obtained by Hg porosimetry showing the variations in meso- (2 – 50 nm) and macro- (>50 nm diameter) porosity in the material. 90 xxvii Figure 39 TGA traces of UiO-66-ndc. Comparison of thermal decomposition by TGA for powdered UiO-66-ndc (black, digitised from Reference 70) to that of monolithUiO-66-ndc (purple) and unfunctionalised monolithUiO-66 (green) (50 – 650 ºC, under N2 atmosphere). 91 Figure 40 Mechanical testing of monolithUiO-66-ndc. Nanoindentation data for monolithUiO-66-ndc showing a, Load (mN) vs. Penetration into the monolith surface (h, nm) for 16 indents. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, as a function of Penetration depth into the monolith surface (h, nm). Mean properties and corresponding errors (inset) were obtained from measurements taken across 16 indents over penetration depths of 250 – 2000 nm, with the aim of eliminating surface defects/tip artefacts. 92 Figure 41 Crystal structure of NU-1000. a, Chemical structure of tbapyH4, precursor to the organic linker, tbapy, for the Zr-MOF NU-1000. b, The crystal structure of NU-1000 demonstrating how each 31 Å hexagonal pore (x) is surrounded by six, 12 Å triangular pores (y). Inset in b, key showing the colours which correspond to each element. 95 Figure 42 monolithNU-1000. Optical image of yellow monolithic MOF, NU-1000, which was cracked under the application of force. 98 Figure 43 Electron microscopy analysis of monolithNU-1000. a, b TEM images of monolithNU-1000 showing sub-5 nm primary MOF particles. c, d, Low and high magnification, respectively, SEM images of monolithNU-1000 showing the smooth surface of the monolith, which is resolved into a densely packed array of MOF NPs upon increased magnification. 99 Figure 44 XRD of NU-1000. Comparison of simulated XRD pattern of NU-1000 (black) generated from its crystal structure, to the experimentally obtained PXRD pattern of monolithNU-1000 (orange). 100 xxviii Figure 45 N2 isotherms for NU-1000. a, Linear N2 adsorption isotherms at 77 K for monolithNU-1000 (filled orange squares) and NU-1000 simulated from the crystal structure (hollow orange squares). b, Adsorption (filled orange squares) and desorption (hollow orange squares) N2 isotherms demonstrating hysteresis during gas uptake and release in the mesoporous-associated pressure range. 101 Figure 46 PSDs for monolithNU-1000. a, Distribution of micro- and mesopore width and b, distribution of mesopore diameter in monolithNU-1000 as obtained from Tarazona NLDFT and BJH model analysis of N2 isotherm data (Figure 45), respectively. 102 Figure 47 Thermal stability of monolithNU-1000. TGA analysis of monolithNU-1000 showing gradual thermal decomposition over the temperature range 50 – 750 °C (under N2 atmosphere). 103 Figure 48 Mechanical properties of monolithNU-1000. Nanoindentation data for monolithNU-1000 showing a, Load (mN) vs. Penetration into the monoliths surface (h, nm) across 16 indents. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, as a function of Penetration depth into the monolith surface (h, nm). Mean properties and corresponding errors (inset) were obtained from measurements taken across 16 indents over penetration depths of 250 – 2000 nm, ensuring elimination of surface defects/tip artefacts. 104 xxix Chapter III| Immobilisation of Nanoparticles in Monolithic Metal-Organic Frameworks Figure 1 Various NP morphologies. Graphic detailing the structural variation which exists amongst nanomaterials. a, 2D graphene nanotube, b, 2D nanorod, c, 3D nanoporous material, d – f, 1D NPs with spherical, cubic and tetrahedron morphologies respectively and g – i, functionalised, core- shell and heterodimeric, respectively, multicomponent 1D NPs. 117 Figure 2 Monolithic MOFs and NP@MOF composites. a, b, Optical images of monolithZIF-8 and SnO2@monolithZIF-8, respectively. c, d, TEM images of cubic SnO2 NPs used to dope monolithZIF-8. e, Representation of how a cubic NP (purple) may be immobilised in a MOF (not to scale): metal atoms/clusters (sky-blue spheres) and organic linkers (navy-blue rods). 120 Figure 3 FIB-SEM images of SnO2@monolithZIF-8 composites. FIB-SEM images of monolithZIF-8 doped with uncapped (a – c) and PVP-capped SnO2 NPs (d – f). 123 Figure 4 Organic capping agents. Chemical structures of organic capping agents applied to the surface-functionalised synthesis of SnO2 NPs. a, OA, b, NaOl, c, CTAB, d, PVP and e, CA. 125 Figure 5 TEM images of SnO2 NPs. a, b, TEM and c, d, HR-TEM images of SnO2-10 showing the presence of polydisperse nanorods. The d-spacing and corresponding hkl crystal face assignments are indicated (c). 128 Figure 6 Photocatalytic dye degradation. Bar chart comparing degradation of MB (%) in aqueous solution after 3 hours solar irradiation (sky blue) and in the absence of light (navy blue) in the presence of different monolithic composites, SnO2-1 – 12@ZIF-8. Error bars represent standard deviation calculated across results taken in triplicate. Line chart (overlaid, in purple) shows variation of SnO2 loading (wt%) for each of the SnO2-X@ZIF-8 samples as measured by ICP-OES. b, Table of p-values showing the 131 xxx statistical likelihood that the dye degradation results of SnO2-2 – 12@ZIF-8 differ significantly from that of reference sample SnO2- 1@ZIF-8. Figure 7 Organic species degradation. Photocatalytic mechanism of organic species degradation as catalysed by SnO2. 134 Figure 8 Dye degradation catalytic mechanism. a, Equation showing how non- fluorescing terephthalic acid is converted to fluorescing 2- hydroxyterephthalic acid in the presence of hydroxyl radicals. b, Multiwavelength fluorescence spectra showing fluorescent intensity of terephthalic acid probe solution varying with time (0 – 180 minutes) exposed to irradiation ( l = 315 nm) in the presence of SnO2@monolithZIF- 8 composite. 135 Figure 9 CO oxidation. Mechanism of CO oxidation catalysed by Au NPs. 138 Figure 10 Photocatalytic hydrogenolysis. Pictographic representation 2-propen-1- ol hydrogenolysis to propene as photocatalysed by Pd/TiO2. 140 Figure 11 TEM images of Au NPs. a, Low and b, high magnification TEM images of undried Au NPs prior to their immobilisation in monolithic ZIF-8. 144 Figure 12 Au NPs in monolithZIF-8. a, TEM and b, HR-TEM images of composite Au@monolithZIF-8 showing high contrast NPs immobilised inside the MOF. Inset in b, FFT diffraction pattern of an Au NP (white box) showing the diffraction spots generated by the [111] crystal plane. 145 Figure 13 Monolithic Au@ZIF-8. Optical image of a monolithic composite material comprised Au NPs, obtained by the modified synthetic procedure, immobilised in monolithZIF-8. 147 xxxi Figure 14 TEM analysis of Au@monolithZIF-8. a, Low and b, high magnification TEM images of Au NPs, prepared by a modified synthetic procedure after immobilisation in monolithic ZIF-8 (Figure 13). c, Histogram showing the particle size distribution of Au NPs in the composite material (a, b). 147 Figure 15 EDX elemental analysis of Au@monolithZIF-8. HAADF-STEM image (left) of Au@monolithZIF-8 showing Au NPs as brighter spots amongst the MOF. White boxes indicate areas selected for EDX elemental mapping (right: A & B). The position of the Au peak is indicated in A (white box, inset). 148 Figure 16 PXRD of doped and undoped monolithZIF-8. Overlaid PXRD patterns of monolithZIF-8 (black) and Au@monolithZIF-8 (red). Indicated is the position of the Au [111] peak referring to the Miller indices, hkl, classification for the crystal plane of Au which generated this reflection in the pattern. Each monolithic composite was gently ground to a powder for PXRD analysis. 149 Figure 17 Thermal stability of doped monolith. TGA traces of monolithZIF-8 (black) and Au@monolithZIF-8 (red) between 50 – 800 °C (under N2 atmosphere). Data for monolithZIF-8 was digitised from Reference 67. 150 Figure 18 Mechanical testing of Au@monolithZIF-8. Mechanical properties of monolithic ZIF-8 doped with Au NPs obtained by nanoindentation. a, Load (mN) vs. Penetration into surface of the monolith (h, nm) for 16 indents. b, c, Hardness (H, GPa) and Young’s Modulus (E, GPa), respectively, as a function of Penetration depth (h, nm) into the monolith surface. Mean properties and corresponding errors (inset in b and c) were obtained from measurements taken across 16 indents over penetration depths of 250 – 2000 nm. Measurements obtained in the sub-250 nm penetration range were excluded, to eliminate errors due to surface defects/tip artefacts. 151 xxxii Figure 19 TEM images of Pd/TiO2 NPs. a, b, Low and higher magnification, respectively, TEM images of bimetallic nanocomposite of Pd and TiO2. 153 Figure 20 PXRD analysis of PdO/TiO2 composite. PXRD pattern of PdO/TiO2 NPs. Indicated (inset) is the Miller indices classification, hkl, for the crystal plane which generated each reflection in the pattern for both TiO2 (red) and PdO (black). 153 Figure 21 Bulk morphology of NPs and monolithic composite. a, b, Optical images of PdO/TiO2 NPs and (PdO/TiO2)@monolithZIF-8, respectively. 155 Figure 22 PXRD of monolithic Pd/TiO2@ZIF-8. Overlaid PXRD patterns of PdO/TiO2@monolithZIF-8 (yellow) and monolithZIF-8 (black). Indicated is the Miller indices classification, hkl, for the crystal planes of the bimetallic dopant that generated each reflection; TiO2 (red) and PdO (black). Each monolithic composite was gently ground to a powder for PXRD analysis. 156 Figure 23 Elemental maps of (PdO/TiO2)@monolithZIF-8. a, HAADF-STEM image of PdO/TiO2 immobilised in monolithic ZIF-8. b, c, d, Elemental maps showing the distribution of Zn, Pd and Ti, respectively, in the composite. 157 Figure 24 TEM images of Au@Pd NPs. a, Low magnification TEM image of Au@Pd NPs and b, HR-TEM image showing the core@shell structure of a single particle (area selected for high magnification imaging indicated in a (white box, inset). 158 Figure 25 Elemental maps of Au@Pd NPs. a, STEM-HAADF electron image of a representative Au@Pd NP. b, c, Corresponding elemental maps showing the distribution of Pd and Au respectively. d, Overlaid elemental map showing the core@shell morphology of the NP. 159 xxxiii Figure 26 TEM analysis of (Au@PdO)/TiO2. a, b, c, TEM images of the trimetallic nanocomposite showing the core-shell Au@PdO NPs dispersed throughout the TiO2 NPs. White boxes in c indicate areas (i, ii, iii) selected for FFT anlaysis. i, ii, iii, artificial FFT diffraction patterns showing the distinct regions of the composite to be comprised of PdO, Au and TiO2, respectively. 160 Figure 27 PXRD analysis of (Au@PdO)/TiO2. PXRD pattern for the nanocomposite comprised of Au@PdO NPs dispersed amongst TiO2 NPs. Indicated is the Miller indices classification, hkl, for the crystal planes which generated each reflection in the pattern: TiO2 (green), PdO (black) and Au (red). 161 Figure 28 Bulk morphology of NPs and monolithic composite. Optical images of a, (Au@PdO)/TiO2 and b, ((Au@PdO)/TiO2)@monolithZIF-8 composite. 162 Figure 29 PXRD analysis of (Au@PdO)/TiO2)@monolithZIF-8. Overlaid PXRD patterns of monolithZIF-8 (black) and (Au@PdO)/TiO2)@monolithZIF-8 (blue). Indicated is the Miller indices classification, hkl, for the crystal planes which generated each reflection in the pattern for the doped trimetallic composite; TiO2 (green), Au (red) and PdO (black). Each monolithic composite was ground to a powder for PXRD. 163 Figure 30 Elemental maps of ((Au@PdO)/TiO2)@monolithZIF-8. a, STEM- HAADF electron image of monolithZIF-8 doped with (Au@PdO)/TiO2. b – e, Elemental maps showing the distribution of Zn, Au, Pd and Ti, respectively, throughout a. 164 Figure 31 HKUST-1. a, b, Cu–Cu paddlewheel dimer and HKUST-1’s crystal structure, respectively, where coloured spheres correspond to the elements Cu (blue), O (red), C (grey) and H (white). In the dimer, a, axial oxygen atoms originate from coordinated water molecules while bridging groups show the orientation of the carboxylate groups from the organic 169 xxxiv linker, 1,3,5-benzenetricarboxylate. c, Volumetric CH4 adsorption isotherm (0 – 70 bar, 298 K) for monolithHKUST-1 (red circles, absolute (filled). The U.S. DOE target for volumetric CH4 storage (263 cm3 (STP) cm–3) is indicted (dashed red line, inset). Data digitised from Reference 89. Figure 32 Pd@HKUST-1 electronic interaction. Representation of electron transfer from immobilised Pd NPs (green spheres) to a surrounding MOF, HKUST-1 (red). Reproduced from Reference 94 with permission from Springer Nature. 170 Figure 33 TEM analysis of Pd NPs. a, b, Low and high magnification, respectively, TEM images of Pd NPs. Inset in b, FFT diffraction pattern generated from a crystalline Pd NP (region used for FFT analysis indicated by dashed box (white)). Diffraction maxima originating from i) [111] and ii) [200] fringes are indicated by white circles in the FFT pattern. 172 Figure 34 Optical images of monoliths. a, monolithic HKUST-1, b, monolithic HKUST-1 with a targeted 5% loading of Pd NPs and c, monolithic HKUST-1 with a targeted 10% loading of Pd NPs. 173 Figure 35 Electron microscope images of Pd@monolithHKUST-1. a, TEM image of Pd@monolithHKUST-1. b, c, Low and high magnification, respectively, STEM images of the composite. Dashed white box (inset in c) shows area selected for electron, Pd and Cu mapping (d – f, respectively). 175 Figure 36 Surface morphology of monolithic MOFs. SEM images showing the surface morphology of monolithic HKUST-1, 5% Pd@HKUST-1 and 10% Pd@HKUST-1 at successively increased magnification (a – c). 176 Figure 37 X-ray tomography of monolithic NP@MOF composite. 3D reconstruction of monolithic Pd@HKUST-1 by X-ray tomography. a – c, 177 xxxv a fragment of monolithic MOF (blue) in different orientations and ai – ci, 3D distribution of Pd NPs throughout the MOF fragment (a – c). Figure 38 XRD patterns of MOF and NP@MOF materials. Simulated XRD patterns generated from the ideal crystal structures of HKUST-1 (red) and Pd NPs (black), compared to the experimental PXRD patterns of monolithHKUST-1 (green), 5% Pd@monolithHKUST-1 (blue) and 10% Pd@ monolithHKUST-1 (purple). Dashed red line (inset) indicates the position of the [111] and [200] reflections in Pd. 178 Figure 39 XPS patterns of Pd doped monoliths. XPS data for 5% Pd@monolithHKUST-1 (blue) and 10% Pd@monolithHKUST-1 (purple). Peaks corresponding to Pd 3d3/2 and 3d5/2 are indicated inset as is the relative ratio of the two. 179 Figure 40 XPS patterns of doped and pure monoliths. Selected XPS data for monolithHKUST-1 (green), 5% Pd@monolithHKUST-1 (blue) and 10% Pd@monolithHKUST-1 (purple) for BE between 930 – 965 eV. Peaks corresponding to Cu 2p1/2, 2p satellite and 2p3/2 are indicated inset. 181 Figure 41 N2 adsorption and PSD. a, Adsorption isotherms showing gravimetric N2 uptake at 77 K for monoliths HKUST-1 (green circles), 5% Pd@HKUST-1 (blue squares), 10% Pd@HKUST-1 (purple diamonds), and theoretical isotherm simulated from the pristine HKUST-1 crystal structure (hollow green circles). b, Semi-logarithmic representation of the isotherm (a) for P/Po < 0.002. c, Meso- and macro-PSD obtained by Hg porosimetry. Inset (red dashed box) indicates the recorded volumes of mesoporosity in the Pd-doped monoliths. 182 Figure 42 NLDFT PSD of doped and pure monoliths. Distribution of micro- and mesopore width across monolith samples: HKUST-1 (green); 5% Pd@HKUST-1 (blue) and 10% Pd@HKUST-1 (purple) as obtained from Tarazona NLDFT analysis of N2 isotherm data (Figure 41a). 184 xxxvi Figure 43 Thermal stability of monolithic materials. TGA traces of monolithic materials (50 – 600 °C, under N2 atmosphere) where data correspond to HKUST-1 (green), 5% Pd@HKUST-1 (blue) and 10% Pd@HKUST-1 (purple). 185 Figure 44 Mechanical testing of Pd@monolithHKUST-1. a, Load (mN) vs. Penetration into surface of the monolith (h, nm) for 16 indents. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, as a function of surface penetration depth (h, nm). Mean properties and corresponding errors (inset in b and c) were obtained from measurements taken from 16 indents over penetration depths of 500 – 2000 nm. Measurements obtained in the sub-500 nm penetration range were excluded, to eliminate errors due to surface defects/tip artefacts. Data correspond to 5% Pd@monolithHKUST-1 (blue squares) and 10% Pd@monolithHKUST-1 (purple diamonds). 187 xxxvii Chapter IV| Practical Applications of Monolithic Materials Figure 1 Potential industrial applications of MOFs. Schematic representation showing various stages of industrial fuel generation and usage that MOFs may be applied to. 200 Figure 2 Gravimetric CH4 adsorption isotherms. a, Comparison of gravimetric CH4 storage capacity of UiO-66 monoliths up to 100 bar (298 K). Colours correspond to UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles). b, Comparison of gravimetric CH4 uptake (0 – 100 bar) in UiO-66_D (green circles) to that of monolithHKUST-1 (white stars) and powdered UiO-66 reported by Zhou et al. (white circles, digitised from Reference 12). c, Comparison of gravimetric CH4 uptake (0 – 100 bar) in powdered UiO-66 (white circles) to that of computationally simulated defect-free microporous UiO-66 (white squares). All isotherms were experimentally obtained or computationally calculated at 298 K. 203 Figure 3 Snapshots and density distributions of CH4 adsorption. a, Snapshots and b, density distributions comparing CH4 (grey) adsorption in mixed micro-/mesoporous UiO-66 at 20 and 80 bar pressure. Colours within the UiO-66 crystal structure correspond to elements C (grey), H (white), Zr (blue) and O (red). 206 Figure 4 Volumetric CH4 storage. a, Comparison of volumetric CH4 storage capacity in UiO-66 monoliths (0 – 100 bar, 298 K). Colours correspond to UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles) and a pressurised storage tank (black crosses). b, Comparison of experimental isotherms for absolute volumetric CH4 uptake in monolithUiO-66_D (green circles) and monolithHKUST-1 (white stars, digitised from Reference 4) to computationally simulated purely microporous/defect-free UiO-66 (white 207 xxxviii squares) at 298 K; the U.S. DOE volumetric CH4 storage target of 263 cm3 (STP) cm–3 (65 bar, 298 K) is indicated by the dashed red line. Figure 5 Gravimetric and volumetric CO2 storage in UiO-66 monoliths. a, b, Comparison of gravimetric and volumetric, respectively, CO2 storage capacity in UiO-66 monoliths (0 – 40 bar, 298 K). Data correspond to UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles) and a standard pressurised storage tank (black crosses). 212 Figure 6 Snapshots and density distributions of CO2 adsorption. a, Snapshots and b, density distributions comparing CO2 (red\grey) adsorption in mixed micro-/mesoporous UiO-66 at 10 and 30 bar pressure. Colours within the UiO-66 crystal structure correspond to elements C (grey), H (white), Zr (blue) and O (red). 214 Figure 7 Optimising gas storage in mixed porosity MOFs. Comparison of volumetric storage capacity of CH4 at 65 bar (blue line) and 100 bar (black line) as well as CO2 at 40 bar (red line) in UiO-66 materials; UiO- 66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares), UiO-66_D (green circles) and single crystal UiO-66 (white squares, computationally simulated from the defect-free crystal structure). All data was collected at 298 K. 215 Figure 8 CH4 adsorption-desorption isotherms for UiO-66 monoliths. Volumetric (cm3 (STP) cm–3) CH4 adsorption (filled markers) and desorption (hollow markers) isotherms for a, UiO-66_A (blue triangles); b, UiO-66_B (red diamonds); c, UiO-66_C (purple squares) and d, UiO- 66_D (green circles) represented as absolute uptake between 0 – 100 bar (298 K). 216 Figure 9 Gas uptake kinetics for UiO-66 monoliths. Decay of pressure in UiO- 66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple 218 xxxix squares) and UiO-66_D (green circles) for a, CH4 at ca. 100 bar and b, CO2 at ca. 40 bar (298 K). Dashed black line (inset) indicates the pressure at which 95% uptake equilibrium was obtained. Figure 10 Equimolar binary isotherms for UiO-66 monoliths. Binary gravimetric isotherms of CO2 (filled marker) and CH4 (hollow marker) uptake in monoliths (0 – 40 bar, 298 K) where a, UiO-66_A (blue triangles); b, UiO-66_B (red diamonds); c, UiO-66_C (purple squares) and d, UiO- 66_D (green circles). 219 Figure 11 Gas mixture adsorption selectivity for UiO-66 monoliths. CO2:CH4 uptake selectivity for monoliths UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles) for a 50:50 gas mixture (0 – 40 bar, 298 K). 220 Figure 12 Low pressure H2 isotherms. a, b, Gravimetric (wt%) and volumetric (g L–1), respectively, linear H2 adsorption isotherms and c, semi-logarithmic adsorption-desorption isotherms (volumetric). Data correspond to monolithHKUST-1 (green circles), 5% Pd@monolithHKUST-1 (blue squares), 10% Pd@monolithHKUST-1 (purple diamonds) and a traditional pressurised storage tank (black line). All isotherms were recorded at 298 K up to a maximum pressure of 1 bar. 226 Figure 13 Comparison of powder and monolith H2 isotherms. Gravimetric (mmol g–1) linear adsorption isotherms for monolithHKUST-1 (green circles), 5% Pd@monolithHKUST-1 (blue squares), 10% Pd@monolithHKUST-1 (purple diamonds) and 24% Pd@powderHKUST-1 (red circles, digitised from Reference 35). All isotherms were recorded at 298 K up to a maximum pressure of 1 bar. 227 Figure 14 Variable temperature H2 isotherms. Semi-logarithmic, gravimetric (mmol g–1) H2 adsorption (filled marker) – desorption (hollow marker) isotherms for monolithHKUST-1 (green), 5% Pd@monolithHKUST-1 (blue) 228 xl and 10% Pd@monolithHKUST-1 (purple) at a, 273 K, b, 283 K and c, 298 K up to a maximum pressure of 1 bar. Figure 15 H/Pd adsorption-desorption isotherms. Comparison of H2 adsorption- desorption semi-logarithmic isotherms (0 – 1 bar) for a, b, 5% Pd@monolithHKUST-1 (blue squares) and 10% Pd@monolithHKUST-1 (purple diamonds), respectively (at 298 K), to c, d, 24% Pd@powderHKUST-1 (red circles, digitised from Reference 35) and pure Pd NPs (black squares, digitised from Reference 35), respectively (303 K), as a function of H atoms per Pd atom. e, Graphical representation of the distinct stages of H2 interaction with Pd under applied pressure. 231 Figure 16 H2/Cu3(btc)2 adsorption-desorption isotherms. Comparison of H2 adsorption-desorption semi-logarithmic isotherms for a, monolithHKUST-1 (green circles) and b, powderHKUST-1 (blue triangles, digitised from Reference 35) as a function of H2 molecules per unit of MOF (Cu3(btc)2). 233 Figure 17 Recalculated H/Pd adsorption-desorption isotherms. Comparison of H2 adsorption-desorption logarithmic isotherms in a, 5% Pd@monolithHKUST-1 (blue squares) and b, 10% Pd@monolithHKUST-1 (purple diamonds) at 298 K to c, pure Pd NPs (black squares, digitised from Reference 35) at 303 K, respectively, as a function of H/Pd. 234 Figure 18 XRD in situ H2 adsorption for 5% Pd@monolithHKUST-1. a, XRD patterns comparing peak positions under vacuum conditions to those under in situ exposure to H2 (0.01, 0.2 and 1.1 bar). Highlighted in pink are the Pd reflections; [111], [200], [220], [311] and [222]. The XRD pattern for defect-free HKUST-1 (simulated from its ideal crystal structure) is provided for comparison (black). b, Semi-logarithmic H2 adsorption-desorption isotherm (0 – 1 bar) where markers (red circles, inset) show the points in the isotherm at which XRD patterns were collected (a). c, Comparison of diffraction angles for each Pd reflection under exposure to different pressures of H2 (a). Data values are provided 237 xli as both absolute (2θ) and relative (Δd (%), compared to the reflections under vacuum) angles. Figure 19 XRD in situ H2 adsorption for 10% Pd@monolithHKUST-1. a, XRD patterns comparing peak positions under vacuum conditions to those under in situ exposure to H2 (0.01, 0.2 and 1.1 bar). Highlighted in blue are the Pd reflections; [111], [200], [220], [311] and [222]. The XRD pattern for defect-free HKUST-1 (simulated from its ideal crystal structure) is provided for comparison (black). b, Semi-logarithmic H2 adsorption-desorption isotherm (0 – 1 bar) where markers (red circles, inset) show the points in the isotherm at which XRD patterns were collected (a). c, Comparison of diffraction angles for each Pd reflection under exposure to different pressures of H2 (a). Data values are provided as both absolute (2θ) and relative (Δd (%), compared to the reflections under vacuum) angles. 238 Figure 20 High pressure H2 adsorption isotherms. a, Gravimetric (wt%) and b, volumetric (g L–1) H2 adsorption isotherms (0 – 100 bar, 298 K). Data correspond to monolithHKUST-1 (green circles), 5% Pd@monolithHKUST-1 (blue squares), 10% Pd@monolithHKUST-1 (purple diamonds) and standard pressurised storage tank (black line). 240 Figure 21 Benchmark H2 adsorption isotherms. a, Gravimetric (wt%) H2 uptake capacity in monolithHKUST-1 (green circles) compared to the benchmark powdered material Ni2(m-dobdc) (yellow triangles, digitised from Reference 34). b, Volumetric (g L–1) uptake capacity of the two materials, calculated from the experimental rb of monolithHKUST-1 and the theoretical crystal density of Ni2(m-dobdc). The expected loss in volumetric uptake capacity (25 – 50% loss) by pelletisation under pressure of powdered Ni2(m-dobdc) is indicated (yellow, inset in b). 243 Figure 22 High pressure H2 adsorption comparison. Comparison between high pressure gravimetric H2 storage capacity for monolithHKUST-1 collected at 245 xlii Sandia National Laboratory California (filled green circles, 298 K) and the U.S. DOE National Renewable Energy Laboratory, Colorado (hollow green circles, 303 K). The theoretical isotherms of the defect-free MOF obtained by GCMC simulation (black circles, 298 K) and Ni2(m-dobdc) (yellow triangles, 298 K, digitised from Reference 34) are also provided. xliii VIII| List of Tables Chapter II| Monolithic Metal-Organic Frameworks Table 1 Experimental conditions for monolithUiO-66 synthesis by washing/drying UiO-66 gel. 32 Table 2 Elemental composition (%) of UiO-66_A – D obtained by ICP-OES compared to the theoretical composition of the ideal crystal structure. 36 Table 3 Physical properties, SBET, Wo and Vtot (gravimetric and volumetric) of monoliths UiO-66_A – D (calculated from the experimentally obtained N2 adsorption isotherms, Figure 11a) compared to the theoretical values calculated for the ideal crystal structure. 43 Table 4 rb of UiO-66 monoliths calculated by Hg porosimetry. Absolute (g cm-3) and relative values (%) compared to the theoretical maximum. 47 Table 5 Experimental conditions for monolithUiO-66-NH2 synthesis by washing/drying UiO-66-NH2 gel. 58 Table 6 Physical properties, SBET, Wo and Vtot, of monoliths UiO-66-NH2_A – C (calculated from the experimentally obtained N2 adsorption isotherms, Figure 24) compared to the theoretical values calculated for the ideal crystal structure. 65 Table 7 Varying volumes (mL) of AA and HCl (37%) used to modulate the synthesis of monolithic UiO-66-ndc. 78 Table 8 Varying volumes (mL) of HCl (37%) used to modulate the synthesis of monolithic UiO-66-ndc for a fixed volume of AA (0.75 mL). 80 xliv Table 9 Physical properties, SBET, Wo and Vtot (calculated from the experimentally obtained N2 adsorption isotherms) and rb of monolithUiO- 66-ndc compared to theoretical values calculated for the ideal structure. 89 Chapter III| Immobilisation of Nanoparticles in Monolithic Metal-Organic Frameworks Table 1 Modifications made to the original SnO2 NP synthesis and variations in resulting NP size and monolithZIF-8 doping level (wt%) as a result. 126 Table 2 Elemental composition (wt%) of monolithic HKUST-1, 5% Pd@HKUST- 1 and 10% Pd@HKUST-1 compared to the theoretical compositions of dehydrated and hydrated HKUST-1. Data obtained by ICP-OES. 174 Table 3 Comparison of Pd and Cu BE, as obtained by XPS (Figure 39 and Figure 40), for Pd NPs, powdered Pd@HKUST-1 and 0%, 5%, 10% Pd doped (by wt%) monolithic HKUST-1. 180 Table 4 Comparison of 2p3/2 and satellite (sat) peak intensity and FWHM for Cu in the XPS data (Figure 40) of monolithic HKUST-1, 5% Pd@HKUST-1 and 10% Pd@HKUST-1. 181 Table 5 Comparison of SBET, Wo, Vtot and ρb of Pd-doped and pure monolithHKUST-1 to the corresponding values in the theoretical MOF’s crystal structure. 183 Chapter IV| Practical Applications of Monolithic Materials Table 1 SBET, ρb and gravimetric and volumetric CH4 uptake (65 and 100 bar, 298K) for monolithUiO-66_A – D compared to simulated defect-free UiO-66. 204 xlv Table 2 Volumetric CH4 adsorption capacities (5, 65 and 100 bar, 298 K) and corresponding working capacities for monolithUiO-66_D compared to the theoretical, defect-free MOF as well as benchmark monolithHKUST-1. 211 Table 3 Comparison of total volumetric H2 storage capacity (g L–1) in monolithHKUST-1, 5% Pd@monolithHKUST-1 and 10% Pd@monolithHKUST-1 to that of a pressurised storage tank at 5, 40 and 100 bar H2 pressure (298 K) as well as that of the corresponding working capacities (5 – 100 bar). 241 xlvi xlvii IX| Methods AFM AFM was used to study indents on monolithic MOF surfaces after nanoindentation. A Dimension ICON pro with RTESPA tip in tapping mode was used. Data were analysed using Nanoscope Analysis_1.9 software. Computational Simulations Computational simulations were performed by M. Aragones-Anglada under the supervision of P. Z. Moghadam. GCMC simulations were employed to obtain N2, CO2, CH4, and H2 adsorption isotherms. Simulations were based on a model utilising Lennard-Jones interactions taken from the Universal Force Field (UFF).1 The interactions2 were described by the TraPPE force field.3 An atomistic representation was used for the MOF from its crystal structure. The simulation cells consisted of 8 (2 × 2 × 2) unit cells for microporous UiO-66, and a single unit cell for all micro-/mesoporous structures, with a Lennard-Jones cut-off radius of 12.8 Å and no tail corrections. For CO2, long-range electrostatic interactions were handled using the Ewald summation technique. Periodic boundary conditions were applied in all three dimensions. For each pressure point, GCMC simulations consisted of 50,000 Monte Carlo cycles to guarantee equilibration, followed by 50,000 production cycles to calculate the ensemble averages. All simulations included insertion/deletion, translation and rotation (for N2 and CO2) with equal probabilities. GCMC simulations were run on 5 models; one corresponding to defect-free microporous UiO-66 and four corresponding to the micro-/mesoporous materials (monolithUiO- 66_A – D). To obtain micro-/mesoporous materials, the microporous UiO-66 supercell (of 8 unit cells) was modified by artificially creating a gap between two purely microporous UiO-66 layers, increasing the simulation cell length. The appropriate mesoporous gap lengths to match the experimental N2 isotherms were 2.75 nm (for monolithUiO-66_A and _B), 2.50 nm (for monolithUiO-66_C) and 2.30 nm (for monolithUiO-66_D). The Python package pyIAST4 was employed to estimate the CO2:CH4 IAST selectivities for monolithUiO-66_A – D at 298 K and at different pressures from experimental pure-component CO2 and CH4 adsorption isotherms by using the BET and the quadratic model fittings for CO2 and CH4 respectively. xlviii Data Digitisation Literature data was extracted from the source by digitisation using OriginLab 2017_9.4 and PlotDigitizer_2.6.8 software. Electron Microscopy SEM images were collected using a TESCAN MIRA3 FEG-SEM and processed with ImageJ_1.51 software. Samples were stuck to an SEM stub using conductive carbon tape and sputter coated with Pt (typically 10 nm thickness) prior to analysis. TEM and STEM-EDX were performed using an FEI Philips Tecnai 20 and FEI Osiris STEM operated in scanning mode (accelerating voltage of 200 keV) with Bruker Super-X detectors operating Hyperspy for STEM-EDX analysis. Primary MOF particles were prepared for analysis by diluting undried MOF gel in acetone and briefly sonicating before coating onto a lacey copper grid. Dried monoliths were prepared for analysis by crushing with a spatula and pressing a lacey copper grid into the resulting powder. NPs were prepared for analysis by dilution in acetone and extensively sonicating prior to drop coating onto a lacey copper grid. Images were processed with ImageJ_1.51 software. FIB-SEM was performed using an FEI Helios Nanolab SEM/FIB Dual Beam at an accelerating voltage of 5 keV. Samples were prepared for analysis by placing on a sticky, conductive carbon tab stuck to a brass specimen stub. FLIM FLIM measurements were collected by N. A. Danaf under the supervision of D. C. Lamb and S. Wuttke. Analysis was performed using a house-built laser scanning confocal microscope equipped with pulsed interleaved excitation and time-correlated single photon counting detection.5 A pulsed laser diode (405 nm) was used for excitation. For the measurements, monoliths were gently ground with a spatula and the powder was suspended in water and vortexed (~ 3 min). The suspension (~ 30 µL) was added to an 8-well LabTek I slide (VWR) and the fragments allowed to sediment. The surface was imaged using a 60x, 1.27 numerical aperture water-immersion objective (Plan Apo IR × 60 WI, Nikon). 100 µm × 100 µm scans were performed at a resolution of 500 pixels × 500 pixels (200 nm pixel size). Magnified regions (30 µm × 30 µm) were collected with a resolution of 500 pixels × 500 pixels (60 nm pixel size). The count rate was kept between 50 and 500 kHz by adjusting the laser power xlix between 2 – 10 µW, as measured at the sample using a slide power meter (S170C-Thorlabs). Image acquisition times of 100 – 200 s ensured the detection of 200 – 1000 photons per pixel, after which the phasor analysis was applied.6 To improve the FLIM analysis, data were spatially smoothed (3 pixel × 3 pixel sliding window). All analysis was performed using PAM software.7 Gas Adsorption Testing Preliminary N2 adsorption isotherms were collected at 77 K using a Micromeritics TriStar II. Prior to analysis, samples were degassed in a vacuum oven at 110 °C for 8 hours. Further degassing was performed after the samples were weighed into analysis tubes using a Micromeritics VacPrep degasser at 110 °C for 8 hours. Sample tubes were fitted with isothermal jackets and filler rods prior to analysis. Higher resolution N2 isotherms (adsorption and desorption) were collected at 77 K using a Micromeritics 3Flex. Prior to analysis, samples were degassed in a vacuum oven at 110 °C for 8 hours. Further degassing was performed after the samples were weighed into analysis tubes using a Micromeritics VacPrep degasser at 110 °C for 8 hours. In situ degas (110 °C, 8 hours) was further performed after sample loading into the instrument, ensuring total evacuation of MOF pores. Sample tubes were fitted with isothermal jackets and filler rods prior to analysis. The experimentally collected N2 isotherms were used to obtain BJH and NLDFT PSDs through the Micromeritics software. SBET was calculated manually using Excel software according to the Rouquerol consistency criteria.8 Wo and Vtot porosity were manually calculated from the N2 isotherms at P/Po = 0.1 and 0.99 respectively. High pressure uptake of both CH4 and CO2 was collected by collaborators at the University of Alicante using a homemade fully automated manometric equipment now commercialized by Quantachrome Instruments (i-sorbHP). Prior to analysis, samples were degassed overnight (120 °C under vacuum) and then again in situ. Data was experimentally obtained as gravimetric excess uptake and manually converted to absolute uptake using a literature procedure which utilises the materials pore volume (as obtained from the N2 adsorption isotherms) as well as the bulk gas density (extracted from the NIST data base of thermophysical properties).9,10 Gravimetric data for CO2 and CH4 storage in a standard pressurised storage tank was collected from the NIST database of thermophysical properties.10 Gravimetric (g g–1) isotherms were converted to volumetric (cm3 (STP) cm–3) using each materials experimental rb (g cm–3) l calculated by Hg porosimetry in addition to the volume of space occupied by the stored gas at STP. H2 adsorption isotherms and in situ XRD analysis were collected by collaborators at Sandia National Laboratories. Data was experimentally obtained as gravimetric excess uptake and manually converted to absolute uptake using a literature procedure which utilises the materials pore volume (as obtained from the N2 adsorption isotherms) as well as the bulk gas density (extracted from the NIST data base of thermophysical properties).9,10 Gravimetric data for H2 storage in a standard pressurised storage tank was collected from the NIST database of thermophysical properties.10 Gravimetric (g g–1 (x 100), wt%) isotherms were converted to volumetric (g L–1) using the rb (g cm–3), experimentally calculated by Hg porosimetry. H/Pd and H2/Cu3(btc) calculations were performed according to work by Kitagawa et al.11 Gravimetric molar H2 uptake (mmol g–1) was converted to atomic H and molecular H2 uptake (mmol) for Pd and Cu, respectively, using the mass of material tested (g) as well as the Pd and Cu content of the composite materials (obtained by ICP-OES). Under the assumption that H2 uptake capacity of monolithHKUST-1 was significant, its H2 adsorption isotherm was interpolated to obtain its uptake at each pressure point for the composite materials isotherm. The monolithHKUST-1 isotherm was then subtracted from the composite materials isotherm after scaling its uptake (mmol g–1) relative to its contribution to the composite’s total composition (as obtained by ICP-OES). Hg Porosimetry Hg porosimetry was run as a service at the University of Alicante. Monolith density at atmospheric pressure was calculated using a PoreMaster-60 GT porosimeter from Quantachrome Instruments. Prior to analysis, samples were degassed overnight under vacuum at 110 °C and again in situ prior to testing. ICP-OES Elemental analysis was run as a service at the Department of Chemistry, Cambridge University. ICP-OES and C, H, and N elemental analyses were performed using a Thermo Scientific iCAP 7400 ICP-OES analyser and an Exeter analytical CE 440 elemental analyser (975 °C) respectively. li Nanoindentation Monoliths were loaded into an epoxy resin and polished flat to a 0.25 μm finish via gradual rotary polishing using sequentially less coarse sand paper. For all monoliths the sand paper was lubricated with water except for monolithHKUST-1 where an oil-based lubricant was used to minimised water exposure. Nanoindentation was subsequently performed using an MTS Nanoindenter XP in an isolation cabinet. Continuous stiffness measurement mode was used to obtain Young’s modulus, Hardness and Load as a function of surface penetration depth up to 2000 nm using a Berkovich diamond tip. Indents were repeated in a 4 x 4 grid (16 indents) per sample. Optical Images Images were recorded on a Samsung Galaxy A5 equip with a 13.1-megapixel Sony Exmor RS IMX135 CMOS sensor camera with a pixel size of 1.1 μm as well as an iPhone8 equip with a 12-megapixel Wide (f/1.8 aperture) camera. Scale bars were obtained by placing a ruler next to the material during imaging. Photocatalytic Testing of SnO2@monolithZIF-8 The testing procedure was replicated from the literature report by Mehta et al.12 SnO2@monolithZIF-8 (300 mg),† was added to a solution of MB in milli-Q water (1.55 x 10–5 M, 25 mL) and irradiated for 3 hours on ice with gentle stirring. Solar radiation was simulated using a 150 W xenon lamp, solar simulator model LSO106, 1 sun illumination, 100 mW cm-2). For control measurements, samples were wrapped in tin-foil to exclude all light. A single wavelength UV-Vis spectrophotometer (BOECO S-22) was used to measure relative absorption of dye solution at 665 nm before and after irradiation. Percentage dye degradation was calculated using the equation: Dye degradation (%) = 31 − concentration (78)concentration (98): × 100 †Due to time constraints, catalytic tests were performed using 300 mg of SnO2@ZIF-8 composite instead of the 400 mg used in the preliminary tests. As such, the degradation results for SnO2-1@ZIF-8 were decreased by 25.0% allowing better comparison to the results obtained using 25.0% less catalyst (SnO2-2 – 12). lii For the photocatalytic mechanism testing, SnO2-2@ZIF-8 (0.25 g) was added to a solution of terephthalic acid (8.31 mg, 0.050 mmol) and sodium hydroxide (8.00 mg, 0.200 mmol) in milli- Q water (30 mL). The resulting mixture was irradiated under simulated solar radiation (150 W xenon lamp, solar simulator model LSO106, 1 sun illumination, 100 mW cm–2) for 3 hours on ice with stirring. Aliquots (3 mL) were collected at 0 min, 30 min, 60 min, 120 min and 180 min. A fluorometer (λexcitation = 315 nm) was used to measure fluorescence intensity of each sample (λmax = 425 nm). Statistical analysis of the experimental photocatalysis data was performed by calculation of the p-values. This is a measure of the statistical likelihood that the photocatalysis results are significantly different to those of the reference sample – SnO2-1. It relies on the assumptions that there is a homogeneity of variance amongst the data sets being compared, that these populations display a normal distribution and that they were sampled independently. The p- value was calculated by the following equations: MSE = S@AB + S@BB2 S@AE@B = F2MSEN df = (N@A − 1) + (N@B − 1) t − value = M@A − M@BS@AE@B Where MSE is the mean standard error, Sn2 is the variance of the repeat tests, N is the number of values in each sample set, n1 is the sample being studied (SnO2-1 – SnO2-12), n2 is the reference sample (SnO2-1), df is the degrees of freedom and Mn is the mean value of photocatalytic dye degradation calculated over each of the data sets (n1 and n2). From the t- value and df, the p-value was computed using a two tailed test. A standard tolerance of 0.05 was used whereby p-values < 0.05 represent samples whose values significantly differ from those of the reference. liii PSD BJH and NLDFT PSDs were calculated by computational analysis of the experimental N2 adsorption/desorption isotherms using automated computational programmes within the Micromeritics software. BJH PSD applies to the meso- and macropore range. For each desorption step in the N2 isotherm, average pore volume is calculated by application of the Kelvin model and t-plot equation to the pore emptying which takes place during desorption. NLDFT applies to the micro- and mesopore range. The critical point for capillary condensation (i.e. minimum pore size at which condensation can occur) is calculated by fitting different integrated adsorption equations to the experimental isotherm to minimise deviation. TGA TGA was collected using a Mettler Toledo/SDTA851 thermobalance with an alumina crucible under N2 atmosphere. XPS XPS was run as a service at the Cavendish Laboratory, Cambridge University. An ESCALAB 250 Xi running Advantage Software was used. XRD PXRD patterns were collected using a PANalytical Empyrean diffractometer with an X’celerator detector (Cu-Kα1 source, λ = 1.5406 Å). Powdered materials were placed onto a zero-background silicon wafer for analysis. Monolith powders were prepared for PXRD analysis by gently crushing with a pestle and mortar before being placed to the silicon wafer. Data were analysed using Origin 2017_9.4 software. Theoretical XRD patters for ideal MOF and NP structures were simulated from literature crystal structures of each material using CrystalMaker 10.3.3 software. X-ray Tomography X-ray tomography was performed by G. Divitini at The Department of Materials Science and Metallurgy, Cambridge University. A ZEISS Xradia 510 Versa was used with a 70 kV X-ray source. liv lv X| References 1. Rappé, A. K., Casewit, C. J., Colwell, K. S., Goddard, W. A. & Skiff, W. M. UFF, A Full Periodic Table Force Field for Molecular Mechanics and Molecular Dynamics Simulations. J. Am. Chem. Soc. 114, 10024–10035 (1992). 2. Martin, M. G. & Siepmann, J. I. Transferable Potentials for Phase Equilibria. J. Phys. Chem. B 102, 2569–2577 (1998). 3. Potoff, J. J. & Siepmann, J. I. Vapor–Liquid Equilibria of Mixtures Containing Alkanes, Carbon Dioxide, and Nitrogen. AIChE J. 47, 1676–1682 (2001). 4. Simon, C. M., Smit, B. & Haranczyk, M. PyIAST: Ideal Adsorbed Solution Theory (IAST) Python Package. Comput. Phys. Commun. 200, 364–380 (2016). 5. Hendrix, J., Schrimpf, W., Höller, M. & Lamb, D. C. Pulsed Interleaved Excitation Fluctuation Imaging. Biophys. J. 105, 848–861 (2013). 6. Digman, M. A., Caiolfa, V. R., Zamai, M. & Gratton, E. The Phasor Approach to Fluorescence Lifetime Imaging Analysis. Biophys. J. 94, 14–16 (2008). 7. Schrimpf, W., Barth, A., Hendrix, J. & Lamb, D. C. PAM: A Framework for Integrated Analysis of Imaging, Single-Molecule, and Ensemble Fluorescence Data. Biophys. J. 114, 1518–1528 (2018). 8. Gómez-Gualdrón, D. A., Moghadam, P. Z., Hupp, J. T., Farha, O. K. & Snurr, R. Q. Application of Consistency Criteria to Calculate BET Areas of Micro- and Mesoporous Metal-Organic Frameworks. J. Am. Chem. Soc. 138, 215–224 (2016). 9. Tian, T. et al. A Sol–Gel Monolithic Metal–Organic Framework with Enhanced Methane Uptake. Nat. Mater. 17, 174–179 (2018). 10. Lemmon, E. W., McLinden, M. O. & Friend, D. G. ‘Thermophysical Properties of Fluid Systems’. in NIST Chemistry WebBook, NIST Standard Reference Database Number 69 (National Institute of Standards and Technology, Gaithersburg MD, 20899, 2018). 11. Li, G. et al. Hydrogen Storage in Pd Nanocrystals Covered with a Metal–Organic Framework. Nat. Mater. 13, 802–806 (2014). 12. Mehta, J. P. et al. Sol-Gel Synthesis of Robust Metal-Organic Frameworks for Nanoparticle Encapsulation. Adv. Funct. Mater. 28, 1705588 (2018). lvi Chapter I General Introduction 2 3 1.0| Context Despite an explosive report from the IPCC warning of the environmental consequences of increased global warming, current temperatures continue to approach the danger zone of 1.5 °C above preindustrial levels. Correspondingly, widespread environmental damage has already been seen, including vast loss of arctic sea ice, increased frequency of heatwaves and forest fires as well as irreversible animal and plant mass extinction.1 Figure 1 shows change point analysis of global surface temperature for a range of different data sets. This demonstrates both a steady global warming trend since the industrial revolution as well as shockingly rapid warming since the 1970s.2 This environmental devastation is widely accepted to correlate with exponentially increased energy consumption since the industrial revolution.3 Atmosphere- warming greenhouse gasses and other pollutants have traditionally been released by conventional energy generation practices e.g. the burning of non-renewable coal and petrol. With the IPCC report warning that critical changes to energy supply and demand must take place immediately, a complete overhaul of energy generation and storage practices is essential if future energy demands are to be sustainably realised. Figure 1| Global warming trends. Change point analysis of global temperature (combined land and ocean) trends (1850 – 2015) for different data sets collected by NOAA (The National Oceanic and Atmospheric Administration), GISTEMP (Goddard’s Global Surface Temperature Analysis collected by The National Aeronautics and Space Administration), Berkeley (Berkeley Earth Surface Temperature), HadCRUT (The Hadley Centre Climate Research Unit) and Cowtan&Way (Revised HadCRUT data). Reproduced from Reference 2 with permission of IOP publishing. a. b. 4 One of the best solutions to mitigate global warming is the use of cleaner gaseous fuels. With abundant natural reserves and reduced greenhouse gas emissions relative to solid coal and liquid petroleum, global usage of NG (primarily composed of CH4) is predicted to increase significantly by 2040.4 Among the potential uses of gaseous fuels, a much sought-after application is in automobile engines. Considering only the European market, the addition of ca. 3 million cars each year increases annual energy consumption by 4%.5 Although NG is itself a potent greenhouse gas, NG combustion engines are estimated to emit 70, 87 and 20% less CO, NOx and CO2, respectively, than gasoline combustion engines. Hence, the careful use of NG as an automobile fuel (with precautions taken to avoid loss of containment) could incur substantial reductions in global emissions, as well as much needed improvements in local air quality in highly populated cities.5 An even more promising energy source is clean, gaseous H2. With a range of environmental production methods available6 as well as an outstanding gravimetric energy density of 120 MJ Kg–1, compared to 47.2 MJ Kg–1 for gasoline, widespread utilisation of this potentially zero- carbon emission gas would revolutionise the energy industry and lessen current oil dependence in what is termed ‘The Future Hydrogen Economy’.7 FCVs, which environmentally convert H2 to electricity, offer further benefits over not just traditional internal combustion engines but also popular battery powered vehicles, including significant enhancements in driving range, with distances exceeding 500 km proposed.8 Yet, the limitations of current gas storage technology remain a significant barrier to widespread gas-fuel usage; the technology needed for this obviously differs drastically from that which underpins the storage of traditional solid and liquid based fossil fuels. From a practical perspective, gas-fuel must be densely stored for feasible on-board usage. Conventionally, CNG is stored at room temperature under high pressures (ca. 180 – 250 bar) in robust, thick-walled steel tanks. The low storage capacity of these tanks, coupled with the high mass of the storage vessels (which contribute at least 90% to the total system mass)9 prevent translation of this technology to NG powered vehicles; both driving range and vehicle payload capacity are significantly limited. Additionally, the onboard storage of highly pressurised flammable-gas tanks presents obvious safety concerns. Low pressure storage of LNG may provide fuel with up to 2.4 times higher energy density than CNG (250 bar). Still, the liquification of high bp NG demands energy intensive cooling to less than 110 K as well as bulky/heavy insulative 5 chambers to maintain cryogenic temperatures, mostly limiting this process to applications in transoceanic shipping.10 To stimulate the development of new technologies that would increase NG viability by reducing the practical, safety and economic barriers to its distribution and usage, the U.S. DOE set an ambitious volumetric NG storage target of 263 cm3 (STP) cm–3 at 65 bar and 298 K, identical to the volumetric capacity of CNG at 250 bar and 298 K.11 Moderate pressures of ca. 65 bar are highly practical, being cost-effectively achieved using inexpensive compressors and requiring lighter storage tanks, though pressures as high as 100 bar can likewise be maintained by light and inexpensive storage tanks.5 Low pressure NG storage offers not only safety benefits but also improvements in large-scale industrial economics. Higher pressures required for liquification can only be achieved using complex and expensive multi-stage compressors.12 Furthermore, the U.S. DOE has collaborated with both the U.S. Council for Automotive Research and a number of energy companies to establish additional targets for on-board H2 storage in light-duty vehicles. The ultimate aim is to achieve a technology which can compete with incumbent vehicle technology by achieving a comparable 300 – 500 mile driving range while simultaneously meeting cost and safety requirements. These ultimate H2 storage targets are highly ambitious, standing at 0.065 g g–1 (6.5 wt%) and 50 g L–1 which crucially cannot be achieved by compression.13 State-of-the-art commercial FCVs produced by companies such as Toyota rely on ambient temperature H2 storage at a mammoth 700 bar,14 yet even under such immense compression, the maximum H2 storage density stands at only 40 g L-1. Hence, these H2 storage targets appear to be far from reach. Moreover, in order to meet current FCV operating specifications the fuel delivery pressure and temperature must range between 12 – 5 bar and 233 – 358 K, respectively, though the temperature/pressure conditions for the storage system may fall significantly outside this range. The U.S. DOE guidelines for on-board H2 storage identify 160 and 430 bar as ‘low’ and ‘moderate’ storage pressures, respectively. Obviously, H2 storage systems which rely on excessively high pressures and/or low temperatures are costly, impractical and energy consuming to achieve/maintain. Clearly, neither of these gas-fuel targets (CH4 or H2) can be achieved by traditional high-pressure compression or low temperature liquification.7 6 2.0| Adsorbed Gas A range of alternative storage techniques for gas-fuels have been proposed with e.g. chemical storage of H2 as metal-hydrides being of particular note.15 Though promising results have been achieved via this process (e.g. inexpensive MgH2 has a 7.6 wt% H2 capacity)16 an inherent pitfall of this technology is the highly exothermic nature of metal hydride formation, and subsequently endothermic H2 fuel regeneration. Heat transfer to/from the metal hydride reactor bed as well as slow dehydrogenation kinetics are significant practical challenges.17 Alternatively, ANG, in which the adsorbate is stored in the internal porosity of an adsorbent material, is regarded as a highly efficient way of densely storing gas.11 Weak physisorption- based storage offers numerous practical advantages over energy intensive metal hydride formation, including fast adsorption/desorption kinetics and high cyclability.18 The strength of the gas-adsorbent interaction in porous materials is dependent on both the size and chemical properties (e.g. polarity) of the adsorbate as well as pore shape/diameter (d) and surface chemistry of the adsorbent.19,20 The Steele function models the potential energy profile of an adsorbing gas molecule in a single pore whereby the minimum interaction potential (𝜙) of the gaseous adsorbate (g) in the model slit-shaped pore (s) is approximated using [Equation 1]: 𝝓𝒐 = 𝟔𝟓 𝝅𝝆𝒔𝜺𝒔𝒈𝝈𝒔𝒈𝟐 ∆ [Equation 1] In this function, the minimum interaction potential (𝜙X) is dependent on molecular lattice density (𝜌Z), potential well depth (𝜀Z] = (𝜀ZZ𝜀]])_^ ) and effective diameter of adsorbent/adsorbate atoms (𝜎Z] = (𝜎ZZ+ 𝜎]])/2)21, the latter two variables being derived from the Lennard-Jones parameters for a surface atom (𝜀ZZ, 𝜎ZZ) and a gas molecule (𝜀]], 𝜎]]) respectively. The fact that 𝜙Xvaries as a function of lattice layer spacing (∆) is of particular note in the design of porous materials for gas adsorption applications. The origin of 𝜀Z] in the Lennard-Jones potential – traditionally used to describe the interaction between porous adsorbents and adsorbates – results in a potential energy profile which varies 7 as a function of adsorbate-adsorbent separation. This classic Lennard-Jones profile is derived from the short-range Pauli repulsion as well as longer range attractive forces such as Van der Waals. On top of these, electrostatic interactions play a crucial role, especially when the adsorbent shows functional groups on the surface and the adsorbates present dipolar or quadrupolar moment. Molecular adsorption occurs at the potential energy minimum of the profile well.22 For an infinitely wide pore, a gas molecule is adsorbed to one pore wall while the attractive potential energy it experiences from the opposing wall of the pore is zero. As pore-diameter (d) is decreased (Figure 2a), the potential fields of opposing pore walls overlap to create an enhanced energy minimum in the middle (Figure 2b).21,23 This enhancement in negative potential energy is manifested significantly for d < 2 nm. A minimum practical d obviously exists, below which repulsive forces between the opposing walls dominate and adsorbing species experience unfavourable positive adsorption energy. As such, gaseous physisorption e.g. of CH4 and H2 in highly microporous materials (d < 2 nm) is particularly energetically favourable as they present both a high density of accessible internal surface sites as well as enhanced gas adsorption properties relative to meso- (2 < d < 50 nm) and macro- (d > 50 nm) pores.5 Amongst known porous materials, highly microporous MOFs stand out as a safe, low energy means of dense gas-fuel storage. Figure 2| Dependence of adsorbate interaction potential on adsorbent pore diameter. a, Model slit-shaped pores of diameters 2.0 nm (sky blue), 1.3 nm (royal blue) and 0.7 nm (navy blue). b, Potential energy profile (𝜙/𝜙o) for a gas molecule (CO2, kinetic diameter, 𝜎 = 0.33 nm) a distance (z) from the centre of model graphite micropores (a). Data digitised from Reference 21. 0.7 nm1.3 nm2.0 nm -2 -1 0 1 2 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 Po te nt ia l P ro fil e (!/ ! o) Distance ("/#) a b 8 3.0| Gas Storage in MOFs MOFs are a diverse family of high porosity, coordination polymers that result from the crystalline self-assembly of metal ions/metal oxide clusters24,25,26 with multi-dentate organic linkers.27 The modular nature of these materials allows high synthetic tuneability, with over 75,600 distinct structures having been added to the Cambridge Structural Database.28 Significantly, the compositional and structural variations between these materials allow comprehensive control over the materials chemical and physical properties. For example, the choice of metal and its degree of saturation influences both the thermal/chemical stability as well as structural topology and thus porosity of the framework. Furthermore, structural defects (i.e. crystalline missing linkers/cluster defects) are also well known to exist in MOFs and can be synthetically controlled to both tune porosity and create open metal sites.29 The inherent porosity of crystalline MOFs corresponds with high surface areas, typically exceeding 1000 m2 g–1. In 2012, Farha et al. developed a novel hexacarboxylate linker to synthesise the MOF NU- 110, which returned a record setting SBET of ~7000 m2 g–1.30 MOFs with large volumes of accessible surface area provide a high density of adsorption sites for gas uptake i.e. CH4 and H2 with one of their most promising applications lying in outstanding gas adsorption.12 Although the application of MOFs to gas adsorption was pioneered by the likes of Yaghi31,32 and Kitagawa,33 significant synthetic advances have since been made, resulting in the tailoring of MOF structures for maximised gas storage capacity. Significantly, a benchmark computational and experimental study of a diverse series of MOFs was recently used to demonstrate the linear relationship between pore volume and H2 capacity. This study successfully demonstrated that Cu-MOF NU-125 exceeds the ultimate U.S. DOE target for gravimetric H2 storage (6.5 wt%), achieving an outstanding storage capacity of 8.5 wt% (0.085 g g–1) under temperature/pressure swing conditions of 77 K / 100 bar (storage) ® 160 K / 5 bar (release).13 Crucially for gas-fuel storage applications, the organic linkers used in MOF construction can be systematically varied in terms of both length (controlling pore sizes and window apertures) and functionality (for tuneable physical properties and host-guest interactions). For example, Zhang et al. combined tri- and bi-topic linkers (Figure 3a, b) to produce a new series of Zn- MOFs with variable sub-4 nm porosity that could be used to study the effect of pore size on CH4 uptake capacity.34 Novel MOF ST-2 was calculated to have a theoretical volumetric CH4 9 working capacity (using the MOF single crystal density) of 289 cm3 (STP) cm–3 (298 K, 5 – 200 bar) (Figure 3c). This capacity not only represents a 31% improvement over the maximum capacity achievable by conventional pressurised gas tanks within the same pressure range but also surpasses the storage capacities achievable using previously record-setting activated carbons.35 Figure 3| ST-2 linkers and NG storage capacity. a, b, Chemical structures of tatb and 2,6- ndc, linkers in the MOF ST-2. c, Volumetric CH4 adsorption isotherm at 298 K for ST-2 (blue circles, digitised from Reference 34). Threshold value of maximum volumetric CNG capacity (250 bar and 298 K, blue line) is indicated. The synthetic possibilities presented by these hierarchical MOFs permits precise tuning of adsorbate-adsorbent interactions to other potential applications, not only in terms of greener fuel storage but also in environmental remediation. For example, pre- and post-combustion capture and storage of CO2 is considered a viable short-term solution to reducing atmospheric CO2 levels while fossil fuel usage persists.36 The MOF NbOFFIVE-1-Ni has demonstrated an outstanding CO2 uptake of 1.3 mmol g–1 under 400 ppm CO2 at 289 K.37 Thence, the sub-3.2 Å diameter square pores of this finely-tuned fluorinated MOF enable efficient capture of trace CO2, which surpasses both traditional amine scrubber solutions and benchmark physical adsorbents. The performance of NbOFFIVE represents a 300% improvement over the reference adsorbent Sr2+-SAPO-34 (a strontium-substituted silico-aluminophosphate molecular sieve).37 Last, the gas storage potential of MOFs can be combined with their molecular sieving tatb 2,6-ndc a COO- COO- N N NHN H N NH COO- COO- -OOC 0 50 100 150 200 250 300 350 400 0 50 100 150 200 b c Pressure (bar) CH 4 up ta ke (c m 3 (S TP ) c m -3 ) 10 capabilities, a result of their tuneable porosity, to purify fuel streams prior to combustion e.g. by the selective removal of contaminating CO2 in NG refinery.38 For example, Long et al. utilised the novel guest templating mechanism of flexible Co(bdp), which forms an expanded clathrate in the presence of CO2, to achieve near exclusive CO2 uptake through size exclusion of CH4 for total NG purification.39 This represents a significant improvement over traditional adsorption materials, such as high surface area carbons,40,41 zeolites42 and mesoporous silica.43,44 11 4.0| Industrially Viable Adsorption Based on the outstanding reports summarised above of MOFs applied to the storage of assorted gaseous fuels, gas-fuel purification and even carbon capture, it is evident that they have the potential for wide-reaching environmental impact in the energy sector. Yet despite their outstanding potential, significant barriers to widespread industrial MOF usage remain. Two of the most important measures of industrial viability in gas storage and separation materials are the gravimetric (g g–1) and volumetric (cm3 (STP) cm–3) capacity, which determine the quantity of gas which can be stored as a function of weight and volume of the adsorbent, respectively. If gravimetric capacity is low, pay-load capacity of the tank is diminished. This reduces fuel economy via the energy-consuming transportation of the heavy on-board tank. If volumetric capacity is low, excessively voluminous tanks are required to store the necessary quantity of fuel for feasible transport; a similar problem is faced in industrial adsorption-based separation colums.45 While substantial progress in the synthesis of MOFs with high gravimetric capacities for CH4, H2 and CO2 has been made, research directed towards achieving practical volumetric capacities has been somewhat overlooked. High gravimetric capacities are achieved in porous materials by maximising void fractions, yet this yields low density frameworks. However, volumetric capacity is maximised by balancing void fraction with high crystal density. This relationship was computationally studied by Snurr et al., who performed the high throughput computational screening of 18,383 MOFs for H2 storage capacity.46 Maximum gravimetric capacity was achieved in materials with void fractions exceeding 0.9 while volumetric capacity was maximised at a void fraction of ~ 0.75. The contrasting topological requirements of these two physical properties represents an inherent limitation to porous materials for feasible large- scale gas storage. A significant issue with many reports of state-of-the-art gas storage MOFs is that their gas storage capacity is recorded gravimetrically and converted to volumetric capacity using the theoretical single crystal density of the adsorbing MOF. Traditionally, however, MOFs have been synthesised as micro- or nanoscale crystals with the resulting bulk material being a loosely packed, fine powder.18 It follows that volumetric capacities calculated from single crystal density do not take into consideration the low packing efficiency of individual MOF crystals, and this has been reported to lead to overcalculation of the practical volumetric gas storage capacity by as much as 300%.47 This overcalculation stems from the inevitable complexity of measuring rb in fine powders. From a practical perspective, even if a MOF with both high 12 gravimetric and theoretical volumetric gas storage capacity is synthesised, a non-optimised powdered material will be rendered incapable of reaching its volumetric gas storage potential. Unprocessed MOFs are further inapplicable to gas storage and separation applications without shaping into pellets, as powders compact during the processes, resulting in pipe blockages and corresponding pressure drops.48 To achieve the wide-ranging potential of MOFs in gas storage and separation, it is clear that powders must be further processed into practical materials. Although an extensive range of MOF shaping techniques have been reported49 including 3D printing,50,51 foaming,52 chemical vapour deposition onto thin films53 and use of supports such as cordierite54,55 or activated carbons,56 the resulting low-density products often perform less well than predicted. Attempts to overcome this without recourse to the use of chemical additives have generally involved powder compaction by the application of pressure (typically 10 – 100 MPa).48 The ultimate aim of this process is to achieve a rb that approaches the MOFs single- crystal density by minimising interparticle space. Dense pellets of a range of MOFs have been reported using this approach. Farrusseng et al. recorded a positive linear relationship between applied pressure and mechanical strength for a range of densified MOF pellets of Cu- and Zr- MOFs.57 Despite achieving 1.8 – 3.4 fold increases in MOF rb compared to the corresponding powders, this synthetic approach caused substantial losses of surface area. For example, exposure of UiO-66-NH2 to 63 – 82 MPa applied pressure resulted in a ca. 80% loss of SBET. This was the result of pressure-induced amorphisation via the cleavage of chemical bonds in the MOF, an inherent consequence of applying mechanical force.58 In a similar vein, the commercially available MOF HKUST-1, was computationally identified by Peng et al. to exceed the U.S. DOE targets for CH4 storage with a volumetric capacity of 270 cm3 (STP) cm-3 (65 bar). Experimentally, however, the powder density of HKUST-1 (0.43 g cm–3) was so much lower than the single crystal density (0.88 g cm–3) that the volumetric capacity of the powder was less than 50% of the theoretical value. Moreover, while the application of high pressures (5 tons) achieved high density (1.10 g cm–3) pelletized disks, the low mechanical stability of these otherwise promising MOF pellets resulted in a 59% loss of pore volume and, correspondingly, no significant increase in volumetric storage capacity was observed.12 Zhou et al. studied the impact of extrusion pressure on the physical properties of highly stable PCN- 250, comprised of Fe3(μ3-O) oxoclusters coordinated by 3,3ʹ,5,5ʹ-azobenzenetetracarboxylate linkers. The application of asymmetric pressure during extrusion was found to incur irreversible phase transformations despite a retention of crystallinity. The resulting PCN-250 analogues (PCN-250ʹ and PCN-250ʹʹ) displayed isomerically flipped N=N bonds in some of the tetratopic 13 linkers. In particular, PCN-250ʹʹ showed both reduced unit-cell volume (1.33%) and gravimetric CH4 uptake (21.4%).59 An alternative approach to free-standing MOF pellets that avoids the application of high pressures involves the addition of chemical binders e.g. polymers such as PVP and PNMA.60,61 These binders act as cohesives, joining individual MOF crystals in e.g. extrusion,62 emulsion templating63 and granulation64 processes. Despite high resulting mechanical strength, which increases ease of usage from a practical perspective, materials obtained in this way must contain a significant percentage of binder; Denayer et al. produced sub-millimetre MIL-53 pellets that comprised 13% PVA.65 Besides the increased synthetic complexity and cost of producing composite materials, binders reduce gas uptake capacity by decreasing the total bulk quantity of MOF in each pellet without achieving substantial gains in total packing efficiency/density. Chang et al. reported the macroscopic granulation processing of Fe-, Zr- and Cr-based MOFs using 5 wt% amorphous-phase mesoporous ρ-alumina as binder.66 Even through the application of relatively moderate binder volumes, rb of the obtained materials were significantly lower than the theoretical crystal densities of the studied MOFs (0.67 g cm–3 vs. 1.20 g cm–3). Additionally, bulky binder molecules (typically long chain polymers) can further block pores, reducing accessible porosity and preventing efficient gas sieving.67 Overall, observed losses of crystallinity and accessible porosity by the traditional processing methods outlined here have, to date, resulted in non-optimised MOF products that cannot achieve their maximum potential for gas-based applications. It stands to reason that the future of MOFs for potential gas-fuel applications lies in the non- trivial densification of pure material without loss of accessible porosity.64 In seeking to solve this challenge, a novel, room-temperature sol-gel synthesis of robust, centimetre-scale monolithic MOF, monolithZIF-8, without the use of chemical binders or applied pressures was recently reported by Fairen-Jimenez and co-workers.47 The resulting transparent, glassy material displayed high rb (1.05 g cm–3), mechanical strength (H = 0.43 ± 0.03 GPa) and SBET (1423 m2 g–1) all of which values compare closely to the corresponding theoretical single crystal values (rb ca. 0.95 g cm–3, H = 0.50 ± 0.02 GPa, SBET = 1630 m2 g–1). The observation that rb exceeds the theoretical maximum was accounted for by the presence of dense/amorphous minor phases within the otherwise crystalline monolithMOF. Excitingly, the generality of this novel synthesis has now been demonstrated through its extension to another MOF, HKUST-1.68 The remarkable physical properties displayed by monolithHKUST-1 (rb = 1.06 g cm–3, SBET = 1288 14 m2 g–1), which approach the MOF’s theoretical maximum (rb = 0.88 g cm–3, SBET = 2014 m2 g– 1), resulted in an outstanding volumetric CH4 uptake capacity of 261 cm3 (STP) cm–3 (65 bar, 298 K). Not only does this represent a ca. 50% improvement over previously reported results for HKUST-1 compacted under a range of pressures,12 but it renders monolithic HKUST-1 prepared by this methodology the first densified material to effectively reach the U.S. DOE’s target for CH4 storage.11 While the development of this densified Cu-based MOF with record-setting volumetric CH4 storage represents a significant breakthrough in the field of powder shaping, novel monolithHKUST-1 is not an industrially optimised material.68 The presence of open Cu(II) sites in this MOF’s coordinatively unsaturated di-copper paddle-wheel-based structure results in a high water affinity. A gradual and irreversible loss of crystal structure through exposure to humidity is observed e.g. under ambient conditions and in the presence of NG.69 Furthermore, the near-ideal packing of the MOF particles may incur a reduced efficiency of gas transport through the exclusively small micropores of the monolithic macrostructure which would not otherwise be observed in a loosely-packed MOF powder. This could have practical implications for the efficiency of gas uptake and release. If MOFs are to be practically applied to gas storage applications, it is clear that this process of synthesising dense and porous monolithic materials must be extended to a wider range of MOFs with precisely tuned properties e.g. air/water stability, mechanical strength, and variable pore size and functionalities. As well as the recent advances in the development of industrially viable MOFs outlined above, significant steps have recently been taken in the development of more complex MOF-based composites. Doping of active metal and metal oxide NPs within MOF pores (NP@MOF) has long been observed to offer synergistic benefits that enhance the physisorptive properties of the host for improved gas storage.70 Of note is work by Kitagawa et al., in which a 74% increase in H2 storage capacity resulted from the immobilisation of 10 nm Pd cubes in HKUST-1 (giving the composite Pd@HKUST-1).71 This enhancement was attributed to the transfer of Pd 4d band electrons to unsaturated Cu 3d – O 2p hybridised bands near the Fermi level.72 The generated holes in the Pd 4d band were then argued to be filled by H 1s electrons during H2 adsorption, and thus the composite material exhibited enhanced H2 adsorption capacity relative to its constituent components. 15 Furthermore, recent advances in MOF synthesis have led to reports of not only powdered NP@MOF composites but novel monolithic materials. Recently, 5 nm SnO2 NPs were successfully immobilised in high density monolithZIF-8 to yield SnO2@monolithZIF-8.73 Crucially, this MOFs physical properties (SBET = 1055 m2 g–1, Vtot = 0.42 cm3 g–1) were not destroyed by doping and remained reasonably similar to that of the original host, monolithZIF-8 (SBET = 1423 m2 g–1, Vtot = 0.54 cm3 g–1).47 The immobilisation of metal-based NPs within these centimetre sized, robust MOF pellets offers an industrially viable solution to the practical problems associated with NP usage (e.g. safety, difficulty of handling, high cost associated with loss of NPs); the unique SnO2@monolithZIF-8 composite demonstrated remarkable recyclability without composite degradation/NP leaching. Additionally, the same study suggests that highly porous MOFs may act as ideal support for active NPs, without inhibiting activity, as reactants appear capable of diffusing through the porous network to reach active NPs within. These reports of NP@MOF and NP@monolithMOF point strongly to the potential of complex, functional MOF- based composites e.g. in industrially viable gas-fuel storage. The above discussed results point towards the numerous industrial applications these recently developed monolithMOF may be applicable towards, including high density gas storage and selective gas adsorption/sieving. Furthermore, the capacity of monolithMOF to host NPs, yielding NP@ monolithMOF, presents an even greater range of potential applications i.e. in catalysis and enhanced gas storage. It is clear that if the full potential of these materials is to be elucidated, the synthetic modularity of MOFs and NPs must be exploited to synthesise a wider array of monolithMOF and NP@monolithMOF composites with differing physical/chemical properties for targeted applications. 16 5.0| References 1. Allen, M. et al. IPCC Special Report: Global Warming of 1.5°C. (2018). 2. Rahmstorf, S., Foster, G. & Cahill, N. Global Temperature Evolution: Recent Trends and Some Pitfalls. Environ. Res. Lett. 12, 054001 (2017). 3. Ummenhofer, C. C. & Meehl, G. A. Extreme Weather and Climate Events With Ecological Relevance: A Review. Phil. Trans. R. Soc. B. 372, 1–13 (2017). 4. BP Energy Outlook, BP p.l.c. (2018). 5. Kumar, K. V., Preuss, K., Titirici, M. M. & Rodríguez-Reinoso, F. Nanoporous Materials for the Onboard Storage of Natural Gas. Chem. Rev. 117, 1796–1825 (2017). 6. Dincer, I. & Acar, C. Innovation in Hydrogen Production. Int. J. Hydrogen Energy 42, 14843–14864 (2017). 7. Berenguer-Murcia, Á., Marco-Lozar, J. P. & Cazorla-Amorós, D. Hydrogen Storage in Porous Materials: Status, Milestones, and Challenges. Chem. Rec. 18, 900–912 (2018). 8. Álvarez Fernández, R., Corbera Caraballo, S., Beltrán Cilleruelo, F. & Lozano, J. A. Fuel Optimization Strategy for Hydrogen Fuel Cell Range Extender Vehicles Applying Genetic Algorithms. Renew. Sustain. Energy Rev. 81, 655–668 (2018). 9. Rowsell, J. L. C. & Yaghi, O. M. Strategies for Hydrogen Storage in Metal-Organic Frameworks. Angew. Chem. Int. Ed. 44, 4670–4679 (2005). 10. Deniz, C. & Zincir, B. Environmental and Economical Assessment of Alternative Marine Fuels. J. Clean. Prod. 113, 438–449 (2016). 11. Gallagher, J. Towards Methane Targets. Nat. Energy 3, 86 (2018). 12. Peng, Y. et al. Methane Storage in Metal − Organic Frameworks: Current Records, Surprise Findings, and Challenges. J. Am. Chem. Soc 135, 11887–11894 (2013). 13. García-Holley, P. et al. Benchmark Study of Hydrogen Storage in Metal-Organic Frameworks under Temperature and Pressure Swing Conditions. ACS Energy Lett. 3, 748–754 (2018). 14. Toyota Mirai Brochure, 2019. (2019). 15. Barthelemy, H., Weber, M. & Barbier, F. Hydrogen Storage: Recent Improvements and Industrial Perspectives. Int. J. Hydrogen Energy 42, 7254–7262 (2017). 17 16. Zhang, X. et al. Enhanced Hydrogen Storage Properties of MgH2 Catalyzed With Carbon-Supported Nanocrystalline TiO2. J. Power Sources 398, 183–192 (2018). 17. Afzal, M., Mane, R. & Sharma, P. Heat Transfer Techniques in Metal Hydride Hydrogen Storage: A Review. Int. J. Hydrogen Energy 42, 30661–30682 (2017). 18. Wu, B. et al. Review on Processing of Metal–Organic Framework (MOF) Materials Towards System Integration for Hydrogen Storage. Int. J. Energy Res. 39, 607–620 (2015). 19. Cychosz, K. A. & Thommes, M. Progress in the Physisorption Characterization of Nanoporous Gas Storage Materials. Engineering 4, 559–566 (2018). 20. Wu, K., Chen, Z., Li, X. & Dong, X. Methane Storage in Nanoporous Material at Supercritical Temperature Over A Wide Range Of Pressures. Sci. Rep. 6, 1–10 (2016). 21. Cui, X., Bustin, R. M. & Dipple, G. Selective Transport of CO2, CH4, and N2 in Coals: Insights From Modeling of Experimental Gas Adsorption Data. Fuel 83, 293–303 (2004). 22. Sullivan, D. E. Van der Waals Model of Adsorption. Phys. Rev. B 20, 3991–4000 (1979). 23. Saha, D., Grappe, H. A., Chakraborty, A. & Orkoulas, G. Postextraction Separation, On- Board Storage, and Catalytic Conversion of Methane in Natural Gas: A Review. Chem. Rev. 116, 11436–11499 (2016). 24. Tsivion, E. & Head-Gordon, M. Methane Storage: Molecular Mechanisms Underlying Room-Temperature Adsorption in Zn4O(BDC)3 (MOF-5). J. Phys. Chem. C 121, 12091– 12100 (2017). 25. Yan, Y. et al. Porous Metal-Organic Polyhedral Frameworks with Optimal Molecular Dynamics and Pore Geometry for Methane Storage. J. Am. Chem. Soc. 139, 13349– 13360 (2017). 26. Kim, A. R. et al. Creating High CO/CO2 Selectivity and Large CO Working Capacity Through Facile Loading of Cu(I) species into an Iron-Based Mesoporous Metal-Organic Framework. Chem. Eng. J. 348, 135–142 (2018). 27. Kirchon, A., Feng, L., Drake, H. F., Joseph, E. A. & Zhou, H.-C. From Fundamentals to Applications: A Toolbox for Robust and Multifunctional MOF Materials. Chem. Soc. Rev. 47, 8611–8638 (2018). 28. Moghadam, P. Z. et al. Development of a Cambridge Structural Database Subset: A 18 Collection of Metal-Organic Frameworks for Past, Present, and Future. Chem. Mater. 29, 2618–2625 (2017). 29. Liu, L. et al. Imaging Defects and Their Evolution in a Metal–Organic Framework at Sub-Unit-Cell Resolution. Nat. Chem. 11, 622–628 (2019). 30. Farha, O. K. et al. Metal-Organic Framework Materials With Ultrahigh Surface Areas: Is The Sky The Limit? J. Am. Chem. Soc. 134, 15016–15021 (2012). 31. Li, H., Eddaoudi, M., Groy, T. L. & Yaghi, O. M. Establishing Microporosity in Open Metal-Organic Frameworks: Gas Sorption Isotherms for Zn(BDC). J. Am. Chem. Soc. 120, 8571–8572 (1998). 32. Eddaoudi, M., Li, H. & Yaghi, O. M. Highly Porous and Stable Metal-Organic Frameworks: Structure Design and Sorption Properties. J. Am. Chem. Soc 122, 1391– 1397 (2000). 33. Kondo, M., Yoshitomi, T., Seki, K., Matsuzaka, H. & Kitagawa, S. Three-Dimensional Framework with Channeling Cavities for Small Molecules. Angew. Chem. Int. Ed. 36, 1725–1727 (1997). 34. Liang, C. C. et al. Engineering of Pore Geometry for Ultrahigh Capacity Methane Storage in Mesoporous Metal-Organic Frameworks. J. Am. Chem. Soc. 139, 13300– 13303 (2017). 35. Casco, M. E. et al. High-Pressure Methane Storage in Porous Materials: Are Carbon Materials in the Pole Position? Chem. Mater. 27, 959–964 (2015). 36. Sreedhar, I., Nahar, T., Venugopal, A. & Srinivas, B. Carbon Capture by Absorption – Path Covered and Ahead. Renew. Sustain. Energy Rev. 76, 1080–1107 (2017). 37. Bhatt, P. M. et al. A Fine-Tuned Fluorinated MOF Addresses the Needs for Trace CO2 Removal and Air Capture Using Physisorption. J. Am. Chem. Soc. 138, 9301–9307 (2016). 38. Trickett, C. A. et al. The Chemistry of Metal–Organic Frameworks for CO2 Capture, Regeneration and Conversion. Nat. Rev. Mater. 2, 1–16 (2017). 39. Taylor, M. K. et al. Near-Perfect CO2/CH4 Selectivity Achieved through Reversible Guest Templating in the Flexible Metal-Organic Framework Co(bdp). J. Am. Chem. Soc. 140, 10324–10331 (2018). 40. Szczęśniak, B., Choma, J. & Jaroniec, M. Gas Adsorption Properties of Graphene-Based 19 Materials. J. Colloid Interface Sci. 514, 801–813 (2018). 41. Sawant, S. Y. et al. Precursor Suitability and Pilot Scale Production of Super Activated Carbon for Greenhouse Gas Adsorption and Fuel Gas Storage. Chem. Eng. J. 315, 415– 425 (2017). 42. Fraccarollo, A., Canti, L., Marchese, L. & Cossi, M. Accurate Evaluation of the Dispersion Energy in the Simulation of Gas Adsorption into Porous Zeolites. J. Chem. Theory Comput. 13, 1756–1768 (2017). 43. Tseng, H. H., Chuang, H. W., Zhuang, G. L., Lai, W. H. & Wey, M. Y. Structure- Controlled Mesoporous SBA-15-Derived Mixed Matrix Membranes for H2 Purification and CO2 Capture. Int. J. Hydrogen Energy 42, 11379–11391 (2017). 44. Ghalei, B., Pournaghshband Isfahani, A., Sadeghi, M., Vakili, E. & Jalili, A. Polyurethane-Mesoporous Silica Gas Separation Membranes. Polym. Adv. Technol. 29, 874–883 (2018). 45. Gómez-Gualdrón, D. A. et al. Understanding Volumetric and Gravimetric Hydrogen Adsorption Trade-Off in Metal-Organic Frameworks. ACS Appl. Mater. Interfaces 9, 33419–33428 (2017). 46. Colón, Y. J., Fairen-Jimenez, D., Wilmer, C. E. & Snurr, R. Q. High-Throughput Screening of Porous Crystalline Materials for Hydrogen Storage Capacity Near Room Temperature. J. Phys. Chem. C 118, 5383–5389 (2014). 47. Tian, T., Velazquez-Garcia, J., Bennett, T. D. & Fairen-Jimenez, D. Mechanically and Chemically Robust ZIF-8 Monoliths With High Volumetric Adsorption Capacity. J. Mater. Chem. A 3, 2999–3005 (2015). 48. Nandasiri, M. I., Jambovane, S. R., McGrail, B. P., Schaef, H. T. & Nune, S. K. Adsorption, Separation, and Catalytic Properties of Densified Metal-Organic Frameworks. Coord. Chem. Rev. 311, 38–52 (2016). 49. Valizadeh, B., Nguyen, T. N. & Stylianou, K. C. Shape Engineering of Metal–Organic Frameworks. Polyhedron 145, 1–15 (2018). 50. Thakkar, H., Eastman, S., Al-Naddaf, Q., Rownaghi, A. A. & Rezaei, F. 3D-Printed Metal-Organic Framework Monoliths for Gas Adsorption Processes. ACS Appl. Mater. Interfaces 9, 35908–35916 (2017). 51. Kreider, M. C. et al. Toward 3D Printed Hydrogen Storage Materials Made with ABS- 20 MOF Composites. Polym. Adv. Technol. 29, 867–873 (2018). 52. Chen, Y. et al. Shaping of Metal-Organic Frameworks: From Fluid to Shaped Bodies and Robust Foams. J. Am. Chem. Soc. 138, 10810–10813 (2016). 53. Stassen, I. et al. Chemical Vapour Deposition of Zeolitic Imidazolate Framework Thin Films. Nat. Mater. 15, 304–310 (2016). 54. Rezaei, F. et al. MOF-74 and UTSA-16 Film Growth on Monolithic Structures and Their CO2 Adsorption Performance. Chem. Eng. J. 313, 1346–1353 (2017). 55. Lawson, S., Hajari, A., Rownaghi, A. A. & Rezaei, F. MOF Immobilization on the Surface of Polymer-Cordierite Composite Monoliths Through In-Situ Crystal Growth. Sep. Purif. Technol. 183, 173–180 (2017). 56. Fernández-Catalá, J., Casco, M. E., Martínez-Escandell, M., Rodríguez-Reinoso, F. & Silvestre-Albero, J. HKUST-1@ACM Hybrids for Adsorption Applications: A Systematic Study of the Synthesis Conditions. Microporous Mesoporous Mater. 237, 74–81 (2017). 57. Dhainaut, J. et al. Systematic Study of the Impact of MOF Densification into Tablets on Textural and Mechanical Properties. Cryst. Eng. Comm. 19, 4211–4218 (2017). 58. Su, Z., Miao, Y. R., Zhang, G., Miller, J. T. & Suslick, K. S. Bond Breakage Under Pressure in a Metal-Organic Framework. Chem. Sci. 8, 8004–8011 (2017). 59. Yuan, S. et al. PCN-250 under Pressure: Sequential Phase Transformation and the Implications for MOF Densification. Joule 1, 806–815 (2017). 60. Yang, Z., Cao, L., Li, J., Lin, J. & Wang, J. Facile Synthesis of Cu-BDC/Poly(N- methylol acrylamide) HIPE Monoliths via CO2-in-Water Emulsion Stabilized by Metal- Organic Framework. Polymer 153, 17–23 (2018). 61. Yusuf, K., Badjah-Hadj-Ahmed, A. Y., Aqel, A. & Alothman, Z. A. Monolithic Metal- Organic Framework MIL-53(Al)-Polymethacrylate Composite Column for the Reversed-Phase Capillary Liquid Chromatography Separation of Small Aromatics. J. Sep. Sci. 39, 880–888 (2016). 62. Zheng, J. et al. Shaping of Ultrahigh-Loading MOF Pellet with a Strongly Anti-Tearing Binder for Gas Separation and Storage. Chem. Eng. J. 354, 1075–1082 (2018). 63. Zhu, H., Zhang, Q. & Zhu, S. Assembly of a Metal–Organic Framework into 3D Hierarchical Porous Monoliths Using a Pickering High Internal Phase Emulsion 21 Template. Chem. Eur. J. 22, 8751–8755 (2016). 64. Chanut, N. et al. Observing the Effects of Shaping on Gas Adsorption in Metal-Organic Frameworks. Eur. J. Inorg. Chem. 2016, 4416–4423 (2016). 65. Finsy, V. et al. Separation of CO2/CH4 Mixtures With the MIL-53(Al) Metal-Organic Framework. Micropor. Mesopor. Mat. 120, 221–227 (2009). 66. Valekar, A. H. et al. Shaping of Porous Metal-Organic Framework Granules Using Mesoporous ρ-alumina as a Binder. RSC Adv. 7, 55767–55777 (2017). 67. Rubio-Martinez, M. et al. New Synthetic Routes Towards MOF Production at Scale. Chem. Soc. Rev. 46, 3453–3480 (2017). 68. Tian, T. et al. A Sol–Gel Monolithic Metal–Organic Framework with Enhanced Methane Uptake. Nat. Mater. 17, 174–179 (2017). 69. Álvarez, J. R. et al. Structure Stability of HKUST-1 Towards Water and Ethanol and Their Effect on its CO2 Capture Properties. Dalton Trans. 46, 9192–9200 (2017). 70. Falcaro, P. et al. Application of Metal and Metal Oxide Nanoparticles@MOFs. Coord. Chem. Rev. 307, 237–254 (2016). 71. Li, G. et al. Hydrogen Storage in Pd Nanocrystals Covered With a Metal-Organic Framework. Nat. Mater. 13, 802–806 (2014). 72. Chen, Y. et al. Electronic Origin of Hydrogen Storage in MOF-Covered Palladium Nanocubes Investigated by Synchrotron X-rays. Commun. Chem. 1, 61 (2018). 73. Mehta, J. P. Incorporation of Metal Oxide Nanoparticles in a Monolithic Metal-Organic Framework for Applications in Photocatalysis. (The University of Cambridge, 2016). 22 Chapter II Monolithic Metal-Organic Frameworks 24 25 1.0| UiO-66 First reported in 2008 by Lillerud et al., UiO-66 is the archetypal zirconium MOF.1 Its structure is comprised of octahedral hexanuclear clusters of Zr6O4(OH)4 (Figure 1a), in which each zirconium atom is 8-coordinated through bridging oxygen centres. The 12-fold coordination of each cluster to adjacent clusters via ditopic bdc linkers (Figure 1b) yields the MOF in a face- centred cubic crystal structure (Figure 1c). The 3D edge-sharing arrangement of microporous octahedral cages (11 Å diameter) which derive from this structure are further face-sharing with tetrahedral cages (8 Å diameter) through triangular windows (5 – 7 Å).2 Figure 1| UiO-66 SBUs and crystal structure. a, Zr6O4(OH)4 cluster. b, bdc. c, Structure of UiO-66 which demonstrates how the metal cluster (a) and linker (b) self-assemble to form the crystalline MOF. Colours correspond to C (purple), H (white), O (sky-blue) and Zr (navy blue). This fascinating MOF has been extensively studied, a consequence of its highly desirable physical and chemical properties. Firstly, the microporous material displays high surface area, with experimentally obtained SBET typically exceeding 1000 m2 g–1.1,3,4 Using GCMC simulations, the theoretical maximum SBET for the perfect crystalline MOF was calculated to be 1644 m2 g–1. Variations amongst experimental samples and the theoretical structure are attributed to defects such as missing linker/cluster defects and the presence of amorphous regions, both of which are well-known and pervasive amongst experimentally synthesised UiO- 66.5,6 Furthermore, this porous material displays exceptional stability, both chemical and thermal, in the presence of air and most solvents up to temperatures of ca. 500 °C. This is 26 attributed to the strong hard-hard coordination of oxophilic zirconium(IV) to carboxylate linkers in this MOF.7 Additionally, while porosity and chemical/mechanical stability are fundamentally competing factors, the 12-fold coordination within UiO-66 yields higher mechanical strength than similarly porous MOFs with lower coordination.8 The ready abundance and relatively low cost of this MOF’s constituent components, coupled with its high porosity, chemical, thermal and mechanical stability, has presented it as an industrially viable candidate for an extensive range of applications including gas-fuel storage,9 pollutant capture,10 adsorption and dissipation of mechanical energy under compression11 and even heterogeneous catalysis.12 The relatively recent development of this novel MOF has also paved the way for a now rapidly expanding MOF family, zirconium MOFs. Within this MOF sub-group, zirconium oxo-clusters are connected to diverse linkers (Figure 2) to yield an extensive range of structures including UiO-66 derivative structures with functionalised linkers (e.g. UiO-66-NH2, UiO-66-F and UiO- 66-SO2),13 isoreticular structures using elongated linkers (e.g. UiO-67 and UiO-68)4 as well as more complex tris- and tetra-topically coordinated structures (e.g. MOF-808 and NU-1000).14 Despite the significant structural variation within this MOF family, Zr-MOFs typically demonstrate highly industrially desirable properties, comparable to the parent MOF UiO-66 – thermal, chemical and mechanical stability is a characteristic feature.15 Coupled with variable pore-sizes and geometries as well as the capacity for both pre- and post-synthetic modification of this broad MOF family,16 numerous applications of Zr-MOFs have been proposed.17 For example, fluorinated UiO-66 analogues (UiO-66-F and UiO-66-CF3) demonstrate a significant enhancement in O2 storage capacity compared to the unfunctionalized MOF, despite a reduction in pore-volume.18 In spite of the abundant studies on both the industrially favourable properties and possible applications of UiO-66, as well as numerous synthetic protocols (even including room temperature19 and water-based syntheses20) for both the parent MOF and its extensive derivative structures, UiO-66 is yet to be commercialised, having remained firmly in the academic sphere. As discussed, this stems from the current technological inability to process the material into robust, dense pellets without damage to the structure’s porosity. A notable report of UiO-66 shaping/processing includes the synthesis of robust, centimetre sized UiO-66- NH2 monoliths by Horcajada et al. who obtained these materials by supercritical CO2 drying of ethanolic MOF NP suspensions.21 While promising mechanical strength was reported, the poor 27 microporosity of the MOF NPs (0.16 cm3 g–1) used to synthesise the materials coupled with the presence of extensive meso- and macro-porosity in the final monoliths, resulted in low density and low surface area macrostructures, unsuitable for gas storage applications. This further highlights the barrier preventing MOF translation to industrial gas storage; an inability to obtain practically useable materials which critically exhibit high density, surface area and mechanical strength. Figure 2| Organic linker molecules used to synthesise zirconium MOFs. a, Linker for the synthesis of monofunctionalised UiO-66-X where X = e.g. H, F, Cl, Br, I, CH3, CF3, NO2, NH2, OH, OCH3, COOH, SO3H. b, Tritopic linker (btc) for the synthesis of MOF-808. c, d, Biphenyl- 4,4´-dicarboxylic acid and triphenyl‐4,4´´‐dicarboxylic acid; organic linkers used to synthesise the isoreticular UiO-66 structures, UiO-67 and UiO-68 respectively. e, 1,3,6,8-tetrakis (p- benzoate)pyrene, the tetratopic organic linker in Zr-MOF NU-1000. COOH COOH X COOH COOH COOH COOH COOH HOOC COOH COOHHOOC COOHHOOC a b c d e 28 1.1| Aims and Objectives The development of a synthetic protocol for robust, densified, monolithic Zr-MOFs without the use of chemical binders or applied pressures would represent a significant advancement in the field of industrial viable MOFs. Thus, the first aim of this research will be to explore the synthesis of prototypical Zr-MOF UiO-66 as a monolith, monolithUiO-66, with the aim of amalgamating the high porosity and surface area of the powdered MOF with the density, chemical stability and mechanical strength of the single crystal material. This research avenue also presents an opportunity to shed light on the poorly understood monolith formation mechanism proposed by Fairen-Jimenez et al.22,23 Through the non-trivial development of a new monolithic MOF and further exploration of the influence of synthetic drying conditions on the physical properties of obtained monolithic materials, the physical/chemical origin of the factors which govern MOF particle interaction during drying may be elucidated. This will not only fundamentally further our understanding of these novel material’s formation mechanism but crucially aids in the design and synthesis of a wider range of monolithic MOFs. Thus, a further aim of this research is to apply the findings from the synthesis of monolithUiO-66 to an extensive range of zirconium MOFs with the aim of expanding the known repertoire of monolithMOFs (which currently comprises ZIF-8 and HKUST-1 exclusively). By exploring the synthesis of more complex UiO-66 derived structures as monoliths e.g. functionalised analogues and non-isostructural Zr-MOFs, the generality of MOFs to be synthesised as monolithic materials may be determined. Extensive characterisation of the physical properties of the obtained materials must be performed. A key aim of this research is to develop monolithic materials which are fundamentally comparable to the single crystal material. Thus, the physical and chemical properties of the materials will be fully elucidated and benchmarked against the corresponding ideal powder/single crystal materials. Furthermore, this extensive characterisation of e.g. porosity and mechanical strength is required to determine the practical viability of MOF materials towards industrial gas storage, a further overall aim of this research (see Chapter IV). 29 1.2| Monolith Synthesis As discussed, the synthesis of high density monolithZIF-822 and monolithHKUST-123 has previously been reported, with a retention of both crystallinity and porosity noted. In these syntheses, MOF NPs were densified during drying to yield centimetre-scale monolithMOFs. It was demonstrated that both mild drying conditions and small primary MOF particles were critical to monolith formation – both monolithZIF-822 and monolithHKUST-123 were synthesised and dried at room temperature. Fairen-Jimenez et al. postulated that mild conditions must be maintained during monolith drying so that inter-particle epitaxial growth is facilitated through reaction with residual precursors without inducing stress at the vapour-liquid meniscus interface which would destroy the gel macrostructure adopted by the primary particles. Yet the formation of strong Zr–O bonds, required for both the synthesis of crystalline UiO-66 and proposed epitaxial primary particle growth, typically demands significantly elevated temperatures. Room temperature synthesis of both the powdered MOF and potential monolith are thus non-trivial. As a preliminary experiment, the sole room temperature synthesis of UiO-66 reported by Farha et al.19 was replicated and followed by a modified work up, with the aim of facilitating monolith formation. This literature procedure was identified as a starting point due to its use of a synthetically isolated, hexanuclear zirconium oxocluster. This highly reactive precursor is capable of forming crystalline UiO-66 under ambient conditions making it both unique amongst known UiO-66 syntheses and applicable for monolith formation under ambient conditions. Firstly, the reactive zirconium oxocluster was synthesised from zirconium(IV) propoxide in DMF at 130 °C. Through the subsequent addition of AA as a modulator, this Zr intermediate self-assembles with bdc at room temperature. The literature workup of the obtained suspension reports washing in DMF and solvent exchanging with acetone before drying at 80 °C to yield UiO-66 as a fine powder. Instead, the work up for the syntheses of monolithZIF-8 and monolithHKUST-1 was applied: primary MOF particles were centrifuged to a pellet, washed in ethanol/methanol and dried under mild conditions at room temperature.22 By combining the experimental procedure for the room temperature synthesis of UiO-66 with the procedure for the room temperature synthesis of monolithic ZIF-8, it was postulated that monolithUiO-66 would be obtained. Upon visual inspection, the dried product was found to be a powder, displaying no resemblance to the shiny and rigid monoliths of ZIF-8 obtained through the same work up. Although the exact mechanism of monolith formation has not been fully elucidated, Fairen- 30 Jimenez et al. have identified primary particle size to be a key requirement for monolith formation. Firstly, the primary MOF particles must be small; for monolithZIF-8 and monolithHKUST-1 they are 70 and 50 nm respectively.22,23 This is to allow efficient compaction during centrifugation, by the reduction of interstitial space between particles. To confirm that the primary particles produced in this experiment met this requirement, SEM was used to measure their size (Figure 3). The particles were found to be 191.0 ± 36.6 nm. Although relatively small by MOF particle standards, this is an order of magnitude larger than the primary particles used for previous monolith syntheses. It is typically advantageous to synthesise MOFs as large single crystals (microscale) which increases the potential for long-range crystalline order. The consequent repetition of the unit cell infers the porosity and high surface area which are ubiquitous to MOFs. Yet, the use of such large primary particles may prevent sufficiently high particle packing for monolith formation. Despite the presence of residual precursor capable of room temperature reaction, poorly packed large particles may not be sufficiently close together during drying for epitaxial growth, consistent with the reported hypothesis of monolith formation. Correspondingly, monolith formation via the preliminary synthesis of significantly smaller primary particles was pursued with the aim of better facilitating densification. Figure 3| SEM images of UiO-66 primary particles. a, b, Low and high magnification, respectively, SEM images of UiO-66 primary particles (191.0 ± 36.6 nm) produced by room- temperature synthesis. a b 31 Bennett et al. reported the solvothermal synthesis of UiO-66 gel (Figure 4a) comprised of ca. 10 nm MOF particles.24 Despite the small particle size, which inherently risks the potential for long-range crystalline order, the particle’s high crystallinity was confirmed by XRD. Crucially, the particle’s capacity to form ca. 200 – 250 µm diameter optically transparent MOF fragments after drying was further demonstrated (Figure 4b). Though orders of magnitude too small to be classed a macroscopic monolith (a key aim of this research), the optically transparent nature of the obtained MOF fragments is indicative of substantial primary particle densification via the minimisation of space at the inter-particle barrier. However, the optical micrograph further demonstrates the presence of large, bubble-like defects throughout the material, suggestive of macroporosity. rb of these MOF fragments was not quantified. However, the reported N2 isotherms indicated that these small fragments were non-optimal for e.g. dense gas storage. Despite high surface area and microporosity, as indicated by significant N2 uptake at pressures below 0.1 bar, the sudden N2 uptake recorded between 0.8 – 1 bar indicates N2 condensation in meso/macropores (Figure 4c). While the small monolith size and low density of the reported material evidently makes it unsuitable for dense gas-fuel storage, it does point towards these particles as a viable synthetic starting point. Figure 4| Literature UiO-66 NPs. a, Optical image of viscous UiO-66 gel comprised of 10 nm primary particles. Inversion of the reaction vessel demonstrates the gel’s viscosity. b, Optical micrograph image of a UiO-66 xerogel fragment (scale bar 100 µm) showing optical transparency. c, N2 adsorption isotherm (0 – 1 bar, 77 K) for UiO-66 gel obtained by drying at 200 ºC in a petri-dish (red diamonds, digitised from Reference 24). a b c 0 400 800 1200 1600 0 0.2 0.4 0.6 0.8 1 P/Po N 2 up ta ke (c m 3 g-1 ) 32 In the present case, the literature synthesis for UiO-66 gel was replicated to obtain 10 nm primary MOF particles in a viscous gel phase. The MOF gel was then washed and dried under four different conditions (Table 1) to obtain various monolithic materials (Figure 5). UiO- 66_A was prepared following the literature procedure reported by Bennett et al., whereby reaction solvent (DMF) and impurities were washed from the UiO-66 primary particles with ethanol before drying at 200 °C. By this method, exclusively sub-millimetre sized MOF fragments were obtained (Figure 5a), consistent with the literature report (Figure 4b). Table 1| Experimental conditions for monolithUiO-66 synthesis by washing/drying UiO-66 gel.24 Washing Procedure Centrifugation Procedure Drying Temperature (°C) UiO-66_A Ethanol (3 × 30 mL) 3 × 10 min* 200 UiO-66_B Ethanol (3 × 30 mL) 3 × 10 min* 30 UiO-66_C DMF (1 × 30 mL) 1 × 10 min* 30 UiO-66_D DMF (1 × 30 mL) 1 x 10 min* + 1 x 180 min** 30 *Centrifugation (5500 rpm) performed after each wash to re-obtain MOF gel as sediment. **Additional 180 min (5500 rpm) centrifugation performed on densified MOF gel after washing in DMF and decanting the supernatant. Since drying temperature was previously found to be critical in the formation of both monolithZIF- 8 and monolithHKUST-1, this was further studied. UiO-66_B was obtained by washing the gel in ethanol under identical conditions to those used for UiO-66_A, centrifuging it into a pellet and instead drying under significantly milder conditions, namely 30 °C. This resulted in a centimetre-sized, opaque monolith, (Figure 5b) supporting the view that the drying temperature is fundamental to controlling the monolith macrostructure. Although the UiO-66_B was both qualitatively robust and macroscopic, its low optical transparency was indicative of poor primary particle densification. In the case of both monolithZIF-8 and monolithHKUST-1, it was reported that slow evaporation of the reaction solvent, 33 coupled with the presence of residual precursors, facilitates primary particle interaction by effectively extending reaction time. Epitaxial growth of the primary particles reduces the optically-visible inter-particle barrier to give a transparent material. According to this view, the use of ethanol to wash UiO-66 particles, which were synthesised in DMF, may quench the reaction whereas washing in DMF may facilitate its continuation. The influence of washing solvent on monolith physical properties was therefore studied by instead washing the primary particles in further DMF before drying at 30 °C; the resulting monolith (UiO-66_C) was optically transparent (Figure 5c). This solvent-induced alteration in optical transparency demonstrates the complex range of solvent-particle interactions that exist. A similar synthesis was used for UiO-66_D (Figure 5d), but with an extended (180 min) centrifugation. This allowed the effects of primary particle densification prior to drying to be explored. These relatively minor changes to the synthetic procedure incurred fascinating changes in the resulting monoliths physical properties. To fully explore the obtained materials macroscopic physical properties, they were extensively characterised. Figure 5| Optical images of UiO-66 monoliths. a, UiO-66 washed in ethanol, dried at 200 °C (UiO-66_A). b, UiO-66 monolith with truncated cone shape, produced by washing primary particles in ethanol and drying at 30 °C (UiO-66_B). c, UiO-66 washed in DMF, dried at 30 °C (UiO-66_C). d, UiO-66 washed in DMF with extended centrifugation, dried at 30 °C (UiO- 66_D). Coloured symbols (bottom-left) in each figure represent the key that will identify each material henceforth. 5 mm5 mm 5 mm 5 mm a b c dUiO-66_C UiO-66_DUiO-66_BUiO-66_A 34 1.3| Monolith Characterisation 1.3.1| Particle Size and Morphology The precursor MOF gel and resulting monolithic materials were studied by TEM (Figure 6a– d). The obtained MOF gel was consistent with the literature report of ca. 10 nm particles in a gelatinous network macrostructure.24 Comparison to TEM images of the monolithic material, UiO-66_D, confirmed that the loose particle network undergoes significant densification during drying (Figure 6c, d). It further demonstrates that despite the optical transparency of this material, the particles do not fully merge together, with distinct particle structures clearly maintained. This was further supported by low magnification SEM (Figure 6e), which showed the smooth surface of the densified monolithic macrostructure. Yet at increased magnification, this was resolved into a homogenous array of densely packed NPs (Figure 6f). These data support the proposed mechanism of monolith formation via primary particle densification. Figure 6| Electron microscope images of UiO-66 gel and monolith. a, b, TEM images of UiO-66 gel. The irregularly shaped MOF primary NPs, ca. 10 nm, adopt a gelatinous network macrostructure. c, d, TEM images of fully dried densified monolith (UiO-66_D). e, f, Low and high magnification, respectively, SEM images of UiO-66_D. a b c d e f 35 1.3.2| Elemental Composition and Structure Elemental mapping by STEM-EDX was used to study the composition of the materials, confirming the homogenous distribution of zirconium, oxygen and carbon throughout the monolith (Figure 7). The chemical composition of the materials was further quantified by ICP- OES (Table 2). The theoretical composition of UiO-66 was calculated from the chemical formula of the pristine MOF (Zr6(OH)4O4(bdc)6). All experimentally obtained samples matched satisfactorily with the theoretical elemental composition and no significant compositional differences were observed between experimental samples (UiO-66_A – D). Experimental deviation from the theoretical elemental composition (calculated from a perfect crystal structure) can result from defects in the crystal structure. In the current case, missing linker defects are suggested by the experimentally observed carbon (%) compositions which were relatively low (-2.5%) compared to the theoretical value. This defect is commonly reported for UiO-66.25 Additionally, the increased hydrogen content of the experimental samples may point to the presence of water in the crystal structure. Although all samples were thoroughly degassed prior to analysis, foreign species in the air can be quickly adsorbed by these highly porous materials during the ICP-OES sample preparation procedure (i.e. weighing under ambient conditions). Figure 7| Elemental maps of monolithic UiO-66. Low magnification STEM electron image of UiO-66_D with area selected for EDX analysis indicated (white box) and corresponding EDX elemental maps showing distribution of Zr (pink), O (blue) and C (orange) throughout. 50 nm Ca rb on Zi rc on iu m 50 nm Ox yg en 50 nm Electron 36 Table 2| Elemental composition (%) of UiO-66_A – D obtained by ICP-OES compared to the theoretical composition of the ideal crystal structure. Composition (%) Theoretical UiO-66_A UiO-66_B UiO-66_C UiO-66_D Zr 32.9 31.2 29.7 31.6 31.7 C 34.6 31.9 31.7 32.0 32.0 H 1.7 2.1 2.2 2.0 2.1 N 0.0 0.0 0.0 0.3 0.0 The monolithic MOFs composition and structure was further characterised by PXRD, Figure 8. Each of the experimentally obtained patterns are consistent with the pattern simulated from the MOFs ideal crystal structure (black). The observed broadening of reflections in each experimental PXRD pattern is a consequence of well-known Scherrer line broadening, which stems from the non-convergence of diffraction peaks in nano-crystallites.26 This was further supported by the earlier discussed TEM and SEM images (Figure 6) which confirmed the monoliths to be comprised of densely packed UiO-66 NPs. 37 Figure 8| UiO-66 XRD patterns. Comparison of simulated XRD pattern for UiO-66 generated from its ideal crystal structure (black), to PXRD patterns of UiO-66 monoliths: UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green), confirming successful synthesis of the MOF in each case. 10 20 30 40 50 In te ns ity (a .u .) 10 20 30 40 50 Angle (2q) 10 20 30 40 50 10 20 30 40 50 10 20 30 40 50 38 1.3.3| Mechanical and Thermal Stability Mechanical properties are crucial in determining the potential industrial viability of monolithic materials towards gas-storage applications; they must be sufficiently robust to withstand the forces experienced during practical usage e.g. frequent impact during tank packing and vehicle transportation. Thus, the mechanical properties of monoliths UiO-66_A – D were characterised via nanoindentation (Figure 9). By this method, application of force by a nanosized indenting needle quantifies material properties including Young’s modulus (E) and Hardness (H). The recorded mechanical properties are a ‘snapshot’ of the mechanical strength in the material and offer an indication of its ability to withstand impact. Despite UiO-66 traditionally exhibiting high mechanical strength (attributable to the presence of 12 hard-hard Zr – O coordination bonds per metal cluster),8 literature mechanical properties vary significantly27 e.g. missing linker defects are reported to reduce mechanical stability whereas the presence of residual modulator within the MOF structure has been reported to increase it.28 Likewise, mechanical properties across materials UiO-66_A – D varied in terms of both Young’s modulus (E = 4.3 – 14.2 GPa) and Hardness (H = 0.11 – 0.48 GPa). The Young’s modulus of UiO-66_A (E = 10.1 ± 0.2 GPa) is consistent with the comparable sub- millimetre MOF samples (E = 10.5 ± 0.5 GPa) previously reported by Bennett et al.24 On the other hand, samples UiO-66_B (E = 6.9 ± 0.4 GPa), UiO-66_C (E = 4.3 ± 0.9 GPa) and UiO- 66_D (E = 14.2 ± 0.2 GPa) varied around this value, highlighting the significance of both washing and drying procedures on macroscopic mechanical properties. The variations in mechanical properties between these materials stems from a combination of differences in both particle packing and MOF composition due to changes in the washing solvent. However, significantly the mechanical properties of all samples reported here are comparable to those of robust monoliths which have previously been reported: monolithZIF-8 (E = 3.6 ± 0.2 GPa, H = 0.43 ± 0.03 GPa)22 and monolithHKUST-1 (E = 9.3 ± 0.3 GPa, H = 0.46 ± 0.03 GPa).23 The materials were further analysed by AFM after nanoindentation analysis (Figure 9a). Streaks in the obtained images are an artefact arising from the ‘dragging’ of contaminants over the materials surface by the microscope tip. This results from the aggressive mechanical polishing in the sample preparation procedure (Methods). In each case, the observed absence of radial cracking about the point of contact further suggests good mechanical stability in each of the materials. 39 Figure 9| Mechanical testing of UiO-66 monoliths. a, Load (mN) vs. Penetration into surface of the monolith (h, nm). The inset in a, 3D rendered AFM images showing the 3D topography of a resulting surface indent. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, as a function of Penetration depth (h, nm) into the monolith surface. Mean properties and corresponding errors (inset b and c) were obtained from measurements taken from 16 indents over penetration depths of 250 – 2000 nm. Measurements obtained in the sub- 250 nm penetration range were excluded, to eliminate errors due to surface defects/tip artefacts. Data correspond to UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green). 0 500 1,000 1,500 2,000 0.0 0.2 0.4 0.6 Ha rd ne ss , H (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 0 2 4 6 8 10 12 14 16 Yo un g's M od ul us , E (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 2,500 0 2 4 6 8 10 Penetration into surface, h (nm) Lo ad (m N) 0 500 1,000 1,500 2,000 2,500 0 5 10 15 20 25 Penetration into surface, h (nm) Lo ad (m N) 0 500 1,000 1,500 2,000 2,500 0 10 20 30 40 50 Penetration into surface, h (nm) Lo ad (m N) 0 500 1,000 1,500 2,000 0 1 2 3 4 5 6 7 Yo un g's M od ul us , E (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Ha rd ne ss , H (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 0 1 2 3 4 5 6 7 8 Yo un g's m od ul us , E (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 0.00 0.05 0.10 0.15 0.20 0.25 Penetration into surface, h (nm) Ha rd ne ss , H (G Pa ) 0 500 1,000 1,500 2,000 0 2 4 6 8 10 12 Yo un g's m od ul us , E (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 0.0 0.1 0.2 0.3 0.4 Ha rd ne ss , H (G Pa ) Penetration into surface, h (nm) a b c H = 0.33 ± 0.02 GPa E = 10.07 ± 0.21 GPa E = 6.85 ± 0.44 GPaH = 0.20 ± 0.01 GPa H = 0.11 ± 0.02 GPa E = 14.19 ± 0.21 GPaH = 0.48 ± 0.04 GPa E = 4.32 ± 0.89 GPa UiO-66_A UiO-66_B UiO-66_C UiO-66_D 0 500 1,000 1,500 2,000 2,500 0 10 20 30 40 50 60 70 Lo ad (m N) Penetration into surface, h (nm) Lo ad (m N) Lo ad (m N) Lo ad (m N) Lo ad (m N) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm)Penetration into surface, h (nm)Penetration into surface, h (nm) Penetration into surface, h (nm) Pen tration i t surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm)Penetration into surface, h (nm)Penetration into surface, h (nm) Ha rd ne ss , H (G Pa ) Ha rd ne ss , H (G Pa ) Ha rd ne ss , H (G Pa ) Ha rd ne ss , H (G Pa ) Yo un g’ s m od ul us , E (G Pa ) Yo un g’ s m od ul us , E (G Pa ) Yo un g’ s m od ul us , E (G Pa ) Yo un g’ s m od ul us , E (G Pa ) 40 Thermal stability is also important in determining the industrial viability of materials. In NG powered vehicles, long term stability of the adsorbent is decisive – consistently elevated temperatures may be expected as a result extensive heat dissipation from inefficient internal combustion engines.29 UiO-66 is the prototypical thermally stable MOF as a result of its exceptionally strong hard-hard Zr(IV) – O bonds.1 TGA was used to compare the thermal stability of monolithic UiO-66_D to UiO-66 powder obtained by a literature procedure over the temperature range 50 – 700 ºC (Figure 10).30 Decomposition temperature (ca. 550 °C) was unchanged between samples, with no alteration in thermal stability apparent as a result of monolith formation. The initial weight loss (%) observed in the powdered sample below 100 °C can be attributed to loss of foreign species (e.g. H2O, CO2, N2) adsorbed by the non-densified, porous sample from the air before the measurement was taken. The greater external surface area of the powdered material may facilitate faster uptake of species into pores than in the corresponding monolith of the same material, which can be expected to display reduced adsorption kinetics. Correspondingly, both the high thermal and mechanical stability of these monolithic materials point strongly to their industrial viability. Figure 10| Thermal stability of monolithic UiO-66. TGA comparison (50 – 700 °C, under N2 atmosphere) of UiO-66_D (green) with a powdered UiO-66 sample (black) prepared according to a literature procedure (Reference 30). 100 200 300 400 500 600 700 40 50 60 70 80 90 100 W ei gh t ( % ) Temperature (oC)Temperature (°C) W ei gh t ( % ) 41 1.3.4| Porosity and Density Experimental characterisation of a MOFs porosity is crucial in determining its relevance towards adsorption applications. Physical properties such as surface area, pore volume, and pore-size are fundamental in tuning the total gas adsorption capacity of a material i.e. availability of adsorption sites and strength of adsorbent-adsorbate interaction.13 Although the crystal structure of defect-free UiO-66 reveals a theoretical SBET of 1644 m2 g–1, reports of experimentally obtained UiO-66 vary significantly – Lillerud et al. first reported UiO-66 in 2008 with SBET = 1187 m2 g–1.1 These disparities can result from defects in the crystal structure such as missing linkers, missing clusters and inclusion of amorphous regions within the MOF.25,31 Furthermore, the synthesis of small MOF particles frequently corresponds with low surface area/porosity as a result of the inherently reduced long-range crystalline order in NPs compared to e.g. microparticles. Yet, Bennett et al. reported the primary NPs used in this study to exhibit high crystallinity with the resulting micro-MOF pieces displaying SBET which varied with primary particle drying procedure (1127 – 1459 m2 g–1).24 The significant impact of varying the washing solvent and drying temperature herein also requires further consideration. The solubility of reaction by-products and impurities varies with solvent, influencing the purity of the product. Residual impurities can occupy vital pore space, reducing accessible porosity and thus the availability of viable gas-adsorption sites.32 Activation, i.e. total evacuation of the porosity, is required for both accurate porosity characterisation and gas storage. While this activation is typically performed via heating under vacuum, the presence of large, high bp solvent molecules within the framework is known to induce pore-collapse via the high surface tension of the evacuating solvent molecules inside the pores.33 Considering the high bp (153 °C) and large kinetic diameter (Dk ~ 5.5 Å) of DMF (both a reactant, and in the case of UiO-66_C – D, a washing solvent) its occupation in the MOFs small pores (8 and 11 Å diameter) is likely. The air-dried monoliths were thus soaked firstly in acetone (bp = 56 °C, Dk = 4.6 Å)34 and subsequently in methanol (bp = 64.7 °C, Dk = 0.38 – 4.1 Å)35 with the aim of facilitating gentle diffusion of residual DMF through the pores and thus its replacement with smaller, low bp solvent molecules prior to vacuum activation. The porosity of the activated monolithic MOFs was subsequently characterised through N2 adsorption studies at 77 K. Firstly, each monolith displayed high N2 uptake for pressures below 0.1 bar, comparable to the isotherm simulated using the crystal structure of defect-free UiO-66 (Figure 11a). Gravimetric and volumetric SBET values (Table 3) were calculated using 42 Rouquerol’s consistency criteria (Appendix, Supplementary Figures 1 – 4).36 These values are consistent with the theoretical maximum, simulated using the crystal structure of defect- free UiO-66 (SBET = 1644 m2 g–1) which was calculated using GCMC simulations as well as the benchmark values reported by Lillerud et al. (SBET = 1187 m2 g–1).1 These combined data support the XRD, ICP and TGA results which characterise the obtained monolithic materials as crystalline UiO-66. This is further supported by Figure 11b which shows a semi-logarithmic plot of the N2 isotherm below 0.01 bar. The two-stage adsorption seen in each material is typical of sequential adsorption into the discrete tetrahedral (8 Å diameter) and octahedral (11 Å diameter) micropores of the MOF. Figure 11c shows the crystal structure of UiO-66, demonstrating how eight tetrahedral pores are arranged to form an octahedral pore in their centre. Figure 11| UiO-66 N2 adsorption isotherms. a, N2 isotherms showing gravimetric N2 uptake at 77 K (0 – 1 bar) for monoliths UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO- 66_C (purple squares) and UiO-66_D (green circles) as well as the theoretical isotherm simulated from the pristine crystal structure (white squares). b, Low pressure semi-logarithmic representation of each N2 isotherm (a) for P/Po < 0.001. c, Crystal structure of UiO-66 where elements correspond to: Zr (navy blue), O (sky blue), C (purple) and H (white). Tetrahedral pore faces are indicated in grey. P/PoP/Po a b 1E-08 1E-06 1E-04 1E-02 0 50 100 150 200 N 2 u pt ak e (S TP ) c m 3 g -1 P/Po 0.0 0.2 0.4 0.6 0.8 1.0 0 200 400 600 800 1,000 1,200 P/Po N 2 u pt ak e (S TP ) c m 3 g -1 N 2 up ta ke c m 3 (S TP ) g -1 N 2 up ta ke c m 3 (S TP ) g -1 c 43 Table 3| Physical properties, SBET, Wo and Vtot (gravimetric and volumetric) of monoliths UiO- 66_A – D (calculated from the experimentally obtained N2 adsorption isotherms, Figure 11a) compared to the theoretical values calculated for the ideal crystal structure. SBET (m2 g–1) Wo* (cm3 g–1) Vtot † (cm3 g–1) Vtot (vol) † (cm3 cm–3) UiO-66_A 1177 0.46 1.62 0.70 UiO-66_B 994 0.39 1.43 0.62 UiO-66_C 1065 0.42 0.81 0.69 UiO-66_D 982 0.38 0.57 0.60 Theoretical 1644 0.49 0.49 0.60 *Obtained at P/Po = 0.1, †Obtained at P/Po = 0.99 Despite the similarity of the experimental N2 adsorption isotherms to those of the theoretical MOF at low pressure (< 0.2 bar), at higher pressures (0.8 – 1.0 bar) significant differences are apparent. The distinctive Type IV shape of the experimental isotherms indicates the presence of mesoporosity within the macrostructures. The observed sudden increase in N2 uptake in this high-pressure range is a consequence of the mesopore filling mechanism – after multi-layer formation, capillary condensation of the gas takes place within the wider cavities. This is further supported by the presence of a hysteresis loop in the adsorption-desorption N2 isotherms (Figure 12). The hysteresis phenomenon has been extensively studied, being typical amongst mesoporous materials below a critical temperature.37 Its origin lies in delayed condensation and evaporation in/from the mesopores, a consequence of the narrower channels which connect the wide cavities.38 The materials meso-/macroporous volume may be quantified by the difference between the micro (Wo) and total (Vtot) pore volumes (Table 3). This confirms that significant volumes of mesoporosity exist in the monoliths UiO-66_A – D, crucially with relative levels of porosity varying significantly between them. 44 Figure 12| N2 desorption isotherms for UiO-66 monoliths. Adsorption (filled markers) and desorption (hollow markers) N2 isotherms collected at 77 K. a, UiO-66_A (blue triangles), b, UiO-66_B (red diamonds), c, UiO-66_C (purple squares) and d, UiO-66_D (green circles). To better elucidate the material’s PSDs, the N2 isotherms were further analysed using Tarazona NLDFT (Figure 13). The observation of significant volumes of porosity below 2 nm width is evidence of the high microporosity of each material, and further supports the calculated high SBET (Table 3) in each case. However, some variation in micro-PSD is clear, suggestive of slight micropore collapse or blocking. This further supports the observed variations in SBET, which ranged from 982 – 1177 m2 g–1 amongst the materials. Since each of the UiO-66 monoliths was obtained from the same microporous primary MOF particles, variations to the microporosity must result from pore collapse or evacuation during the differing drying procedures utilised to obtain each material. Since ICP-OES elemental analysis indicated near identical compositions in each case, it can be deduced that differences in SBET and thus microporosity are the result of minor levels of micropore collapse. This is further supported by the calculated Wo (Table 3) which show a suggestive 17% loss of microporosity in UiO-66_D compared to UiO-66_A. 0.0 0.2 0.4 0.6 0.8 1.0 0 200 400 600 800 1,000 1,200 N 2 u pt ak e (S TP ) c m 3 g- 1 P/Po 0.0 0.2 0.4 0.6 0.8 1.0 0 200 400 600 800 1,000 N 2 u pt ak e (S TP ) c m 3 g- 1 P/Po 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 500 600 N 2 u pt ak e (S TP ) c m 3 g- 1 P/Po 0.0 0.2 0.4 0.6 0.8 1.0 0 50 100 150 200 250 300 350 400 N 2 u pt ak e (S TP ) c m 3 g- 1 P/Po a b c d P/Po P/Po P/Po P/Po N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 45 Furthermore, the presence of porosity with width exceeding 2 nm diameter in the NLDFT PSD further indicates the presence of small volumes of mesoporosity in each material. The meso- PSD was calculated (Figure 14) by application of the BJH model to each of the N2 desorption isotherms (Figure 12). This supported the previous data which indicated, not only the presence of mesoporosity in each monolith in the 2 – 25 nm range, but also the differences in relative volume between them. The mesopore volumes follow the trend UiO-66_A > UiO-66_B > UiO- 66_C > UiO-66_D. Figure 13| NLDFT PSDs of UiO-66 monoliths. Distribution of micro- and mesopore width across monolith samples: UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO- 66_D (green), as obtained from Tarazona NLDFT model analysis of N2 isotherm data (Figure 11a). Microporous Microporous 10 100 0.00 0.05 0.10 0.15 0.20 0.25 Pore Width (Å) Pore Width (Å) Pore Width (Å) MesoporousMesoporous dV /d W P or e Vo lu m e (c m 3 g- 1 . Å) Pore Width (Å) Microporous 10 100 0.00 0.05 0.10 0.15 0.20 0.25 10 100 0.00 0.05 0.10 0.15 0.20 0.25 Mesoporous dV /d W P or e Vo lu m e (c m 3 g- 1 . Å) 10 100 0.00 0.05 0.10 0.15 0.20 0.25 dV /d W P or e Vo lu m e (c m 3 g- 1 . Å) MesoporousMicroporous dV /d W P or e Vo lu m e (c m 3 g- 1 . Å) Pore width (Å) dV /d W (c m 3 g-1 . Å ) dV /d W (c m 3 g-1 . Å ) dV /d W (c m 3 g-1 . Å ) dV /d W (c m 3 g-1 . Å ) Pore width (Å) Pore width (Å)Pore width (Å) 46 Figure 14| BJH PSDs of UiO-66 monoliths. Distribution of mesopore diameter across monolith samples: UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green), as obtained from BJH model analysis of the N2 isotherm data (Figure 12). These data establish the importance of the materials showing both variable meso- and macroporosity. Although larger pore volumes are generally associated with enhanced adsorption capacities on a gravimetric basis, they also incur lower volumetric adsorption. To further confirm the differences in porosity and thus density amongst the monoliths, Hg porosimetry was utilised (Figure 15). When each materials ρb was quantified, as expected, those with greater meso- and macropore volumes showed lower densities (Table 4). Thus, ρb monolithic UiO-66_D > UiO-66_C > UiO-66_B > UiO-66_A. Notably, ρb of monolithUiO-66_D (1.05 g cm–3) approaches the theoretical maximum crystal density of UiO-66 (1.24 g cm–3). This relatively high density may be postulated to derive from the efficient packing of small primary particles within the monolithic macrostructure, with the extraneous inter-particle space, which reduces ρb in e.g. powdered materials/pellets being minimized. 39,40 For example, Dhainut et al. previously reported the densification of UiO-66 powder at 18 MPa to obtain a UiO-66 pellet with 0.43 g cm–3 density.41 This pressure was selected as a compromise between maximising industrially significant physical properties, such as pellet density and mechanical strength, while minimising compressive loss of SBET. Crucially, UiO-66_D exhibited not only exceptional ρb but also high SBET and relatively low but significant mesopore volume. b 0 100 200 300 400 0.00 0.01 0.02 0.03 0.04 0.05 dV /d D cm 3 g -1 , Å Pore Diameter (Å) dV /d D Po re v ol um e (c m 3 g-1 . Å ) Pore diameter (Å) 47 Figure 15| Hg porosimetry PSDs. PSDs as obtained by Hg porosimetry showing extensive meso- (2 – 50 nm diameter) and macro- (> 50 nm diameter) porosity. Colours correspond to UiO-66_A (blue), UiO-66_B (red), UiO-66_C (purple) and UiO-66_D (green). The defect-free structure of UiO-66 is purely microporous, though crystalline defects (i.e. missing linkers/missing clusters) are known to physically manifest as mesopores. However, in the current case the observation that both meso- and macroporosity vary amongst monoliths prepared from the same crystalline primary MOF particles suggests that the origin of this variable porosity may be physical rather than chemical. Hence, additional porosity can be postulated to arise from interstitial space external to the MOFs crystal structure; it is between the primary particles of which each monolith is comprised. This observation of additional porosity is comparable to the isotherms obtained by Bennett et al. in the microscale MOF fragments.24 Yet the mesoporosity in those materials was uncontrollably high making it inapplicable for high density gas storage. Alterations to the synthetic work up offer the novel ability to finely tune the level of macroscopic mesoporosity and thus density. Since this observation is unique amongst these monolithic MOFs, elucidation of the physical origin of these variations is fundamental to our understanding of monolith formation. Table 4| rb of UiO-66 monoliths calculated by Hg porosimetry. Absolute (g cm–3) and relative values (%) compared to the theoretical maximum. UiO-66_A UiO-66_B UiO-66_C UiO-66_D Theoretical ρb (g cm–3) 0.430 (35%) 0.434 (35%) 0.848 (69%) 1.051 (85%) 1.237 (100%) Pore diameter (nm) 101 102 103 104 105 0.0 0.4 0.8 1.2 1.6 2.0 Pore diameter (nm) Macro Di ffe re nt ia l in tru sio n (m l g -1 ) Meso Di ffe re nt ia l i nt ru sio n (H g) m l g -1 48 1.3.5| Fluorescent Lifetime Imaging Microscopy FLIM utilises spatially resolved fluorescence lifetime decays to investigate a material’s nanoscopic environment.42 This provides additional information compared to traditional spatial mapping by e.g. SEM, such as the influence of morphology and defects on the materials physicochemical properties.43 To further elucidate the synthetic origins of the monolithic MOFs different textural properties and investigate their sub-structural morphologies, UiO-66_A – D were studied by FLIM. The monoliths were first gently ground into smaller pieces with a spatula such that the substructures could be monitored on the microscope. The obtained FLIM images (Figure 16a, b) of the auto-luminescing monoliths reveals them to be comprised of smaller particle aggregates, each with a distinct fluorescent lifetime. UiO-66_A showed high homogeneity, being comprised of only quazispherical aggregates with uniform fluorescent lifetimes of ~ 4.7 ns (orange). This differs from UiO-66_B, _C and _D, where two distinct aggregate morphologies co-exist within the same sample; both small, circular particles and larger rod-like particles. The circular particles (red-orange) exhibit a longer autofluorescence lifetime (~ 4.5 ns, orange) than the larger aggregates (blue-green) whose fluorescent decay is more rapid (~ 3 ns (blue)). The distinct morphologies of the aggregates of which each UiO-66 monolith appears to be comprised, decisively effect the fluorescence lifetime; large rod-like aggregates exhibit a faster fluorescence lifetime decay than small, circular particles. The phasor approach was used to further analyse the FLIM data (Figure 16c). This model graphically translates fluorescence lifetime into Fourier space where mono-exponential decays fall on an arc of radius 0.5; long lifetime components are located near the origin (0, 0) and short lifetime components near (1, 0). Since multi-exponential decays comprise a weighted vector of the constituent phasors, all decay pathways in phasor space must lie within the semi-circle.44,45 In the current case, each material, UiO-66_A – D, occupies phasor space within the arc, indicating multicomponent exponential decay pathways. In its phasor plot, a single population was recorded for UiO-66_A whereas plots for UiO-66_B – D appear heterogeneous, showing at least two distinct populations. This can be correlated with their biphasic morphology as described above and supports the proposed relationship between monolith textural properties and synthetic parameters. Since all monoliths were obtained from the same primary particles (ca. 10 nm), any changes in photo-physical properties must be caused by the way these identical particles interact with each other under the different 49 drying conditions. UiO-66_B, _C and _D showed a statistically greater prevalence of larger rod like aggregates within the monolith macrostructure than UiO-66_A. To incur the differences in fluorescent lifetime observed between these different sized aggregates, they may not be comprised of completely discrete particles (as they would be in powdered MOF) and instead an interaction on the molecular level is suggested. These data reinforce our understanding of the mechanism by which monoliths are formed, indicating that alterations to the synthetic procedure do influence the interaction of primary particles. Figure 16| FLIM studies of UiO-66 monoliths. a, b, FLIM images of UiO-66_A – D showing the aggregates that comprise each monolith. White dashed boxes (a, inset) indicate the areas selected for high magnification imaging (b). Colours correspond to excitation lifetime (see upper colour bar). c, 2D histogram phasor plots generated from FLIM images of each monolith. Colours correspond to the frequency of occurrence (see lower colour bar). An average of four images was used to generate each plot (Appendix, Supplementary Figures 5 – 8). 50 1.4| Conclusions The mechanism of dense monolith formation in the case of monolithZIF-8 and monolithHKUST-1 has previously been proposed by Fairen-Jimenez et al.22,23 They postulated that slow evaporation of the reaction solvent, coupled with the presence of residual precursors, facilitated primary particle interaction by effectively extending reaction time into the drying stage. As such, the small primary particles densified by centrifugation were proposed to undergo inter- particle epitaxial growth, accounting for the resulting polycrystalline monoliths which exhibited both ρb and mechanical properties comparable to the single crystal material. MOF epitaxy is the process by which consecutive layers of MOF are grown from a surface by the gradual addition of its constituent components i.e. metal/metal oxide clusters and organic linkers. Epitaxial growth of MOFs has been extensively studied, most notably in the production of SURMOFs. These surface anchored MOFs are grown by the epitaxial growth of the crystalline MOF on a target substrate which is functionalised with e.g. an organic molecule that mimics the chemical properties of the linker and hence facilitates MOF growth.46 In the current case, the UiO-66 primary particle can be regarded as the surface from which further MOF growth takes place. Hence, interstitial space is minimised by the epitaxial growth of the MOF into the void space between adjacent primary particles, the result of which may be chemical linking of previously distinct particles. The materials UiO-66_A – D obtained in this study are comparable to monolithZIF-8 and monolithHKUST-1, being likewise comprised of densely packed MOF NPs. The significant differences in physical properties between UiO-66_A – D, as outlined above, can be rationalised by the simple synthetic modifications made during the MOF drying stage. This offers a significant insight into the monolith formation mechanism. Firstly, the observation that washing solvent alters the optical properties of the materials is highly suggestive. UiO-66_A and UiO-66_B are opaque while UiO-66_C and UiO-66_D display optical transparency. The only synthetic difference between these materials is the use of ethanol or DMF to wash the primary particles prior to drying. Fairen-Jimenez et al. reported similar optical transparency for monolithZIF-8 and monolithHKUST-1, both of which were synthesised and washed in ethanol. This transparency was proposed to originate from the reduced electron contrast between the primary particles via the minimisation of the inter- particle barrier. Based on the results herein, it is not the use of ethanol which facilitates this but rather the use of a solvent which permits continued reaction during drying. In this case, the UiO-66 primary particles were synthesised in DMF. By this logic, the use of ethanol as the 51 washing solvent quenches the reaction during drying whereas the use of DMF facilitates its continuation. At 153 °C, DMF has a significantly higher bp than ethanol (78 °C). The slow evaporation of DMF at 30 °C enables, not only favourable reaction conditions, but also a greater timeframe over which drying, and thus the extended reaction, can take place. This is particularly significant considering that the formation of strong Zr–O bonds in UiO-66 is slow at room temperature, making extended reaction times further beneficial. The significance of drying temperature is also highlighted by comparing the size of macroscopic monoliths UiO-66_A and UiO-66_B. Although both materials were washed in ethanol, the rapid drying of UiO-66_A (200 °C) resulted in micron-sized fragments whereas UiO-66_B (30 °C) was obtained as a robust centimetre sized monolith. Since higher temperatures induce rapid solvent loss, the time frame over which inter-particle epitaxial growth can take place is reduced when using ethanol. Additionally, high temperatures induce fast removal of solvent from the interstitial spaces between primary particles causing stress at the vapour–liquid meniscus interface. This may destroy the gelatinous macrostructure to prevent monolith formation. These observations of extended primary particle interaction were also supported by the FLIM results. The micron sized aggregates observed by FLIM were significantly larger than the ca. 10 nm primary particles used in this study. The fact that these large aggregates showed distinct fluorescence confirms them to be not just physically compacted primary particles, but chemically connected. For this chemical interaction to take place, the particles may undergo epitaxial inter-particle growth during drying. UiO-66_A, quenched by washing in ethanol and rapid drying at 200 °C, showed very small and homogenous particles. Alternatively, UiO-66_B, UiO-66_C and UiO-66_D, which were all dried at 30 °C, showed less homogenous and often elongated aggregates. As a consequence of the reduced drying temperatures, the extended drying time appeared to facilitate aggregate formation. UiO-66_D showed the largest aggregates, supporting the hypothesis that both usage of a solvent that facilitates reaction continuation and extended/mild drying conditions encourage primary particle epitaxial growth. The observed variations in mesoporosity and density were also supported by this mechanism. The near identical densities of highly mesoporous UiO-66_A (ρb = 0.430 g cm –3) and UiO- 66_B (ρb = 0.434 g cm –3), dried at 200 °C and 30 °C respectively, suggest that while macroscopic monolith size is dependent on drying temperature, ρb is not. In contrast, high 52 density monolithUiO-66_C (ρb = 0.834 g cm –3) demonstrated a significant reduction in interstitial space/mesoporosity, with the only synthetic difference being washing solvent (i.e. DMF). Again, the slow evaporation of this solvent during drying enables both primary particle densification via the preservation of the gelatinous particle macrostructure as well as continued primary particle interaction, controlling both the density and porosity of the monolith. Finally, at 1.05 g cm–3, UiO-66_D reveals the highest ρb amongst the synthesised materials. In this case, the extended centrifugation period applied prior to drying evidently facilitated better primary particle compaction to minimise meso-/macroporosity. These observations are especially significant when the high SBET and microporosity of each monolith is considered. Pore collapse in materials obtained through traditional densification procedures (e.g. applied pressures) renders them unsuitable for prospective physisorptative applications (e.g. dense gas storage).47 Crucially the density was varied from 35 – 85% of the theoretical maximum, with high retention of microporosity. This differs from previous reports of monolithic MOFs where monolithZIF-8 and monolithHKUST-1 reported exclusively microporous character. The novel capacity to tune PSD in the same MOF, UiO-66, makes the materials reported herein a highly promising means of experimentally studying the influence of PSD on gas adsorption capacity. For the first time, the capability to synthetically tune the density of pure, monolithic MOFs has been demonstrated without significant collapse/blocking of microporosity. Furthermore, the physical origin of this monolith formation has been elucidated through extensive materials characterisation. This expands not only our understanding of the monolith formation mechanism but also advances the capabilities of synthetic MOF design to include tuneable, non-crystalline porosity external to the MOFs defined structure with a high level of experimental control. This first report of a densified monolithic Zr-MOF, displaying physical properties (i.e. rb, SBET, Vtot) comparable to the single crystal material is highly promising. Zr- MOFs in general are much sought after, a consequence of their desirable properties which include high thermal, chemical and mechanical stability. Consequently, a logical question is raised by these results: Can the generality of the synthetic procedure be demonstrated via translation to further monolithZr-MOFs? 53 2.0 | UiO-66-NH2 By applying mono-functionalised bdc linkers to the original synthesis of UiO-66, Lillerud et al. subsequently reported a number of isoreticular analogues; UiO-66-NH2, UiO-66-NO2 and UiO-66-Br.48 Not only was the characteristic thermal and chemical stability of the parent Zr- MOF retained despite functionalisation, but H/D exchange experiments were further utilised to demonstrate that functionalities localised within the MOFs internal structure remained chemically accessible via its porosity. Consequently, the known Zr-MOF family was rapidly expanded to include a wide range of UiO-66-X analogues by applying numerous functionalised linkers (Figure 17).49 As a consequence of the parent MOFs ability to withstand extensive functionalisation, coupled with the relative ease of such functionalisation, the unique chemical and physical properties offered by UiO-66-X materials have been extensively studied.16,49,50 Of note is the aminated analogue, UiO-66-NH2, which offers potential in diverse applications i.e. as a photocatalyst in reactions such as CO2 reduction51,52 and oxidative coupling53 as well as selective adsorption. For example, its liquid phase adsorptive properties have been extensively studied for e.g. water purification, with promising results in the selective removal of toxic cationic dyes,54 fluoride55 and heavy metals.56 Furthermore, it has demonstrated great promise in gas phase adsorption including removal of gaseous pollutants (NO2,57 NH3,58 CNCl,58 Cl259 and HCl59) in addition to enhanced storage of CO2 relative to unfunctionalised UiO-66.13 However, as with other MOFs, the significant potential of this otherwise promising material is limited by its inability to be processed into a functional material without damage to its desired physical properties. Peterson et al. reported the post-synthetic processing of UiO-66-NH2 powder into sub-mm sized granules under the application of pressure.59 However under the minimum pressure used in the study, 5000 psi (ca. 350 bar), the surface area of the unprocessed powder (SBET = 1074 m2 g–1) underwent significant reduction (-23%), resulting in a 35% loss of total pore volume, while the granules obtained under the maximum pressure, 50000 psi (ca. 3450 bar) exhibited a massive 45% loss of SBET. 54 Figure 17| Functionalised precursors to UiO-66-X. a, b, Chemical structures of mono- and di- functionalised terephthalic acid derived precursors, respectively, which have previously been used to synthesise correspondingly functionalised UiO-66-X; X = H, F, Cl, Br, I, CH3, CF3, NO2, NH2, OH, OCH3, COOH, SO3H etc. COOH COOH X COOH COOH X X a b 55 2.1| Aims and Objectives The aim of this section is to explore the generality of the monolith synthesis procedure, reported in Chapter II, Section 1.0, towards a wider range of zirconium MOFs by extending it to functionalised UiO-66-NH2. The physical properties of obtained materials will be fully characterised by a range of analytical techniques and benchmarked against both the single crystal properties of UiO-66-NH2 as well as monolithUiO-66. In particular the novel capability of the developed monolith procedure to synthetically tune mesoporosity within the macrostructure will be studied. 56 2.2| Monolith Synthesis With the aim of synthesising monolithic UiO-66-NH2 comparable to the unfunctionalized monolithUiO-66, a synthetic procedure for an equivalent MOF gel was first targeted. In synthetic MOF chemistry, generic swapping of organic linker molecules does not necessarily correspond to successful formation of the desired crystalline MOF. This stems from variations in both the linker solubility and affinity for the metal node/cluster which are inherently highly variable amongst the wide range of organic molecules used as linkers in MOFs. In the current case, the presence of the polar amine group in 2-aminobenzene-1,4-dicarboxylic acid is expected to enhance this organic molecule’s solubility in the reaction solvent, polar DMF, compared to unfunctionalized 1,4-benzene dicarboxylic acid. This may be expected to alter the rate of MOF crystallisation and precipitation, which in turn would alter MOF primary particle size or crystallinity. This is critical in obtaining the desired MOF NPs which must be simultaneously small enough for monolith formation while maintaining high surface area/porosity. As a preliminary experiment, the discussed synthesis of UiO-66 gel (Chapter I, Section 1.0) was replicated, replacing only the 1,4-benzene dicarboxylic acid linker with a molar equivalent quantity of 2-aminobenzene-1,4-dicarboxylic acid. Despite the altered solubility of the amine- functionalised linker, UiO-66-NH2 was yielded as a viscous yellow gel (Figure 18a). The high viscosity of the material was comparable to UiO-66 gel (Figure 4a), which likewise did not flow upon inversion of the reaction vessel. These physical observations are indicative of the presence, once again, of small MOF NPs in the desired gelatinous macrostucture arrangement. Figure 18| Optical and TEM images of UiO-66-NH2 gel. a, Optical image of UiO-66-NH2 gel. b, c, Low and high magnification, respectively, TEM images UiO-66-NH2 gel. a b c 57 The UiO-66-NH2 gel was analysed by TEM (Figure 18b, c), confirming it to comprise ca. 10 nm gelatinous MOF NPs, macroscopically arranged into a loosely packed network structure. This result is critical as primary particle size, a central factor in monolith formation, was maintained despite functionalisation of the MOF linker. However, of note is the presence of a secondary phase, visible by TEM as darker spots within the primary particles (Figure 18c). Similar defects were previously observed in the primary MOF particles used to synthesise monolithHKUST-1.23 This was attributed to the presence of dense, non-crystalline defects within the crystalline primary particles, originating from the high speed synthesis used to obtain the small particles needed for monolith formation. As discussed, the presence of the amine group in the UiO-66-NH2 linker increases this reagent’s solubility in polar DMF relative to that of bdc, the UiO-66 linker. This may facilitate more rapid UiO-66-NH2 primary particle nucleation, the result of which is an increase in defect prevalence. Despite these minor defects, the obtained MOF gel was highly comparable to that obtained in the unfunctionalized UiO-66 gel in terms of particle size and macroscopic particle arrangement. In this case, simple linker exchange can be used to obtain a macro- and microscopically comparable functionalised MOF gel without further modification to the synthetic procedure. The obtained gel was dried under a range of conditions with the aim of replicating the varied physical properties of the UiO-66 monoliths. The drying conditions used for the synthesis of UiO-66_B – D (Table 1) were thus applied to the UiO-66-NH2 gel (Table 5), yielding UiO-66- NH2_A – C respectively (Figure 19). The harsh drying conditions for UiO-66_A were disregarded based on the undesirably small, sub-millimeter size of the MOF fragments it formed. Figure 19| Optical images of amine functionalised monoliths. Optical images of UiO-66- NH2 monoliths corresponding to a, UiO-66-NH2_A, b, UiO-66-NH2_B and c, UiO-66-NH2_C. 10 mm UiO-66-NH2_A UiO-66-NH2_B 10 mm 10 mm UiO-66-NH2_C b c a 58 Table 5| Experimental conditions for the synthesis of monolithUiO-66-NH2 from MOF gel. Washing Procedure Centrifugation Drying Temperature UiO-66-NH2_A Ethanol (3 × 30 mL) 3 × 10 min* 30 °C UiO-66-NH2_B DMF (1 × 30 mL) 1 × 10 min* 30 °C UiO-66-NH2_C DMF (1 × 30 mL) 1 × 10 min* + 1 × 180 min** 30 °C *Centrifugation (5500 rpm) performed after each wash to re-obtain MOF gel as sediment. **Additional 180 min (5500 rpm) centrifugation performed on densified MOF gel after washing in DMF and decanting the supernatant. Each of the obtained materials, UiO-66-NH2_A – C, (Figure 19) displayed qualitatively comparable physical properties to their UiO-66_B – D counterparts (Figure 5). Firstly, each aminated material was synthesised as robust macroscopic fragments, typically exceeding several millimetres in size. Furthermore, the optical transparency of each monolith is of significant note, being comparable to monoliths of UiO-66 obtained under identical experimental conditions. UiO-66_B and UiO-66-NH2_A were both obtained by washing the MOF gel in ethanol and drying at 30 °C. Both monolithic materials show lower optical transparency than monoliths obtained by washing in DMF i.e. UiO-66_C – D and UiO-66- NH2_B – C. This is consistent with the proposed monolith formation mechanism, as discussed extensively in Chapter I, Section 1.0. 59 2.3| Monolith Characterisation 2.3.1| Morphology and Crystal Structure The fully dried UiO-66-NH2 monoliths were studied by TEM (Figure 20). The results showed that the gelatinous network structure adopted by the ca. 10 nm diameter primary particles (Figure 18b, c) was densified via the drying process. These combined observations (particle size, gelatinous macrostructure and densification upon drying) are consistent with comparable observations in UiO-66 gel and monolithUiO-66 (Figure 6), supporting the proposed mechanism of monolith formation via primary MOF NP densification. Figure 20| TEM images of monolithic UiO-66-NH2. a, b, Low and high magnification, respectively, TEM images of dried UiO-66-NH2_C. The crystal structure of the monoliths was further studied by PXRD, characterising the materials as crystalline UiO-66-NH2 in each case (Figure 21). The observation of wide peaks in the XRD patterns is a consequence of line broadening, characteristic of the nano-sized particles (as supported by TEM, Figure 18 and Figure 20). This is further consistent with the broad diffraction peaks observed in unfunctionalized monolithUiO-66 (Figure 8). a b 60 Figure 21| XRD patterns of UiO-66-NH2 monoliths. Comparison of simulated XRD pattern for UiO-66-NH2 generated from its ideal crystal structure (black, Reference 59), to PXRD patterns experimentally obtained for the monolithic MOFs: UiO-66-NH2_A (green), UiO-66- NH2_B (blue) and UiO-66-NH2_C (purple). 10 20 30 40 50 10 20 30 40 50 10 20 30 40 50 10 20 30 40 50 UiO-66-NH2_Simulated UiO-66-NH2_C UiO-66-NH2_B In te ns ity (a .u .) UiO-66-NH2_A Angle (2q)Angle (2q) In te ns ity (a .u .) 61 2.3.2| Thermal and Mechanical Stability The thermal stability of monolithic UiO-66-NH2 was characterised by TGA. Analysis of UiO- 66-NH2_C (Figure 22, purple line) demonstrates high thermal stability, with complete decomposition taking place at ca. 500 °C. This is characteristic of thermally stable Zr-MOFs. Both this decomposition temperature and the shape of the trace, being a gradual slope, are near identical to literature reports of powdered UiO-66-NH2 (Figure 22, black line).60 This supports the materials characterisation as the amine functionalised analogue of UiO-66. Furthermore, the thermal stability can be compared to analogous UiO-66_D. While the ultimate decomposition at ca. 500 °C is similar, a consequence of their comparably strong Zr–O bonds, the overall shape of the traces differs. The more gradual decomposition slope in UiO-66-NH2 corresponds to a slightly reduced thermal stability for the aminated linker. Figure 22| Thermal stability of UiO-66-NH2. TGA traces comparing the thermal decomposition of UiO-66-NH2_C (purple), powdered UiO-66-NH2 (black, digitised from Reference 60) and UiO-66_D (green) over the temperature range 50 – 650 °C (under N2 atmosphere). The mechanical properties of the obtained UiO-66-NH2 monoliths were quantified by nanoindentation (Figure 23). The obtained values for Hardness (H = 0.38 – 0.61 GPa) and Young’s modulus (E = 9.6 – 13.7 GPa) varied amongst the monolithic materials, highlighting the significance of experimental drying conditions. The average mechanical properties for individual materials were fairly homogenous across the repeated nano-indents, as indicated by the small error bars in each case (Figure 23b, c), as well as the small spread of Load vs. 50 60 70 80 90 100 50 150 250 350 450 550 650 Temperature (oC) W ei gh t ( % ) 62 Penetration curves (Figure 23a). Notably, the robust mechanical properties are highly comparable to those of UiO-66_A – D (H = 0.11 – 0.48 GPa, E = 4.2 – 14.3 GPa) which likewise demonstrated variation between materials, likely a consequence of alterations in primary particle packing. Furthermore, the macroscopic mechanical properties of each UiO-66-NH2 are similar to those of robust and industrially viable monolithZIF-8 (E = 3.57 ± 0.22 GPa, H = 0.43 ± 0.03 GPa)22 and monolithHKUST-1 (E = 9.3 ± 0.3 GPa, H = 0.46 ± 0.03 GPa).23 This demonstrates the potential of the developed procedure for monolithZr-MOF synthesis to yield a range of industrially significant materials. 63 Figure 23| Mechanical testing of monolithUiO-66-NH2. Nanoindentation data for UiO-66- NH2_A (green), UiO-66-NH2_B (blue) and UiO-66-NH2_C (purple) showing a, Load (mN) vs. Penetration into the monolith surface (h, nm) for 16 indents. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, plotted as a function of Penetration depth into the monolith surface (h, nm). Mean properties and corresponding errors (inset) were obtained from measurements taken across 16 indents over penetration depths of 250 – 2000 nm. Measurements obtained in the sub-250 nm penetration range were excluded with the aim of eliminating errors due to surface defects/tip artefacts. UiO-66-NH2_C UiO-66-NH2_B 0 500 1,000 1,500 2,000 0 2 4 6 8 10 12 Yo un g' s M od ul us , E (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 2,500 0 10 20 30 40 50 Lo ad (m N ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 0.0 0.1 0.2 0.3 0.4 0.5 H ar dn es s, H (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 0 3 6 9 12 15 Yo un g' s M od ul us , E (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 2,500 0 10 20 30 40 50 60 70 80 Lo ad (m N ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 0.0 0.2 0.4 0.6 0.8 Penetration into surface, h (nm) H ar dn es s, H (G Pa ) 0 500 1,000 1,500 2,000 0 2 4 6 8 10 12 14 Yo un g' s M od ul us , E (G Pa ) Penetration into surface, h (nm) 0 500 1,000 1,500 2,000 2,500 0 10 20 30 40 50 Penetration into surface, h (nm) Lo ad (m N ) 0 500 1,000 1,500 2,000 0.0 0.1 0.2 0.3 0.4 0.5 H ar dn es s, H (G Pa ) Penetration into surface, h (nm) H = 0.61 ± 0.06 GPa E = 13.7 ± 0.2 GPa H = 0.38 ± 0.02 GPa E = 9.6 ± 0.1 GPa H = 0.40 ± 0.02 GPa E = 9.6 ± 0.1 GPa a b c Lo ad (m N ) UiO-66-NH2_A Lo ad (m N ) Lo ad (m N ) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm)Penetration into surface, h (nm) H ar dn es s, H (G Pa ) H ar dn es s, H (G Pa ) H ar dn es s, H (G Pa ) Yo un g’ s m od ul us , E (G Pa ) Yo un g’ s m od ul us , E (G Pa ) Yo un g’ s m od ul us , E (G Pa ) 64 2.3.3| Porosity and Density The porosity of the activated monolithic MOFs was extensively characterised through N2 adsorption studies at 77 K, enabling comparison with both theoretical, defect-free UiO-66-NH2 and monolithUiO-66. Firstly, the simulated crystal structure of defect-free UiO-66-NH2 reveals a maximum SBET of 921 m2 g–1. This is notably lower than the theoretical maximum of defect free UiO-66 (maximum SBET = 1644 m2 g–1), an inherent consequence of the MOFs amine functionality, which protrudes into the MOFs micropores.48 This reduces both Vtot and SBET relative to the unfunctionalized analogue. Despite the observed presence, by TEM, of dense non-crystalline regions in the primary particles of this aminated MOF (Figure 18), SBET of the monoliths approached the theoretical maximum (Table 6). This implies that the secondary, dense phase is a minor defect compared to the abundant crystalline MOF. Furthermore, the calculated SBET varies to some extent amongst the obtained monoliths. This is comparable to monoliths of unfunctionalized UiO-66 in which surface area was observed to vary as a function of drying conditions (Table 3). Again, analogously to UiO-66_D, UiO-66-NH2_C demonstrates a slight reduction in SBET and Wo compared to both the theoretical maximum and to the other experimental monolithMOFs in the series. Since both UiO-66_D and UiO-66-NH2_C were prepared via extended centrifugation, this result supports the hypothesis that a slight collapse of microporosity takes place as a consequence of this synthetic protocol. Figure 24| N2 adsorption isotherms for UiO-66-NH2. Linear N2 adsorption isotherms (collected at 77 K) for monoliths UiO-66-NH2_A (green diamonds), UiO-66-NH2_B (blue squares) and UiO-66-NH2_C (purple circles) compared to the theoretical adsorption isotherm for defect-free UiO-66-NH2 (red dots). 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 500 600 700 800 900 P/Po N 2 up ta ke (S TP ) c m 3 g -1 P/Po N 2 up ta ke (S TP ) c m 3 g– 1 65 Table 6| Physical properties (SBET, Wo and Vtot) of monoliths UiO-66-NH2_A – C (calculated from the experimentally obtained N2 adsorption isotherms, Figure 24) compared to the theoretical values calculated for the ideal crystal structure. *Calculated according to Rouquerol’s consistency criteria36 (Appendix, Supplementary Figures 9 – 11). †Obtained at P/Po = 0.1 ‡Obtained at P/Po = 0.99. The comparable nature of the SBET data for each monolithic MOF is a consequence of their having been synthesised from identical primary UiO-66-NH2 particles with equivalent microporosity. However, the differing N2 uptakes between 0.8 – 1.0 P/Po in their Type IV isotherms (Figure 24) indicates these macrostructures to also contain differing volumes of mesoporosity, comparable to monolithUiO-66. This view is further supported by the presence of a hysteresis loop in the adsorption-desorption N2 isotherms (Figure 25) – the origin of which was extensively discussed earlier (Chapter I, Section 1.0). The difference between the materials Wo and Vtot values (Table 6) demonstrates that UiO-66-NH2_A exhibits the greatest volume of mesoporosity, while UiO-66-NH2_B and UiO-66-NH2_C show relatively reduced mesopore volumes. These results are again consistent with the variations in mesoporosity observed in UiO-66_B – D. This supports the proposed hypothesis, as discussed earlier, regarding the effect of solvent and temperature on the primary particle interaction during drying and consequently the density and mesoporosity of the obtained material. SBET* (m2 g–1) Wo† (cm3 g–1) Vtot‡ (cm3 g–1) UiO-66-NH2_A 841 0.33 1.23 UiO-66-NH2_B 822 0.32 0.56 UiO-66-NH2_C 665 0.26 0.43 Theoretical 921 0.33 0.38 66 Figure 25| N2 adsorption-desorption isotherms for monolithUiO-66-NH2. N2 adsorption (filled markers) and desorption (hollow markers) isotherms (recorded at 77 K) for monolithic MOFs a, UiO-66-NH2_A (green diamonds), b, UiO-66-NH2_B (blue squares) and c, UiO-66-NH2_C (purple circles). As was the case for monolithUiO-66, the differences in porosity amongst the obtained monolithUiO- 66-NH2 were further quantified by applying Tarazona NLDFT (Figure 26) to the N2 isotherms (Figure 24). By this model the PSDs were visualised to confirm, as expected, extensive microporosity (sub-2 nm). The observed alterations in Wo succinctly demonstrate that modification to the synthetic drying procedure influences not only the macrostructure adopted by the primary particles during drying but also the crystalline micropores of the particles themselves. This further supports the calculated variations in SBET amongst the monoliths (Table 6). The PSDs further indicate this microporosity to be extensive, with relative volumes of mesoporosity (2 – 25 nm) being comparably small for each material. 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po 0.0 0.2 0.4 0.6 0.8 1.0 0 200 400 600 800 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po a b 0.0 0.2 0.4 0.6 0.8 1.0 0 50 100 150 200 250 300 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po P/Po P/Po P/Po N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 c N 2 up ta ke (S TP ) c m 3 g– 1 N 2 up ta ke (S TP ) c m 3 g– 1 N 2 up ta ke (S TP ) c m 3 g– 1 P/Po P/Po P/Po 67 Figure 26| NLDFT PSDs of UiO-66-NH2 monoliths. Distribution of micro- and mesopore width across monolithic MOFs: UiO-66-NH2_A (green), UiO-66-NH2_B (blue) and UiO-66- NH2_C (purple), as obtained from Tarazona NLDFT analysis of the N2 isotherm data (Figure 24). BJH analysis (Figure 27) of the desorption isotherms (Figure 25) shows near identical dimensions (ca. 50 Å diameter) and volumes (ca. 0.015 cm3 g–1) of mesoporosity in UiO-66- NH2_B and UiO-66-NH2_C while UiO-66-NH2_A exhibits significantly larger pores (< 150 Å diameter). For each material, both the size and volume of mesoporosity are near identical to that recorded in their unfunctionalized analogues; UiO-66_B – D. a 10 100 0.00 0.05 0.10 0.15 0.20 MesoporosityMicroporosity dV /d W (c m 3 g -1 . Å ) Pore width (Å) 10 100 0.00 0.05 0.10 0.15 0.20 MesoporosityMicroporosity dV /d W (c m 3 g -1 . Å ) Pore width (Å) 10 100 0.00 0.05 0.10 0.15 0.20 MesoporosityMicroporosity dV /d W (c m 3 g -1 . Å ) Pore width (Å) b 0.00 0.01 0.02 0 50 100 150 200 250 300 dV /d D (c m 3 g -1 .Å ) Pore diameter (Å) a b c 68 Figure 27| BJH PSDs of UiO-66-NH2 monoliths. Distribution of mesopore diameter across monolith samples: UiO-66-NH2_A (green), UiO-66-NH2_B (blue) and UiO-66-NH2_C (purple), as obtained from BJH model analysis of the N2 isotherm data (Figure 25). a 10 100 0.00 0.05 0.10 0.15 0.20 MesoporosityMicroporosity dV /d W (c m 3 g -1 . Å ) Pore width (Å) 10 100 0.00 0.05 0.10 0.15 0.20 MesoporosityMicroporosity dV /d W (c m 3 g -1 . Å ) Pore width (Å) 10 100 0.00 0.05 0.10 0.15 0.20 MesoporosityMicroporosity dV /d W (c m 3 g -1 . Å ) Pore width (Å) b 0.00 0.01 0.02 0 50 100 150 200 250 300 dV /d D (c m 3 g -1 .Å ) Pore diameter (Å)Pore diameter (Å) dV /d D (c m 3 g– 1 . Å) 69 2.4| Conclusions A range of monolithUiO-66-NH2 were successfully synthesised by facile modification of the monolithUiO-66 synthetic procedure, reported in Chapter I, Section 1.0, by exchange of the bdc linker with an aminated analogue. Not only was a comparable functionalised gel of MOF particles obtained but these particles maintained the appropriate nanoscale dimensions for robust monolith formation. Identical drying procedures to those used for UiO-66_B – D were applied to the amine-functionalised gel and the resulting monoliths, UiO-66-NH2_A – C, were studied by TEM, XRD, TGA, nanoindentation and N2 adsorption studies. Robust mechanical properties, thermal stability, high crystallinity and porosity of each novel UiO-66-NH2 monolith were recorded. Besides the obvious potential applications this high density, porous monolith offers in dense gas storage, this pelletised material presents a promising practical avenue by which to achieve the selective gas adsorption that this aminated MOF is recognised for.59 Porous materials must be pelletised prior to packing into gas separation columns to prevent in situ compaction which could cause pressure drops or blockages.61 Furthermore, the pelletised nature of this material further enhances its industrial applicability towards water purification, an area in which it has demonstrated great proof of concept potential if low practical viability.54–56 Wheatley et al. previously demonstrated the ability of monolithZIF-8 to host photocatalytic SnO2 NPs and established the composites capacity to efficiently degrade MB, a toxic water pollutant.62 In this paper, the commercial benefits of monolithic materials, compared to powdered analogues, were discussed in terms of their high recyclability and ease of use. Furthermore, the distinctive optical transparency of monolithic UiO-66-NH2 materials may further offer substantial potential towards light driven applications such as photocatalytic reactions.52,53 Finally, as with monolithUiO-66, the novel capability to vary PSD in the UiO-66-NH2 monoliths through synthetic variation of the gels drying conditions was replicated. This result is highly promising as it not only demonstrates the generality of the developed synthetic procedure but further supports the proposed monolith formation mechanism. The trends in variation of macroscopic physical properties, i.e. mesoporosity, amongst the monolithUiO-66-NH2, were near identical to the trends observed amongst the monolithUiO-66 analogues. In particular, the relationship between primary particle interaction during drying and washing solvent/drying temperature is highlighted. These results are critical to suggesting that the novel procedure for 70 monolithic Zr-MOFs synthesis developed herein can be easily modified by linker alteration. An extensive range of functionalised UiO-66-X analogues have been reported (e.g. X = H, F, Cl, Br, I, CH3, CF3, NO2, NH2, OH, OCH3, COOH, SO3H) with unique chemical and physical properties. The developed procedure for monolithic Zr-MOF synthesis thus enables selection and synthesis of different monoliths with different functionalities based on the target application. The aim of this section was to determine the generality of the developed synthetic protocol for robust Zr-monolithic MOF formation. While this was successfully demonstrated, the promising positive results obtained by facile linker exchange raise further questions. For example, to what extent does monolith formation permit more complex functionalisation of the bdc linker? 71 3.0| UiO-66-ndc Besides the gas storage, capture and sieving applications for which MOFs are well known, the potential of these extensively tuneable structures towards a much broader range of applications has been studied, from drug delivery and biological macromolecule encapsulation63 to selective catalysis and supercapacitors.17 For example, chemical sensing is an application of significant interest.64 These porous materials offer numerous benefits over traditional sensing materials; analytes can be concentrated within the porosity and furthermore the pore size and functionality can be selectively tailored for the study of specific analytes. In particular, the unique ability of MOFs to combine metal ions with aromatic and conjugated p systems in a crystalline arrangement, offers a number of luminescent pathways including organic linker emission and ligand/metal charge transfer.65 Since luminescence emission can be both enhanced and quenched by the MOFs interaction with a wide range of target analytes (including pH changes, small ionic species and large organic molecules), MOF-derived luminescent sensors have seen increasing interest over recent years.66 For example, Liu et al. demonstrated the highly selective luminescent quenching of a lanthanide based MOF in the presence of Fe3+ and Ce3+ cations, as well as acetone. This is relevant from an environmental perspective e.g. for water contamination testing.67 Yet, as with gas storage, the biggest problem with the practical application of luminescent MOFs lies in their traditional morphology as bulk microcrystalline powders. The significant light scattering that powdered MOFs exhibit diminishes detectable photo-efficiency.68 Although there has been extensive research into the development of luminescent MOFs on the molecular level, little success has been found in the processing of functional, optically transparent MOF materials with minimised scattering. For example, non-crystalline MOF- glasses are rendered inapplicable due to their low microporosity, despite being optically transparent.69 The previously reported monolithic MOFs, monolithZIF-822 and monolithHKUST-1,23 as well as the materials developed herein, monolithUiO-66 and monolithUiO-66-NH2, all display optical transparency, a rare feature amongst bulk MOF materials. Despite their glassy, transparent appearance, porosity and surface area comparable to their powdered, microporous counterparts is retained. This combination of porosity and optical transparency in a bulk material presents a unique opportunity for the synthesis of functional fluorescing materials with minimised light scattering for high efficiency chemical sensing. 72 Since the first reports of UiO-66 and its monosubstituted derivative structures, a more extensive range of derivative MOFs have also been reported e.g. containing ortho, meta and para disubstituted terephthalic acid linkers (Figure 28). Ortho di-substituted structures are particularly relevant to luminescent applications as they facilitate the inclusion of larger, conjugated p systems. Naphthalene-1,4-dicarboxylic acid (Figure 28d) is a disubstituted terephthalic acid derivative which has been applied to the synthesis of functionalised UiO-66. Light sensitised by its 10p aromatic ring system, UV absorbing UiO-66-ndc is a strongly luminescent MOF. Recently, Allendorf et al. reported the synthesis of UiO-66-ndc nanocrystals with the aim of minimising light scattering.68 While particles with diameter half that of the incident light wavelength maximise scattering, this rapidly decreases for smaller particles.70 Accordingly, Allendorf reported that 35 nm diameter UiO-66-ndc particles displayed optical transmittance up to 2 orders of magnitude greater than the 250 nm diameter particles of the same MOF. Considering these results, the ca. 10 nm crystalline particles which make up optically transparent monolithUiO-66 and monolithUiO-66-NH2 present an interesting research avenue for the synthesis of luminescent, monolithic MOFs for e.g. chemical sensing applications. Figure 28| Di-functionalised terephthalic acid derived structures. Chemical structures of a, ortho, b, meta and c, para di-substituted terephthalic acid molecules used to synthesise functionalised UiO-66 derivative structures. d, Chemical structure of ndcH2. COOH COOH X X COOH COOH XX COOH COOH X X a b c d COOH COOH 73 3.1| Aims and Objectives The aim of this part of the study is to further demonstrate the generality of the monolithZr-MOF synthesis by extending it to more complex di-functionalised UiO-66 derivatives. In particular the photosensitised organic linker, naphthalene-1,4,-dicarboxylic acid, will be used to synthesise UiO-66-ndc as a gel of primary MOF NPs. The effect of linker functionalisation on the physical properties (i.e. crystallinity, porosity and particle size) of the obtained MOF will be further studied by a range of characterisation techniques. The ultimate aim is to determine if the functionalised MOF can be obtained as an optically transparent monolithic MOF, monolithUiO-66-ndc, with porosity and crystallinity comparable to the bulk powder of the same material. This may render it of interest in luminescence applications. 74 3.2| Monolith Synthesis 3.2.1| Preliminary Synthesis As a preliminary reaction (UiO-66-ndc_1), the terephthalic acid linker used for the synthesis of UiO-66 gel was swapped with a molar equivalent quantity of naphthalene-1,4-dicarboxylic acid. This was also identical to the procedure previously used to synthesise aminated MOF gel, UiO-66-NH2. However, in the current case it failed to produce a gelatinous suspension and instead resulted in a fine white powder. As expected, when the Zr-monolithMOF drying procedure (used for UiO-66_C and UiO-66-NH2_C) was applied to this powder, a glassy monolithic material was not obtained. Rather, a compacted powder was seen (Figure 29a), which was easily broken up into fine particles. Both observations are consistent with UiO-66-ndc_1 being synthesised as large, individual primary particles rather than the small, gelatinous NPs known to achieve dense packing and monolith formation. This was confirmed by observation of ca. 50 nm diameter primary particles under TEM (Figure 29b). Figure 29| UiO-66-ndc powder. a, Optical and b, TEM images showing the macroscopic and microscopic morphology of UiO-66-ndc powder, respectively. The material was synthesised by equimolar exchange of the terephthalic acid linker applied to UiO-66 gel synthesis with its naphthalene functionalised analogue. Primary particle size is determined by the rate of crystallisation and thus consideration of the chemical properties of linker molecules is key to controlling particle morphology. The influence of this functionalisation on the rate of reaction can be considered in terms of pKa of the linker. For example, the monocarboxylic acid benzoic acid has a weakly acidic pKa value of 4.19. Yet 10 mm a b 75 the addition of a conjugated phenyl ring i.e. 1-napthoic acid, significantly increases acidity, reducing pKa to 3.70. The 10p electron system delocalised over the aromatic naphthalene system exhibits a greater inductive withdrawal effect on the carboxylic acid group than the 6p system in benzoic acid. This weakens the O–H bond in the carboxylic acid by reducing electron density as well as stabilising the resulting conjugate base. Comparable inductive effects are expected in the dicarboxylic acid counterparts, terephthalic acid and naphthalene-1,4- dicarboxlyic acid, used in this study. Increased stability of the organic linkers conjugate base can be understood to reduce rate of reaction, and thus increase particle size, by considering the crystallisation process of these zirconium MOFs. As a first step, the zirconium precursor (ZrOCl2.8H2O) reacts to form a hexanuclear cluster ([Zr6O4(OH)4]12+), a crucial SBU for the MOF (Figure 30a). Each cluster is stabilised by the coordination of 12 carboxylate ions (RCOO–), yielding Zr6O4(OH)4(RCOO)12 (Figure 30b). Coordination of dicarboxylate molecules results in bridges between different metal clusters; the foundation of MOF crystallisation and polymerisation (Figure 30c). Figure 30| MOF crystallisation. a, Metal cluster [Zr6O4(OH)4]12+, an SBU in UiO-66 formation. b, Coordination of twelve organic linkers around the zirconium cluster (a). c, Cross- linking of two zirconium hexaclusters by a bidentate organic linker. Black dashed line shows the symmetry incurred by crystal growth. Coloured spheres in the chemical structures represent C (purple), H (white), O (sky blue) and Zr (navy blue). a b c 76 Disordered/amorphous MOF NPs can nevertheless be formed if the crystallisation process outlined above is too fast. Rapid nucleation of MOFs will result in small, defective particles with reduced long-range order. For this reason, chemical modulators are often added to the reaction mixture, allowing control over the rate of reaction and thus crystallinity/particle size. In the synthesis of UiO-66 gel, AA can be utilised as a coordination modulator. This monocarboxylic acid competes with the organic linker, terephthalic acid, by reversibly coordinating to the zirconium hexaclusters. The coordination of the monocarboxylic acid temporarily terminates the structure, preventing inter-cluster cross-linking (Figure 30c). Thus, the reduced rate of MOF crystallisation, by the competition between the mono- and di- carboxylic acid substrates, decreases the rate of nucleation so that large crystals are slowly grown instead of there being rapid formation of smaller, amorphous particles. HCl can be further utilised for protonation modulation. Besides aiding the initial formation of the ([Zr6O4(OH)4]12+cluster, this acid pushes the organic acid equilibrium towards its protonated state, further slowing the rate of crystallisation by reducing modulators/linkers exchange. By carefully tuning the concentration of these modulators in the reaction mixture, the size and crystallinity of MOF particles can be precisely controlled.3 As discussed already, by changing the organic linker from bdc to more acidic naphthalene-1,4- dicarboxylic acid, the primary particle size was observed to significantly increase despite the same concentrations of modulator being present in the reaction. This may be accounted for by considering how the altered chemical properties of the linker influence its ability to compete with the modulators. In addition to the chemical properties discussed above, as a consequence of the enhanced inductive withdrawal effect of naphthalene ring on the carboxylate group, the strength of coordination to the hard Zr(IV) metal centres in the hexanuclear cluster may be reduced. The rate of linker exchange may thus be enhanced, reducing the linkers ability to compete against the AA modulator. Furthermore, the bulky naphthalene ring may exert some degree of steric hindrance, further slowing the organic molecules ability to coordinate to the crystallising framework. If the particle size in UiO-66-ndc is to be sufficiently reduced to facilitate monolith synthesis, the concentration of modulators in the reaction mixture must be tuned to enable reduction of the particle size without compromising crystallinity. 77 3.2.2| Modulator Study A study was performed in which the volume of the modulators AA and HCl was systematically varied (Table 7). In all syntheses, the modulator volume was reduced relative to the original procedure (UiO-66-ndc_1), yielding a non-flowing gel. This indicates that in all cases, primary particle size was reduced by the decreased modulator volume in the reaction solution. The previously developed monolith drying procedure (i.e. for UiO-66_C) was applied to each of the obtained gels. Despite each synthesis producing primary particles small enough to form a gelatinous suspension, monolithic materials were not obtained in all cases. This demonstrates that while small primary particles are required to form a gelatinous suspension, it does not guarantee that a monolithic material will be formed from said gel. Visual comparison of the obtained materials (Figure 31) was used as a rapid, qualitative means of determining monolithicity. Firstly, AA volume was significantly reduced by 50%, while maintaining the original volume of HCl. However visual observation of the dried product, UiO-66-ndc_2, again confirmed synthesis of a non-transparent powder. Further AA reduction to only 0.75 mL (62.5%) yielded similar results. UiO-66-ndc_3 was obtained as white pellets with only a slight qualitative improvement relative to UiO-66-ndc_1; the obtained pellets demonstrated marginally more robust mechanical properties. From these initial modifications to the synthetic procedure, it was evident that substantial deviation from the original synthesis was required. Figure 31| Optical images of UiO-66-ndc. Optical images of UiO-66-ndc_1 – 10 synthesised using variable volumes of modulator in the reaction mixture (see Table 7). 78 Table 7| Varying volumes (mL) of AA and HCl (37%) used to modulate the synthesis of monolithic UiO-66-ndc. Modulator Volume (mL) Monolithic MOF AA HCl UiO-66-ndc_1 2 1.5 UiO-66-ndc_2 1 1.5 UiO-66-ndc_3 0.75 1.5 UiO-66-ndc_4 1 1 UiO-66-ndc_5 0.75 1 UiO-66-ndc_6 0.5 1 UiO-66-ndc_7 0.25 1 UiO-66-ndc_8 0.25 0.5 UiO-66-ndc_9 0.5 0.5 UiO-66-ndc_10 0.75 0.75 While diminished modulator volume in the synthesis would appear to reduce particle size, as evidenced by the obtained gels, it should not be decreased too far. The primary purpose of a modulator is to slow MOF growth sufficiently to allow crystalline assembly of the SBUs. Instead of further reducing AA volume, UiO-66-ndc_2 and _3 were instead repeated with reduced volumes of HCl from 1.5 mL to 1.0 mL in each case (UiO-66-ndc_4 and _5 respectively). Rather than competing against the organic linker for zirconium cluster coordination sites, as AA does, HCl acts as a protonation modulator. Crystal growth is slowed by inhibited dissociation of the organic acid (e.g. ndcH2) into its conjugate base (ndc2-). Visual comparison of UiO-66-ndc_4 and _5 to _2 and _3, respectively, revealed the striking influence that HCl volume has on the properties of the obtained materials. A significant improvement in monolithic properties was observed, with both robust materials displaying the shiny surfaces synonymous with monolith formation despite their opaque white colour. For the synthesis of robust and optically transparent monolithUiO-66-ndc, comparable to monolithHKUST-1, monolithZIF- 8, monolithUiO-66 and monolithUiO-66-NH2, the volume of both AA and HCl must be carefully tuned in the reaction mixture. 79 Based on these results, the HCl volume was maintained at 1 mL while AA volume was further reduced to 0.5 and 0.25 in UiO-66-ndc_6 and UiO-66-ndc_7, respectively. In this case however, the optical transparency was reduced, with the materials exhibiting a more powder-like morphology. This suggests that a threshold AA volume exists, above and below which the optical properties of the material are diminished. Comparison of UiO-66-ndc_5 and UiO-66- ndc_7 succinctly demonstrates that while reduced volumes of modulator should decrease particle size, it does not necessarily correspond to improvements in synthesising monolithic materials. To confirm this, for each of the previous AA volumes, HCl volume was decreased to 0.5 mL (UiO-66-ndc_8 and UiO-66-ndc_9) yielding no significant improvements in the optical transparency of the product. This could stem from alterations to the particle morphology, crystallinity and missing linker defects, which are also influenced by modulator volume. Finally, for a low volume of HCl, AA volume was again increased. UiO-66-ndc_10 (0.75 mL each of AA and HCl) showed significant visual improvements over previous samples made with both greater and smaller volumes of AA. The benefit of reduced HCl volume was clearly seen by comparing shiny and slightly opaque UiO-66-ndc_10 to more powder-like UiO-66- ndc_3, which uses an equivalent volume of AA and twice the HCl volume. Based on the results above, 0.75 mL was identified as the optimum AA volume for the synthesis, in terms of optical properties, with both increased and reduced volumes of it yielding materials with reduced transparency as well as less robust mechanical properties. A further modulator study was therefore performed with AA maintained at 0.75 mL throughout while HCl was sequentially reduced (Table 8). The results of this modulator study are immediately apparent from visual comparison of the obtained materials (Figure 32). Decreasing HCl volume from 1.5 mL (UiO-66-ndc_3) to 0.5 mL (UiO-66-ndc_12), leads to sequentially more transparent materials; UiO-66-ndc_12 was a monolithic, yet optically cloudy material. Marginal further reduction of HCl to 0.4 mL achieved a completely transparent and glassy-looking material, UiO-66-ndc_13. This material was comparable to previously reported MOF monoliths i.e. monolithHKUST-1,23 monolithZIF-8,22 monolithUiO-66 (Chapter I, Section 1.0) and monolithUiO-66-NH2 (Chapter I, Section 1.0). Further reduction of HCl volume, 0.2 and 0 mL in UiO-66-ndc_14 and UiO-66-ndc _15, respectively, incurred enhancements in the product material’s yellow colour. Farha et al. postulated that HCl may play an additional role in the synthesis of UiO-66 by neutralising the basic impurities found in the solvent DMF.4 The yellow colour of the monoliths synthesised with low HCl volume could thus stem from non-neutralised 80 impurities trapped within the porosity. Alternatively, the crystallinity of the material may have been so significantly reduced by the low modulator volumes, that the colour arises from incompletely coordinated naphthalene-1,4-dicarboxylic acid, which is itself yellow (see discussion of crystallinity below). Table 8| Varying volumes (mL) of HCl (37%) used to modulate the synthesis of monolithic UiO-66-ndc for a fixed volume of AA (0.75 mL). Modulator Volume (mL) Monolithic MOF AA HCl UiO-66-ndc_3 0.75 1.5 UiO-66-ndc_5 0.75 1 UiO-66-ndc_10 0.75 0.75 UiO-66-ndc_11 0.75 0.6 UiO-66-ndc_12 0.75 0.5 UiO-66-ndc_13 0.75 0.4 UiO-66-ndc_14 0.75 0.2 UiO-66-ndc_15 0.75 0 81 Figure 32| Optical images of UiO-66-ndc. Optical images comparing UiO-66-ndc monoliths synthesised with variable volumes of HCl modulator for a fixed volume of AA modulator (see Table 8). 82 3.2.3| Modulator Influence on MOF Physical Properties Through systematic variation of the volumes of coordination and protonation modulator in the reaction mixture, the ability to visually tune the physical properties of the materials was demonstrated. This stems from alterations in the MOF crystallisation and growth during solvothermal synthesis. To further understand the influence of the modulator on the MOF primary particles, each undried MOF gel was further studied by TEM (Figure 33). As discussed, at ca. 50 nm the primary particle size of gel UiO-66-ndc_1 (Figure 29) was much greater than the ca. 10 nm primary particles observed in gel UiO-66 (Figure 6). This is consistent with its inability to produce a glassy monolith. As with previous materials, the gelatinous nature (which facilitates close packing) and non-uniform shape of the synthesised MOF particles prevented accurate measurement of their size. However, by comparison of the particles produced over the entire modulator study (Figure 33), it is apparent that reduced modulator volumes firstly correspond to a general reduction in particle size as well as a reduction in uniformity of the particle shape, relative to UiO-66-ndc_1. For example, UiO-66- ndc_10 (AA = 0.75mL, HCl = 0.75 mL) exhibited a particle size between 20 – 30 nm. Direct comparison of UiO-66-ndc_10 to UiO-66-ndc_13 (AA = 0.75 mL, HCl = 0.4 mL) is informative. These materials were identically synthesised except for a reduced HCl volume in UiO-66-ndc_10. The obtained ca. 20 nm particles of UiO-66-ndc_13 were both more densely packed and further adopted a more gelatinous macrostructure than those of UiO-66-ndc_10, with individual particles less apparent – the boundaries between particles appear blurred. Finally, UiO-66-ndc_14 and UiO-66-ndc_15 also exhibited a very gelatinous appearance, comparable to UiO-66-ndc_13, with densely packed sub-10 nm particles. However, unlike UiO-66-ndc_13, the dark spots previously observed in monolithHKUST-1 and monolithUiO-66- NH2, are more apparent amongst these samples. This dense secondary phase has previously been identified as amorphous regions of MOF, which corresponds well with the reduced crystallinity expected from using low volumes of modulator when synthesising these particles. As discussed, the amorphous phase may account for the dark yellow colour observed in some of the materials, due to the increased prevalence of incompletely coordinated yellow naphthalene-1,4-dicarboxylic acid in these non-crystalline regions. 83 Figure 33| TEM images of UiO-66-ndc. TEM images showing the size and morphology of primary MOF particles of UiO-66-ndc_1 – 15 synthesised using a range of different modulator volumes (Table 7 and Table 8). To further correlate the changes in crystal structure with the modulator volume, each dried UiO- 66-ndc sample was gently crushed to a powder and studied by PXRD (Figure 34). UiO-66- ndc_1, synthesised with the largest volumes of modulator showed high crystallinity with clear diffraction peaks. Wide peaks are consistent with Scherrer line broadening, which is both consistent with the NP size of the primary particles (as observed by TEM) as well with previous crystalline MOF NPs (i.e. UiO-66 and UiO-66-NH2). Despite significant reductions in AA and HCl volume, the majority of the materials obtained in the modulator study showed comparable diffraction patterns, with no apparent amorphisation; an inherent risk of reducing modulator in the synthesis. Notable exceptions to this were UiO-66-ndc_14 and _15 which showed a significant loss of crystalline diffraction. This may corelate with their reduced primary particle size as well as their increased prevalence of dense, amorphous regions recorded by TEM. UiO-66-ndc_1 UiO-66-ndc_2 UiO-66-ndc_3 UiO-66-ndc_4 UiO-66-ndc_5 UiO-66-ndc_6 UiO-66-ndc_7 UiO-66-ndc_8 UiO-66-ndc_9 UiO-66-ndc_10 UiO-66-ndc_11 UiO-66-ndc_12 UiO-66-ndc_13 UiO-66-ndc_14 UiO-66-ndc_15 84 Figure 34| PXRD patterns of UiO-66-ndc. PXRD patterns for UiO-66-ndc_1 – 15 synthesised with variable volumes of modulator (Table 7 and Table 8). Finally, the porosity of materials synthesised in the modulator study was characterised by N2 adsorption isotherms (Figure 35a). For this preliminary work, low-resolution isotherms were obtained by recording N2 uptake between 0.01 – 0.9 bar. While this method does not permit accurate calculation of SBET or PSD, it does allow direct comparison of porosity over large sample sizes. In the current case, comparison of the isotherms obtained for samples UiO-66- ndc_1 – 14 shows that microporosity remained fairly consistent amongst the materials. This can be inferred from the N2 uptake of each material at P/Po = 0.1, which ranged between 140 – 190 cm3 (STP) g–1 (Figure 35b). The error in these experimental N2 adsorption measurements is non-negligible, a consequence of the Tristar equipment on which they were collected (see Methods) which suffers weighing errors, undesirable collection of free-space measurements prior to isotherm analysis and an inability to perform in situ degasses. Hence, this experimentally recorded variation of ± 25 cm3 g–1 amongst UiO-66-ndc_1 – 14 may be considered reasonable experimental error. Furthermore, the N2 uptake of each material at this pressure is fairly consistent with the theoretical isotherm (Figure 35a, red line), simulated from 5 15 25 35 45 Angle (2!) In te ns ity (a .u .) UiO-66-ndc _1 UiO-66-ndc_5 UiO-66-ndc _3 UiO-66-ndc _2 UiO-66-ndc _4 UiO-66-ndc _10 5 15 25 35 45 UiO-66-ndc _6 UiO-66-ndc _9 UiO-66-ndc _12 UiO-66-ndc _14 UiO-66-ndc _15 Angle (2!) Angle (2!) UiO-66-ndc _7 UiO-66-ndc _8 5 15 25 35 45 UiO-66-ndc _11 UiO-66-ndc _13 85 the ideal crystal structure. These results are significant in that they demonstrate a retention of porosity despite the reduced modulator volumes. Again, the exception to this is UiO-66-ndc_15, in which a clear drop in N2 uptake at P/Po = 0.1 is apparent. This loss of microporosity correlates with both the TEM images (Figure 33) which showed dense, non-crystalline, regions as well as the XRD pattern (Figure 34), which also indicated low crystallinity. Of final note are the alterations in mesoporosity amongst these monolithic UiO-66-ndc materials. These are suggested by variations in N2 uptake above 0.8 bar (Figure 35a). Since each material was synthesised using the same gel drying procedure (previous applied to UiO- 66_C), these variations in mesoporosity are again consistent with changes in the packing of the primary particles. Such alterations in packing efficiency may result from the differing particle sizes and morphologies observed across the various materials. Changes in porosity can also result from differing densities of defects e.g. missing linker/cluster defects incurred by the variable quantities of modulator applied to each synthesis. Considering the results of the modulator study as a whole, it is apparent firstly that a powdered MOF can be converted to an optically transparent monolithic material by altering only the volume of modulators in the primary particle reaction. Furthermore, significant changes to the visual appearance of the resulting monolith occurred with only small changes in the modulator volume, suggesting that careful tuning is required for material optimisation. Despite the significant reductions in particle size incurred by modulator variation, crystallinity and microporosity were reasonably well maintained amongst most of the monolithic materials. Considering optical transparency, crystallinity and N2 adsorption, UiO-66-ndc_13 (henceforth monolithUiO-66-ndc) was selected as the optimised material from this part of the study. It was subsequently characterised in full. 86 Figure 35| Adsorption isotherms of UiO-66-ndc. a, N2 adsorption isotherms (77 K) for UiO- 66-ndc materials synthesised with variable volumes of modulator. Marker colours correspond to materials in the key (b). Red line, inset, shows the computationally simulated N2 adsorption isotherm for defect-free UiO-66-ndc, digitised from Reference 71. b, Table showing the colour key for each material as well as the N2 uptake capacity (cm3 (STP) g–1) of each at P/Po = 0.1. P/Po N 2 up ta ke , c m 3 (S TP ) g -1 a b 87 3.3| Monolith Characterisation 3.3.1| Morphology The monolithic material, monolithUiO-66-ndc, was characterised by SEM (Figure 36). The textural properties of this material are highly comparable to previous results for monolithUiO-66 (Figure 6). At low magnification the surface appears smooth and homogenous, while at increased magnification the array of densely packed MOF NPs of which is it comprised are resolved. Figure 36| SEM images of monolithUiO-66-ndc. SEM images of a monolith fragment with magnification increased from a – d, showing the smooth monolith surface to be comprised of densely packed MOF NPs. a b c d 88 3.3.2| Porosity and Density High resolution N2 adsorption and desorption isotherms were collected for the monolithic material allowing complete characterisation of its porosity (Figure 37 and Table 9). Firstly, at 714 m2 g–1, the SBET of the monolithic material was very close to the theoretical value simulated from the ideal crystal structure (747 m2 g–1). The monoliths Wo value (0.28 cm3 g–1) is also close to the theoretical value (0.24 cm3 g–1). These results demonstrate that despite the low concentrations of modulator used to synthesise the MOF primary particles, the crystalline monolith showed high porosity and surface area, close to the theoretical maximum. Both SBET and Wo of UiO-66-ndc were lower than that of UiO-66 (Table 3). This is a consequence of the bulky naphthalene ring in the organic linker in the former which intrudes into the MOFs porosity, reducing the pore volume relative to that of the unfunctionalized parent MOF. Figure 37| N2 isotherms and PSDs of monolithUiO-66-ndc. a, Adsorption (filled markers) and desorption (hollow markers) N2 isotherm of monolithUiO-66-ndc (77 K) compared to the adsorption isotherm simulated from the ideal crystal structure (red dots). b, Distribution of micro- and mesopore diameter in monolithUiO-66-ndc as obtained from Tarazona NLDFT model analysis of N2 isotherm data (a). c, Distribution of mesopore diameter in monolithUiO-66-ndc as obtained from BJH model analysis of the N2 isotherm data (a). 0.00 0.05 0.10 0.15 0.20 0.25 5 50 500 P/Po N 2 up ta ke , c m 3 (S TP ) g -1 dV /d W (c m 3 g- 1 .Å) Pore width (Å) Micro Meso Pore width (Å) dV /d D (c m 3 g-1 .Å) a b c 0 50 100 150 200 250 300 350 400 450 0 0.2 0.4 0.6 0.8 1 0.00 0.01 0.01 0.02 0.02 0 100 200 300 89 The desorption isotherm for monolithUiO-66-ndc demonstrates hysteresis in the mesoporous pressure range (Figure 37a) which was previously observed in both monolithUiO-66 and monolithUiO-66-NH2. The micro-/meso-PSD was also obtained by NLDFT analysis of the N2 isotherm (Figure 37b). Data confirm the presence of large volumes of crystalline microporosity (< 2 nm diameter) as well as indicating the presence of residual mesoporosity (2 – 50 nm diameter). A similar conclusion was obtained by applying the BJH model to the isotherms, showing extensive narrow mesopore volume at < 10 nm diameter (Figure 37c). Table 9| Physical properties, SBET, Wo and Vtot (calculated from the experimentally obtained N2 adsorption isotherms, Figure 37a) and rb of monolithUiO-66-ndc compared to the theoretical values calculated for the ideal crystal structure. *Calculated according to Rouquerol’s consistency criteria (Appendix, Supplementary Figure 12). †Calculated at P/Po = 0.1. ‡Calculated at P/Po = 0.99. §Experimentally obtained by Hg porosimetry. monolithUiO-66-ndc was also characterised by Hg porosimetry, with results confirming the absence of macroporosity in the material (Figure 38). Consistent with the BJH PSD (Figure 37c), a substantial mesopore volume of diameter was recorded. Furthermore, by this characterisation method, the material was found to have a rb of 0.92 g cm–3. The theoretical bulk density of UiO-66-ndc (1.43 g cm–3) is higher than that of UiO-66 (1.24 g cm–3) as a result of its bulky naphthalene functional group, which reduces pore volumes and increases the materials mass. In the current case, the experimental rb being 35.7% lower than the theoretical maximum is indicative of its abundant mesopore volume (Table 9). Each of these data are consistent with previously discussed results observed for the comparably synthesised mixed micro-/mesoporous monolithic materials, monolithUiO-66 and monolithUiO-66-NH2. SBET* (m2 g–1) Wo† (cm3 g–1) Vtot‡ (cm3 g–1) rb§ (g cm–3) monolithUiO-66-ndc 714 0.28 0.64 0.92 Theoretical 747 0.24 0.24 1.43 90 Figure 38| PSD of monolithUiO-66-ndc. PSD of monolithUiO-66-ndc obtained by Hg porosimetry showing the variations in meso- (2 – 50 nm) and macro- (> 50 nm diameter) porosity in the material. Pore diameter (nm) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1E+00 1E+01 1E+02 1E+03 1E+04 1E+05 Di ffe re nt ia l I nt ru sio n (H g) m l g -1 Meso Macro 91 3.3.3| Thermal and Mechanical Stability The thermal stability of monolithUiO-66-ndc was studied by TGA up to 650 °C (Figure 39). Minor weight loss was observed until about 450 °C, after which temperature the material rapidly decomposed. This behaviour is characteristic of high stability Zr-MOFs. Comparison with a literature TGA trace for the powdered material shows a comparable decomposition pattern. The difference in final weight (%) between the traces is accounted for by the reference decomposition (black line) being run in the presence of air whereas the experimentally collected decomposition for monolithUiO-66-ndc (purple line) was performed under N2 atmosphere.71 These data are further comparable to results for unfunctionalised monolithUiO-66, with negligible difference seen in trace shape and decomposition temperature range. Figure 39| TGA traces of UiO-66-ndc. Comparison of thermal decomposition by TGA for powdered UiO-66-ndc (black, digitised from Reference 71) to that of monolithUiO-66-ndc (purple) and unfunctionalised monolithUiO-66 (green) (50 – 650 ºC, under N2 atmosphere). Finally, the mechanical properties of the monolithUiO-66-ndc were studied by nanoindentation (Figure 40). The material displayed moderately robust mechanical properties with average values of E = 5.9 ± 1.3 GPa and H = 0.16 ± 0.07 GPa. These fall within the range previously observed amongst UiO-66_A – D (E = 4.2 – 14.3 GPa, H = 0.11 – 0.48 GPa) and, in particular, are close to values for UiO-66_C (E = 4.32 ± 0.89 GPa, H = 0.11 ± 0.02 GPa), which was experimentally obtained under the same drying conditions i.e. washed in DMF, dried at 30 °C. Notably, the mechanical properties recorded amongst the repeated indents (x16 over the surface 0 20 40 60 80 100 50 150 250 350 450 550 650 Temperature (°C) W ei gh t ( % ) monolithUiO-66-ndc powderUiO-66-ndc monolithUiO-66 92 of the material) were very consistent, as indicated by the small error bars in the penetration curves (Figure 40b, c). This suggests high homogeneity amongst different areas of the synthesised material. However, both the Hardness and Young’s modulus graphs demonstrate a substantial increase in mechanical strength at increased surface penetration. Superficial surface defects often result in reduced mechanical strength at low penetration depths which is non- representative of the bulk material. Yet, this normally plateaus in the bulk material at depths in excess of ca. 250 nm. This was the case for previous monoliths (UiO-66 (Figure 9) and UiO- 66-NH2 (Figure 23)), where an initial increase in mechanical strength was followed by a plateau. In the current case, however, the results suggest that either the material was damaged to some extent by the harsh polishing process required for nanoindentation, or that the material is not homogenous throughout. Despite this variation in mechanical strength, the material appears essentially robust and comparable to previously studied monolithic Zr-MOFs. Figure 40| Mechanical testing of monolithUiO-66-ndc. Nanoindentation data for monolithUiO-66- ndc showing a, Load (mN) vs. Penetration into the monolith surface (h, nm) for 16 indents. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, as a function of Penetration depth into the monolith surface (h, nm). Mean properties and corresponding errors (inset) were obtained from measurements taken across 16 indents over penetration depths of 250 – 2000 nm, with the aim of eliminating surface defects/tip artefacts. 0 500 1000 1500 2000 2500 0 5 10 15 20 25 30 35 0 500 1000 1500 2000 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0 500 1000 1500 2000 0 2 4 6 8 10 Penetration into surface (nm) Penetration into surface (nm) Penetration into surface (nm) Lo ad (m N ) H ar dn es s, H (G Pa ) Yo un g’ s M od ul us , E (G Pa ) H = 0.16 ± 0.07 GPa E = 5.9 ± 1.3 GPa a b c 93 3.4| Conclusions Through modification of the synthetic procedure for UiO-66 gel, a comparable gel of UiO-66- ndc was synthesised and dried to produce an optically transparent material. Unlike previously discussed UiO-66-NH2, extensive modification to the synthetic protocol was required to achieve this result. Substitution of the terephthalic acid linker used in UiO-66 for naphthalene- 1,4-dicarboxylic acid incurred a five-fold increase in primary particle size under the same synthetic conditions. Both the coordination modulator (AA) and protonation modulator (HCl) were found to play critical roles in controlling the particle size and morphology. By decreasing the modulator’s concentration, a range of different MOF gels were obtained. This demonstrates that not all MOF gels can be dried to yield monolithic materials. Indeed, most of these gels were dried to yield opaque, powdered materials. Through a modulator study, the optimum coordination modulator (AA) volume was identified to be 0.75 mL (-63% with respect to the original volume). Using both greater and smaller volumes of this modulator incurred a reduction in the visual monolithic properties of the obtained material. Furthermore, protonation modulator HCl was found to play a key role in the optical transparency of the material. This was optimised to 0.4 mL (-73% volume), with greater volumes yielding cloudy white monoliths and lesser volumes yielding yellowed/amorphous materials. Despite the significant reduction in modulator concentration required to synthesise the transparent monolithUiO-66-ndc, the optimised material displays high crystallinity, SBET and porosity, with values obtained being comparable to the theoretical maxima. Although UiO-66- ndc is a purely microporous MOF, the presence of mesoporosity in the monolith was recorded by PSD analysis of the N2 adsorption isotherms. This is comparable again to the results already observed in variably mesoporous monolithUiO-66 and monolithUiO-66-NH2. Considered together, these results are highly significant, demonstrating the generality of the developed procedure for generating monolithic Zr-MOFs with chemical and physical properties akin to those of the corresponding single crystal materials. Thus, prototypical Zr-MOF, UiO-66, as well as mono- and now di-substituted derivative structures have been synthesised for the first time as robust monolithic materials without using chemical binders or applied pressures. However, Zr-MOFs make up an extensive chemical family of which functionalised UiO-66 analogues are but a small part. Can other Zr-MOFs, with significantly altered organic linkers be synthesised as comparable monoliths? 94 4.0| NU-1000 Thus far in this work, the zirconium MOF UiO-66 and two of its isostructural derivatives, UiO- 66-NH2 and UiO-66-ndc, have been successfully synthesised as monolithic materials with chemical and physical properties comparable to their respective theoretical structures. Besides these Zr-MOFs, which are based on bidentate linkers, structures based on comparably more complex organic molecules have also been produced as powders. These have been based on, for example, tri- and tetra-carboxylate linkers.7 The synthesis of Zr-MOFs with diverse organic linkers is a key foundation of MOF synthesis, where it is critical for the design of materials for targeted applications; it enables variation of pore geometries, volumes, functionalities and cluster coordination.15 For example, the tetragonal linker btc is used to synthesise MOF-808.17 The unsaturated 6-coordinate zirconium hexaclusters within this framework result in its increased catalytic activity towards i.e. the degradation of chemical warfare agents, relative to more saturated, 12-coordinate UiO-66.14 The points discussed above make it significantly interesting to explore the synthesis of monolithZr-MOFs that are non-isostructural to the UiO-66 family. Considering known Zr-MOFs, the NU series developed at Northwestern University are perhaps some of the most topologically interesting. Amongst the first of the Northwestern MOFs was NU-1000 (Figure 41a).72 This archetypal NU MOF’s tetracarboxylate linkers enforce a csq topology in which wide, 31 Å, hexagonal channels are connected by 10 x 8 Å windows to 12 Å triangular channels (Figure 41b). The high Vtot (1.4 cm3 g–1) and SBET (2280 m2 g–1) of this MOF lend it to a number of potential applications including catalysis, adsorption and separation.73 Furthermore, using the organic linker for NU-1000 as a template, a family of isoreticular MOFs have further been developed through the addition of longer organic moieties before the carboxylate functional groups. Thus the NU-1101 – NU-1104 series incur ultra-high surface area and pore volumes i.e. an outstanding SBET of ~ 6500 m2 g–1 was reported for NU-1103 which utilised an extended tetracarboxylate linker, namely 4,4',4'',4'''-(pyrene-1,3,6,8-tetrayltetrakis(benzene-4,1-diyl)) tetrakis(ethyne-2,1-diyl)tetrabenzoate.74 As a consequence of its outstanding chemical and physical properties, in addition to the relative ease of synthesising the non-commercially available linker compared to the more complex derivative NU MOFs, the parent MOF NU-1000 remains of significant interest e.g. for a number of possible gas-adsorption applications.75 It exhibits a promising H2 working capacity 95 of 5.2 wt% and 30 g L–1 (5 – 100 bar, 77 K).76 In this respect, the synthesis of this MOF as a crystalline monolith would make a fascinating study of potential densified materials for practical gas storage. Figure 41| Crystal structure of NU-1000. a, Chemical structure of tbapyH4, precursor to the organic linker, tbapy, for the Zr-MOF NU-1000. b, The crystal structure of NU-1000 demonstrating how each 31 Å hexagonal pore (x) is surrounded by six, 12 Å triangular pores (y). Inset in b, key showing the colours which correspond to each element. a b x y C H O Zr COOH COOHHOOC HOOC a b x y C H O Zr COOH COOHHOOC HOOC 96 4.1| Aims and Objectives Thus far in this study, it has only been determined that Zr-MOFs isoreticular to UiO-66 can be synthesised as robust, porous and optically transparent monoliths. To further understand the potential of such MOFs to be produced as monoliths, the synthetic protocol developed for monolithZr-MOF will be applied to a material with a significantly different structural geometry. By exploring the synthesis of a MOF comprised of non-isostructural tetragonal linkers, NU- 1000, the synthetic capability of the developed procedure to be extended to a wider range of structures will be elucidated. Characterisation of primary MOF particles as well as any resulting materials will be performed to further understand the implications of such significant alterations to the linker on the chemical and physical properties of the MOF. In this respect, the crystallinity and porosity are of particular interest. NU-1000 displays outstanding gas adsorption properties, an outcome of its high surface area and considerable pore volume.72 These highly desirable properties are incurred by the MOFs wide pores, enforced by the long tetragonal linkers of which the MOF is comprised. Yet it has been frequently observed that the synthesis of MOFs containing longer organic linkers with high crystallinity is synthetically more complex than synthesising equivalent materials with small linkers; linker solubility is reduced, and formation of the 3D crystalline phase may become disfavoured.77 In the current case, the synthesis of gelatinous NU-1000 NPs, small enough to densify as a monolith, yet large enough to exhibit long range crystalline order and thus microporosity poses a significant synthetic challenge. 97 4.2| Monolith Synthesis The synthesis of crystalline NU-1000 is often reported to require harsher synthetic conditions i.e. higher temperatures and stronger modulators, than that of UiO-66.72 The bulky organic linker, tbapyH4, typically displays relatively low solubility in the reaction mixture while sterics further hinder its assembly into the crystalline structure. In the current case, it is not feasible to perform a simple ligand exchange in the Zr-MOF gel reaction, as was the case for UiO-66-NH2 and UiO-66-ndc (with modulator tuning). NU-1000 NPs were synthesised as part of a wider study within the research group into the effect of modulators on the particle size of this MOF with a view to drug delivery for cancer therapy. It was noted, as part of this wider study, that one of the NU-1000 samples was obtained as a viscous gel that bore comparison to the UiO-66, UiO-66-NH2 and UiO-66-ndc gels previously discussed in this work. The NU-1000 gel in question was synthesised at 140 °C, using the same metal source as UiO-66 (zirconium(IV) oxychloride octahydrate). However, due to the low solubility of tbapyH4, a significantly lower linker concentration (2.5 mg mL–1 vs. 40 mg mL–1) was used. Furthermore, the strongly acidic modulator TFA (pKa = 0.3), rather than weaker AA (pKa = 4.8), was used to modulate the synthesis. TFA is often applied to NU-1000 synthesis as it better inhibits tbapyH4 deprotonation and more strongly coordinates to the Zr(IV) metal clusters, facilitating slow crystalline growth of the MOF.78 The obtained NU-1000 MOF gel was washed in DMF and dried at 30 °C, yielding NU-1000 as a shiny yellow monolith with a semi-transparent appearance (Figure 42). Comparably to previous monolithic Zr-MOFs, under application of force the material did not crumble to powder but cracked with retention of monolithicity. The material was fully characterised to determine if the physical properties of this monolithic MOF were consistent with that of the ideal MOF crystal structure. 98 Figure 42| monolithNU-1000. Optical image of yellow monolithic MOF, NU-1000, which was cracked under the application of force. 99 4.3| Characterisation 4.3.1| Particle Size and Morphology The dried material was studied by TEM (Figure 43a, b), enabling observation of the primary particle packing. The gelatinous NPs retained their small size (ca. 10 nm) but showed substantial densification as a result of drying. The darker patches in these TEM images suggest the present of dense, amorphous phases which have likewise been observed in monolithHKUST- 1,23 monolithUiO-66-NH2 (Figure 18) and monolithUiO-66-ndc (Figure 33). The observations were further supported by SEM (Figure 43c, d) which again shows the smooth monolithic surface to be comprised of densely packed smaller particles. Figure 43| Electron microscopy analysis of monolithNU-1000. a, b TEM images of monolithNU- 1000 showing sub-5 nm primary MOF particles. c, d, Low and high magnification, respectively, SEM images of monolithNU-1000 showing the smooth surface of the monolith, which is resolved into a densely packed array of MOF NPs upon increased magnification. 100 4.3.2| Crystallinity and Porosity The crystallinity of monolithNU-1000 was studied by PXRD (Figure 44). The obtained pattern showed extremely wide reflections, the shape of which cannot be purely attributed to Scherrer line broadening. While low angle major reflections (2q = 7.6 and 5.6), consistent with the simulated structure, are apparent the broad nature of these reflections masks any lower intensity peaks. These broad peaks may be linked to the presence of a secondary amorphous phase within the material – as suggested by the presence of dark spots in the TEM images (Figure 43a, b). This result is consistent with a low crystallinity material. Figure 44| XRD of NU-1000. Comparison of simulated XRD pattern of NU-1000 (black) generated from its crystal structure, to the experimentally obtained PXRD pattern of monolithNU- 1000 (orange). This material’s low crystallinity was further indicated by the appearance of its N2 isotherm (Figure 45a). While some degree of microporosity is indicated by N2 adsorption at P/Po < 0.1, the experimental uptake in the monolith was substantially lower than the theoretical maximum in the perfect crystalline MOF. Correspondingly, the SBET of monolithNU-1000 is 1063 m2 g–1 (calculated according to Rouquerol’s consistency criteria; Appendix, Supplementary Figure 13) while the theoretical maximum for the defect-free MOFs crystal structure stands at 2280 m2 g–1. The slight rise in uptake in the experimental material between P/Po 0.2 – 0.4 is characteristic of crystalline NU-1000, which contains larger, 30 Å, hexagonal mesopores in addition to its 12 Å triangular micropores (Figure 41b). This allows condensation of N2 in the 10 20 30 40 10 20 30 40 In te ns ity (a .u .) Angle (2q) 101 wider pores to occur at a higher pressure than in narrow pores and suggests that while the material displayed low bulk crystallinity, some longer-range crystallinity was present within the otherwise amorphous material. Finally, the presence of non-crystalline mesoporosity was suggested by the increased N2 adsorption by the experimental material at P/Po > 0.8. Such behaviour is not predicted by the simulated isotherm. This was again supported by the hysteresis loop observed over this pressure range in the N2 adsorption – desorption isotherm (Figure 45b), as was extensively discussed for previous mixed micro-/mesoporous monolithic MOFs (Chapter I, Section 1.0). Figure 45| N2 isotherms for NU-1000. a, Linear N2 adsorption isotherms at 77 K for monolithNU- 1000 (filled orange squares) and NU-1000 simulated from the crystal structure (hollow orange squares). b, Adsorption (filled orange squares) and desorption (hollow orange squares) N2 isotherms demonstrating hysteresis during gas uptake and release in the mesoporous-associated pressure range. The porosity of the material was further characterised by PSD analysis of the adsorption and desorption N2 isotherms. The NLDFT PSD (Figure 46a) showed the presence of both the 12 Å and 30 Å pores expected for crystalline NU-1000 (Figure 41).78 However, the presence of smaller micropores (below 12 Å) with high pore volume may be accounted for as non- crystalline micropores or partially collapsed micropores79 consistent with the semi-amorphous regions suggested by XRD (Figure 44). Data also indicated the mesoporous volume to be small relative to the micropore volume. By the BJH model (Figure 46b), a mesoporous range 0 200 400 600 800 0 0.2 0.4 0.6 0.8 1 0 200 400 600 800 1000 1200 0 0.2 0.4 0.6 0.8 1 P/PoP/Po N 2 up ta ke cm 3 (S TP ) g -1 N 2 up ta ke cm 3 (S TP ) g -1 a b 102 between 40 – 100 Å was observed, consistent with the N2 isotherms enhanced gas uptake at P/Po > 0.8. The material’s low crystallinity was further supported by Hg porosimetry, where the rb was recorded to be 0.997 g cm–3. The exceptionally large pore volume of the theoretical, crystalline MOF structure predicts a crystal density of 0.486 g cm–3.72 The density of the monolithic material being twice the theoretical maximum is symptomatic of its low crystallinity and extensive presence of a high density, amorphous phase. Figure 46| PSDs for monolithNU-1000. a, Distribution of micro- and mesopore width and b, distribution of mesopore diameter in monolithNU-1000 as obtained from Tarazona NLDFT and BJH model analysis of N2 isotherm data (Figure 45), respectively. 0 100 200 300 400 0.000 0.004 0.008 0.012 0.016 dV /d D cm 3 g -1 , Å Pore Diameter (Å) 10 100 0.00 0.02 0.04 0.06 0.08 0.10 Mesoporosity dV /d W (c m 3 g -1 . ? ) Pore width (? ) Microporosity Pore idth (Å) dV /d W (c m 3 g-1 . Å ) a b dV /d D (c m 3 g-1 . Å ) dV /d W (c m 3 g-1 . Å ) 103 4.3.3| Thermal and Mechanical Stability The thermal stability of monolithNU-1000 was studied by TGA over the temperature range 50 – 750 ºC (Figure 47). The literature reports of powder NU-1000 demonstrate the high thermal stability which is synonymous with Zr-MOFs; rapid decomposition occurs only after 500 ºC.72 Yet monolithNU-1000 demonstrates an unexpected primary decomposition at 300 ºC followed by a secondary weight loss at 500 ºC. This may correlate with loss of solvent trapped in the porosity or rapid onset decomposition of the primary amorphous phase, as a consequence of incomplete coordination. The smaller decomposition peak at 500 ºC, matches the expected decomposition of crystalline Zr-MOF. Figure 47| Thermal stability of monolithNU-1000. TGA analysis of monolithNU-1000 showing gradual thermal decomposition over the temperature range 50 – 750 °C (under N2 atmosphere). Finally, the mechanical properties of the monolithic material were determined by nanoindentation. The material demonstrated high homogeneity, with little variation between repeated indents, as demonstrated by the narrow spread of the load displacement curves (Figure 48a) as well as small error bars in the Hardness (Figure 48b) and Young’s modulus (Figure 48c) curves. The mechanical stability of this MOF has not been substantially reported in the literature, however, the obtained values of H (0.31 ± 0.03 GPa) and E (6.8 ± 0.3 GPa) are comparable to those recorded for previously obtained, industrially viable monolithic MOFs e.g. monolithUiO-66 (Figure 9), monolithUiO-66-ndc (Figure 40) and monolithZIF-8.22 This suggests that despite the apparent semi-amorphous nature of the obtained material, its displays reasonably robust physical properties. 50 60 70 80 90 100 50 250 450 650 W ei gh t ( % ) Temperature (oC) 104 Figure 48| Mechanical properties of monolithNU-1000. Nanoindentation data for monolithNU- 1000 showing a, Load (mN) vs. Penetration into the monoliths surface (h, nm) across 16 indents. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, as a function of Penetration depth into the monolith surface (h, nm). Mean properties and corresponding errors (inset) were obtained from measurements taken across 16 indents over penetration depths of 250 – 2000 nm, ensuring elimination of surface defects/tip artefacts. 0 500 1000 1500 2000 0 2 4 6 8 Penetration into surface, h (nm) Yo un g' s M od ul us , E (G Pa ) 0 500 1000 1500 2000 0.0 0.1 0.2 0.3 0.4 H ar dn es s, H (G Pa ) Penetration into surface, h (nm) 0 500 1000 1500 2000 2500 0 10 20 30 40 Penetration into surface, h (nm) Lo ad (m N ) H = 0.31 ± 0.3 GPa E = 6.8 ± 0.3 GPa Lo ad (m N) Ha rd ne ss , H (G Pa ) Yo un g’s m od ul us , E (G Pa ) Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm) a b c 105 4.4| Conclusions With the aim of extending the known monolithZr-MOF family beyond purely that of UiO-66 derived structures, the syntheses of monolithNU-1000 was targeted due to its substantially dissimilar structure. NU-1000 was first synthesised as a gel of ca. 10 nm particles and subsequently dried to yield a robust, macroscopic material which was visually comparable to the Zr-MOF monoliths of UiO-66, UiO-66-NH2 and UiO-66-ndc. However, upon further characterisation of this material, it was found to express low crystallinity and porosity. The PXRD pattern indicated low crystallinity of the structure, which idea was further supported by its low N2 adsorption reading; SBET was -53% with respect to the theoretical value. Likewise, Hg porosimetry revealed a density twice the theoretical maximum. The dense, amorphous phase required to explain these data was observed visually as darker regions in TEM images. Its inclusion had further ramifications, namely that the materials thermal decomposition was uncharacteristically low for a Zr-MOF. These results demonstrate that while NU-1000 was successfully obtained as a robust, macroscopic material which could reasonably be described as monolithic, its physical properties were non-optimised and were not comparable to the theoretical crystalline material. Hence, it differed substantially from previous monoliths of UiO-66, UiO-66-NH2 and UiO-66-ndc reported herein. The low crystallinity of the produced NU-1000 sample may stem from the greater synthetic complexity incurred by the use of a larger organic linker than was the case for the UiO-66 MOFs. The large size of tbapyH4 reduces its solubility to make the rapid synthesis of NPs less favourable. The increased unit cell diameters incurred by the long linkers also influences the materials capacity to be synthesised as crystalline NPs. For the various UiO-66 derived monoliths, gel phases containing ca. 10 nm primary particles were required to achieve dense particle packing (Chapter I, Sections 1.0 – 3.0). While such small MOF particles are not common in the literature, for these UiO-66 derived structures, NPs of such dimensions could be readily synthesised while maintaining crystallinity/porosity in the resulting macrostructure due to their small unit cell parameters – for UiO-66 the cubic unit cell has dimensions a = 20.8 Å. This facilitates crystalline order in 10 nm UiO-66 particles via multiple repeats of the comparatively small 2.1 nm unit cell. In contrast, for NU-1000, the longer linkers incur unit cell dimensions of a, b = 39.4 Å, c = 16.5 Å. This doubling of the unit cell to ca. 4 nm means that long range crystalline order cannot be easily achieved in ca. 10 nm NPs. 106 The observations reported here highlight the fact that, while a wide range of Zr-MOFs may be obtained as gels and thence monoliths, the synthesis of highly crystalline materials using large/complex linkers may not be facile. With this inherent structural barrier to crystalline NU- 1000 NP synthesis in mind, modification of the synthesis to improve the material’s bulk physical properties has not been pursued. Furthermore, the linker for this MOF is not commercially available, and must be synthesised under inert conditions in a multistage reaction. The synthetic complexity and expense of this MOF significantly reduces its industrial viability. Finally, while the high surface area and large pore volume of NU-1000 offers promising gravimetric gas storage capacity, the inverse relationship between pore-volume and crystal- density, inherently reduces this materials potential for high density gas storage. While the results herein serve to highlight the limitations of the designed Zr-MOF monolith synthesis towards MOFs with excessively large unit cells, there remain a vast number of promising MOFs with relatively short linkers which have not yet been explored for monolith synthesis i.e. MOF- 808, UiO-67 and Zr-FUM.7 107 5.0| References 1. Cavka, J. H. et al. A New Zirconium Inorganic Building Brick Forming Metal Organic Frameworks with Exceptional Stability. J. Am. Chem. Soc. 130, 13850–13851 (2008). 2. Kolokolov, D. I. et al. Probing the Dynamics of the Porous Zr Terephthalate UiO-66 Framework Using 2H NMR and Neutron Scattering. J. Phys. Chem. C. 116, 12131– 12136 (2012). 3. Morris, W. et al. Role of Modulators in Controlling the Colloidal Stability and Polydispersity of the UiO-66 Metal–Organic Framework. ACS Appl. Mater. Interfaces 9, 33413–33418 (2017). 4. Katz, M. J. et al. A Facile Synthesis of UiO-66, UiO-67 and Their Derivatives. Chem. Commun. 49, 9449–9451 (2013). 5. Vermoortele, F. et al. Synthesis Modulation as a Tool to Increase The Catalytic Activity of Metal-Organic Frameworks: The Unique Case of UiO-66(Zr). J. Am. Chem. Soc. 135, 11465–11468 (2013). 6. Shearer, G. C. et al. Tuned To Perfection: Ironing Out The Defects in Metal-Organic Framework UiO-66. Chem. Mater. 26, 4068–4071 (2014). 7. Bai, Y. et al. Zr-Based Metal-Organic Frameworks: Design, Synthesis, Structure, and Applications. Chem. Soc. Rev. 45, 2327–2367 (2016). 8. Wu, H., Yildirim, T. & Zhou, W. Exceptional Mechanical Stability of Highly Porous Zirconium Metal-Organic Framework UiO-66 and Its Important Implications. J. Phys. Chem. Lett. 4, 925–930 (2013). 9. Bambalaza, S. E., Langmi, H. W., Mokaya, R., Musyoka, N. M. & Khotseng, L. E. Compaction of a Zirconium Metal–Organic Framework (UiO-66) for High Density Hydrogen Storage Applications. J. Mater. Chem. A 6, 23569–23577 (2018). 10. Zhang, X. et al. Adsorption / Desorption Kinetics and Breakthrough of Gaseous Toluene for Modified Microporous-Mesoporous UiO-66 Metal Organic Framework. J. Hazard. Mater. 366, 140–150 (2019). 11. Miao, Y., Su, Z. & Suslick, K. S. Energy Storage During Compression of Metal − Organic Frameworks. J. Am. Chem. Soc. 13, 4667–4670 (2017). 12. Dhakshinamoorthy, A., Santiago-Portillo, A. & Asiri, A. M. Engineering UiO-66 Metal 108 Organic Framework for Heterogeneous Catalysis. Chem. Cat. Chem. 11, 899–923 (2019). 13. Cmarik, G. E., Kim, M., Cohen, S. M. & Walton, K. S. Tuning the Adsorption Properties of UiO-66 Via Ligand Functionalization. Langmuir 28, 15606–15613 (2012). 14. Moon, S. Y., Liu, Y., Hupp, J. T. & Farha, O. K. Instantaneous Hydrolysis of Nerve- Agent Simulants With a Six-Connected Zirconium-Based Metal-Organic Framework. Angew. Chemie 54, 6795–6799 (2015). 15. Wang, H. et al. Topologically Guided Tuning of Zr-MOF Pore Structures for Highly Selective Separation of C6 Alkane Isomers. Nat. Commun. 9, 1745 (2018). 16. Marshall, R. J. & Forgan, R. S. Postsynthetic Modification of Zirconium Metal-Organic Frameworks. Eur. J. Inorg. Chem. 2016, 4310–4331 (2016). 17. Connolly, B. M. et al. From Synthesis to Applications : Metal – Organic Frameworks for an Environmentally Sustainable Future. Curr. Opin. Green Sustain. Chem. 12, 47– 56 (2018). 18. Piscopo, C. G. et al. Positive Effect of the Fluorine Moiety on the Oxygen Storage Capacity of UiO-66 Metal – Organic Frameworks. New. J. Chem. 40, 8220–8224 (2016). 19. DeStefano, M. R., Islamoglu, T., Garibay, S. J., Hupp, J. T. & Farha, O. K. Room Temperature Synthesis of UiO-66 and the Thermal Modulation of Densities of Defect Sites. Chem. Mater. 29, 1357–1361 (2017). 20. Hu, Z. et al. Modulator Effects on the Water-Based Synthesis of Zr/Hf Metal-Organic Frameworks: Quantitative Relationship Studies between Modulator, Synthetic Condition, and Performance. Cryst. Growth Des. 16, 2295–2301 (2016). 21. Solla, E. L., Yot, P. G. & Horcajada, P. A Robust Monolithic Metal–Organic Framework with Hierarchical Porosity. Chem. Commun. 54, 13088–13091 (2018). 22. Tian, T., Velazquez-Garcia, J., Bennett, T. D. & Fairen-Jimenez, D. Mechanically and Chemically Robust ZIF-8 Monoliths With High Volumetric Adsorption Capacity. J. Mater. Chem. A 3, 2999–3005 (2015). 23. Tian, T. et al. A Sol–Gel Monolithic Metal–Organic Framework with Enhanced Methane Uptake. Nat. Mater. 17, 174–179 (2018). 24. Bueken, B. et al. Gel-Based Morphological Design of Zirconium Metal-Organic Frameworks. Chem. Sci. 8, 3939–3948 (2017). 109 25. Wu, H. et al. Unusual and Highly Tunable Missing-Linker Defects in Zirconium Metal- Organic Framework UiO-66 and Their Important Effects on Gas Adsorption. J. Am. Chem. Soc. 135, 10525–10532 (2013). 26. Howarth, A. J. et al. Best Practices for the Synthesis, Activation, and Characterization of Metal–Organic Frameworks. Chem. Mater. 29, 26–39 (2017). 27. Burtch, N. C., Heinen, J., Bennett, T. D., Dubbeldam, D. & Allendorf, M. D. Mechanical Properties in Metal-Organic Frameworks: Emerging Opportunities and Challenges for Device Functionality and Technological Applications. Adv. Mater. 30, 1704124 (2017). 28. Thornton, A. W., Babarao, R., Jain, A., Trousselet, F. & Coudert, F.-X. Defects in Metal– Organic Frameworks: A Compromise Between Adsorption and Stability? Dalton Trans. 45, 4352–4359 (2016). 29. Chen, X. et al. Study of Different Cooling Structures on the Thermal Status of an Internal Combustion Engine. Appl. Therm. Eng. 116, 419–432 (2017). 30. Orellana-Tavra, C. et al. Amorphous Metal–Organic Frameworks for Drug Delivery. Chem. Commun. 51, 13878–13881 (2015). 31. Zhou, C. et al. Metal-Organic Framework Glasses With Permanent Accessible Porosity. Nat. Commun. 9, 5042 (2018). 32. Lincke, J. et al. A Novel Copper-Based MOF Material : Synthesis, Characterization and Adsorption Studies. Microporous Mesoporous Mater. 142, 62–69 (2011). 33. Farha, O. K. & Hupp, J. T. Rational Design, Synthesis, Purification, and Activation of Metal - Organic Framework Materials. Acc. Chem. Res. 43, 1166–1175 (2010). 34. Gao, C., Shi, Q. & Dong, J. Adsorptive Separation Performance of 1-butanol onto Typical Hydrophobic Zeolitic Imidazolate. Cryst. Eng. Comm. 18, 3842–3849 (2016). 35. Elshof, J. E., Abadal, C. R., Sekuli, J., Chowdhury, S. R. & Blank, D. H. A. Transport Mechanisms of Water and Organic Solvents Through Microporous Silica in the Pervaporation of Binary Liquids. Microporous Mesoporous Mater. 65, 197–208 (2003). 36. Go, D. A., Moghadam, P. Z., Hupp, J. T., Farha, O. K. & Snurr, R. Q. Application of Consistency Criteria To Calculate BET Areas of Micro- And Mesoporous Metal − Organic Frameworks. J. Am. Chem. Soc. 14, 215–224 (2016). 37. Zeng, Y. et al. On the Hysteresis of Adsorption and Desorption of Simple Gases in Open End and Closed End Pores. Chem. Eng. Sci. 158, 462–479 (2017). 110 38. Cychosz, K. A. & Thommes, M. Progress in the Physisorption Characterization of Nanoporous Gas Storage Materials. Engineering 4, 559–566 (2018). 39. Zhu, H., Zhang, Q. & Zhu, S. Assembly of a Metal–Organic Framework into 3D Hierarchical Porous Monoliths Using a Pickering High Internal Phase Emulsion Template. Chem. Eur. J. 22, 8751–8755 (2016). 40. Nandasiri, M. I., Jambovane, S. R., McGrail, B. P., Schaef, H. T. & Nune, S. K. Adsorption, Separation, and Catalytic Properties of Densified Metal-Organic Frameworks. Coord. Chem. Rev. 311, 38–52 (2016). 41. Dhainaut, J. et al. Systematic Study of the Impact of MOF Densification into Tablets on Textural and Mechanical Properties. Cryst. Eng. Comm. 19, 4211–4218 (2017). 42. Becker, W. Fluorescence Lifetime Imaging - Techniques and Applications. J. Microsc. 247, 119–136 (2012). 43. Schrimpf, W., Ossato, G., Hirschle, P., Wuttke, S. & Lamb, D. C. Investigation of the Co-Dependence of Morphology and Fluorescence Lifetime in a Metal-Organic Framework. Small 12, 3651–3657 (2016). 44. Schrimpf, W. et al. Chemical Diversity in a Metal–Organic Framework Revealed by Fluorescence Lifetime Imaging. Nat. Commun. 9, 1647 (2018). 45. Digman, M. A., Caiolfa, V. R., Zamai, M. & Gratton, E. The Phasor Approach to Fluorescence Lifetime Imaging Analysis. Biophys. J. 94, 14–16 (2008). 46. Gliemann, H. & Wöll, C. Epitaxially Grown Metal-Organic Frameworks. Mater. Today 15, 110–116 (2012). 47. Peng, Y. et al. Methane Storage in Metal − Organic Frameworks: Current Records, Surprise Findings, and Challenges. J. Am. Chem. Soc. 135, 11887–11894 (2013). 48. Kandiah, M. et al. Synthesis and Stability of Tagged UiO-66 Zr-MOFs. Chem. Mater. 22, 6632–6640 (2010). 49. Biswas, S. & Van Der Voort, P. A General Strategy For The Synthesis Of Functionalised UiO-66 Frameworks: Characterisation, Stability and CO2 Adsorption Properties. Eur. J. Inorg. Chem. 2154–2160 (2013). 50. Demir, H., Walton, K. S. & Sholl, D. S. Computational Screening of Functionalized UiO-66 Materials for Selective Contaminant Removal from Air. J. Phys. Chem. C 121, 20396–20406 (2017). 111 51. Wang, X., Zhao, X., Zhang, D., Li, G. & Li, H. Microwave Irradiation Induced UiO-66- NH2 Anchored on Graphene with High Activity for Photocatalytic Reduction of CO2. Appl. Catal. B Environ. 228, 47–53 (2018). 52. Sun, D. et al. Studies on Photocatalytic CO2 Reduction Over NH2-UiO-66(Zr) and its Derivatives: Towards a Better Understanding of Photocatalysis on Metal-Organic Frameworks. Chem. Eur. J. 19, 14279–14285 (2013). 53. Liu, R. et al. Atmospherical Oxidative Coupling of Amines by UiO-66-NH2 Photocatalysis Under Milder Reaction Conditions. Catal. Commun. 124, 108–112 (2019). 54. Chen, Q. et al. Selective Adsorption of Cationic Dyes by UiO-66-NH2. Appl. Surf. Sci. 327, 77–85 (2015). 55. Lin, K. Y. A., Liu, Y. T. & Chen, S. Y. Adsorption of Fluoride to UiO-66-NH2 in Water: Stability, Kinetic, Isotherm and Thermodynamic Studies. J. Colloid Interface Sci. 461, 79–87 (2016). 56. Saleem, H., Rafique, U. & Davies, R. P. Investigations on Post-Synthetically Modified UiO-66-NH2 for the Adsorptive Removal of Heavy Metal Ions from Aqueous Solution. Microporous Mesoporous Mater. 221, 238–244 (2016). 57. Peterson, G. W., Mahle, J. J., Decoste, J. B., Gordon, W. O. & Rossin, J. A. Extraordinary NO2 Removal by the Metal-Organic Framework UiO-66-NH2. Angew. Chemie 55, 6235–6238 (2016). 58. Peterson, G. W., Decoste, J. B., Fatollahi-Fard, F. & Britt, D. K. Engineering UiO-66- NH2 for Toxic Gas Removal. Ind. Eng. Chem. Res. 53, 701–707 (2014). 59. Browe, M. A., Napolitano, A., DeCoste, J. B. & Peterson, G. W. Filtration of Chlorine and Hydrogen Chloride Gas by Engineered UiO-66-NH2 Metal-Organic Framework. J. Hazard. Mater. 332, 162–167 (2017). 60. Garibay, S. J. & Cohen, S. M. Isoreticular Synthesis and Modification of Frameworks With The UiO-66 Topology. Chem. Commun. 46, 7700–7702 (2010). 61. Finsy, V. et al. Separation of CO2/CH4 Mixtures With the MIL-53(Al) Metal-Organic Framework. Micropor. Mesopor. Mat. 120, 221–227 (2009). 62. Mehta, J. P. et al. Sol-Gel Synthesis of Robust Metal-Organic Frameworks for Nanoparticle Encapsulation. Adv. Funct. Mater. 28, 1705588 (2018). 112 63. Wu, M. X. & Yang, Y. W. Metal–Organic Framework (MOF)-Based Drug/Cargo Delivery and Cancer Therapy. Adv. Mater. 29, 1606134 (2017). 64. Kreno, L. E. et al. Metal - Organic Framework Materials as Chemical Sensors. Chem. Rev. 112, 1105–1125 (2012). 65. Zhao, S. N., Wang, G., Poelman, D. & Van Der Voort, P. Luminescent Lanthanide MOFs: A Unique Platform for Chemical Sensing. Materials 11, 1–26 (2018). 66. Zhang, Y. et al. Luminescent Sensors Based on Metal-Organic Frameworks. Coord. Chem. Rev. 354, 28–45 (2018). 67. Zhang, Q. et al. Multifunctional Ln-MOF Luminescent Probe for Efficient Sensing of Fe3+, Ce3+ and Acetone. ACS Appl. Mater. Interfaces 10, 23976–23986 (2018). 68. Wang, T. C. et al. Get The Light Out: Nanoscaling MOFs for Luminescence Sensing and Optical Applications. Chem. Commun. 55, 4647–4650 (2019). 69. Zhao, Y., Lee, S. Y., Becknell, N., Yaghi, O. M. & Angell, C. A. Nanoporous Transparent MOF Glasses with Accessible Internal Surface. J. Am. Chem. Soc. 138, 10818–10821 (2016). 70. Cox, A. J., DeWeerd, A. J. & Linden, J. An Experiment to Measure Mie and Rayleigh Total Scattering Cross Sections. Am. J. Phys. 70, 620–625 (2002). 71. Butova, V. V. et al. Modulator Effect in UiO-66-NDC (1,4-naphthalenedicarboxylic acid) Synthesis and Comparison with UiO-67-NDC Isoreticular Metal-Organic Frameworks. Cryst. Growth Des. 17, 5422–5431(2017). 72. Mondloch, J. E. et al. Vapor-Phase Metalation by Atomic Layer Deposition in a Metal- Organic Framework. J. Am. Chem. Soc. 135, 10294–10297 (2013). 73. Wang, T. C. et al. Scalable Synthesis and Post-Modification of a Mesoporous Metal- Organic Framework Called NU-1000. Nat. Protoc. 11, 149–162 (2016). 74. Wang, T. C. et al. Ultrahigh Surface Area Zirconium MOFs and Insights into the Applicability of the BET Theory. J. Am. Chem. Soc. 137, 3585–3591 (2015). 75. Vissers, G. O. et al. Heats of Adsorption of N2, CO, Ar, and CH4 Versus Coverage on the Zr-Based MOF NU-1000: Measurements and DFT Calculations. J. Phys. Chem. C 123, 6586–6591 (2019). 76. García-Holley, P. et al. Benchmark Study of Hydrogen Storage in Metal-Organic 113 Frameworks under Temperature and Pressure Swing Conditions. ACS Energy Lett. 3, 748–754 (2018). 77. Schaate, A. et al. Modulated Synthesis of Zr-Based Metal-Organic Frameworks: From Nano to Single Crystals. Chem. Eur. J. 17, 6643–6651 (2011). 78. Islamoglu, T. et al. Revisiting the Structural Homogeneity of NU-1000, a Zr-Based Metal-Organic Framework. Cryst. Eng. Comm. 20, 5913–5918 (2018). 79. Suh, M. P., Park, H. J., Prasad, T. K. & Lim, D. W. Hydrogen Storage in Metal-Organic Frameworks. Chem. Rev. 112, 782–835 (2012). 114 Chapter III Immobilisation of Nanoparticles in Monolithic Metal-Organic Frameworks 116 117 1.0| Monolithic SnO2@ZIF-8 Nanotechnology refers to the manipulation and structural control of materials on the nanoscale (1 – 100 nm) with the aim of producing innovative materials with improved or novel properties. This term encompasses materials with a diverse range of morphologies but as a requirement must have at least one nanoscale structural feature1 (Figure 1) e.g. 0D NPs (all dimensions on the nanoscale),2 1D nanowires/nanotubes (one dimension greater than nanoscale),3,4 2D nanosheets (two dimensions greater than the nanoscale),5 3D nanostructured/nanoporous materials (all dimensions greater than the nanoscale but bulk material is comprised of nanoscale features)6 etc. This diverse set of materials affords a wide range of valuable physical properties and as such possess an unquantifiable number of potential applications including gas storage and sensing7, drug delivery8 and even aeronautical engineering.9 Figure 1| Various NP morphologies. Graphic detailing the structural variation which exists amongst nanomaterials. a, 2D graphene nanotube, b, 2D nanorod, c, 3D nanoporous material, d – f, 1D NP with spherical, cubic and tetrahedron morphologies respectively and g – i, functionalised, core-shell and heterodimeric, respectively, multicomponent 1D NPs. a b c d e f g h i 118 The use of 0D NPs as highly active and tuneable catalysts has garnered substantial industrial interest.10,11 Since catalysis is utilised in 90% of all industrial reactions,12 manufacturers are focused on obtaining more efficient and cost-effective procedures for achieving this. NPs represent a cornerstone in catalysis due to their intrinsically high activity as well as their potential for significant morphological and compositional variation.13 As the size of a material is decreased, the ratio of active surface area to inactive bulk is increased and, as such, 0D NPs (with all dimensions on the nanoscale) often demonstrate high catalytic activity.14 They are also highly tuneable, with the capability for extensive variation in both composition (monometallic, bimetallic, trimetallic etc.) and tuneable morphology (cubic, triangular, spherical, rods etc.) (Figure 1).15,16 This enables significant control over reaction selectivity.17 For example Narayanan and El-Sayed compared the ability of cubic, octahedral and tetrahedral Pd NPs to catalyse an electron transfer reaction between hexacyanoferrate (III) and thiosulfate ions.18 It was demonstrated that rate of reaction was dependant on the percentage of the NPs constituent atoms located on the corner and edges. Thus, the more surface-exposed morphology of the tetrahedral NPs incurred the highest catalytic activity for the studied reaction. Despite these many useful properties, a wide range of problems are associated with cost- effective and large-scale usage of 0D NPs, greatly limiting their industrial applicability.19 Since NPs are inherently small (< 100 nm), they can be difficult to recover after a reaction, with technologically demanding large-scale centrifugation being the best choice to do so. Since they are often produced from expensive materials e.g. rare earth metals, it is essential that they can be efficiently recovered for cost effective re-use. Even if successfully recovered, Van der Waals forces can cause NPs to agglomerate, reducing the availability of active surface to target reactants, the result of which is gradual activity loss.20 Finally, since the toxicity of NPs is not fully understood and many are also produced from hazardous materials, they must be completely removed from any reaction product/waste effluent to avoid risk to human health or environmental damage.21,22 These combined problems (recoverability, stability, and toxicity) are significant drawbacks to what are otherwise excellent catalysts. As a means of countering these barriers to practical usage, NP catalysts are often stabilised by deposition on the surface of macroscopic support materials. For example, cordierite honeycombs, carbon supports (i.e. graphene and carbon nanotubes) as well as polymeric membranes and fibres have demonstrated the capability to immobilise NPs.23–26 Yet this method of sol-immobilisation is inherently flawed as the NP surface directly attached to the support 119 material is rendered catalytically inaccessible. Furthermore, the cost and stability of the support material must be also considered for industrial catalytic applications. It is apparent that an economic method of immobilising catalytic NPs in a stable and easily recoverable material without reducing their activity is much needed. The immobilisation of NPs in porous MOFs (NP@MOF) is an example of just such a composite material. The porosity of the porous MOF host may permit reactants to reach the NPs, while simultaneously acting as a barrier to NP aggregation. Numerous reports of functional powdered NP@MOF systems are available, with the resulting systems combining the molecular sieving properties of MOFs with the high activity of NPs.27–29 While NP aggregation is prevented by this immobilisation, recoverability of powdered NP@MOF composites remains problematic. Recently, Wheatley et al. reported the incorporation of photocatalytic SnO2 NPs in monolithic ZIF-8 (Figure 2).30 This was achieved by forming a colloidal, ethanolic suspension of SnO2 NPs under sonication. The colloid was subsequently used as the reaction solution for ZIF-8 primary particle synthesis, thereby immobilising the NPs in the growing MOF. ZIF-8 is comprised of a network of large pores (11.6 Å diameter) interconnected by smaller windows (3.4 Å diameter) in a sodalite topology.31 Since the utilised particles were significantly larger (5 nm)30 than the MOFs small pores, they were not localised inside the frameworks channels. Instead, the MOF was grown around them, surrounding the NPs in a porous material that allowed guest molecules to reach their active surface for reaction but prevented NP leaching (Figure 2e). Hence, NPs immobilised within this nonporous material remained catalytically accessible. Crucially, ZIF-8 has a flexible structure due to the swinging motion of the 2- methylimidazolate linkers, allowing guest molecules i.e. H2O to fully penetrate the structure and enter the internal cavities.32,33 Correspondingly, SnO2@monolithZIF-8 displayed photocatalytic activity towards the degradation of MB (even after 10 catalytic cycles) with reduced agglomeration and improved NP recoverability; it represented a truly recyclable photocatalyst. However, this proof-of-concept material required further optimisation. Firstly, the active NP dopant comprised a relatively low portion of the total composite (2 wt%). Increased doping levels may further enhance the rate of catalytic dye degradation. Additionally, FIB-SEM was used to study the NP distribution throughout the monolith. In this case, significant agglomeration of the NP into micro-scale aggregates was observed. This may have further reduced the availability of active NP surface in the composite material, limiting catalytic rate. 120 Figure 2| Monolithic MOFs and NP@MOF composites. a, b, Optical images of monolithZIF- 8 and SnO2@monolithZIF-8, respectively. c, d, TEM images of cubic SnO2 NPs used to dope monolithZIF-8. e, Representation of how a cubic NP (purple) may be immobilised in a MOF (not to scale): metal atoms/clusters (sky-blue spheres) and organic linkers (navy-blue rods). 20 nm50 nm c d a b 1 cm 1 cm e 121 1.1| Aims and Objectives There is significant scope for further study and optimisation of the novel SnO2@monolithZIF-8 composite reported by Wheatley et al.35 Firstly, the effect of capping the dopant NPs will be explored with the aim of increasing its loading level into the host as well as reducing agglomeration. It is well known that surface functionalisation of NPs by bulky organic capping agents can prevent NP aggregation.34 This is achieved by the physical barrier incurred by the organic molecules, which prevents direct contact between different NP surfaces. Not only do the capping agents prevent agglomeration, but they further alter NP solubility, modifying their capacity to form a stable colloid which may also affect MOF doping. Finally, the addition of surface capping agents to the NP synthesis is well known to influence NP size and morphology. Capped NPs should be further studied by TEM to fully elucidate the influence of the modified synthetic procedure. Changes in SnO2 size, morphology and surface functionality will inherently alter the availability of active sites on the NP surface. Thus, the catalytic activity of SnO2@monolithZIF-8 doped with capped NPs towards the degradation of toxic dye MB will be studied. The results will be benchmarked against the original report by Wheatley et al. and correlated with NP morphology and doping level in the MOF. Furthermore, while the mechanism of dye- degradation by free SnO2 NPs has previously been studied, this has not been explored in the case of NPs immobilised in a MOF and hence requires verification. A key aim of this research is to further our understanding of the proof-of-concept material, SnO2@monolithZIF-8, elucidating the synthetic conditions which enable control over NP dispersion and doping level in the monolithic support. Ultimately this may aid in the synthesis of more catalytically active materials towards e.g. water purification. This is also an important step towards the synthesis of NP@monolithMOF in general, which is interesting both from an exploratory synthetic perspective as well as a practical one – offering the potential for numerous tuneable and industrially viable composite materials, tailored to different catalytic applications. 122 1.2| Preliminary Composite Characterisation In the proof of concept SnO2@monolithZIF-8 synthesis by Wheatley et al.,35 preliminary work investigating NPs surface capping-agents was undertaken. The aim of this study was to determine if the NP loading could be enhanced and agglomeration reduced by surface capping. Through ICP-OES analysis, it was shown that SnO2 NPs post-synthetically capped with PVP10K achieved a five-fold increase in monolith loading (by wt%) relative to the uncapped counterparts (i.e. 10.4% vs. 2.0%), despite the same mass of isolated NPs being applied to the synthesis. This may be attributed to the exclusion of large NP aggregates from the composite; such large agglomerates cannot form the stable colloid required for immobilisation in the growing MOF. This is further supported by the observation of white powder below the dried monolith in the centrifuge tube i.e. uncapped, dopant NP aggregates sink below the smaller primary MOF particles during the centrifugation stage of the synthesis. The observed increase in doping level may thus result from reduced NP agglomeration incurred by the PVP capping. The improved loading could also be due to increased solubility of the PVP-capped NPs in ethanol (the solvent utilised during MOF primary particle synthesis). Due to time constraints, no further analysis on this improved PVP-capped SnO2@monolithZIF-8 system was performed by Wheatley et al. As a start to this part of the current project, FIB-SEM analysis of the system containing PVP- capped SnO2 was undertaken and compared to the previously analysed system containing uncapped NPs (Figure 3). This technique allows direct observation of the spatial distribution and agglomeration of SnO2 NPs within the monolithic MOF. Due to the higher atomic mass of 119Sn than 65Zn, SnO2 NPs display stronger backscattering of incident electrons during SEM and thus appear as brighter regions in the acquired images. In both cases, SnO2 NPs were observed both on the monolithMOFs external surface (Figure 3a, d) and its internal structure exposed by etching (Figure 3b, c, e, f). Assessment of the FIB-SEM images from each of the samples at the same magnification (Figure 3b, e) firstly demonstrated a qualitative increase in NPs prevalence within the PVP capped material. This further supports the ICP-OES results, as discussed above. Large agglomerates were seen in each of the studied composites i.e. monolithZIF-8 doped with unfunctionalised and PVP-functionalised SnO2 NPs. Due to the non-uniform nature of these 3D aggregates, their size was difficult to accurately quantify. It is clear, however, that despite 123 the use of PVP as a capping agent, agglomeration of NPs in the MOF remained prevalent. This suggests that the enhanced MOF doping for the capped NPs stems from their improved solubility in the colloidal ethanolic suspension prepared for NP@MOF synthesis, rather than a significant reduction in aggregation. Figure 3| FIB-SEM images of SnO2@monolithZIF-8 composites. FIB-SEM images of monolithZIF-8 doped with uncapped (a – c) and PVP-capped (d – f) SnO2 NPs. 124 1.3| Tuning NP Surface Functionality Based on the results of this preliminary study, which suggest the significance of NP surface functionality on the doping level of the MOF, a wider selection of capping agents was applied to the NP synthesis. The aim was to identify the functionalised NP which could be best loaded into the MOF, with the aim of enhancing the resulting composite’s photocatalytic activity. Rather than post-synthetically capping fully isolated particles, capping agents were directly added to the reaction mixture during solvothermal SnO2 synthesis. The widespread consequences of directly capping NPs in situ are well known. Firstly, capping agents have been extensively employed as a means of controlling NP shape. Through selective interaction with crystal faces, crystal growth can be directed to influencing particle morphology.36 It is widely accepted that for NPs, since catalytic activity shows morphological dependence (due to different crystal faces having different catalytic activities),37 addition of capping agents to the mixture may demonstrate interesting consequences for photocatalytic activity. Secondly, by capping during synthesis a permanent steric barrier to agglomeration is formed as soon as particle nucleation occurs. This means that agglomeration may be reduced relative to post- synthetically capped NPs, which can become irreversibly agglomerated prior to capping (as seen by FIB-SEM analysis of post-synthetically PVP capped SnO2@monolithZIF-8, Figure 3). As such, a list of non-toxic and moderately priced organic capping agents was compiled from the literature, in line with the project’s aim of industrially viable catalysis for water purification (Figure 4). Based on the idea that NP shape is related to catalytic activity, a further literature search for examples of unusually shaped SnO2 NPs was conducted. Most reported particles were too large to be promising photocatalysts – an inherent feature of the reduced availably of active surface. The octahedral SnO2 NPs reported by Han et al. are one such example.38 This material was produced in a comparable hydrothermal reaction to that used for cubic particles in the preliminary studies, but with the further addition of HCl. They theorised that this octahedral morphology was a direct result of the presence of Cl– ions selectively interacting with the high energy [211] crystal faces, slowing crystal growth in this reaction to achieve kinetic shape control. However, the dimensions (200 nm by 300 nm) of these octahedral particles likely renders them too large for immobilisation in the monolith; previous observations of the monoliths inability to immobilise NP aggregates suggests that a NP size limit exists. The 125 previous observations regarding the large size of these particles was accounted for by Jiang et al., who studied the size dependence of SnO2 particles on pH, demonstrating that at increased pH the size of synthesised NPs was reduced.39 Based on this, a reaction with NaCl as the additive was undertaken with the aim of cost-effectively introducing Cl– ions to the reaction mixture while maintaining the alkaline pH required for small size. Figure 4| Organic capping agents. Chemical structures of organic capping agents applied to the surface-functionalised synthesis of SnO2 NPs. a, OA, b, NaOl, c, CTAB, d, PVP and e, CA. Samulski et al. reported a hydrothermal synthesis of SnO2 nanorods identical to the preliminary synthetic procedure for nanocubes, but with temperature reduced from 200 °C to 150 °C.40 This synthesis was replicated, as immobilisation of nanorods with such an unusual asymmetric shape (17.0 ± 4.0 nm by 3.4 ± 0.6 nm rods) would be an interesting tool with which to determine if larger NPs can be successfully immobilised in monolithZIF-8. However, a higher rate of photocatalysis was not anticipated as these nanorods are expected to present a reduced availability of active surface area compared to those with smaller, cubic morphology. Finally, SnO2 NPs produced in the preliminary studies were fully isolated by drying under vacuum prior to re-dispersion in ethanol under sonication for immobilisation in the MOF.35 This is a logical synthetic step to take as it allows accurate measurement of the quantity of NPs to be immobilised, leading to reproducibility in doping levels of the final composite material. However, NP agglomeration under drying is a persistent problem due to the high thermodynamic favourability of reducing exposed high energy surface. Agglomeration may be a b c d e 126 limited by elimination of this drying stage in the synthetic procedure. An example composite containing undried NPs (SnO2-2) was therefore produced to compare with composites containing SnO2 NPs fully isolated (by drying) prior to immobilisation i.e. SnO2-1. It was supposed that NP loading could be increased, and agglomeration reduced by this cost-effective and time-saving synthetic modification. Table 1| Modifications made to the original SnO2 NP synthesis and variations in resulting NP size and monolithZIF-8 doping level (wt%) as a result. Synthetic modification† Size‡ (nm) Loading§ (wt%) SnO2-1 No modification 5.0 ± 1.1 2.0 SnO2-2 NPs not dried 4.6 ± 1.1 23.5 SnO2-3 OA added 4.8 ± 1.2 2.9 SnO2-4 CTAB added 3.9 ± 1.0 3.8 SnO2-5 CA added 4.5 ± 1.1 3.5 SnO2-6 NaOl added 4.4 ± 0.9 10.7 SnO2-7 PVP10K added 4.5 ± 1.0 6.7 SnO2-8 NaCl (1 equivalent) added 4.6 ± 0.8, 9.3 ± 3.5 14.3 SnO2-9 NaCl (2 equivalent) added 4.5 ± 0.9, 12.4 ± 5.7 3.4 SnO2-10 NaCl (5 equivalent) added 5.5 ±2.2, 15.9 ± 7.8 11.3 SnO2-11 Reduced temperature (150 °C) 4.9 ± 2.1 7.1 SnO2-12 NaCl (2 equivalents) added and reduced temperature (150 °C) 4.5 ± 1.6 10.9 †For all organic additives, a 1:1 molar ratio to Sn was utilised. Unless indicated otherwise, all samples were synthesised at 200 °C. All samples were dried under vacuum except SnO2-2 which was re-dispersed to a white suspension in absolute ethanol (40 mL) after washing. Full details of synthetic modifications are provided (Appendix, Experimental). ‡NP size directly measured from TEM images (Appendix, Supplementary Figure 14), with means and standard deviations calculated over 200 particles. §SnO2 doping level of each SnO2@monolithZIF-8 was quantified by ICP-OES. 127 The NPs synthesised by these modified procedures were studied by TEM (Appendix, Supplementary Figure 14), allowing measurement of their shape and size (Table 1). Qualitatively, despite in situ surface functionalisation, substantial aggregation was observed in each of the observed SnO2 NPs, a common feature amongst nanomaterials. Furthermore, the image resolution was lessened in some of the samples. This was attributed to the formation of carbon residue by the decomposition of organic capping agents under the high energy electron beam. The TEM imaging further revealed no significant changes to NP shape or size (Table 1) for any of the SnO2 samples prepared with in situ organic capping agent; all particle sizes fall within the standard deviation range of sample SnO2-1. Hence, it is the result of altered NP solubility, rather than size/morphology, which is of interest. In each material (including the unmodified reaction), nanorods were occasionally observed as a secondary morphology amongst the primary nanocubes; it is estimated that nanorods comprise ca. 1% of the total NP population. For the purposes of recording nanocube dimensions (Table 1), this minor nanorod morphology was excluded. However, for SnO2-8 – 10, the inclusion of sequentially larger quantities of NaCl was observed to substantially enhance this nanorod prevalence. In these samples, nanorods substantially became the dominant morphology, and hence the occasional minor nanocube morphology was excluded from the nanorod size dimension measurements. This supports the theory set out by Zheng et al. regarding the capability of Cl– ions to influence SnO2 morphology. It can be clearly seen that as the quantity of NaCl (and therefore [Cl–]) in the reaction mixture is increased, both the length and standard deviation of the nanorods increases (Table 1). This phenomenon is highlighted by the TEM images of SnO2-10 in which a wide range of nanorod sizes were observed within the same sample (Figure 5), the result of which was large deviations around the average values (Table 1). Measurement of d-spacing in HR-TEM images demonstrates the exposure of low energy [110] and [101] crystal faces, with growth through the latter direction being dominant (Figure 5c). Although these observations of altered SnO2 size in the presence of NaCl support the utility of Cl– ions in controlling the NP’s morphology, the isolated product did not reveal the expected high energy [211] crystal faces.38 It should further be noted that the expected nanorod morphology40 was also not obtained for samples produced at the reduced temperature of 150 °C (previous NPs were synthesised at 200 °C). Instead polydisperse cubic NPs were obtained, mixed with short nanorods as a minor secondary morphology. This is consistent with the La 128 Mer model of NP formation, which suggests that lower reaction temperatures lead to slower bursts of nucleation.41 Thus, nucleation and growth stages of different particles occur simultaneously, causing polydispersity. Figure 5| TEM images of SnO2 NPs. a, b, TEM and c, d, HR-TEM images of SnO2-10 showing the presence of polydisperse nanorods. The d-spacing and corresponding hkl crystal face assignments are indicated (c). Despite producing no significant morphology changes (except by the addition of NaCl) all samples were immobilised in monolithZIF-8, allowing variation in monolith doping levels with NP surface functionality to be studied by ICP-OES. In this case, significant variation was observed. Each SnO2@monolithZIF-8 was prepared via the addition of the same quantity of SnO2 NPs (Table 1 and Methods). Hence, differences in monolith loading recorded by ICP-OES are presumably due to the size-selective exclusion of large agglomerates from the monoliths (as discussed earlier). During centrifugation, heavier agglomerates may drop to the bottom of the centrifuge tube preventing monolith formation from taking place around them. When dried monoliths were obtained, un-encapsulated NPs were often found underneath. The doping level 129 may therefore be linked to how stable and homogenous the ethanolic NP suspension is, meaning that both NP solubility in ethanol (the reaction solvent) and agglomeration are central. In the proof of concept work, a 2.0% doping level of SnO2@monolithZIF-8 was achieved.35 However, in SnO2-2@monolithZIF-8 this was increased to 23.4%. This represents a significant achievement, demonstrating that MOF doping can be enhanced through a synthetic modification that is both free and time saving e.g. simply removing the dopants drying stage. It also demonstrates that monolithic MOFs can be doped to significantly higher loadings than previously thought possible, while retaining monolithicity. The resulting monolith was visually identical (large shiny fragments) to less doped samples, but displayed a stronger white colour, suggesting that the high loading did not inhibit monolith formation. The enhanced loading can be attributed to reduced agglomeration of NPs in sample SnO2-2 as no excluded NP powder was found under those monoliths – that is, all added NPs appeared to have been immobilised. For samples capped with organic surfactants, no doping level greater than that obtained in SnO2-2@ZIF-8 was obtained, with the highest loading being achieved by SnO2-6@ZIF-8 (NaOl capped) at 10.7%. However, this is still significantly higher than SnO2-1@ZIF-8. Many factors must be taken into consideration when accounting for variation in monolith doping level by NPs of different surface functionality e.g. variations in chain length altering NP agglomeration, affinity of surfactant to the NPs, surfactant solubility in ethanol, stability of surfactant during hydrothermal conditions used for NP synthesis etc. As such, it may not be possible to fully rationalise the observed differences in SnO2 wt% across all samples, with a combination of these factors most likely responsible. For example, an interesting difference in NP loading was observed between samples SnO2-3@ZIF-8 (OA capped, 2.9%) and SnO2- 6@ZIF-8 (NaOl capped, 10.7%). Such a large difference was unexpected as NaOl is merely the sodium salt of OA. This suggests that the salt may display a higher affinity for the SnO2 particles surface, providing better capping, lowering agglomeration, and thus improving uptake. The main result is to observe which of the capping agents have increased NP wt% in the monolith relative to the preliminary work and apply these findings to improving NP@monolithMOF synthesis in the future. 130 1.4| Photocatalytic Dye Degradation SnO2@ZIF-8 composites (Table 1) were screened as catalysts for water purification by testing their ability to degrade the toxic dye MB under simulated solar irradiation. A catalytic testing method identical to that used in the preliminary work was utilised.35 Simulation of the UV region of the solar spectrum is often inexact and hence not comparable between different light simulating devices. Likewise, it is not a true representation of how photocatalytic materials will act under true solar irradiation. Since SnO2 is a wide-band gap semi-conductor that adsorbs significantly in the UV region, the same solar-simulator to the preliminary studies was also utilised, allowing reasonable comparison between results. It has been previously demonstrated that monolithic ZIF-8 displays no significant photocatalytic activity and this background check was not repeated.35 It was presumed that all decreases in MB concentration during testing were the result of catalytic dye degradation by the SnO2 NPs. The potential effects of dye adsorption to the MOFs external surface were discounted as the negligible decreases in MB concentration in the presence of only monolithZIF-8 suggested this to be insignificant.35 Concentration of dye in the solution was measured by UV-vis adsorption at 665 nm. Percentage dye degradation was therefore calculated using UV-vis absorption measurements from before simulated solar irradiation and after three hours simulated solar irradiation. Results of the dye degradation tests (Figure 6) show several points of interest. Firstly, rate of catalysis did not correlate to SnO2 wt% in the monolith. For example, at 23.4% SnO2-2@ZIF- 8 possessed by far the highest NP loading (Table 1). Although this material did show a statistically significant improvement (P < 0.05, Figure 6b) in photocatalysis relative to that of SnO2-1@ZIF-8 (2.0 wt% NP doping), it also achieved lower mean dye degradation than substantially lesser doped materials i.e. SnO2-4@ZIF-8. A viable explanation for this may be that accessibility of the active surface to the substrate (water) is diffusion limited. The MOFs small pores limit access of water to the active SnO2 NPs surface, meaning that NP loading is not the limiting factor in catalysis but rather the rate of water diffusion. A second factor for consideration is that higher loading does not correlate to a greater availability of active catalyst. Capping agents were incorporated into NP syntheses with the aim of reducing agglomeration and increasing their solubility in ethanol. Although the targeted higher loading was achieved for many of the samples, this may not have translated directly to an increase in dye degradation due to the detrimental effects of the capping agents on catalysis. By passivating the NPs surface with a bulky organic molecule, providing a steric barrier to agglomeration, access of water 131 molecules to the active surface may also have been inhibited.13 For example, SnO2-4@ZIF-8 (3.8% SnO2 loading) and SnO2-6@ZIF-8 (10.7% SnO2 loading) displayed extremely different loadings but achieved near identical mean dye degradations (46.8% and 46.1% respectively). This could be due to differences in ability of the different capping agents to hinder the approach of water molecules to the NPs surface e.g. CTAB is a long chain alkane while NaOl is of a similar length but contains an alkene unit. This may reduce flexibility, with the surfactant having fewer degrees of freedom, inhibiting diffusion of water to the active surface. Different catalytic performance may also be dependent on the differences in hydrophobicity of utilised capping agents, affinity of each surfactant for the NPs surface and their respective abilities to prevent agglomeration. Figure 6| Photocatalytic dye degradation. a, Bar chart comparing degradation (%) of MB in aqueous solution after 3 hours solar irradiation (sky blue) and in the absence of light (navy blue) in the presence of different monolithic composites, SnO2-1 – 12@ZIF-8. Error bars represent standard deviation calculated across results taken in triplicate. Line chart (overlaid, in purple) shows variation of SnO2 loading (wt%) for each of the SnO2@ZIF-8 samples as measured by ICP-OES. b, Table of p-values showing the statistical likelihood that the dye degradation results of SnO2-2 – 12@ZIF-8 differ significantly from that of reference sample SnO2-1@ZIF-8. 0 5 10 15 20 25 0 10 20 30 40 50 60 Sn O 2 -1@ ZIF -8 Sn O 2 -2@ ZIF -8 Sn O 2 -3@ ZIF -8 Sn O 2 -4@ ZIF -8 Sn O 2 -5@ ZIF -8 Sn O 2 -6@ ZIF -8 Sn O 2 -7 @ ZIF -8 Sn O 2 -8@ ZIF -8 Sn O 2 -9@ ZIF -8 Sn O 2 -10 @ ZIF -8 Sn O 2 -11 @ ZIF -8 Sn O 2 -12 @ ZIF -8 Dy e de gr ad at io n (% ) N anoparticle loading in M O F (w eight % ) SnO2-2 SnO2-3 SnO2-4 SnO2-5 SnO2-6 SnO2-7 SnO2-8 SnO2-9 SnO2-10 SnO2-11 SnO2-12 P-value 0.04246 0.07918 0.00009 0.45890 0.00270 0.00770 0.04559 0.29296 0.02271 0.83465 0.02572 a b 132 Furthermore, it is expected that increasing nanorod length should decrease catalytic activity; a result of the decreased availability of active surface area associated with increased aspect ratio of the asymmetric particles. Yet two of the composites containing nanorods (SnO2-8@ZIF-8 and SnO2-10@ZIF-8) showed statistically increased rates of catalysis while changes in photocatalysis for the other nanorod doped material, SnO2-9@ZIF-8, were calculated to be insignificant (Figure 6b). Although, these three materials are not easily comparable due to their different NP loadings levels (Table 1), their overall efficiencies as catalytic composites remain comparable to the reference point sample SnO2-1. For example, SnO2-10@ZIF-8 showed a MB degradation of only 33.0% despite its high catalyst loading (11.3% SnO2). This is only a minor improvement over the 29.6% degradation achieved by SnO2-1@ZIF-8, which had a much lower loading (2.0%). These results are accounted for by the expected reduction in active SnO2 surface area upon utilising long nanorods instead of small nanocubes. For catalysis, availability of active surface area to the reactant, water, appears to be more important than overall NP loading (wt%). These observations are highlighted by consideration of the significance of each of these experimental dye degradation results. P-values were calculated (see Methods) for each of the samples to determine the statistical likelihood that alterations in photocatalytic dye degradation differ significantly from those of the original composite, SnO2-1@ZIF-8. Although samples SnO2-3@ZIF-8, SnO2-5@ZIF-8, SnO2-9@ZIF-8 and SnO2-11@ZIF-8 all showed greater doping levels of SnO2 (Table 1), their photocatalytic dye degradation results exceed the standard p-value tolerance of 0.05 and hence are statistically insignificant. This is particularly interesting for SnO2-11@ZIF-8 which shows a substantial increase in doping level (7.1 wt%) relative to that of SnO2-1@ZIF-8 (2.0 wt%) and yet displays a statistically insignificant change in photocatalytic dye degradation. Furthermore, each of the remaining composites show statistically significant changes in dye degradation (with p < 0.05) despite several of these showing SnO2 wt% substantially lower than that of statistically insignificant SnO2-11@ZIF-8. These results highlight that while statistically significant improvements in the ability of the SnO2@ZIF-8 composite to photo-catalytically degrade MB were achieved by altering the NP synthesis conditions and monolith loading procedure, the results cannot be rationalised by the doping level (wt%) of each composite. Overall, the most significant result of the catalytic testing was observed for SnO2-4@ZIF-8 (synthesised in the presence of CTAB), which showed a statistically significant 53.0% increase 133 in dye degradation relative to SnO2-1@ZIF-8 and a p-value of only 0.00009. This demonstrates the dependence of catalytic activity on NP surface functionality. This sample also showed significant promise as a photocatalyst due to high rates of dye degradation despite low NPs loading of the MOF (3.8%). Significant variation was observed between samples and, for future catalyst design, capping agents must be selected with care so that catalytic activity may be maximised. 134 1.5| Mechanism of Catalysis To further understand the variations in photocatalytic dye degradation observed between the samples, the mechanism of dye degradation must be considered. There is substantial evidence that the ability of semiconductor SnO2 NPs to degrade toxic organic pollutants stems from the photocatalytic generation of radical species. Photoinduced electron (e-) – hole (h+) pairs in the NPs interact with water molecules, generating active radicals which subsequently break down organic species to CO2 and H2O. This is summarised by the reaction mechanism proposed by Houas et al. (Figure 7).42 Figure 7| Organic species degradation. Photocatalytic mechanism of organic species (R) degradation as catalysed by SnO2. However, it has not yet been confirmed that this catalytic mechanism is maintained for SnO2 NPs when immobilised in the hydrophobic MOF ZIF-8. The catalytic dye degradation mechanism was thus studied using a fluorescing probe. In the presence of hydroxyl radicals, non-fluorescing terephthalic acid is converted to 2-hydroxyterephthalic acid (Figure 8a). Under exposure to 315 nm UV light, the fluorescence of 2-hydroxyterephthalic acid can be detected as intense emission at 445 nm. Thus, the photocatalytic evolution of hydroxyl radicals by the immobilised SnO2 NPs was studied by including terephthalic acid in aqueous solution SnOB + ℎ𝑣 → eE + hf i) Photoexcitation OB + eE → OB∙E ii) Oxygen Reduction HBO + hf → Hf + OH∙ iii) Hydroxyl Radical Generation OB∙E + Hf → HOB∙ iv) Oxygen Radical Neutralisation 2HOB∙ + HBOB + OB v) Peroxide Formation HBOB + eE → OH∙ + OHE vi) Active Hydroxyl Radical Production R + OH∙ → Rj∙ + HBO vii) Dye Oxidation 135 with the catalytic monolith and recording how the reaction solutions fluorescence intensity varies with solar exposure time (Figure 8b). Figure 8b reveals amplified emission at λmax = 445 nm, corresponding to an increased concentration of 2-hydroxyterephthalic acid in the reaction solution with extended exposure time to solar radiation. The increased fluorescence intensity demonstrates continued generation of •OH radicals and the mechanism of dye degradation by the NPs appears to be unmodified by their immobilisation in the MOF. SnO2@monolithZIF-8 is capable of MB degradation despite the organic species being larger than the porous MOFs small windows (3.4 Å). The degradation relies on diffusion of water molecules (Dk 2.68 Å) through the monolith’s pores, where inactive H2O is converted to active •OH on the NPs surface. Generated radicals diffuse back out of the MOF where they breakdown organic species, e.g. MB, in the surrounding solution. This supports the mechanism for dye degradation by the semiconductor SnO2 being unchanged despite immobilisation in the MOF. Figure 8| Dye degradation catalytic mechanism. a, Equation showing how non-fluorescing terephthalic acid is converted to fluorescing 2-hydroxyterephthalic acid in the presence of hydroxyl radicals. b, Multiwavelength fluorescence spectra showing fluorescent intensity of terephthalic acid probe solution varying with time (0 – 180 minutes) exposed to irradiation ( l = 315 nm) in the presence of SnO2@monolithZIF-8. 375 395 415 435 455 475 495 515 535 555 180 minutes 120 minutes 60 minutes 30 minutes 0 minutes Wavelength (nm) In te ns ity (a .u .) a b terephthalic acid 2-hydroxyterephthalic acid COOH OH COOH COOH COOH OH 136 1.6| Conclusions Using the preliminary work by Wheatley et al. as a foundation,35 the immobilisation of cubic SnO2 NPs in monolithic ZIF-8 was further studied. The surface functionality of the NPs was varied by including a selection of industrially viable (i.e. safe and low-cost) organic capping agents in the NP synthesis. Although this failed to achieve meaningful changes in particle size or morphology, significant differences in NP doping level in the MOF were observed by ICP- OES; a significant five-fold increase in doping (by wt%) was achieved by the sodium oleate- capped NPs (10.7% loading) relative to the unfunctionalized counterparts from the original synthesis (2.0% loading). It is postulated that the observed differences in MOF loading stem from the differently functionalised NPs capabilities to form stable colloidal ethanolic suspensions in which the growing MOF primary particles can assemble around them. The most significant improvement in MOF doping was achieved by using ‘undried NPs’ – i.e. the SnO2 particles were not fully isolated by solvent evaporation prior to immobilisation but rather maintained as an ethanolic suspension. This synthetic alteration achieved an outstanding 23.5% loading, most likely by avoiding the irreversible aggregation incurred by particle drying. Finally, the inorganic salt NaCl was applied to the NP synthesis demonstrating its ability to favour the formation of asymmetric nanorods over that of nanocubes. The immobilisation of these large nanorods in the monolithic MOF indicates that a wider range of nanoarchitectures may be immobilised for other potential applications. Despite achieving significant improvements in monolith doping level, variations in loading were not observed to correlate with differences in photocatalytic dye degradation. The highest rate of toxic dye degradation was observed for CTAB-capped NPs immobilised in monolithic ZIF-8. Despite achieving only a 3.4% loading in the MOF, this composite demonstrated a statistically significant 53% improvement in photocatalytic activity relative to the original composite material. This catalytic activity was also higher than other composite materials demonstrating substantially elevated NP loading levels. This inconsistency between active NP loading level and observed catalytic activity indicates that other physical factors determine the material’s ability to degrade the dye. For example, NP aggregation, and pore blocking (by excessive NP doping and organic capping agents) may alter the availability of catalytically active surface. 137 Furthermore, the mechanism of photocatalysis was explored by probing the generation of active radicals from water. This supports the proposed mechanism whereby bulky MB is degraded outside the monolith by radicals generated on the dopant SnO2 NPs surface within the MOFs internal porosity. This result may provide an explanation for the apparent inability to improve photocatalytic capacity of the material proportionately to its NP doping; the rate of photocatalysis is inherently limited by the rate of water diffusion through the small pores of ZIF-8. This is a major pitfall of using monolithic MOFs to immobilise active NPs for potential catalytic applications – depending on the substrates size, its approach to active NP surface is kinetically limited by the host MOF. The size of the MOFs pores, relative to the target substrate, is clearly a factor for significant consideration when designing composite NP@monolithMOF materials for catalysis. For example, the synthesis of NP@MOF composites for potential gas- phase applications is of interest as the kinetics of small gaseous molecule diffusion may be enhanced relative to liquid phase diffusion. 138 2.0| Monolithic NP@ZIF-8 Due to their myriad attractive physical and chemical properties, the noble metals (Au, Pd, Ag etc.) are a highly-valued class of material. While the bulk material’s high stability, which makes them resistant to oxidation and corrosion, is well-known, their nanoscale counterparts exhibit more reactive physical properties. This is the result of quantum confinement effects and increased availability of reactive surface atoms on the nanoscale.43 As such, noble metal NPs are utilised in various academic fields including optics44 and medicine.45 Of significant note is their contribution to the field of catalysis where they have distinguished themselves as catalysts for e.g. pollutant degredation,46 organic oxidation reactions47,48 and selective hydrogenation.49 Since NPs are inherently difficult to recover and yet some of the most interesting properties of these expensive materials can be found on the nanoscale, recyclability must be increased if these materials are to be cost-effectively utilised. By designing active noble metal-based NPs for target applications and post-synthetically immobilising in monolithic MOFs, the economic feasibility and practical industrial applicability of the expensive materials may be vastly increased. Industrially applicable reactions which utilise noble metal NPs were selected with the aim of immobilising the catalysts in thermally/chemically stable, monolithic ZIF-8. Figure 9| CO oxidation. Mechanism of CO oxidation catalysed by Au NPs. Firstly, carbon monoxide (CO) oxidation catalysis was selected as a likely research avenue. Due to modern society’s significant fossil fuel dependency, oxidation of CO (a toxic by-product of incomplete combustion e.g. from auto exhaust) is a process essential for reducing emissions and improving already poor air quality.50,51 Noble metal NPs have demonstrated high efficiency in the catalysis of this reaction, with small Au NPs in particular (sub-5 nm) reporting high 1 Au + CO + O2 Reactants 2 Au–O2 + CO O2 Adsorption 3 CO–Au–O2 CO Adsorption 4 CO–Au–2O O2 Dissociation 5 CO2–Au–O Oxidation of CO 6 Au–O + CO2 CO2 Desorption 139 activity, even at room temperature. While this observed activity is still a topic of study, the current literature consensus is that these sub-5 nm Au NPs show the highest activity due to their increased prevalence of unsaturated surface sites relative to inactive bulk.52 This facilitates the adsorption and dissociation of molecular O2, producing unstable atomic O which subsequently reacts with CO to yield CO2 (Figure 9).53 In particular, as the Au NPs diameter is decreased an exponential increase in activity is recorded; reports indicate that 2 – 4 nm particles exhibit activity 2 orders of magnitude greater than ca. 20 nm particles.54 Since these small Au NPs present both a greater challenge for post-catalytic recovery as well as a greater tendency to agglomerate, these expensive NPs are ideal candidates for monolith immobilisation. As catalysis can be performed at room temperature52–54 involving only NPs already known to be small enough for immobilisation in the monolith, synthesis of Au NPs (<< 5 nm) and their subsequent immobilisation in ZIF-8 to yield Au@monolithZIF-8 may present a novel and interesting catalytic composite for this reaction. Hydrogenolysis is another example of an industrially significant reaction which can be catalysed using noble metal NPs. For example, propene (C3H6) is a simple molecule of high commercial value due to its utility as both a precursor for more complex chemicals (e.g. propylene oxide and polypropene) as well as its necessity in plastic production.55 It can be obtained from fossil fuels or, in a more environmentally sensitive way, derived from biomass by conversion of glycerol to 2-propen-1-ol followed by hydrogenolysis to propene.56,57 It is this hydrogenolysis reaction that is of interest due to the practical difficulty associated with splitting the strong C–O bond while leaving the more reactive C=C bond intact.58,59 Of note is the recent work by Naka et al. who reported the room temperature, photocatalytic conversion of 2-propen- 1-ol to propene with high selectivity using a bimetallic composite of Pd/TiO2 NPs (Figure 10).60 A cooperative catalytic role between the two NP components was proposed whereby semiconductor TiO2 was photosensitised by incident radiation to generate excited electrons which may be transferred to the Pd NPs dispersed amongst the metal oxide. Naka further proposed that the charge-separated electrons react with protons in solution to form H atoms adsorbed at the Pd surface. It is the reaction of H atoms with 2-propen-1-ol (protonated in solution to give CH2=CHCH2OH2+) which yields the target propene with only water as the by- product. 140 Figure 10| Photocatalytic hydrogenolysis. Pictographic representation of 2-propen-1-ol hydrogenolysis to propene, as photocatalysed by Pd/TiO2. Due to the low costs and ‘green’ nature of this low temperature reaction, poor recyclability of the expensive noble metal based nanocatalyst may be the biggest cost on the industrial scale. Synthesis of Pd/TiO2@monolithZIF-8 composites for photocatalytic hydrogenolysis could provide a novel solution to this problem. As was the case with SnO2@monolithZIF-8, the optically transparent nature of the monolithic MOF makes it a favourable host for photocatalytic NPs.35 Furthermore, the monoliths ability to host more complex NPs systems (i.e. bimetallic composites) has not been fully explored and hence it is of general synthetic interest to determine the extent to which multicomponent composite NPs can be immobilised. Finally, in hydrogenolysis the reaction of 2-propen-1-ol must take place directly at the Pd NP surface. It is of further interest to determine if the target reactant is capable of efficiently diffusing through ZIF-8’s small (3.4 Å) pore windows; this MOF is well known to be capable of adsorbing species greater than its pore window size due to the swinging motion of its imidazolate linkers.33 The altered diffusion kinetics of the substrate through the host MOFs pores may have implications for the rate of catalysis. Thus, immobilisation of the catalyst makes for an interesting study of the monoliths capability to host functional catalysts for a wider range of catalytic reactions. Finally, CH4 gas is emitted from several industrial processes including agriculture and waste treatment in addition to fuel loss from incomplete NG combustion.61 Yet the capability of this gas to reap environmental havoc is often overlooked; although CH4 has a lower atmospheric residence time (14.4 years) than CO2 (230 years), its molar global warming potential is 3.7 times higher.62 The applicability of monometallic noble metal NPs as active catalysts for total CH4 oxidation (yielding less damaging CO2 and water) is well known. Furthermore, partial oxidation of this gas can yield CH3OH, an industrially valuable material in itself. However, OH Pd/TiO2 H+ hv 141 elevated temperatures are typically required for reasonable conversion e.g. expensive, monometallic Pd NPs require temperatures in the range 300 – 700 °C.63–65 Notably, a decrease in activity is observed at temperatures higher than 700 °C due to conversion of PdO to metallic Pd (a less active catalyst).66 At room temperature and under oxygen-rich conditions, Pd typically exists as its stable Pd(II) oxide (PdO). The lower activity of metallic Pd has been established multiple times e.g. by McCarty et al. who demonstrated its negative apparent activation energy: the rate of oxidation increased during cooling from 750 – 450 °C due to the re-oxidation of the Pd metal at lower temperatures.66 The difference in activity between PdO and Pd was attributed to a higher activation energy of oxygen reduction when chemisorbed on metallic Pd rather than the corresponding oxide. Since pure monolithZIF-8 has previously demonstrated thermal decomposition at ca. 600 °C 67 with the doped NP@monolithZIF-8 systems decomposing at even lower temperatures (ca. 450 – 500 °C),67 a NP@ZIF-8 system for recyclable CH4 oxidation catalysis is only feasible if the immobilised particles can demonstrate high activity at significantly lower temperatures. Often, catalytic NPs for CH4 oxidation are used in conjunction with a support structure whose primary purpose is to increase dispersal of the expensive catalyst. However, increased catalytic activity has also been observed when a metal oxide support (e.g. TiO2, Al2O3 etc.) is used.61 Such supports can facilitate oxygen transport (an essential process in oxidation reactions) within the multicomponent material.68 For this reason, Pd/TiO2 nanocomposites are a promising catalyst for high temperature CH4 oxidation, though not suitable for incorporation into monolithZIF-8, as a low temperature catalyst is still required for this. Despite initially high catalytic activity, monometallic systems of noble metals (supported and unsupported) are not able to maintain high conversion rates over extended periods of time. It has been observed that the addition of a second metal to the catalyst increases its stability, prolonging activity, and catalyst lifetime.69,70 As well as increased stability, higher catalytic activity can also be achieved at lower reaction temperatures if bimetallic systems are used e.g. Pd/Au (a nanocomposite of both Au and Pd NPs).71 It has been proposed that the bimetallic system stabilises the Pd catalyst and increases activity by a combination of electronic effects in addition to providing mixed metal active sites.69 Recently, Hutchings et al. reported the highly efficient catalytic partial oxidation of CH4 to CH3OH using Pd/Au NPs on a TiO2 support.71 TiO2 is known to produce surface-stabilized peroxo or hydroperoxo species, yet experiments using only TiO2 as the catalyst indicated that in the absence of Au/Pd it cannot oxidize CH4. 142 Since reasonable conversion, selectivity and turnover frequency were achieved at temperatures less than 100 °C, a composite of Pd, Au and TiO2 immobilised in monolithZIF-8 may provide a highly promising, recyclable, catalyst for this reaction. 143 2.1| Aims and Objectives The aim of this part of the project is to expand the known NP@monolithZIF-8 systems beyond SnO2 to include a wider range of NPs. This will include the synthesis and post-synthetic immobilisation of more compositionally, structurally and morphologically diverse nanocomposites of mono-, bi-, and trimetallic materials. This is firstly of interest from a purely synthetic perspective. The capabilities of monolithZIF-8 have not yet been fully explored and it is critical to determine if this material can host NPs in general or if the SnO2 immobilisation reported by Wheatley et al. was a unique case.35 The NPs chosen for immobilisation will target industrially significant reactions where a recyclable catalyst is most useful. In particular noble metals will be immobilised due to their combined high cost and activity, which makes them practically useful but limits their economic feasibility. Au@monolithZIF-8 for CO oxidation will be synthesised where the NPs are very small i.e. 2 – 4 nm. Secondly, the synthesis of a bimetallic nanocomposite of Pd/TiO2 will be targeted as a hydrogenolysis catalyst, as will trimetallic nanocomposite containing Au, Pd and TiO2 for CH4 oxidation. A range of characterisation techniques will further be employed to determine to what extent the components of these multicomponent catalysts can be immobilised simultaneously in the monolithic MOF and the effect that this immobilisation has on the host material’s crystallinity. Finally, the catalytic activity of synthesised monolithic composites towards each reaction will be tested to elucidate the influence of the NPs immobilisation on activity. This should offer insight into the potential of monolithic MOFs to host catalysts for a wide range of reactions, rather than just the photocatalytic water purification which has been studied to date. 144 2.2| Au@monolithZIF-8 2.2.1| Preliminary Material Synthesis Sub-5 nm Au NPs were first synthesised according to the literature procedure reported by Murphy et al.72 with the aim of post-synthetically immobilising the isolated product in monolithic ZIF-8. By this method, a Au precursor salt (HAuCl4.3H2O) was rapidly reduced to metallic Au in aqueous solution by the injection of NaBH4 (a strong reducing agent). This induced burst-nucleation of the Au to yield 4.9 ± 1.1 nm particles, as confirmed by TEM (Figure 11). Figure 11| TEM images of Au NPs. a, Low and b, high magnification TEM images of undried Au NPs prior to their immobilisation in monolithic ZIF-8. Since small Au NPs have a well-known tendency to aggregate and grow via uncontrollable Ostwald ripening (i.e. receiving atoms from dissolving smaller particles)73 they were post- synthetically capped by the addition of long-chain polymer, PVP10K. The capped colloid was finally concentrated to 2.5% of its original volume by solvent removal under reduced pressure, before being re-diluted in absolute ethanol to achieve an ethanolic suspension comparable to the ethanolic SnO2 colloids previously applied to NP@monolithZIF-8 synthesis (see Chapter III, Section 1.0). The Au NPs were immobilised in ZIF-8 to yield Au@monolithZIF-8 as a dark purple material. Elemental analysis by ICP-OES confirmed this material to be comprised of 4.1% Au. TEM analysis was repeated, allowing the NP’s size after immobilisation in the monolithic MOF to be studied (Figure 12). Firstly, this confirmed that the NPs had been successfully immobilised – the dense Au particles were apparent as high contrast spots in the lower density Zn-MOF host. Furthermore, clear 2.4 Å lattice fringes corresponding to the [111] planes of a b 100 nm 25 nm 145 metallic Au can be seen in the high magnification images (Figure 12b). The NPs showed only a small amount of agglomeration, with reasonable dispersity and separation between individual particles. This can be attributed to a combination of their PVP capping in addition to the absence of NP drying prior to immobilisation, a common cause of agglomeration.74 However, the images indicate that the NPs increased considerably in both size and polydispersity, with a post- immobilisation diameter of 6.0 ± 1.6 nm. The increase in NP size between initial synthesis and immobilisation is likely to have occurred in the isolation stage of the particle synthesis. During this stage, when significant solvent was removed from the colloid, the increased proximity of the individual particles may have facilitated their growth via Ostwald ripening. Furthermore, the absence of a washing stage in this literature reaction means that residual unreacted Au precursor salt may remain in the solution and may further contribute to the particle’s growth. Figure 12| Au NPs in monolithZIF-8. a, TEM and b, HR-TEM images of composite Au@monolithZIF-8 showing high contrast NPs immobilised inside the MOF. Inset in b, FFT diffraction pattern of an Au NP (white box) showing the diffraction spots generated by the [111] crystal plane. The combined results of this preliminary study are promising, indicating that Au NPs can achieve a high level of doping in monolithZIF-8. However, Au NPs with diameter below 5 nm diameter are required to achieve high catalytic activity for low temperature CO oxidation; an exponential increase in catalytic activity is observed as Au NP size is decreased below 5 nm.54 This is due to the increased concentration of active, unsaturated surface sites in smaller NPs compared to inactive, saturated sites which exist in the bulk. The synthetic procedure was thus modified extensively with the aim of immobilising more active, 2 – 4 nm NPs in the monolithic host. 146 2.2.2| Composite Synthesis Optimisation A procedure for the synthesis of Au seed particles, as reported by Fischer et al.,75 was modified extensively (see Appendix, Experimental) with the aim of immobilising monodisperse sub-5 nm Au NPs in the monolithic MOF. For the optimised synthetic procedure, a 17.5% molar increase in reducing agent (NaBH4) was applied relative to the original. This was increased to encourage more rapid burst nucleation of the Au. The La Mer model of particle growth suggests that rapid nucleation should facilitate the synthesis of both smaller and more monodisperse NPs.16 Furthermore, the surface capping agent (PVP10k) was added to the reaction solution before the reducing agent with the aim of capping the NPs as they formed, rather than post- synthetically (as was the case in the preliminary synthesis). This may further help to limit NP growth and prevent aggregation/Ostwald ripening. Additionally, after the addition of NaBH4, the reaction was mixed for only 20 s before being quenched by its addition to a large volume of ice-cold acetone. This was performed with the aim of permitting the Au seeds only a short growth window before full isolation via flocculation. The acetone was cooled in ice to further ensure that all NP growth would be immediately halted. Subsequently, the precipitated NPs were collected under centrifugation and re-dispersed in ethanol before immediate immobilisation in the monolith. The washing stage was included with the aim of removing the reaction solution, and thus any unreacted precursor (HAuCl4.3H2O) which might otherwise permit continued NP growth. This differs from the lengthy isolation process of the preliminary synthesis, as discussed earlier, in which the reaction solution was mixed for 30 minutes at room temperature and slowly isolated by solvent removal under vacuum. Such treatment permitted the previously small primary particles to grow beyond the target 5 nm threshold (Figure 12). Immobilisation of the Au NPs obtained by the modified reaction yielded Au@monolithZIF-8 as a shiny and robust monolithic composite (Figure 13). When cracked open, the material broke into fragments, rather than crumbling to a powder. This is comparable to previous qualitative observations regarding monolithZIF-8, SnO2@monolithZIF-8 (Chapter III, Section 1.0) and monolithZr-MOF (Chapter II). Furthermore, the material’s internal composition visually matched its external appearance in terms of colour, indicating the (dark red) Au NPs to be immobilised throughout the bulk, rather than existing only on the material’s external surface. 147 Figure 13| Monolithic Au@ZIF-8. Optical image of a composite material comprised of Au NPs, obtained by the modified synthetic procedure, immobilised in monolithZIF-8. As with the preliminary study, the composite material was analysed by TEM imaging. (Figure 14a, b). Again, the immobilised NPs were visible as dark spots amongst the lower contrast MOF. However, the modified synthesis yielded a composite in which the immobilised NPs displayed far better dispersity throughout the host MOF with almost no aggregation apparent. Critically, the size of the obtained Au NPs was significantly reduced (2.6 ± 1.0 nm) relative to the preliminary experiments. These monodisperse particles (Figure 14c) appear ideal for the target application of CO oxidation, being well within the most active particle size range. This novel material, composed of highly active and well dispersed Au seeds permanently immobilised within a porous framework, therefore presents high catalytic potential from an industrial perspective. Figure 14| TEM analysis of Au@monolithZIF-8. a, Low and b, high magnification TEM images of Au NPs, prepared by a modified synthetic procedure after immobilisation in monolithic ZIF- 8 (Figure 13). c, Histogram showing the particle size distribution of Au NPs in the composite material (a, b). 1 cm 0 10 20 30 40 Pe rc en ta ge (% ) 0.0 - 1.0 1.0 - 2.0 2.0 - 3.0 3.0 - 4.0 4.0 - 5.0 > 5.0 Particle size (nm) c 148 2.2.3| Characterisation The elemental composition of the material was studied by STEM-EDX (Figure 15). Here, Au NPs appeared as bright spots; their high atomic mass better scatters incident electrons than the low-density Zn-MOF. While the Au NPs in the current case were too small to permit accurate elemental mapping, selected area EDX spectra were obtained. In Figure 15, the EDX spectra generated from two different composite areas are compared. The region containing bright NP spots (A) shows a clear Au peak (3d5/2) which region B is missing. This strongly supports the characterisation of the observed 2.6 ± 1.0 nm particles within the MOF as that of Au NPs. This result was corroborated by ICP-OES analysis in which a 0.9% Au loading (by wt%) was recorded. This is significantly lower than the 4.1% loading observed in the preliminary material and may stem from the incomplete collection of such small Au particles by centrifugation prior to immobilisation. Furthermore, the NPs growth time was minimised, which may have prevented incomplete reaction of the Au precursor salt. Figure 15| EDX elemental analysis of Au@monolithZIF-8. HAADF-STEM image (left) of Au@monolithZIF-8 showing Au NPs as brighter spots amongst the MOF. White boxes indicate areas selected for EDX elemental mapping (right: A and B). The position of the Au 3d5/2 peak is indicated in A (white box, inset). The material’s composition was further studied by PXRD (Figure 16), which confirmed it to comprise crystalline ZIF-8, comparable to the pure monolithic MOF. This suggests that the MOF doping by the small NPs did not deter its crystallisation. The presence of the Au was 149 indicated by the addition of a broad peak at 2q ~38 which was not previously observed in the MOFs diffraction pattern. The width of this [111] Au reflection is indicative of Scherrer line broadening, consistent with the TEM results which showed the small 2.6 ± 1.0 nm diameter of the Au NPs. Figure 16| PXRD of doped and undoped monolithZIF-8. Overlaid PXRD patterns of monolithZIF- 8 (black) and Au@monolithZIF-8 (red). Indicated is the position of the Au [111] peak referring to the Miller indices, hkl, classification for the crystal plane of Au which generated this reflection in the pattern. Each monolithic composite was gently ground to a powder for PXRD analysis. Since a key aim of this project was to develop catalytic composite materials which are industrially viable, the prepared material’s thermal stability was assessed by TGA (Figure 17). While pure monolithZIF-8 displayed high thermal stability with rapid decomposition only above 600 °C, the doped material displayed a more gradual decomposition starting at around 400 °C and finishing by 500 °C. It is apparent that the material’s thermal stability was significantly reduced as a consequence of its NP doping. This may stem from the presence of PVP capping agent, which is known to decompose at 380 °C. The decomposition of this organic material may encourage more rapid decomposition of its host MOF by the formation of e.g. reactive radicals during combustion. Likewise, the greater overall weight loss of the composite material compared to monolithZIF-8 is correlated with this organic decomposition while residual weight is attributable to the presence of non-combustible inorganic matter i.e. metal and metal oxide residue. 10 20 30 40 50 Au111 Angle (2θ) In te ns ity (a .u .) 150 Figure 17| Thermal stability of Au doped monolith. TGA traces of monolithZIF-8 (black) and Au@monolithZIF-8 (red) between 50 – 800 °C (under N2 atmosphere). Data for monolithZIF-8 was digitised from Reference 67. The material’s mechanical stability was studied by nanoindentation (Figure 18). While the pure monolith has previously returned H = 0.42 ± 0.03 GPa, the Au doped material exhibited a 20% reduction in stability to 0.34 ± 0.07 GPa. The pure monolith’s Young’s modulus (E = 3.57 ± 0.22 GPa) was also reduced by 18% to 2.94 ± 0.56 GPa. However, the opposite phenomenon was reported by Wheatley et al. in the case of SnO2 immobilisation in monolithZIF-8, where a 10% enhancement in both E and H was observed following doping.35 They attributed this to the enhanced packing efficiency of primary particles in the composite material. Furthermore, significantly greater variation in the obtained results was observed over the series of 16 indents than in the pure monolith; the Load vs. Penetration curves (Figure 18a) were notably spread which was not observed in the original report.35 Likewise, the error bars in the H (Figure 18b) and E (Figure 18c) were much wider, suggesting a reduction in overall material homogeneity. While these results demonstrate that the monolith doping process has rendered the composite material less mechanically robust, this does not mean it is not industrially viable; the calculated values remain within the range of values associated with robust ZIFs.76 However, these results do highlight the importance of careful consideration of the consequences that material doping/modification can have on its overall viability. 20 40 60 80 100 50 250 450 650 Temperature (oC) W ei gh t ( % ) 151 Figure 18| Mechanical testing of Au@monolithZIF-8. Mechanical properties of monolithic ZIF- 8 doped with Au NPs, as obtained by nanoindentation. a, Load (mN) vs. Penetration into surface of the monolith (h, nm) for 16 indents. b, c, Hardness (H, GPa) and Young’s Modulus (E, GPa), respectively, as a function of Penetration depth (h, nm) into the monolith surface. Mean properties and corresponding errors (inset in b and c) were obtained from measurements taken across 16 indents over penetration depths of 250 – 2000 nm. Measurements obtained in the sub- 250 nm penetration range were excluded, to eliminate errors due to surface defects/tip artefacts. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 0 500 1000 1500 2000 0.0 0.1 0.2 0.3 0.4 0.5 0 500 1000 1500 2000 Penetration into surface, h (nm) Penetration into surface, h (nm) Ha rd ne ss , H (G Pa ) Yo un g’ s M od ul us , E (G Pa ) H = 0.34 ± 0.07 GPa E = 2.94 ± 0.56 GPa b c 0 500 1000 1500 2000 0 5 10 15 20 25 30 35 a Penetration into surface, h (nm) Lo ad (m N) 152 2.3| PdO/TiO2@monolithZIF-8 2.3.1| NP Synthesis A bimetallic nanocomposite of Pd/TiO2 was synthesised through medication to a procedure for Pt/TiO2; the Pt precursor salt was swapped for an equimolar equivalent of Pd salt.30 By this method, Pd NPs were first fully synthesised under reflux before the Ti precursor, Ti(OBu)4, was injected to achieve rapid TiO2 NP synthesis. The bimetallic composite was subsequently isolated and further calcined at 500 °C to ensure complete removal of organic species as well as well as maximising crystallinity of the metal oxide NPs. This high temperature isolation stage of the synthesis prevented the use of undried NPs for immobilisation in the MOF, a method found to improve doping in the case of SnO2 (Chapter III, Section 1.0). The nanoscopic structure of the calcined NPs was first visualised by TEM (Figure 19). The expected composite morphology was observed; smaller high contrast NPs (Pd, 6.7 ± 2.6 nm) dispersed throughout inhomogeneous, larger particles with lower contrast (TiO2). This is comparable to the previously reported Pt/TiO2 nanocomposite from which this synthesis was derived.30 The smaller, high contrast NPs can likely be assigned as Pd due to its atomic mass being higher than that of TiO2, an inherent result of the TEM imaging. The composite was synthesised by first isolating the Pd NPs followed by the subsequent addition of the Ti precursor and hence, some of the small Pd NPs may be located within the larger metal oxide NPs i.e. Pd@TiO2. However, the clear presence of darker Pd NPs on the edges of the lighter TiO2 (Figure 19) demonstrates that at least some of the Pd is not encapsulated and hence its surface remains accessible. The composition consisting of a major TiO2 component, with a minor Pd phase distributed throughout, was further supported by ICP-OES analysis. Here, a Pd loading of 2.7 wt% was recorded within the TiO2-based nanocomposite. 153 Figure 19| TEM images of Pd/TiO2 NPs. a, b, Low and higher magnification, respectively, TEM images of bimetallic nanocomposite of Pd and TiO2. PXRD was used to further elucidate the crystal structure of the bimetallic nanocomposite (Figure 20) with the resulting trace showing two distinct phases. The crystallinity of the anatase form of TiO2 is suggested by its sharp reflections, attributable to the material’s lengthy calcination at 500 °C. Further reflections originating from the Pd component are also visible, the positions of which indicate this to exist as Pd(II) oxide. This is again accounted for by the composite’s air-based calcination, which oxidises the metal. The lower intensity of the PdO component, relative to the TiO2, is consistent with its lower fraction in the nanocomposite (as indicated by ICP-OES and TEM (Figure 19) analysis). Figure 20| PXRD analysis of PdO/TiO2 composite. PXRD pattern of PdO/TiO2 NPs. Indicated (inset) is the Miller indices classification, hkl, for the crystal plane which generated each reflection in the pattern for both TiO2 (red) and PdO (black). 10 20 30 40 50 60 70 80 101 Angle (2θ) In te ns ity (a .u .) 110 103 105 211 101 004 200 204 116 220 202 215 TiO2 PdO 154 These data demonstrate that the target material, a PdO/TiO2 nanocomposite, was successfully synthesised. By TEM, the Pd component appeared well distributed throughout its TiO2 support and the obtained NPs of each component further appeared to be sufficiently small for immobilisation in the monolithic MOF, monolithZIF-8. No further peaks or impurities were observed in the PXRD patterns and both components displayed high crystallinity. 155 2.3.2| Monolithic Composite Synthesis and Characterisation The PdO/TiO2 nanocomposite was dispersed in ethanol under extensive sonication to produce a colloid and the procedure for SnO2@monolithZIF-8 was replicated using instead the bimetallic dopant. The immobilisation of the brown NPs (Figure 21a) in the otherwise optically transparent monolith, yielded a doped composite with a consistent brown colour throughout (Figure 21b). This matches previous observations of monolith doping i.e. SnO2@monolithZIF-8 and Au@monolithZIF-8, suggesting immobilisation of the NPs in the host. Figure 21| Bulk morphology of NPs and monolithic composite. a, b, Optical images of PdO/TiO2 NPs and (PdO/TiO2)@monolithZIF-8, respectively. PXRD of the monolithic composite demonstrated a triphasic structure (Figure 22). Firstly, comparison to pure ZIF-8 shows that each of the expected reflections corresponding to the MOF remain and that no loss of crystallinity occurred as a result of the doping; reflections remain sharp and their position has not shifted. Additional reflections corresponding to both TiO2 and PdO are visible amongst the MOF peaks suggesting that both components of the bimetallic nanocomposite were successfully encapsulated by the monolithic MOF. This was further supported by ICP-OES analysis which recorded an elemental Pd loading of 0.7 wt%. Under the assumption that all elemental Pd exists in its (+II) oxidation state i.e. PdO, the monolithic composite was calculated to comprise 0.8% PdO. Highly stable TiO2 could not be adequately dissolved for this elemental analysis, preventing direct quantification. However, the monolithic composite materials composition was further recorded to be 85.3% ZIF-8, leaving ca. 14.0% of the mass unaccounted for. This can reasonably be used to approximate the loading of TiO2 within the monolith. These data suggest a higher PdO:TiO2 ratio than that observed in the pre- synthesised PdO/TiO2 NPs. Firstly, this indicates that the bi-metallic nanocomposite comprised ba 156 an inhomogeneous distribution of PdO throughout the TiO2 i.e. not all of the TiO2 NPs being equally associated with Pd. This is consistent with the synthetic procedure used to obtain the bi-metallic composite; Pd NPs were grown on pre-synthesised TiO2. As discussed, TEM demonstrated that the TiO2 NPs were both polydisperse and in-homogeneous in terms of their size and morphology (Figure 19). Since NP size is linked to its available surface area, large TiO2 NPs (and closely packed agglomerates of smaller NPs) inherently present a diminished accessible surface area on which Pd NPs may nucleate. Hence, the largest TiO2 NPs and aggregates can be expected to display a lower Pd:TiO2 ratio than that of comparably small NPs. Secondly, some quantity of added dopant appeared to have been excluded from the monolith. It may be reasonable to suppose that, as was the case in the synthesis of SnO2@monolithZIF-8 (Chapter III, Section 1.0), large dopant aggregates (i.e. those with comparably low Pd:TiO2 ratio) sedimented out of the colloidal suspension during synthesis and thus were excluded from the growing monolith. This may account for the total Pd:TiO2 ratio in the doped monolith being increased relative to that of the original nanocomposite. Figure 22| PXRD of monolithic PdO/TiO2@ZIF-8. Overlaid PXRD patterns of PdO/TiO2@monolithZIF-8 (yellow) and monolithZIF-8 (black). Indicated is the Miller indices classification, hkl, for the crystal planes of the bimetallic dopant that generated each reflection; TiO2 (red) and PdO (black). Each monolithic composite was gently ground to a powder for PXRD analysis. 10 20 30 40 50 60 70 Angle (2θ) In te ns ity (a .u .) 101 101 TiO2 PdO 200 105 211 204 157 Finally, the elemental distribution throughout the multicomponent composite was mapped by STEM-EDX (Figure 23). STEM-HAADF imaging (Figure 23a) shows the presence of high- intensity regions distributed throughout the lower contrast host. EDX mapping (Figure 23b – d) confirms this to correspond to both Pd and Ti within the Zr-MOF. As expected, the Ti signal is more intense than that of Pd, a consequence of its higher fraction in the bimetallic nanocomposite. The inhomogeneous distribution of these elements throughout the host is indicative of dopant aggregation. This is comparable to FIB-SEM observations of SnO2@monolithZIF-8 (Chapter III, Section 1.0), where micron sized aggregates were observed. These combined results demonstrate that both components of a bimetallic PdO/TiO2 nanocomposite were immobilised in monolithZIF-8 with a high doping level, while further maintaining crystallinity and monolithicity of the host MOF. Figure 23| Elemental maps of (PdO/TiO2)@monolithZIF-8. a, HAADF-STEM image of PdO/TiO2 immobilised in monolithic ZIF-8. b, c, d, Elemental maps showing the distribution of Zn, Pd and Ti, respectively, in the composite. a b c dZn Pd Ti 250 nm 250 nm 250 nm 250 nm 158 2.4| ((Au@PdO)/TiO2)@monolithZIF-8 2.4.1| NP Synthesis A synthetic procedure for a trimetallic nanocomposite of Au, Pd and TiO2 was developed by combining different literature reports for its components. Firstly, the aqueous synthesis of core@shell Au@Pd NPs by Yamauchi et al.77 was replicated and the morphology of the obtained NPs was studied by TEM (Figure 24). Here, slightly polydisperse NPs of 61.1 ± 18.6 nm diameter with an inhomogeneous quasi-spheroidal morphology were observed. Although this is somewhat larger than the 20 nm particles claimed in the original paper, a limited number of high magnification TEM images were provided there and furthermore a particle size range was not reported.77 Despite the observed polydispersity in our newly experimentally obtained material, these particles were sufficient for the preliminary catalytic tests targeted in the early stage of the project. It is of general interest to determine if large NPs can be successfully immobilised in the monolith as the current understanding of NP@monolithMOF encapsulation is limited. HR-TEM imaging reveals the NPs to comprise high contrast cores encapsulated in lower contrast shells. This suggests dense Au to be surrounded by less dense Pd i.e. Au@Pd, the target morphology. STEM-EDX elemental mapping was further performed (Figure 25), showing the presence of both Au and Pd in the NPs. When the signal maps for these two elements are overlaid, the Au core and Pd shell is apparent (Figure 25d), confirming the morphology to be Au@Pd. Figure 24| TEM images of Au@Pd NPs. a, Low magnification TEM image of Au@Pd NPs and b, HR-TEM image showing the core@shell structure of a single particle (area selected for high magnification imaging indicated in a (white box, inset). a b 50 nm 10 nm 159 Figure 25| Elemental maps of Au@Pd NPs. a, STEM-HAADF electron image of a representative Au@Pd NP. b, c, Corresponding elemental maps showing the distribution of Pd and Au, respectively. d, Overlaid map showing the core@shell morphology of the NP. After verification of the initial core@shell target structure, a trimetallic system was pursued. This was achieved by using the aqueous colloidal suspension of pre-prepared Au@Pd NPs as the reaction solution for TiO2 NP synthesis. A hot-injection reaction of Ti precursor, identical to the earlier discussed procedure for Pd/TiO2, was performed thereby dispersing the core@shell NPs amongst the crystallising TiO2. As with the Pd/TiO2 synthesis, the obtained nanocomposite material was calcined (500 °C) to ensure complete removal of organic impurities as well as high crystallinity of the TiO2 component. The calcined nanocomposite’s morphology was first studied by TEM. Low magnification images (Figure 26a, b) revealed the same non-homogenous particles of TiO2 that were previously observed in the Pd/TiO2 composite (Figure 19). In the current case, amongst these non-homogeneous TiO2 particles are higher contrast spots which, upon increased magnification, are revealed to be the core@shell NPs (Figure 26c). These Au@Pd NPs appear to be well dispersed; many individually dispersed particles are visible as well as occasional larger aggregates. TEM data further demonstrates that these Au@Pd NPs have maintained their original morphology (Figure 24), despite exposure to the harsh reaction conditions required for TiO2 calcination i.e. using artificial diffraction pattern generation (FFT) of the lattice fringes observed in the HR-TEM image. The d-spacings obtained from these artificial reflections clearly show the presence of three distinct crystalline phases in the composite. The Pd component was found to exist in its Pd(II) oxide state (Figure 26ci); hence, the core-shell Au@Pd was in fact Au@PdO. Further distinct Au (Figure 26cii) and TiO2 (Figure 26ciii) regions were also recorded. ICP-OES analysis further corroborated the TEM results by a b c d 50 nm 50 nm 50 nm 50 nm Pd Au 160 confirming the presence (by wt%) of both Au; 19.0% and Pd; 8.5% amongst the TiO2. Under the assumption that all Pd exists in its (+II) oxidation state, this corresponds to 9.8 wt% PdO. Figure 26| TEM analysis of (Au@PdO)/TiO2. a, b, c, TEM images of the trimetallic nanocomposite showing the core-shell Au@PdO NPs dispersed throughout the TiO2 NPs. White boxes (inset) indicate areas (i, ii, iii) selected for FFT analysis. i, ii, iii, artificial FFT diffraction patterns showing the distinct regions of the composite to be comprised of PdO, Au and TiO2, respectively. Finally, the crystallinity of the tricomponent nanocomposite material was also studied by PXRD (Figure 27) with the obtained results matching the FFT diffraction patterns generated from the HR-TEM images (Figure 26ci – iii). The nanocomposite was recorded to comprise three distinct crystalline phases, namely Au, PdO and TiO2 (anatase), verifying the isolated elemental regions of the composite as well as the oxidation states of both the Pd and Ti components. These combined data confirm the synthesis of a composite nanomaterial with the targeted composition and morphology for CH4 oxidation catalysis; (Au@PdO)/TiO2. [101] [111] [220] [200] i) PdO ii) Au iii) TiO2 [101] i iii ii a b c 1 µm 300 nm 20 nm 161 Figure 27| PXRD analysis of (Au@PdO)/TiO2. PXRD pattern for the nanocomposite comprised of Au@PdO NPs dispersed amongst TiO2 NPs. Indicated is the Miller indices classification, hkl, for the crystal planes which generated each reflection in the pattern: TiO2 (green), PdO (black) and Au (red). 10 20 30 40 50 60 70 80 Angle (2θ) In te ns ity (a .u .) TiO2 PdO Au101 200 105 211 204 101 110 111 200 220 311 116 220 215 162 2.4.2| Monolithic Composite Synthesis and Characterisation The synthetic procedure for monolithic SnO2@ZIF-8 was adapted to immobilise the (Au@PdO)/TiO2 NPs described above. The immobilisation of the grey coloured nanocomposite (Figure 28a) in the otherwise optically transparent MOF, monolithZIF-8, yielded a comparably grey coloured material (Figure 28b). The monolithic properties appeared to have been maintained despite the NP immobilisation, as indicated by the obtained large composite fragments which displayed a shiny appearance. The grey colour was consistent throughout the composite, perfunctorily implying macroscopically homogenous dopant dispersion throughout the host. Figure 28| Bulk morphology of NPs and monolithic composite. Optical images of a, (Au@PdO)/TiO2 and b, ((Au@PdO)/TiO2)@monolithZIF-8. PXRD was used to study the different crystalline constituents of the doped composite material (Figure 29). Comparison to the diffraction pattern for pure monolithZIF-8 shows that all the expected reflections for this MOF are present with no changes in peak width or position. This suggests that the MOF maintained its crystalline structure despite immobilisation of the NPs. Furthermore, a number of additional reflections are present in the obtained pattern. These well- defined peaks are consistent with the diffraction pattern recorded for the trimetallic nanocomposite prior to immobilisation (Figure 27). ICP-OES analysis was used to quantify this doping; Au and Pd comprised 3.8% and 1.9% by weight, respectively, of the composite. Again, TiO2 could not be quantified due to an inability to dissolve this high stability metal oxide for elemental analysis. However, this component’s presence is apparent from its reflections in the PXRD pattern. a b 163 Figure 29| PXRD analysis of (Au@PdO)/TiO2)@monolithZIF-8. Overlaid PXRD patterns of monolithZIF-8 (black) and (Au@PdO)/TiO2)@monolithZIF-8 (blue). Indicated (inset) is the Miller indices classification, hkl, for the crystal planes which generated each reflection in the pattern for the doped trimetallic composite; TiO2 (green), Au (red) and PdO (black). Each monolithic composite was ground to a powder for PXRD. Finally, the elemental distribution of the dopant NPs components throughout the host MOF was mapped by STEM-EDX (Figure 30). Data confirm the presence of Zn, Au, Pd and Ti. Furthermore, the Ti appears well dispersed throughout the host monolith while several distinct areas expressing intense Au and Pd signal are apparent and furthermore appear to overlap. Hence, the crystalline composite nanomaterial of Au@PdO dispersed on TiO2 NPs was successfully immobilised in monolithZIF-8 without loss of MOF crystallinity or macroscopic monolithic appearance; ((Au@PdO)/TiO2)@monoltithZIF-8 was achieved. 10 20 30 40 50 60 70 80 Angle (2θ) In te ns ity (a .u .) TiO2 PdO Au 101 101 111 200 200 105 211 220 311 164 Figure 30| Elemental maps of ((Au@PdO)/TiO2)@monolithZIF-8. a, STEM-HAADF electron image of monolithZIF-8 doped with (Au@PdO)/TiO2. b – e, Elemental maps showing the distribution of Zn, Au, Pd and Ti, respectively, throughout a. c Au Pd Ti 250 nm250 nm Zna b d e 250 nm 250 nm 250 nm 165 2.5| Conclusions A key research aim of this part of the project was to investigate the capabilities of monolithZIF-8 to host a range of catalytic NPs beyond that of previously explored SnO2 (Chapter III, Section 1.0). In particular, the immobilisation of highly active, but expensive, noble metal NPs was explored with the aim of increasing the industrial viability and economic potential of these catalysts through robust immobilisation and correspondingly simpler catalyst recovery. In the first system probed, monometallic Au NPs were immobilised in the monolith of ZIF-8. While the preliminary synthesis was successful in yielding the first example of Au NPs doping this monolith, the previously sub-5 nm particles demonstrated uncontrollable Ostwald ripening, surpassing the target diameter. The synthetic procedure was optimised to prevent NP growth by reducing the reaction time, functionalising the NP surface with organic capping agent (PVP) and quenching the NP growth by rapid flocculation. Through these synthetic modifications, Au@monolithZIF-8 was synthesised with a 1% doping level of 2.6 ± 1.0 nm Au NPs. The crystallinity and monolithicity of the host were maintained despite its doping. The permanent immobilisation of such small, typically unstable Au NPs in an industrially viable and porous host appears to be unique in the literature. Such small NPs are expected to show high activity towards CO oxidation54 and furthermore comparable NPs have demonstrated high activity towards this reaction when immobilised in powdered ZIF-8. For example, Zhu et al. reported a reduced energy barrier to CO oxidation for Au NPs immobilised in the MOF compared to free particles suggesting an enhancement in activity as a consequence of the doping.78 Yet in the current case, when the monolithic composites catalytic activity towards CO oxidation was experimentally tested by collaborators at Stony Brook University (USA), negligible activity was detected. This somewhat unexpected result may be accounted for by a number of practical considerations. Firstly, bulky organic polymer (PVP10k) was used to functionalise the NPs surface. This was an essential synthetic requirement to prevent uncontrollable aggregation and growth of the NPs, such as was observed in the preliminary syntheses. However, capping agents may passivate the particles by hindering the approach of target substrates towards the active surface. Secondly, the organic molecules on the NP’s surface may further block the MOFs pores, reducing rate of diffusion through the crystalline, porous network. Finally, at 1% doping, the low loading of active NPs in the MOF may have contributed to the poor detectable activity. 166 With the aim of determining if more morphologically complex NPs can dope the same host, a bimetallic composite of PdO/TiO2 was synthesised and fully characterised. Subsequently, this nanocomposite was resuspended as a colloid and successfully encapsulated in monolithZIF-8. XRD was used to demonstrate that both of the NP components were immobilised to some extent and that high crystallinity of all three constituents of Pd/TiO2@monolithZIF-8 was achieved. The novel materials capacity to photocatalytically catalyse the hydrogenolysis of 2-propen-1-ol to propene was studied. Again, however, this NP@monolithMOF composite displayed an absence of activity towards this reaction. Since, non-immobilised PdO/TiO2 NPs have previously demonstrated high activity towards this reaction, it is reasonable to again suppose that the catalytic activity has been inhibited by its immobilisation in the host MOF. It has previously been reported that small organic and inorganic molecules, including propene (ca. 4 Å kinetic diameter), are capable of entering ZIF-8’s narrow porosity (3.4 Å windows) due to the swinging motion of the imidazolate linkers.79 However, the close fit may limit the rate of reactant diffusion towards the catalytic surface of the NPs inside the MOF, preventing efficient catalysis. For this reason, catalytic reactions in which the target substrate is a large organic molecule which must directly approach the NPs surface may not be applicable to catalysts immobilised in a porous host unless the hosts pore-size can be substantially increased. As discussed in Chapter I, while the capacity to synthesise a range of monolithic MOFs with different pore sizes has been demonstrated, preliminary results suggest that the synthesis of crystalline MOFs with organic linkers substantially larger than that of ZIF-8, and hence wide porosity i.e. monolithNU-1000, appears non-trivial. Finally, a trimetallic composite comprised of core@shell Au@PdO dispersed on a TiO2 support was synthesised and immobilised in monolithZIF-8. The encapsulation of all three components in the host was demonstrated by analysis of the obtained monolithic material. This result demonstrates the monolith’s capacity to host a wide range of morphologically varied and multicomponent nanocatalysts. However once again, when applied to CH4 oxidation catalysis, no activity was recorded. Considering the results of this project as a whole, several key findings are apparent. Firstly, a wide range of NPs with different sizes, compositions and morphologies can be immobilised in monolithic ZIF-8 without compromising the hosts monolithicity or crystallinity. This is a significant result as it offers an industrially viable means of achieving economic and environmental catalyst immobilisation for a wide range of reactions. Yet the inability of each 167 of the three composite materials (Au@monolithZIF-8, PdO/TiO2@monolithZIF-8 and Au@PdO/TiO2@monolithZIF-8) to achieve any measurable catalytic activity towards CO oxidation, alkylation or CH4 oxidation, respectively, suggests a considerable limitation of using monolithMOFs to host catalytic NPs. Despite computational simulations indicating that each of these substrates are capable of penetrating the MOFs porosity, the diffusion rate through the micropores (< 2 nm) appears to be the limiting factor in the catalytic reaction.79 Additionally, the organic surface capping agents often applied to NP synthesis may further reduce the accessibility of the active surface by blocking pores and further slowing reactant diffusion. Some of the most promising potential applications of noble metal NPs are the catalysis of otherwise energetically demanding organic reactions. The negative results from the alkylation of 2-propen-1-ol using PdO/TiO2@monolithZIF-8 suggest that even relatively small organic molecules cannot efficiently diffuse through this MOFs porosity. Furthermore, even for faster gas phase oxidation of the small molecule CH4 using ((Au@PdO)/TiO2)@monolithZIF-8, negligible activity was recorded. These observations can be compared to the results for photocatalytic water purification using SnO2@monolithZIF-8 (Chapter III, Section 1.0). Significant enhancements in NP doping level did not correlate with proportional improvements in catalytic activity. It is apparent that nanocatalysts, no matter how active, may be passivated to some extent by their immobilisation in the monolithic host. Hence, although compositionally diverse NP@monolithMOF composites can be obtained, development of functional materials appears to be less trivial. Careful consideration of the best NP@MOF combination for each application is vital in the development of novel functional materials. One viable research avenue may lie in the utilisation of the MOFs defined and narrow porosity as a confined reaction chamber. An alternative approach may be to pursue applications which are not kinetically limited i.e. high-density gas storage is an area in which monolithic MOFs have already demonstrated outstanding results and diffusion kinetics is of less practical concern than in rate- dependant catalytic applications. 168 3.0| Monolithic Pd@HKUST-1 HKUST-1 is perhaps the most studied Cu-MOF.80 In this structure, Cu(II) ions combine with tricarboxylate linkers and water (when the MOF is hydrated) to adopt a characteristic ‘copper– copper paddlewheel’ dimer arrangement (Figure 31a).81 These SBUs further assemble to a cubic topology of wide 9 x 9 Å square micropores (Figure 31b), the result of which is a high theoretical SBET that exceeds 2000 m2 g–1. Furthermore, activation of the MOF results in a dehydrated crystal structure with coordinatively unsaturated Cu(II) sites that are available for a range of catalytic processes as well as enhanced gas storage. However, a consequence of its active, open metal sites is a high affinity for water, adsorption of which results in gradual loss of crystal structure i.e. under ambient conditions. Despite its air-sensitivity, this fascinating MOF has been extensively researched for numerous likely applications including heterogeneous oxidation catalysis,82 new antibacterial materials83 and chemical sieving.84 In particular, its gas adsorption properties have been comprehensively studied, including for CO2,85 H286 and CH4 storage.87 Despite displaying outstanding computational and experimental results for gas storage, HKUST-1 has traditionally suffered low industrial viability as a result of its powdered morphology, which is typical amongst MOFs. As discussed (Chapter I), the low packing efficiency of powdered MOFs reduces their bulk volumetric gas storage capacity under practical application. Yet, attempts to densify HKUST-1, while retaining its sought-after physical properties have often fallen short. For example, Kim et al. studied the post-synthetic densification of HKUST-1 powder both with and without PVA binder.88 Under the application of 25 bar pressure, a 35% loss of pore volume (relative to the original powdered MOF) was observed while the binder was observed to further reduce accessible surface area to only 55% of the original. A proportional loss of CO2 adsorption capacity was recorded for the pelletised materials, highlighting the detrimental effects of the densification process. In 2019, Tian et al. demonstrated the generality of their sol-gel monolith procedure by extending the previously reported synthesis of pure monolithZIF-867 to monolithHKUST-1.89 The centimetre scale material’s robust mechanical properties, high SBET and rb presented it as an ideal candidate for industrial gas storage. Correspondingly, it displayed a benchmark 259 cm3 (STP) cm–3 CH4 storage capacity (65 bar, 298 K), representing a 50% improvement over the 169 previous benchmark and making it the first densified material to reach the U.S. DOE target for CH4 uptake (Figure 31c) Figure 31| HKUST-1. a, b, Cu–Cu paddlewheel dimer and HKUST-1’s crystal structure, respectively, where coloured spheres correspond to the elements Cu (blue), O (red), C (grey) and H (white). In the dimer, a, axial oxygen atoms originate from coordinated water molecules while bridging groups show the orientation of the carboxylate groups from the organic linker, btc. c, Volumetric CH4 adsorption isotherm (0 – 70 bar, 298 K) for monolithHKUST-1 (red circles, absolute (filled). The U.S. DOE target for volumetric CH4 storage (263 cm3 (STP) cm–3) is indicted (dashed red line, inset). Data digitised from Reference 89. Besides the promising physical/chemical properties demonstrated by the pure material, HKUST-1 has further been applied to the synthesis of composite materials. For example, powderNP@HKUST-1 has incorporated Ag, Fe3O4 and CuO, demonstrating encouraging results for hydrolytic pollutant degradation, biodiesel production and electrochemical reactions respectively.90–92 In particular, the immobilisation of NPs in HKUST-1 for the enhanced storage of H2 gas has garnered extensive attention. One key example of this is the benchmark paper by Kitagawa et al., who reported the novel immobilisation of Pd nanocubes (10 nm) in HKUST-1 (Figure 32). This yielded a powdered MOF composite which exhibited both a 74% increase in H2 storage capacity of the Pd at 1 bar as well as improved adsorption kinetics relative to that of the pure NPs.93 The electronic origin of this altered sorption capacity was subsequently a b 0 50 100 150 200 250 300 0 20 40 60 c Pressure (bar) CH 4 up ta ke cm 3 (S TP ) c m -3 170 elucidated using high energy synchrotron radiation.94 They demonstrated that the nanocomposite material is more than a physical mixture of two separate components but rather an electronically interacting amalgamation of the two; data indicated that an electronic Pd–Cu– O interaction takes place at the NP/MOF interface with charge transfer from Pd 4d band to the hybridized Cu 3d – O 2p band in the MOF (Figure 32). The holes incurred in the Pd 4d band enhance the H2 sorption capacity – they are filled by 1s electrons of atomic H via Pd–H formation. Figure 32| Pd@HKUST-1 electronic interaction. Representation of electron transfer from immobilised Pd NPs (green spheres) to a surrounding MOF, HKUST-1 (red). Reproduced from Reference 94 with permission from Springer Nature. Palladium nanoparticle Pd@HKUST-1 Enhanced H2 uptake a b 171 3.1| Aims and Objectives The promising reports of gas storage in HKUST-1 and NP@HKUST-1 outlined above, coupled with the report of benchmark monolithHKUST-1 earlier this year, present a timely opportunity. The immobilisation of NPs in these sol-gel monolithic MOFs has not yet been extended beyond monolithZIF-8. The study of NP@monolithHKUST-1 would further demonstrate the generality of the monolithic MOFs capability to immobilise NPs, crucially expanding the repertoire of monolith hosts to include this fascinating Cu-MOF. This is critical for the design of tuneable NP@monolithMOF materials, tailored to specific applications where e.g. differing pore sizes/geometries, chemical and physical properties of the host may be required. Specifically, the immobilisation of Pd NPs in this dense/porous monolithic material may yield a composite which displays both the industrially valued physical properties imbued by the host (mechanical strength, centimetric scale macrostructure etc.) as well as enhanced H2 storage capacity by the dopant. The key objective in this case is to determine to what degree NPs can be immobilised in the monolith. The doping level in addition to the dispersion/aggregation of particles throughout the host are areas which will require significant study. Furthermore, the effect that the foreign NP species exert on the MOF must be elucidated. For gas storage, it is essential that the synthesis of the crystalline MOF, with high surface area and accessible porosity, is not impaired by the addition of the dopant NPs. SBET, Vtot, PSD and rb, as well as other physical properties, e.g. thermal and mechanical stability, will be benchmarked against the undoped monolith. Furthermore, while the electronic interaction between Pd and Cu in the powdered composite has been extensively studied by Kitagawa et al., the nature and extent of this interaction must be verified for the analogous monolithic material. 172 3.2| Monolithic Composite Synthesis Cubic Pd NPs (8.9 ± 1.8 nm) were synthesised according to the literature procedure by Kitagawa et al. (Figure 33a).93 The experimentally obtained NPs displayed the same monodispersity and absence of aggregation as those originally reported. HR-TEM imaging (Figure 33b) of individual NPs demonstrates their crystalline lattice fringes. Generation of an artificial diffraction pattern by FFT analysis of these fringes shows diffraction maxima whose measured d-spacings suggest the [111] and [200] planes of metallic Pd. Figure 33| TEM analysis of Pd NPs. a, b, Low and high magnification, respectively, TEM images of Pd NPs. Inset in b, FFT diffraction pattern generated from a crystalline Pd NP (region used for FFT analysis indicated (dashed white box, inset). Diffraction maxima originating from i) [111] and ii) [200] fringes are indicated in the FFT pattern. In the preliminary study of monolithic SnO2@ZIF-8 (Chapter III, Section 1.0), a significant twelve-fold enhancement in NP loading was achieved by the immobilisation of undried SnO2 NPs, relative to the dried NPs from the original procedure.35 It is well known that drying NPs routinely incurs irreversible aggregation to form large clusters.74 It was postulated that undried particles were more capable of forming a stable colloidal ethanolic suspension due to their reduced aggregation. The suspended NPs were thus more efficiently immobilised by the growing MOF than the dried NPs during synthesis. Consequently, prior to immobilising Pd nanocubes in monolithHKUST-1, the particles were not dried. Instead the washed particles (isolated by centrifugation) were re-suspended in ethanol and the concentration was determined by recording the dry mass of a known colloid volume. The volumes of the colloid required to i) [111] 0.23 nm ii) [200] 0.19 nm i iia b 173 achieve a monolith with 5% and 10% NP loading by weight were then diluted to form the reaction solution for the primary MOF particle synthesis. This was identical to the procedure used for NP immobilisation in monolithZIF-8 – dopant particles were suspended in the MOF reaction solution and the growing MOF was subsequently assembled around them.67 Besides this inclusion of dopant NPs in the early stage of MOF synthesis, the remaining synthetic procedure was identical to the literature procedure for monolithHKUST-1.89 By this method, centimetre scale monoliths of Pd@HKUST-1 were obtained, comparable to the original undoped monolith reported by Tian et al. (Figure 34).89 The obtained materials exhibited the glassy, shiny appearance which has become synonymous with monolithic MOFs. Yet the blue colour of HKUST-1 was changed to black; the colour of the pre-synthesised Pd colloid. The materials were fully characterised to elucidate the NP immobilisation in the monolithic MOF in addition to any physical changes in the MOFs structure as a result. Figure 34| Optical images of monoliths. a, monolithic HKUST-1, b, monolithic HKUST-1 with a targeted 5% loading of Pd NPs and c, monolithic HKUST-1 with a targeted 10% loading of Pd NPs. 5 mm 5 mm 5 mm a b cmonolithHKUST-1 5% Pd@monolithHKUST-1 10% Pd@monolithHKUST-1 174 3.3| Characterisation 3.3.1| Composition The Pd doping level of the experimentally obtained composites was quantified by ICP-OES (Table 2). Firstly, the Pd loading of the doped monoliths was revealed to be relatively high at 5% and 10% by weight respectively. This significantly exceeds the proof of concept 2% SnO2 loading reported for monolithZIF-8.35 This high loading can be attributed to the colloidal suspension of ‘undried’ NPs applied to the composite synthesis, consistent with previously discussed observations (Chapter III, Section 1.0). Furthermore, the results suggest that the doping level can be varied with a high level of experimental control, not otherwise observed for the case of SnO2 NP immobilisation. There, a quantity of dopant was excluded from the monolith during drying. The current finding is critical in the design and reproducible synthesis of composite materials for particular applications. Furthermore, each of the MOFs experimentally obtained in this study displayed an elemental composition consistent with the hydrated structure, Cu3(btc)2(H2O)3. This was likewise reported in the original work by Tian et al.89 While all the experimental materials were activated (forming Cu3(btc)2) through prolonged heating under vacuum, theoretically removing all foreign species from the porosity, even brief exposure of this hygroscopic material to water in the air (e.g. under ambient conditions) enables re-hydration. The recorded trace quantities of nitrogen may be residual nitrate ions from the Cu precursor used for the MOF synthesis (see Appendix, Experimental). Table 2| Elemental composition (wt%) of monolithic HKUST-1, 5% Pd@HKUST-1 and 10% Pd@HKUST-1 compared to the theoretical compositions of dehydrated and hydrated HKUST- 1. Data obtained by ICP-OES. Composition (%)* C H N Cu Pd HKUST-1† 36.0 1.0 0.0 31.0 0.0 HKUST-1 hydrated† 33.0 2.0 0.0 29.0 0.0 monolithHKUST-1 33.0 2.2 0.2 28.4 0.0 5% Pd@ monolithHKUST-1 32.1 2.0 0.0 27.1 5.3 10% Pd@ monolithHKUST-1 31.0 1.7 0.2 26.1 10.2 *Data obtained by ICP-OES. †Theoretical elemental compositions for hydrated (Cu3(btc)2(H2O)3) and de-hydrated Cu3(btc)2 MOF assuming defect-free crystal structures. 175 3.3.2| Morphology and Crystal Structure The obtained monolithic materials were first studied by TEM. Pd NPs were clearly observed as high contrast 10 nm cubes (Figure 35a) mixed amongst the low contrast Cu-MOF. The observation of small dark spots in the surrounding MOF is consistent with original observations reported by Tian et al. who characterised these spots as dense amorphous regions of MOF.89 This observation is further comparable to the dense spots observed in monolithUiO-66-NH2, monolithUiO-66-ndc and monolithNU-1000 (Chapter II, Sections 2.0 – 4.0) The darker appearance of the cubic NPs stems from the inherent contrast mechanism of TEM imaging whereby denser materials (e.g. Pd) more efficiently prevent incident electrons from reaching the detector than do lower density materials (e.g. the porous MOF). To further support the material’s elemental composition, STEM-EDX analysis was used to elementally map a selected area (Figure 35b – f and Appendix, Supplementary Figure 15). This confirmed that the nanocubes were comprised of Pd and were immobilised in the copper-based host. Visually, the nanocubes displayed reasonable dispersion throughout the MOF with only slight clustering in certain regions. However, distinct particle agglomeration/direct contact between particles was not apparent. Figure 35| Electron microscope images of Pd@monolithHKUST-1. a, TEM image of Pd@monolithHKUST-1. b, c, Low and high magnification, respectively, STEM images of the composite. Dashed white box (inset in c) shows area selected for electron, Pd and Cu mapping (d – f, respectively). 25 nm 25 nm25 nm CuPd a b c d e f 176 The 3D surface morphologies of the dried monoliths were compared by SEM (Figure 36). For each monolithic material, the expected smooth and homogenous surfaces were observed at low magnification while at further enhanced magnification, they were resolved into densely packed NP arrays. These SEM observations are consistent with those of the original monolithHKUST-1 material89 as well as other monolithic materials reported herein; monolithUiO-66, monolithUiO-66- NH2, monolithUiO-66-ndc and monolithNU-1000 (Chapter II). For the Pd-doped material, high contrast spots are visible, attributable to dopant on the composite’s surface. Crucially, the absence of distinct surface features or defects indicates that the monolith formation mechanism (i.e. dense primary particle packing followed by epitaxial primary particle growth) was not disturbed by the inclusion of the foreign NP species. This is further consistent with SEM observations of the comparably doped monolith, SnO2@ZIF-8, by Wheatley et al.35 Figure 36| Surface morphology of monolithic MOFs. SEM images showing the surface morphology of monolithic HKUST-1, 5% Pd@HKUST-1 and 10% Pd@HKUST-1 at successively increased magnification (a – c). m on ol ith H K U ST -1 10 % P d@ H K U ST -1 5% P d@ H K U ST -1 177 Finally, the monolith’s internal structure was studied by X-ray tomography. Through this technique, the location of the composite’s various components was mapped, enabling reconstruction in 3D space (Figure 37a – c). Through thresholding of the obtained results, the recorded signal attributable to the bulk monolith was eliminated. This permited visualisation of the Pd NP distribution throughout the host monolith (Figure 37ai – ci). The results corroborate the discussed TEM observations (Figure 35); dopant NPs are not homogenously distributed throughout the internal structure of the monolith. While some areas display high dopant saturation, other regions are more sparsely doped. This may stem from the unavoidable use of centrifugation in monolith synthesis; dense aggregates of NP are selectively pulled towards the base of the centrifuge tube with lighter MOF particles being displaced to the top. This may suggest the presence of a dopant gradient throughout the material. The critical finding, however, is that the NPs are conclusively located within the monolith, rather than on its external surface. This first example of NP immobilisation in monolithHKUST-1 extends the known repertoire of NP@monolithMOF and suggests that a wider range of composite materials may be developed/tuned for different prospective applications. Figure 37| X-ray tomography of monolithic NP@MOF composite. 3D reconstruction of monolithic Pd@HKUST-1 by X-ray tomography. a – c, A fragment of monolithic MOF (blue) in different spatial orientations and ai – ci, 3D distribution of Pd NPs throughout the MOF fragment (a – c). Pd @ m on ol ith H K U ST -1 Pd in sid e m on ol ith 178 Consistent with the above results, PXRD of the synthesised materials characterised them as composites of both Pd and HKUST-1 (Figure 38). The crystallinity of the MOF is confirmed by comparison to the theoretical diffraction pattern, simulated from the ideal crystal structure. The slight broadening of the diffraction peaks in all samples is consistent with previously discussed diffraction patterns of monolithic MOFs (monolithUiO-66, monolithUiO-66-NH2 and monolithUiO-66-ndc) where Scherrer line broadening resulting from the nanoscale dimensions of the primary MOF particles was observed (Chapter II). Reflections attributable to Pd NPs ([111] and [200]) are also visible in the doped composites, again with distinct line broadening. The intensity of the Pd peaks was further observed to vary amongst the materials, supporting the ICP-OES results, which showed corresponding variations in NP loading. Figure 38| XRD patterns of MOF and NP@MOF materials. Simulated XRD patterns generated from the ideal crystal structures of HKUST-1 (red) compared to the experimental PXRD patterns of Pd NPs (black), monolithHKUST-1 (green), 5% Pd@monolithHKUST-1 (blue) and 10% Pd@monolithHKUST-1 (purple). Dashed red line (inset) indicates the position of the [111] and [200] reflections in metallic Pd. 10 20 30 40 50 In te ns ity (a .u .) HKUST-1 simulated 20 % Pd@monoHKUST-1 10 % Pd@monoHKUST-1 Pd nanoparticles monoHKUST-1 Angle (2q) 20 onolithHKUST-1 10 onolithHKUST-1 monolithHKUST-1 Angle (2θ) In te ns ity (a .u .) [111] [200] 179 3.3.3| XPS Finally, the composites were studied by XPS. This spectroscopic technique was previously employed by Kitagawa et al. to probe the electronic interaction between Cu and Pd in powdered Pd@HKUST-1.93 They disclosed the altered electronic states of the NPs immobilised in the MOF relative to the original Pd. The increased BE of Pd 3d3/2 and 3d5/2 coupled with the comparably decreased binding of 2p1/2 and 2p3/2 in Cu was accredited to the electronic interaction of the two i.e. electron transfer from Pd to Cu. These results were used to account for the enhanced H2 uptake in the composite by postulating that the transfer of electron density from Pd increases its density of 4d band holes. When diatomic H2 gas is chemisorbed by Pd to form Pd–H, these 4d band holes are filled and thus increased band holes can be inferred to enhance H2 storage capacity. These results were subsequently corroborated by high energy synchrotron X-ray analysis of the materials.94 In the current case, the BE of Pd 3d3/2 and 3d5/2 in the monolithic composite was again observed to increase relative to the original NPs (Figure 39). Furthermore, the recorded BEs were near identical to the literature values obtained for the powdered composite,93 suggesting a comparable electronic interaction in each case (Table 3). Figure 39| XPS patterns of Pd doped monoliths. XPS data for 5% Pd@monolithHKUST-1 (blue) and 10% Pd@monolithHKUST-1 (purple). Peaks corresponding to Pd 3d3/2 and 3d5/2 are indicated (black arrows, inset) as is the relative ratio of the two. 344 342 340 338 336 334 332 330 6,000 6,500 7,000 7,500 344 342 340 338 336 334 332 330 9,000 10,000 11,000 12,000 In te ns ity (c /s ) Binding energy (eV) Pd3d3/2 Pd3d5/2 Pd3d3/2 Pd3d5/2 Pd3d3/2: Pd3d5/2 0.95 Pd3d3/2: Pd3d5/2 0.94 180 Table 3| Comparison of Pd and Cu BE, as obtained by XPS (Figure 39 and Figure 40), for Pd NPs, powdered Pd@HKUST-1 and 0%, 5%, 10% Pd doped (by wt%) monolithic HKUST-1. *Data from literature report by Kitagawa et al. (Reference 93). For peaks in the same XPS dataset that are attributable to Cu (Figure 40), no shift in BE comparable to that noted for Pd was observed (Table 3). Kitagawa et al. previously reported a decrease in Cu 2p1/2 and 2p3/2 BE for their doped material; a result of increased electron density via donation from the dopant NPs. Yet in the current case, the Cu 2p1/2 and Cu 2p3/2 BE in the doped monoliths are near comparable to those recorded for undoped monolithHKUST-1. This is surprising given the discussed shift of Pd BE peaks in the same materials. However, the broadness and intensity of these unshifted Cu peaks was noted to change as a result of the doping (Table 4). For example, the FWHM of Cu 2p3/2 was reduced by ca. 57% in each of the doped monoliths. Furthermore, a reduction in the 2p satellite peaks relative intensity was also recorded in each case. These observations (narrowing of 2p3/2 and reduced intensity of the Cu2+ satellite peak) are consistent with partial reduction of the Cu atoms (Cu2+ ® Cu+) and support the donation of electron density from the dopant Pd to the host Cu atoms in the composites. BE (eV) Pd 3d3/2 Pd 3d5/2 Cu 2p1/2 Cu 2p3/2 Pd NPs * 339.4 334.2 HKUST-1* 954.8 934.7 monolithHKUST-1 955.1 935.5 Pd@HKUST-1* 340.6 335.7 952.9 933.3 5% Pd@monolithHKUST-1 341 335.7 955.0 935.0 10% Pd@monolithHKUST-1 341 335.7 955.3 935.3 181 Table 4| Comparison of 2p3/2 and satellite (sat) peak intensity and FWHM for Cu in the XPS data (Figure 40) of monolithic HKUST-1, 5% Pd@HKUST-1 and 10% Pd@HKUST-1. Peak Intensity Cu 2p3/2 FWHM Cu 2p3/2 (c/s) Cu 2p sat (c/s) Cu 2psat: Cu 2p3/2 (ratio) monolithHKUST-1 4.9 40027 33901 0.85 5% Pd@monolithHKUST-1 2.7 54696 40082 0.73 10% Pd@monolithHKUST-1 2.8 33577 25543 0.76 Figure 40| XPS patterns of doped and pure monoliths. XPS data for monolithHKUST-1 (green), 5% Pd@monolithHKUST-1 (blue) and 10% Pd@monolithHKUST-1 (purple) for BE between 930 – 965 eV. Peaks corresponding to Cu 2p1/2, 2p satellite and 2p3/2 are indicated (black arrows, inset). 965 960 955 950 945 940 935 930 20,000 25,000 30,000 965 960 955 950 945 940 935 930 30,000 40,000 50,000 965 960 955 950 945 940 935 930 20,000 25,000 30,000 35,000 40,000 Binding energy (eV) In te ns ity (c /s ) Cu2psatellite Cu2psatellite Cu2psatellite Cu2p3/2 Cu2p3/2 Cu2p3/2Cu2p1/2 Cu2p1/2 Cu2p1/2 182 3.3.4| Porosity and Density The porosity of the doped and undoped monoliths was studied by recording their N2 adsorption isotherms (Figure 41a). Each of the monolithic MOFs showed reasonably high N2 uptake in the microporous pressure range (P/Po < 0.1), with calculated SBET ranging between 1400 – 1550 m2 g–1 (Table 5). As expected, Pd doping of 5% or 10% (by wt) incurred proportional reductions in composite SBET. This originates from the reduced gravimetric porosity of the material as microporous MOF is replaced by non-porous and dense Pd NPs. It is promising, however, that no further loss of porosity was observed for the doped materials, except the mass replaced by Pd. This suggests that the formation of crystalline MOF with accessible porosity was not inhibited despite the presence of foreign Pd particles in the reaction solution during MOF assembly. This is further supported by consideration of the semi-logarithmic adsorption isotherm (Figure 41b) which shows comparable isotherm shapes in each case, though differing total N2 uptake capacities. This differs from the results of SnO2@monolithZIF-8, where the composite’s SBET was diminished by ~25% despite low doping levels of only 2% by weight.35 Wheatley et al. attributed this to pore blocking by the non-porous dopant phase. Figure 41| N2 adsorption and PSD. a, Adsorption isotherms showing gravimetric N2 uptake at 77 K for monoliths HKUST-1 (green circles), 5% Pd@HKUST-1 (blue squares), 10% Pd@HKUST-1 (purple diamonds), and theoretical isotherm simulated from the pristine HKUST-1 crystal structure (hollow green circles). b, Semi-logarithmic representation of the isotherm (a) for P/Po < 0.002. c, Meso- and macro-PSD obtained by Hg porosimetry. Inset (red dashed box) indicates the recorded volumes of mesoporosity in the Pd-doped monoliths. 0 100 200 300 400 500 600 0 0.2 0.4 0.6 0.8 1 Pressure (bar) N 2 up ta ke (S TP ) c m 3 g- 1 1E+01 1E+02 1E+03 1E+04 1E+05 0.0 0.5 1.0 1.5 Pore diameter (nm) Di ffe re nt ia l i nt ru sio n (H g) m l g -1 MacroMeso a b Pressure (bar) Pore diameter (nm) N 2 up ta ke (S TP ) c m 3 g– 1 Di ffe re nt ia l i nt ru sio n (H g) m l g –1 a b c 0 100 200 300 400 500 600 0 0.2 0.4 0.6 0.8 1 Pressure (bar) N 2 up ta ke (S TP ) c m 3 g- 1 1E+01 1E+02 1E+03 1E+04 1E+05 0.0 0.5 1.0 1.5 Pore diameter (nm) Di ffe re nt ia l i nt ru sio n (H g) m l g -1 MacroMeso a b Pressure (bar) N 2 up ta ke (S TP ) c m 3 g– 1 0 50 100 150 200 250 300 350 400 1E-7 1E-6 1E-5 1E-4 1E-3 183 It is of note that the surface area of monolithHKUST-1 is lower than the theoretical maximum for the defect-free MOF (Table 5). This is very common in experimentally obtained HKUST-1; the chemical sensitivity of the material, which amorphises under exposure to water, means that any exposure of the material to air will diminish its surface area. The original monolith procedure reported by Tian et al. must be performed by ambient drying (i.e. in air), an inherent pitfall of the sol-gel synthesis procedure.89 Yet, in the current case the obtained SBET (1,557 m2 g–1) was significantly enhanced relative to the original report for monolithHKUST-1 (1,193 m2 g– 1). This result may be attributed to the experimental precautions taken to safeguard porosity in the current case, including activation immediately after air drying, purging the sample tubes with N2 and subsequent storage in a vacuum desiccator. Table 5| Comparison of SBET, Wo, Vtot and ρb of Pd-doped and pure monolithHKUST-1 to the corresponding values in the theoretical MOFs crystal structure. SBET* (m2 g–1) Wo† (cm3 g–1) Vtot‡ (cm3 g–1) rb§ (g cm–3) Theoretical HKUST-1 2014 0.75 0.80 0.88 monolithHKUST-1** 1193 0.51 0.52 1.06 monolithHKUST-1 1557 0.59 0.64 1.07 5% Pd@monolithHKUST-1 1468 0.56 0.62 1.06 10% Pd@monolithHKUST-1 1408 0.54 0.58 1.04 *Calculated according to Rouquerol’s consistency criteria (Appendix, Supplementary Figures 16 – 18). †Obtained at P/Po = 0.1. ‡Obtained at P/Po = 0.99. §Quantified using Hg porosimetry. **Data from Reference 89. The rb of the monolithic MOFs was further quantified by Hg porosimetry (Table 5). Firstly, at 1.07 g cm–3 the density of the experimentally obtained monolithHKUST-1 was near identical to the original literature report of the same material.89 In each case, experimental rb exceeds the 184 theoretical crystal density of HKUST-1 (0.80 g cm–3), which Tian attributed to the presence of amorphous/dense secondary phases in the MOF. This was supported by TEM (Figure 35) which showed dark/dense spots, with comparable dense phases also observed in the current case. For the materials doped with denser Pd (12.02 g cm–3), rb would be expected to increase proportionately to the doping level. Yet experimentally, rb was relatively unchanged amongst the doped materials (Table 5). To further elucidate this unexpected result, the PSD as obtained by Hg porosimetry (Figure 41b) was studied. While monolithHKUST-1 displayed the expected purely microporous PSD, unanticipated mesoporous volumes were recorded for each of the doped materials. The presence of this wide porosity in the macrostructure counteracts the increased density attributable to the presence of Pd, with the result being a retention of rb despite doping. However, this mesoporosity was not observed by NLDFT analysis (Figure 42) of the N2 isotherms (Figure 41a) which suggested the PSD to be exclusively microporous in each case. These conflicting results may be accounted for by marginal damage incurred to the materials during transit. Samples were shipped to Spain for Hg porosimetry characterization and may have become cracked or exposed to air, which could account for the observed mesopores. Furthermore, these analytical techniques use differing models to calculate PSD, each of which is but a reasonable approximation of the true structure. Figure 42| NLDFT PSDs of doped and pure monoliths. Distribution of micro- and mesopore width across monolith samples: HKUST-1 (green); 5% Pd@HKUST-1 (blue) and 10% Pd@HKUST-1 (purple) as obtained from Tarazona NLDFT analysis of N2 isotherm data (Figure 41a). 10 100 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 dV /d W (c m 3 g -1 .Å ) Pore width (Å) 10 100 0.0 0.1 0.2 0.3 0.4 0.5 0.6 dV /d W (c m 3 g -1 . Å ) Pore width (Å) 10 100 0.0 0.1 0.2 0.3 0.4 0.5 0.6 dV /d W (c m 3 g- 1 . Å ) Pore width (Å) 185 3.3.5| Thermal and Mechanical Stability The thermal stability of the monoliths reported herein were characterised by TGA over the temperature range 50 – 600 °C (Figure 43). No change in thermal stability was observed as a result of the doping, with each of the materials showing rapid decomposition only at temperatures exceeding 300 °C. These results are consistent with both the literature report of monolithHKUST-1 by Tian et al.89 as well as previous literature reports of the powdered MOF.95 Data suggest that the immobilization of NPs within the monolithic MOF does not influence its thermal stability. This differs from the case of SnO2-doped monolithZIF-8 where Wheatley et al. observed a significant reduction in thermal stability (from 600 °C down to 450 °C).35 While moderate thermal stability is characteristic of HKUST-1, it is notably lower than that of the monolithZr-MOFs previously discussed (Chapter II). This finds its origin in the reduced bond enthalpy of Cu(II)–O relative to Zr(IV)–O. In all samples a small, decrease in mass was observed at lower temperatures, corresponding to loss of adsorbed foreign species within the MOF porosity (e.g. moisture from the air). Figure 43| Thermal stability of monolithic materials. TGA traces of monolithic materials (50 – 600 °C, under N2 atmosphere) where data correspond to HKUST-1 (green), 5% Pd@HKUST-1 (blue) and 10% Pd@HKUST-1 (purple). Temperature (oC) W ei gh t ( % ) 0 20 40 60 80 100 50 150 250 350 450 550 186 Finally, the mechanical properties of the doped monolithic MOFs were characterised by nanoindentation (Figure 44). Wheatley et al. previously compared mechanical properties for monolithZIF-8 and SnO2@monolithZIF-8.35 They reported a ca. 10% improvement in both H and E as a result of the doping process, which they attributed to enhanced monolith packing efficiency by virtue of the presence of the dopant. However, in the current case the 5% Pd@monolithHKUST-1 (H = 0.15 ± 0.01 GPa, E = 4.7 ± 0.2 GPa) and 10% Pd@monolithHKUST- 1 (H = 0.13 ± 0.04 GPa, E = 2.6 ± 0.3 GPa) monoliths each showed a substantial reduction in mechanical strength relative to the original monolithHKUST-1 (H = 0.46 ± 0.03 GPa, E = 9.3 ± 0.3 GPa).89 The high doping levels (5% and 10% Pd by weight) appear to be sufficiently great to disturb monolith formation. This idea is further supported by the reduced H and E for the 10% loaded monolith relative to the 5% loaded material. However, comparably to Wheatley et al., an increase in mechanical variation amongst the repeated indents was observed. This is demonstrated by the wide spread of the Load vs. Penetration curves (Figure 44a) as well as larger error bars in both the Hardness and Young’s modulus curves (Figure 44b, c). This must originate from reduced homogeneity of the composite monolith by the NP doping. This observation is further consistent with the X-ray tomography results which demonstrated the inhomogeneous spatial distribution of dopant throughout the host (Figure 37); this was postulated to be a density gradient incurred by the centrifugation stage of the monolith synthesis. Despite the reduced mechanical strength of the Pd-doped monoliths, they remain moderately robust, with mechanical properties comparable to industrially viable monolithZIF-8 (H = 0.42 ± 0.04 GPa, E = 3.7 ± 0.2 GPa). 187 Figure 44| Mechanical testing of Pd@monolithHKUST-1. a, Load (mN) vs. Penetration into surface of the monolith (h, nm) for 16 indents. b, c, Hardness (H, GPa) and Young’s modulus (E, GPa), respectively, as a function of surface penetration depth (h, nm). Mean properties and corresponding errors (inset in b and c) were obtained from measurements taken from 16 indents over penetration depths of 500 – 2000 nm. Measurements obtained in the sub-500 nm penetration range were excluded, to eliminate errors due to surface defects/tip artefacts. Data correspond to 5% Pd@monolithHKUST-1 (blue squares) and 10% Pd@monolithHKUST-1 (purple diamonds). 0 1 2 3 4 5 6 0 500 1000 1500 2000 0 3 6 9 12 15 18 0 500 1000 1500 2000 2500 0 0.03 0.06 0.09 0.12 0.15 0.18 0.21 0 500 1000 1500 2000 E = 4.65 ± 0.24 GPaH = 0.15 ± 0.01 GPa Penetration into surface, h (nm) Penetration into surface, h (nm) Penetration into surface, h (nm) Lo ad (m N ) Ha rd ne ss (G Pa ) M od ul us (G Pa ) 0 2 4 6 8 10 12 14 16 18 20 0 500 1000 1500 2000 2500 Lo ad (m N) Penetration into surface, h (nm) 0 0.5 1 1.5 2 2.5 3 3.5 0 500 1000 1500 2000 0 0.05 0.1 0.15 0.2 0.25 0 500 1000 1500 2000 H = 0.13 ± 0.04 GPa E = 2.58 ± 0.33 GPa Penetration into surface, h (nm) Penetration into surface, h (nm) Ha rd ne ss (G Pa ) M od ul us (G Pa ) a b c 188 3.4| Conclusions The synthesis of new NP@monolithMOF composite materials was explored by studying the novel immobilisation of Pd NPs in monolithic HKUST-1. Firstly, 10 nm Pd nanocubes were synthesised according to the literature procedure reported by Kitawaga et al.93 Instead of coating the obtained NPs with HKUST-1 to obtain a powdered composite (as Kitagawa reported), they were instead immobilised in monolithic HKUST-1.89 The literature procedure for preparing monolithHKUST-1 was repeated but with a colloidal NP suspension included in the MOF primary particle reaction, as per the immobilisation procedure developed for SnO2@monolithZIF-8. Both 5% and 10% loading of the Pd NPs (by weight) into the monolith was demonstrated by ICP-OES, indicating a high level of synthetic control over the doping level. The resulting novel materials were fully characterised by an extensive range of analytical techniques to demonstrate the inclusion of the Pd NPs within the macroscopic, crystalline MOF structure. XPS was further used to demonstrate the electronic interaction of Pd and Cu in the composites. Here, the donation of electron density from Pd to Cu was supported by alterations in electron BE. This corroborates the previous work regarding powdered Pd@HKUST-1 and may have interesting implications for the H2 storage capacity of the obtained materials. Overall, the synthesis of a new composite monolith, Pd@HKUST-1, has been demonstrated. The material exhibited a wide range of industrially valuable physical properties (high rb, porosity, stability and a monolithic macrostructure). From a purely synthetic perspective, this novel material furthers our understanding of monolith’s capabilities to host NPs, extending the known composites beyond NP@monolithZIF-8 to include a wider range of monolithic hosts. Practically, the combined high rb and SBET of the doped materials offers significant potential in volumetric H2 storage. In particular, the effects of Pd loading level on the gas-fuel storage capacity is an area that also merits further exploration. 189 4.0| References 1. Pokropivny, V. V. & Skorokhod, V. V. Classification of Nanostructures by Dimensionality and Concept of Surface Forms Engineering in Nanomaterial Science. Mater. Sci. Eng. C 27, 990–993 (2007). 2. Bakueva, L. et al. PbS Quantum Dots With Stable Efficient Luminescence in the Near- IR Spectral Range. Adv. Mater. 16, 926–929 (2004). 3. Iijima, S. & Ichihashi, T. Single-Shell Carbon Nanotubes of 1-nm Diameter. Nature 363, 603–605 (1993). 4. Vayssieres, L. Growth of Arrayed Nanorods and Nanowires of ZnO From Aqueous Solutions. Adv. Mater. 15, 464–466 (2003). 5. Boulesbaa, A. et al. Ultrafast Charge Transfer and Hybrid Exciton Formation in 2D/0D Heterostructures. J. Am. Chem. Soc. 138, 14713–14719 (2016). 6. Liu, X. et al. 3D Hierarchically Porous ZnO Structures and Their Functionalization by Au Nanoparticles for Gas Sensors. J. Mater. Chem. 21, 349–356 (2011). 7. Das, S. & Jayaraman, V. SnO2: A Comprehensive Review on Structures and Gas Sensors. Prog. Mater. Sci. 66, 112–255 (2014). 8. Horcajada, P. et al. Porous Metal-Organic-Framework Nanoscale Carriers as a Potential Platform for Drug Delivery and Imaging. Nat. Mater. 9, 172–178 (2010). 9. Njuguna, J. & Pielichowski, K. Polymer Nanocomposites for Aerospace Applications: Properties. Adv. Eng. Mater. 5, 769–778 (2003). 10. Bell, A. The Impact of Nanoscience on Heterogeneous. Science. 299, 1688–1691 (2003). 11. Stark, W. J., Stoessel, P. R., Wohlleben, W. & Hafner, A. Industrial Applications of Nanoparticles. Chem. Soc. Rev. 98, 2035–2044 (2015). 12. de Vries, J. G. & Jackson, S. D. Homogeneous and Heterogeneous Catalysis in Industry. Catal. Sci. Technol. 2, 2009 (2012). 13. Astruc, D., Lu, F. & Aranzaes, J. R. Nanoparticles as Recyclable Catalysts: The Frontier Between Homogeneous and Heterogeneous Catalysis. Angew. Chem. Int. Ed. 44, 7852– 7872 (2005). 14. Hermanek, M., Zboril, R., Medrik, I., Pechousek, J. & Gregor, C. Catalytic Efficiency of Iron(III) Oxides in Decomposition of Hydrogen Peroxide: Competition Between the 190 Surface Area and Crystallinity of Nanoparticles. J. Am. Chem. Soc. 129, 10929–10936 (2007). 15. Kluenker, M. et al. Controlling the Morphology of Au–Pd Heterodimer Nanoparticles by Surface Ligands. Inorg. Chem. 57, 13640–13652 (2018). 16. Tao, A. R., Habas, S. & Yang, P. Shape Control of Colloidal Metal Nanocrystals. Small 4, 310–325 (2008). 17. Narayanan, R. & El-Sayed, M. A. Shape-Dependent Catalytic Activity of Platinum Nanoparticles in Colloidal Solution. Nano Lett. 4, 1343–1348 (2004). 18. Cheong, S., Watt, J. D. & Tilley, R. D. Shape Control of Platinum and Palladium Nanoparticles for Catalysis. Nanoscale 2, 2045–2053 (2010). 19. Bystrzejewska-Piotrowska, G., Golimowski, J. & Urban, P. L. Nanoparticles: Their Potential Toxicity, Waste and Environmental Management. Waste Manag. 29, 2587– 2595 (2009). 20. Min, Y., Akbulut, M., Kristiansen, K., Golan, Y. & Israelachvili, J. The Role of Interparticle and External Forces in Nanoparticle Assembly. Nat. Mater. 7, 527–538 (2008). 21. Vandebriel, R. J. & De Jong, W. H. A Review of Mammalian Toxicity of ZnO Nanoparticles. Nanotechnol. Sci. Appl. 5, 61–71 (2012). 22. Crosera, M. et al. Nanoparticle Dermal Absorption and Toxicity: A Review of the Literature. Int. Arch. Occup. Environ. Health 82, 1043–1055 (2009). 23. Zhang, J. et al. TiO2/T-PVA Composites Immobilized on Cordierite: Structure and Photocatalytic Activity for Degrading RhB Under Visible Light. Water. Air. Soil Pollut. 224, 1555–1564 (2013). 24. Wildgoose, G. G., Banks, C. E. & Compton, R. G. Metal Nanoparticles and Related Materials Supported on Carbon Nanotubes: Methods and Applications. Small 2, 182– 193 (2006). 25. Campisi, S., Chan-Thaw, C. E., Wang, D., Villa, A. & Prati, L. Metal Nanoparticles on Carbon Based Supports: The Effect of the Protective Agent Removal. Catal. Today 278, 91–96 (2016). 26. Sarkar, S., Guibal, E., Quignard, F. & SenGupta, A. K. Polymer-Supported Metals and Metal Oxide Nanoparticles: Synthesis, Characterization, and Applications. J. 191 Nanoparticle Res. 14, 715–739 (2012). 27. Falcaro, P. et al. Application of Metal and Metal Oxide Nanoparticles@MOFs. Coord. Chem. Rev. 307, 237–254 (2016). 28. Zhang, W. et al. Controlled Incorporation of Nanoparticles in Metal-Organic Framework Hybrid Thin Films. Chem. Commun. 50, 4296–4298 (2014). 29. Liu, Y. & Tang, Z. Multifunctional Nanoparticle@MOF Core-Shell Nanostructures. Adv. Mater. 25, 5819–5825 (2013). 30. Mehta, J. P. Incorporation of Metal Oxide Nanoparticles in a Monolithic Metal-Organic Framework for Applications in Photocatalysis. (The University of Cambridge, 2016). 31. Lu, G. et al. Imparting Functionality to a Metal-Organic Framework Material by Controlled Nanoparticle Encapsulation. Nat. Chem. 4, 310–316 (2012). 32. Moggach, S. A., Bennett, T. D. & Cheetham, A. K. The Effect of Pressure on ZIF-8: Increasing Pore Size with Pressure and the Formation of a High-Pressure Phase at 1.47 GPa. Angew. Chem. Int. Ed. 48, 7087–7089 (2009). 33. Fairen-Jimenez, D. et al. Opening The Gate: Framework Flexibility in ZIF-8 Explored by Experiments and Simulations. J. Am. Chem. Soc. 133, 8900–8902 (2011). 34. Koczkur, K. M., Mourdikoudis, S., Polavarapu, L. & Skrabalak, S. E. Polyvinylpyrrolidone (PVP) in Nanoparticle Synthesis. Dalton Trans. 44, 17883–17905 (2015). 35. Mehta, J. P. et al. Sol-Gel Synthesis of Robust Metal-Organic Frameworks for Nanoparticle Encapsulation. Adv. Funct. Mater. 28, 1705588 (2018). 36. Wiley, B., Sun, Y., Mayers, B. & Xia, Y. Shape-Controlled Synthesis of Metal Nanostructures: The Case of Silver. Chem. Eur. J. 11, 454–463 (2005). 37. Burda, C., Chen, X., Narayanan, R. & El-Sayed, M. A. Chemistry and Properties of Nanocrystals of Different Shapes. Chem. Rev. 105, 1025–1102 (2005). 38. Han, X. et al. Synthesis of Tin Dioxide Octahedral Nanoparticles with Exposed High- Energy {221} Facets and Enhanced Gas-Sensing Properties. Angew. Chem. Int. Ed. 48, 9180–9183 (2009). 39. Jiang, L. et al. Size-Controllable Synthesis of Monodispersed SnO2 Nanoparticles and Application in Electrocatalysts. J. Phys. Chem. B 109, 8774–8778 (2005). 192 40. Cheng, B., Russell, J. M., Shi, W., Zhang, L. & Samulski, E. T. Large-Scale, Solution- Phase Growth of Single-Crystalline SnO2 Nanorods. J. Am. Chem. Soc. 126, 5972–5973 (2004). 41. LaMer, V. K. & Dinegar, R. H. Theory, Production and Mechanism of Formation of Monodispersed Hydrosols. J. Am. Chem. Soc. 72, 4847–4854 (1950). 42. Houas, A. et al. Photocatalytic Degradation Pathway of Methylene Blue in Water. Appl. Catal. B Environ. 31, 145–157 (2001). 43. Li, J. et al. Engineering Noble Metal Nanomaterials for Environmental Applications. Nanoscale 7, 7502–7519 (2015). 44. Huang, X. & El-Sayed, M. A. Gold Nanoparticles: Optical Properties and Implementations in Cancer Diagnosis and Photothermal Therapy. J. Adv. Res. 1, 13–28 (2010). 45. Prabhu, S. & Poulose, E. K. Silver Nanoparticles: Mechanism of Antimicrobial Action, Synthesis, Medical Applications, and Toxicity Effects. Int. Nano Lett. 2, 32 (2012). 46. Pradeep, T. & Anshup. Noble Metal Nanoparticles for Water Purification: A Critical Review. Thin Solid Films 517, 6441–6478 (2009). 47. Mazumder, V. & Sun, S. Oleylamine-Mediated Synthesis of Pd Nanoparticles for Catalytic Formic Acid Oxidation. J. Am. Chem. Soc. 131, 4588–4589 (2009). 48. Xia, B. Y., Wu, H. Bin, Wang, X. & Lou, X. W. Highly Concave Platinum Nanoframes with High-Index Facets and Enhanced Electrocatalytic Properties. Angew. Chem. Int. Ed. 52, 12337–12340 (2013). 49. Chen, G. et al. Interfacial Electronic Effects Control the Reaction Selectivity of Platinum Catalysts. Nat. Mater. 15, 564–571 (2016). 50. Shelef, M. & McCabe, R. W. Twenty-Five Years After Introduction of Automotive Catalysts: What Next? Catal. Today 62, 35–50 (2000). 51. Xie, X., Li, Y., Liu, Z.-Q., Haruta, M. & Shen, W. Low-Temperature Oxidation of CO Catalysed by Co3O4 Nanorods. Nature 458, 746–749 (2009). 52. Janssens, T. V. W., Carlsson, A., Puig-Molina, A. & Clausen, B. S. Relation Between Nanoscale Au Particle Structure and Activity for CO Oxidation on Supported Gold Catalysts. J. Catal. 240, 108–113 (2006). 193 53. Remediakis, I. N., Lopez, N. & Nørskov, J. K. CO Oxidation on Gold Nanoparticles: Theoretical Studies. Appl. Catal. A 291, 13–20 (2005). 54. Lopez, N. et al. On the Origin of the Catalytic Activity of Gold Nanoparticles for Low- Temperature CO Oxidation. J. Catal. 223, 232–235 (2004). 55. Eisele, P. & Killpack, R. Propene. in Ullmann’s Encyclopedia of Industrial Chemistry 1–13 (Wiley, 2005). 56. Liu, Y. et al. From Glycerol to Allyl Alcohol: Iron Oxide Catalyzed Dehydration and Consecutive Hydrogen Transfer. Chem. Comm. 46, 1238–1240 (2010). 57. Ruppert, A. M., Weinberg, K. & Palkovits, R. Hydrogenolysis Goes Bio: From Carbohydrates and Sugar Alcohols to Platform Chemicals. Angew. Chem. Int. Ed. 51, 2564–2601 (2012). 58. Bhandari, R., Pacardo, D. B., Bedford, N. M., Naik, R. R. & Knecht, M. R. Structural Control and Catalytic Reactivity of Peptide-Templated Pd and Pt Nanomaterials for Olefin Hydrogenation. J. Phys. Chem. C 117, 18053–18062 (2013). 59. Gao, Y. et al. Catalytic Asymmetric Epoxidation and Kinetic Resolution: Modified Procedures Including in Situ Derivatization. J. Am. Chem. Soc. 109, 5765–5780 (1987). 60. Saito, S. et al. A Route to Propylene from Renewable Allyl Alcohol by Photocatalytic Transfer Hydrogenolysis. Catal. Sci. Technol. 4, 4093–4098 (2014). 61. Escandón, L. S., Ordóñez, S., Vega, A. & Díez, F. V. Oxidation of Methane Over Palladium Catalysts: Effect of the Support. Chemosphere 58, 9–17 (2005). 62. Lashof, D. A. & Ahuja, D. R. Relative Global Warming Potentials of Greenhouse Gas Emissions. Nature 344, 529–531 (1990). 63. Farrauto, R. J. Low Temperature Oxidation of Methane. Science. 337, 659–660 (2012). 64. Abbasi, R., Huang, G., Istratescu, G. M., Wu, L. & Hayes, R. E. Methane Oxidation Over Pt, Pt: Pd, and Pd Based Catalysts: Effects of Pre-Treatment. Can. J. Chem. Eng. 93, 1474–1482 (2015). 65. Lyubovsky, M. R. & Pfefferle, L. D. Complete Methane Oxidation over Pd Catalyst Supported on α-Alumina. Influence of Temperature and Oxygen Pressure on the Catalyst Activity. Catal. Today 47, 29–44 (1999). 66. McCarty, J. G. Kinetics of PdO Combustion Catalysis. Catal. Today 26, 283–293 (1995). 194 67. Tian, T., Velazquez-Garcia, J., Bennett, T. D. & Fairen-Jimenez, D. Mechanically and Chemically Robust ZIF-8 Monoliths With High Volumetric Adsorption Capacity. J. Mater. Chem. A 3, 2999–3005 (2015). 68. Ciuparu, D., Bozon-Verduraz, F. & Pfefferle, L. Oxygen Exchange Between Palladium and Oxide Supports in Combustion Catalysts. J. Phys. Chem. B 106, 3434–3442 (2002). 69. Persson, K., Ersson, A., Jansson, K., Iverlund, N. & Järås, S. Influence of Co-Metals on Bimetallic Palladium Catalysts for Methane Combustion. J. Catal. 231, 139–150 (2005). 70. Osman, A. I. et al. A Bimetallic Catalyst on a Dual Component Support for Low Temperature Total Methane Oxidation. Appl. Catal. B Environ. 187, 408–418 (2016). 71. Ab Rahim, M. H. et al. Oxidation of Methane to Methanol With Hydrogen Peroxide Using Supported Gold-Palladium Alloy Nanoparticles. Angew. Chem. Int. Ed. 52, 1280– 1284 (2013). 72. Jana, N. R., Gearheart, L. & Murphy, C. J. Seeding Growth for Size Control of 5-40 nm Diameter Gold Nanoparticles. Langmuir 17, 6782–6786 (2001). 73. Bastús, N. G., Comenge, J. & Puntes, V. Kinetically Controlled Seeded Growth Synthesis of Citrate-Stabilized Gold Nanoparticles of up to 200 nm: Size Focusing Versus Ostwald Ripening. Langmuir 27, 11098–11105 (2011). 74. Rahman, I. A., Vejayakumaran, P., Sipaut, C. S., Ismail, J. & Chee, C. K. Effect of the Drying Techniques on the Morphology of Silica Nanoparticles Synthesized via Sol-Gel Process. Ceram. Int. 34, 2059–2066 (2008). 75. Rösler, C. et al. Encapsulation of Bimetallic Nanoparticles into a Metal-Organic Framework: Preparation and Microstructure Characterization of Pd/Au@ZIF-8. Eur. J. Inorg. Chem. 2014, 5514–5521 (2014). 76. Burtch, N. C., Heinen, J., Bennett, T. D., Dubbeldam, D. & Allendorf, M. D. Mechanical Properties in Metal-Organic Frameworks: Emerging Opportunities and Challenges for Device Functionality and Technological Applications. Adv. Mater. 30, 1704124, (2017). 77. Wang, L. & Yamauchi, Y. Strategic Synthesis of Trimetallic Au@Pd@Pt Core-Shell Nanoparticles from Poly(vinylpyrrolidone)-Based Aqueous Solution toward Highly Active Electrocatalysts. Chem. Mater. 23, 2457–2465 (2011). 78. Dou, L. et al. Structures and Electronic Properties of Au Clusters Encapsulated ZIF-8 and ZIF-90. J. Phys. Chem. C 122, 8901–8909 (2018). 195 79. Krokidas, P., Castier, M., Moncho, S., Brothers, E. & Economou, I. G. Molecular Simulation Studies of the Diffusion of Methane, Ethane, Propane, and Propylene in ZIF- 8. J. Phys. Chem. C 119, 27028–27037 (2015). 80. Chui, S. S., Lo, S. M., Charmant, J. P. H., Orpen, A. G. & Williams, I. D. A Chemically Functionalizable Nanoporous Material [Cu3(TMA)2(H2O)]n. Science. 283, 1148–1151 (1999). 81. Prestipino, C. et al. Local Structure of Framework Cu(II) in HKUST-1 Metal-Organic Framework: Spectroscopic Characterization Upon Activation and Interaction with Adsorbates. Chem. Mater. 18, 1337–1346 (2006). 82. Xiao, Y. et al. Ultrathin 2D Cu-Porphyrin MOF Nanosheets as Heterogeneous Catalysts for Styrene Oxidation. Mater. Chem. Front. 3, 1580–1585 (2019). 83. Ren, X. et al. Copper Metal-Organic Framework Loaded on Chitosan Film for Efficient Inhibition of Bacteria and Therapy of Local Infection. Nanoscale 11, 11830–11838 (2019). 84. Wang, X. et al. Amino-Functionalized Cu-MOF for Efficient Purification of Methane from Light Hydrocarbons and Excellent Catalytic Performance. Inorg. Chem. Front. 6, 1152–1157 (2019). 85. Álvarez, J. R. et al. Structure Stability of HKUST-1 Towards Water and Ethanol and Their Effect on its CO2 Capture Properties. Dalton Trans. 46, 9192–9200 (2017). 86. Lin, K. S., Adhikari, A. K., Ku, C. N., Chiang, C. L. & Kuo, H. Synthesis and Characterization of Porous HKUST-1 Metal Organic Frameworks for Hydrogen Storage. Int. J. Hydrogen Energy 37, 13865–13871 (2012). 87. Peng, Y. et al. Methane Storage in Metal − Organic Frameworks: Current Records, Surprise Findings, and Challenges. J. Am. Chem. Soc. 135, 11887–11894 (2013). 88. Kim, J., Kim, S. H., Yang, S. T. & Ahn, W. S. Bench-Scale Preparation of Cu3(BTC)2 by Ethanol Reflux: Synthesis Optimization and Adsorption/Catalytic Applications. Microporous Mesoporous Mater. 161, 48–55 (2012). 89. Tian, T. et al. A Sol–Gel Monolithic Metal–Organic Framework with Enhanced Methane Uptake. Nat. Mater. 17, 174–179 (2018). 90. Li, Y. et al. Significant Enhancement in Hydrolytic Degradation of Sulfur Mustard Promoted by Silver Nanoparticles in the Ag NPs@HKUST-1 Composites. J. Hazard. 196 Mater. 358, 113–121 (2018). 91. Xie, W. & Wan, F. Basic Ionic Liquid Functionalized Magnetically Responsive Fe3O4@HKUST-1 Composites Used for Biodiesel Production. Fuel 220, 248–256 (2018). 92. Murinzi, T. W., Clement, T. A., Chitsa, V. & Mehlana, G. Copper Oxide Nanoparticles Encapsulated in HKUST-1 Metal-Organic Framework for Electrocatalytic Oxidation of Citric Acid. J. Solid State Chem. 268, 198–206 (2018). 93. Li, G. et al. Hydrogen Storage in Pd Nanocrystals Covered with a Metal–Organic Framework. Nat. Mater. 13, 802–806 (2014). 94. Chen, Y. et al. Electronic Origin of Hydrogen Storage in MOF-Covered Palladium Nanocubes Investigated by Synchrotron X-rays. Commun. Chem. 1, 61 (2018). 95. Mu, B. & Walton, K. S. Thermal Analysis and Heat Capacity Study of Metal-Organic Frameworks. J. Phys. Chem. C 115, 22748–22754 (2011). Chapter IV Practical Applications of Monolithic Materials 198 199 1.0| Gas Storage and Separation in monolithUiO-66 As discussed in Chapter II, Section 1.0, UiO-66 was synthesised for the first time as a robust monolithic MOF with rb and SBET comparable to those of its theoretical single crystal structure. As a consequence of their desirable physical properties (i.e. high thermal, chemical and mechanical stability), such monolithic Zr-MOFs offer numerous potential industrial applications i.e. high-density NG (CH4) storage.1 A much sought-after application of this fuel is in automobile engines, yet current gas storage technology does not permit widespread usage.2 The low fuel capacity, high mass and safety risks associated with highly pressurised storage tanks (up to 250 bar) prevent NG powered vehicles. Furthermore, liquified NG (which requires cooling to < 110 K) is too costly and energy intensive to maintain for such commonplace applications.3 Adsorption of NG in porous MOFs, at lower pressures and ambient temperatures is regarded as a highly efficient means of achieving practical and efficient gas fuel storage. As discussed, even traditionally synthesised powdered MOFs with high gravimetric gas storage capacity have failed to make the leap into industry; their poor powder packing efficiency limits volumetric gas storage capacity. Furthermore, these MOFs cannot be densified by traditional pelletisation techniques which both collapse and block their porosity. The novel sol-gel synthesis of high density, high surface area monolithHKUST-1 (rb = 1.06 g cm–3, SBET = 1288 m2 g–1) by Tian et al. resulted in an outstanding volumetric CH4 capacity of 259 cm3 (STP) cm–3 (65 bar, 298 K).4 This was the first report of a shaped/densified MOF which could reach the U.S. DOE target for NG storage; 263 cm3 (STP) cm–3 at 65 bar and 298 K. However, the low chemical stability of this hygroscopic Cu(II) MOF under ambient conditions renders it industrially non-optimised from a practical perspective.5 In the current case, CH4 storage in high stability, high surface area, densified monolithUiO-66 presents a number of practical benefits over benchmark monolithHKUST-1. Furthermore, a capability to synthesise UiO-66 with a range of different PSDs by the inclusion of tuneable, non-crystalline porosity external to the MOFs defined structure was demonstrated without significant collapse/blocking of microporosity. This exists as interstitial space between the microporous primary particles of which the monoliths are comprised and has significant industrial implications. Pore-width is well known to affect the onset pressure at which gas- condensation occurs in an adsorbent.6 This may influence the materials adsorption isotherm shape with repercussions for both its total adsorption capacity as well as its working capacity. 200 Additionally, the tuneable mesoporosity alters the monolith’s density, a key factor in determining its volumetric gas storage capacity. Moreover, the potential of monolithUiO-66 as an industrially multifunctional material presents numerous other research avenues (Figure 1) – the capacity of this MOF to adsorb numerous other gasses is well known.7,8 In particular CO2 adsorption is a topical research avenue for both environmental remediation and fuel refinery. Pre- and post-combustion capture and storage of CO2 is considered a viable short-term solution to reducing atmospheric CO2 levels while fossil fuel usage persists.9 Additionally, the gas storage potential of this MOF can be combined with its molecular sieving capabilities, to purify fuel streams prior to combustion e.g. by the selective removal of contaminating CO2 in NG refinery. Again, unprocessed MOFs are further inapplicable to gas sieving without shaping into pellets, as powders compact during the processes, resulting in pipe blockages and corresponding pressure drops.10 Figure 1| Potential industrial applications of MOFs. Schematic representation showing various stages of industrial fuel generation and usage that MOFs may be applied to. Fuel Refinery Fuel Transport Waste Capture Selective Gas Separation Gaseous Fuel Storage Adsorption of Pollutants 201 1.1| Aims and Objectives The ultimate aim of this chapter is to determine the potential of monolithUiO-66 in real-world, industrial, gas-phase applications. Firstly, the CH4 adsorption capacity of each of monolithUiO- 66_A – D will be recorded to determine their NG storage capacities. Both the gravimetric and volumetric capacities will be compared to elucidate the influence of the different monoliths varied physical properties (i.e. SBET, rb, PSD etc.) on their CH4 storage capacity. The results will be benchmarked against both record-setting powdered MOFs and the current benchmark, monolithHKUST-1.4 Furthermore, the applicability of the monolithic UiO-66 materials towards a wider range of potential industrial gas-phase applications will also be explored. Firstly, the capture and storage of other gases, namely CO2 will be studied. The material’s gravimetric and volumetric capacities for this gas will be recorded and compared in terms of the different material’s bulk physical properties. Additionally, from the experimental isotherms of CH4 and CO2 adsorption, the capacities of monolithUiO-66 samples to selectively separate these gases will be calculated and its capability to further purify NG determined. Moreover, the overall industrial viability of these materials will be explored by considering a number of physical parameters which influence practical viability. The kinetics of gas adsorption, which determines the rate of storage tank re-fuelling, will be recorded as well as each material’s working capacity – the accessible volume of gas stored in the material under real working conditions. In all cases, the experimental results validity will be tested through computational experiments. Not only will the theoretical adsorption isotherms of the defect-free MOF be computationally generated by GCMC simulations, but the adsorption sites of CH4 and CO2 in the mixed micro- /mesoporous MOFs will be mapped to better understand the influence of this altered PSD on its gas adsorption properties. 202 1.2| CH4 Storage 1.2.1| Gravimetric CH4 Storage The CH4 storage capacities of the UiO-66 monoliths UiO-66_A – D were studied at 298 K between 0 – 100 bar. Adsorption and desorption isotherms are always experimentally obtained as excess gravimetric uptake, Nklm. This is defined as the amount of gas present in the system which was directly adsorbed by the material being studied. However, the true measure of gas storage capacity is absolute uptake, Nnop, the total number of gas molecules in the system i.e. those adsorbed by the adsorbent in addition to bulk gas molecules which would otherwise have been present in its absence. Hence, excess was converted to absolute uptake using [Equation 2]:11 this adds the bulk gas which should be present in the materials porosity, using each MOFs Vtot, as well as the density of the gas (ρ, g cm–3) at each data point in the experimental isotherm: 𝐍𝐚𝐛𝐬 = 𝐍𝐞𝐱𝐜 + 𝛒𝐕𝐭𝐨𝐭 [Equation 2] The isotherms showing the absolute gravimetric CH4 storage capacity of each UiO-66 monolith up to 100 bar pressure are displayed in Figure 2a. Firstly, it is apparent that the gravimetric capacities do not vary substantially between the different materials. This is because gravimetric adsorption capacity (g g–1) is correlated with the availability of surface area (m2 g–1) to which the gas can be adsorbed. In the current case, the comparable SBET of the materials (Table 1) incur similar gravimetric CH4 uptake capacities at each pressure. The modest differences between the isotherms can be correlated well with the minor differences in SBET exhibited by each sample. Hence, Nabs (at each pressure point) for UiO-66_A (SBET = 1177 m2 g–1) > UiO- 66_C (SBET = 1065 m2 g–1) > UiO-66_B (SBET = 994 m2 g–1) > UiO-66_D (SBET = 982 m2 g–1). Secondly, UiO-66 is expected to exhibit a traditional Type I isotherm shape i.e. rapid uptake of gas at low pressure due to adsorption in its microporosity followed by a plateau in uptake at higher pressures as a result of micropore saturation. Just such an isotherm was reported by Zhou et al. for the powdered MOF (Figure 2b).12 The modest absolute CH4 uptake they reported (0.11 g g–1 at 300 K, 63 bar)12 was the result of micropore saturation at ca. 30 bar pressure and further increases in gas pressure would not be expected to yield significant improvements in 203 gas storage capacity. This expected plateau matches the Type I theoretical isotherm (Figure 2c) obtained by GCMC simulations obtained from the defect-free and purely microporous material. In contrast, each UiO-66 monolith (UiO-66_A – D) displays a Type II CH4 adsorption isotherm. The rapid gas adsorption in the micropores is not followed by a plateau but rather by continued gas uptake at high pressure. Correspondingly, the recorded gravimetric capacities of monolithUiO-66_A – D (0.18 – 0.14 g g–1 at 298 K, 65 bar) far exceed the computationally determined/theoretical maximum value (0.10 g g–1 at 298 K/65 bar). These Type II isotherms may be attributed to the condensation of gas in each material’s mesopores. Unlike purely microporous materials, mixed micro-/mesoporous materials can exhibit further gas uptake once the micropores have been filled, though gas condensation in the wider mesoporous cavities. This leads to enhanced gas uptake at high pressures and an overall enhancement in gas uptake capacity. This observation is analogous to the observed mesoporous peak at higher pressures in the N2 isotherms of each of these materials (Chapter II, Section 1.0). Figure 2| Gravimetric CH4 adsorption isotherms. a, Comparison of gravimetric CH4 storage capacity of UiO-66 monoliths up to 100 bar (298 K). Colours correspond to UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles). b, Comparison of gravimetric CH4 uptake (0 – 100 bar) in UiO-66_D (green circles) to that of monolithHKUST-1 (white stars) and powdered UiO-66 (white circles, digitised from Reference 12). c, Comparison of gravimetric CH4 uptake (0 – 100 bar) in powdered UiO-66 (white circles) to that of computationally simulated defect-free microporous UiO-66 (white squares). All isotherms were experimentally obtained or computationally calculated at 298 K. 0 20 40 60 80 100 0.0 0.1 0.2 0.3 CH 4 u pt ak e (S TP ) g g- 1 Pressure (bar) CH 4 up ta ke (S TP ) g g -1 Pressure (bar) 0 20 40 60 80 100 0.00 0.05 0.10 0.15 0.20 0.25 CH 4 u pt ak e (S TP ) g g- 1 Pressure (bar) CH 4 up ta ke (S TP ) g g -1 Pressure (bar) a b c 0 0.02 0.04 0.06 0.08 0.1 0.12 0 10 20 30 40 50 60 CH 4 up ta ke (g g -1 ) Pressure (bar) 0 20 40 60 80 100 0. 0.1 0.2 0.3 CH 4 u pt ak e (S TP ) g g- 1 Pressure (bar) CH 4 up ta ke (S TP ) g g -1 Pressure (bar) 0 20 40 60 80 100 0. 0 0. 5 0.10 0.15 0.20 0.25 CH 4 u pt ak e (S TP ) g g- 1 Pressure (bar) CH 4 up ta ke (S TP ) g g -1 Pressure (bar) a b CH 4 up ta ke (g g -1 ) CH 4 up ta ke (g g -1 ) Pressure (bar)Pressure (bar) a 204 Figure 2b directly compares the isotherm shapes of monolithic UiO-66_D to powdered UiO- 66. The overlap of the two isotherms at low pressure is the result of their identical microporosity (they are both crystalline UiO-66) while the divergence of the two at higher pressures demonstrates their differing volumes of mesoporosity. The significant consequences of gas condensation in the monolith’s unique mesoporosity are further highlighted by comparison with benchmark material monolithHKUST-1. While this high SBET material displays significantly greater CH4 storage capacity than UiO-66 at low pressure, its purely microporous structure results in saturation at ca. 40 bar with poor uptake above this pressure. The continued CH4 uptake of mixed micro-/mesoporous UiO-66_D results in it exceeding its theoretical maximum for CH4 adsorption, to surpass that of monolithHKUST-1 at 100 bar. Table 1| SBET, ρb and gravimetric and volumetric CH4 uptake (65 and 100 bar, 298K) for UiO- 66_A – D compared to theoretical UiO-66, simulated from the ideal crystal structure. SBET ρb‡ CH4 65 barɤ CH4 100 barɤ m2 g–1 g cm–3 g g–1 cm3 cm–3 g g–1 cm3 cm–3 UiO-66_A 1177 0.43 0.18 109 0.25 153 UiO-66_B 994 0.43 0.16 98 0.23 142 UiO-66_C 1065 0.85 0.17 202 0.24 289 UiO-66_D 982 1.05 0.14 211 0.20 296 Simulation 1644 1.24 0.10 178 0.12 200 ‡Quantified by Hg porosimetry. ɤTotal absolute uptake. 205 1.2.2| CH4 Adsorption Simulations To further explore the unprecedented enhancement in gas storage of these monolithUiO-66 relative to the powdered analogue, CH4 adsorption in mixed micro-/mesoporous UiO-66 was simulated by artificial generation of a mesopore between two microporous crystals (Figure 3). The artificially generated mesopores, which were earlier modelled to match the experimental and computational N2 isotherms (Chapter II, Section 1.0), were again utilised for this simulation. The results confirm that while the MOF micropores adsorb CH4 at 20 bar, little adsorption occurs in the mesoporous cavity below this pressure. Gas molecules recorded in the mesopores at this pressure are likely to be unabsorbed bulk gas which would otherwise have been found in this area in the absence of the adsorbent. Yet at 80 bar, sizable quantities of gas condense in the modelled mesopore. This phenomenon of enhanced gas uptake by mesopore condensation can be accounted for by the Steel function (Chapter I) which considers the relationship between the potential energy profile of an adsorbing gas molecule and the lattice spacing of the adsorbent.13 This uses a Lennard-Jones profile, which takes into account both short-range Pauli repulsion and long range, attractive Van der Waals forces. Hence, the attractive potential fields of opposing pore walls overlap to create an enhanced energy minimum in the middle which is manifested significantly for sub-2 nm micropores. As such, gaseous physisorption e.g. of CH4 in micropores is particularly energetically favourable at low pressure as they present both a high density of accessible internal surface sites as well as enhanced gas adsorption properties relative to meso- (2 < d < 50 nm) and macro- (d > 50 nm) pores. Higher pressures are required to achieve gas storage at the adsorption sites provided by mesopores, accounting for the simulated and experimental results in the mixed micro-/mesoporous materials. 206 Figure 3| Snapshots and density distributions of CH4 adsorption. a, Snapshots and b, density distributions comparing CH4 (grey) adsorption in mixed micro-/mesoporous UiO-66 at 20 and 80 bar pressure. Colours within the UiO-66 crystal structure correspond to elements; C (grey), H (white), Zr (blue) and O (red). 20 bar 20 bar80 bar 80 bar a b 20 bar 20 bar80 bar 80 bar 207 1.2.3| Volumetric CH4 Storage As discussed, volumetric gas storage capacity is a more industrially significant measure of a material’s practical relevance towards NG storage. While mesoporosity was observed to significantly enhance gravimetric CH4 storage capacity, it is further expected to reduce volumetric capacity as a result of the decreased rb it incurs. Hence, monolith ρb (g cm–3) was used to convert gravimetric (g g–1) to volumetric (cm3 (STP) cm–3) uptake (Figure 4).14 Figure 4| Volumetric CH4 storage. a, Comparison of absolute volumetric CH4 storage capacity in UiO-66 monoliths (0 – 100 bar, 298 K). Colours correspond to UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles) and a pressurised storage tank (black crosses). b, Comparison of experimental isotherms for absolute volumetric CH4 uptake in monolithUiO-66_D (green circles) and monolithHKUST-1 (white stars, digitised from Reference 4) to computationally simulated purely microporous/defect-free UiO- 66 (white squares) at 298 K; the U.S. DOE volumetric CH4 storage target of 263 cm3 (STP) cm–3 (65 bar, 298 K) is indicated (dashed red line, inset). Firstly, each monolithic material, UiO-66_A – D, revealed a significant improvement in volumetric gas storage relative to an empty tank at the same pressure (Figure 4a). However, unlike the gravimetric results, which were similar between the different monoliths, it is clear that significant differences in volumetric capacity exist. This results from their greatly differing rb (the key physical parameter in determining volumetric gas storage capacity). The 0 20 40 60 80 100 0 50 100 150 200 250 300 CH 4 u pt ak e (S TP ) c m 3 c m 3 Pressure (bar) CH 4 up ta ke (S TP ) c m 3 cm -3 Pressure (bar) CH 4 up ta ke (S TP ) c m 3 cm -3 0 20 40 60 80 100 0 50 100 150 200 250 300 Pressure (bar) a b 208 negligible effects of gravimetric SBET (m2 g–1) between the UiO-66 samples appeared to have little effect on the total uptake capacity relative to the substantial effect of differing rb. Both volumetric CH4 uptake capacity and rb followed the trend: UiO-66_D ≈ UiO-66_C >>> UiO- 66_B ≈ UiO-66_A. Samples UiO-66_D and UiO-66_C showed exceptional CH4 uptake capacities of 202 and 211 cm3 (STP) cm–3 respectively at 65 bar. These data are comparable to previously reported benchmark uptakes in industrially non-viable powdered MOFs such as NU-111 (206 cm3 (STP) cm–3) and PCN-14 (230 cm3 (STP) cm–3).15 Furthermore, the results are significantly higher than the theoretical maximum (Table 1) and can be attributed to these materials finely tuned combination of high rb, maintained microporosity and residual mesoporosity. Their high SBET facilitate physisorptative uptake at low pressures while mesopores undergo internalised gas condensation at high pressures. Conversely, if mesoporosity is too high (as was the case for UiO-66_A – B), the material is not sufficiently dense for high volumetric storage. A careful balance between SBET, PSD and rb is therefore needed for optimised gas uptake. Overall, the densest monolith, UiO-66_D, showed the greatest CH4 uptake amongst the monoliths with 211 and 296 cm3 (STP) cm–3 at 65 and 100 bar, respectively (Table 1). ρb for powdered UiO-66 was not reported by Zhou et al., 12 preventing comparison to the monolithUiO-66 in terms of volumetric capacity. However, the computationally obtained gravimetric isotherm from the defect-free microporous MOF (demonstrated to closely match the experimental results for UiO-66 powder; Figure 2c) was converted into a volumetric isotherm using the MOFs theoretical single-crystal density (Table 1). Figure 4b compares the volumetric CH4 uptake of UiO-66_D (the best performing monolith in the current study) to both this theoretical isotherm and that of benchmark monolithHKUST-1.4 Again, the simulated volumetric CH4 storage capacity of theoretical microporous UiO-66 matches the isotherm for UiO-66_D up to ca. 40 bar. The fact that these isotherms overlap at lower pressures, while diverging at higher pressures, supports the hypothesis that enhanced gas uptake occurs through high pressure CH4 condensation in the synthetically introduced mesoporous cavities. The exceptional CH4 uptake of benchmark densified material monolithHKUST-1 (259 cm3 (STP) cm–3, 65 bar) has previously been reported to approach the U.S. DOE targets of 263 cm3 (STP) cm3 at 65 bar.4 Volumetric CH4 uptake isotherms in monolithHKUST-1 and UiO-66_D can be compared over the 0 – 100 bar pressure range (Figure 4b). Unlike monolithHKUST-1, the gradual 209 slope of monolith UiO-66_D isotherm does not permit it to reach the U.S. DOE targets (red line) at 65 bar. However, gas uptake by monolith UiO-66_D shows no sign of plateauing at high pressures, allowing it to overtake that of plateaued monolithHKUST-1 at 100 bar. Even at this high pressure, no indication of pore saturation is observed. This, interestingly, suggests that at further elevated pressures, UiO-66_D may be capable of additional CH4 storage. 210 1.2.4| Working Capacity As described earlier (Chapter I), one of the most important practical engineering parameters for an optimal NG adsorbent is the working capacity. This is the uptake at maximum storage pressure minus the uptake at the minimum release pressure (typically ca. 5 bar) – i.e. the accessible volume of gas in the storage system.16 At the U.S. DOE’s target pressure of 65 bar, UiO-66_D shows a working capacity of 172 cm3 (STP) cm–3 using its real rb (Table 2). This can be compared to the theoretical benchmark ca. 200 cm3 (STP) cm–3 working capacity of powdered UTSA-76a, which, under packing, would be expected to display a 25 – 50% reduction in volumetric capacity, reducing it to 100 – 150 cm3 (STP) cm–3.17 This means that, at 172 cm3 (STP) cm–3, densified UiO-66_D surpasses this benchmark significantly. Furthermore, by assuming a Langmuir-fitting and using single-crystal density, the volumetric working capacity (5 – 100 bar) of UTSA-76a will be 236 cm3 (STP) cm–3. This theoretical maximum value would be expected to fall by 25 – 50% to 118 – 177 cm3 (STP) cm–3 in a standard, compacted pellet. In contrast, already densified UiO-66_D shows a volumetric working capacity of 261 cm3 (STP) cm–3 over the same pressure range. When comparing with the 185 cm3 (STP) cm–3 (5 – 65 bar) working capacity of the chemically unstable monolithHKUST-1, the results of monolithUiO-66_D are 10% lower (Table 2). Although 65 bar is considered an optimal storage pressure, being easily obtained by low cost, single-stage compressors, higher pressures are increasingly being considered both industrially viable and safe. For example, the Toyota Miari FCV utilises a H2 fuel tank stored under 700 bar pressure.18 In the current case, a maximum pressure of 100 bar, though above the U.S. DOE target, represents significantly milder storage conditions than the 250 bar commonly applied to industrially compressed gas. Yet, even at these higher pressures, the monolithHKUST-1 working capacity increases only up to 235 cm3 (STP) cm–3 (5 – 100 bar) and hence still cannot reach the U.S. DOE target. This result is commonly observed amongst Type I microporous MOFs which saturate at low pressure; substantial increases in gas uptake are prevented even at substantially elevated pressures. Interestingly, after ca. 90 bar, monolithUiO-66_D overtakes monolithHKUST-1 and reaches a volumetric working capacity of 261 cm3 (STP) cm–3 (5 – 100 bar). This outstanding result is a consequence of the material’s finely tuned physical properties i.e. the unique combination of weak CH4 interaction at low pressure and its Type II isotherm character, which enhances gas uptake at high pressure. Due to its non-crystalline mesoporosity, UiO-66_D does not saturate even at 100 bar, representing an 11% improvement in working capacity 211 relative to benchmark densified monolithHKUST-1 (5 – 100 bar). Furthermore, this significant improvement in volumetric working CH4 storage capacity has been achieved with an air-stable monolith. The working capacity of 261 cm3 (STP) cm–3 (5 – 100 bar) for air-stable monolithUiO- 66_D is the highest recorded for a conformed MOF over this pressure range. Table 2| Volumetric CH4 adsorption capacities (5, 65 and 100 bar, 298 K) and corresponding working capacities for monolithUiO-66_D compared to the theoretical, defect-free MOF as well as benchmark monolithHKUST-1. CH4 uptake cm3 (STP) cm–3 Working capacity cm3 (STP) cm–3 5 bar 65 bar 100 bar 5 – 65 bar 5 – 100 bar UiO-66_D 40 212 301 172 261 UiO-66† 61 178 200 117 139 monolithHKUST-1* 74 259 309 185 235 † Theoretical gas adsorption data collected by GCMC simulation of the defect-free MOFs CH4 adsorption isotherm (298 K, 0 – 100 bar). *Data from Reference 4. 212 1.3| CO2 Storage The different monolith’s abilities to store pollutant CO2 was explored. Excess gravimetric isotherms were again collected (0 – 40 bar, 298 K) and data were converted to absolute gravimetric and volumetric results (Figure 5). Energetically demanding higher pressures were not applied to these adsorption studies, as CO2 liquefies under further compression. Comparable to the results for CH4 storage, the recorded gravimetric uptake capacities did not vary substantially across UiO-66_A – D (Figure 5a). Again, this stems from the similar availability of adsorptive surface area in each of the monolithic materials (Table 1) with the small observed differences in gravimetric uptake correlating with relatively minor differences in SBET between them. Figure 5| Gravimetric and volumetric CO2 storage in UiO-66 monoliths. a, b, Comparison of gravimetric and volumetric, respectively, CO2 storage capacity in UiO-66 monoliths (0 – 40 bar, 298 K). Data correspond to UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO- 66_C (purple squares) and UiO-66_D (green circles) and a standard pressurised storage tank (black crosses). The Type II isotherms, previously observed for CH4 storage in the monolithUiO-66, were again apparent. This is highlighted by the high CO2 uptake across the monoliths, which ranges from 0.43 – 0.54 g g–1 at 298 K, 30 bar. For comparison, Zhou et al. reported CO2 uptakes of only 0.38 g g–1 at 300 K, 30 bar in the powdered analogue of the same MOF.12 The superior gas 0 10 20 30 40 0 50 100 150 200 250 300 CO 2 u pt ak e (S TP ) c m 3 c m -3 Pressure (bar) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 10 20 30 40 CO 2 u pt ak e (S TP ) g g -1 Pressure (bar) CO 2 up ta ke g g -1 CO 2 up ta ke c m 3 (S TP ) c m -3 Pressure (bar) Pressure (bar) a b 213 adsorption capacity of the mixed micro-/mesoporous monoliths (Type II isotherm) compared to the purely microporous powdered material (Type I isotherm), is again the result of gas condensation in the wide mesopores. To confirm that the origin of this enhanced gas adsorption capacity was the result of condensation in the monolith’s mesopores, the location of adsorbed species in the mixed micro- /mesoporous MOF was again computationally simulated (Figure 6). Identically to both N2 and CH4 adsorption, CO2 was observed to occupy exclusively the micropores at low pressure while rapidly condensing in the mesopores at elevated pressures. Conversion of the gravimetric (g g–1) CO2 adsorption data to volumetric (cm3 (STP) cm–3) using the experimental rb of monolithUiO-66_A – D (Table 1) resulted in comparable trends to those seen for CH4 adsorption; Nabs CO2 (cm3 (STP) cm–3) UiO-66_D ≈ UiO-66_C >> UiO-66_B ≈ UiO-66_A (Figure 5b). Furthermore, the volumetric capacity of each of the studied materials represents a significant improvement over that of a pressurised tank. UiO-66_D (284 cm3 (STP) cm–3 CO2 at 40 bar, 298 K) in particular shows an excellent ca. 500% improvement over a conventional storage tank (48 cm3 (STP) cm–3 CO2 at 40 bar, 298 K) in addition to a 50% improvement over commercial adsorbent zeolite 13X (ca. 190 cm3 (STP) cm–3 at 40 bar, 298 K).19 Powdered MOF-177 has previously demonstrated an outstanding CO2 storage capacity of 323 cm3 (STP) cm–3 (40 bar, 298 K).19 While this exceeds the storage capacity of monolithUiO- 66_D, this theoretical maximum was calculated using the single crystal density of MOF-177. As is typical of MOFs, this low-density powdered adsorbent is not optimised for industrial usage and significant reductions in experimental volumetric capacity would be expected under practical working conditions. Contrarily, monolithUiO-66_D represents a viable industrial means of achieving CO2 storage with substantial benefits over pressurised storage systems, commercial adsorbents and top performing powdered MOFs. 214 Figure 6| Snapshots and density distributions of CO2 adsorption. a, Snapshots and b, density distributions comparing CO2 (red\grey) adsorption in mixed micro-/mesoporous UiO- 66 at 10 and 30 bar pressure. Colours within the UiO-66 crystal structure correspond to elements C (grey), H (white), Zr (blue) and O (red). In each case, the studied gases (N2, CH4, CO2) demonstrated a capacity to condense in each of the monolith’s unique non-crystalline mesoporosity, the results of which was Type II isotherms and enhanced gas uptake relative to that of the purely microporous material. By this method, a volumetric capacity exceeding the theoretical maximum can be achieved. Yet this conflicts with the observation that the densest monoliths exhibit the highest volumetric capacity (comparison of UiO-66_A and UiO-66_D, which both have Type II isotherms, highlights this). In order to achieve the maximum volumetric gas storage capacity, the ratio of micro:meso porosity in a material must be finely tuned. Figure 7 plots the volumetric CH4 and CO2 adsorption capacities of monoliths UiO-66_A – D as well as the theoretical/defect free MOF against their mesopore:micropore ratio. This clearly demonstrates the existence of a maximum performance in terms of gas-uptake capacity. 10 bar 10 bar30 bar 30 bar a b 10 bar 10 bar30 bar 30 bar 215 Figure 7| Optimising gas storage in mixed porosity MOFs. Comparison of volumetric storage capacity of CH4 at 65 bar (blue line) and 100 bar (black line) as well as CO2 at 40 bar (red line) in UiO-66 materials; UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO- 66_C (purple squares), UiO-66_D (green circles) and single crystal UiO-66 (white squares, computationally simulated from the defect-free crystal structure). All data were collected at 298 K. While the inclusion of mesoporosity in the MOF enhances its gravimetric gas storage capacity, it inherently sacrifices the material’s rb. Hence, excessively mesoporous UiO-66_A and UiO- 66_B yield low density structures which cannot achieve high density gas storage. Alternatively, purely microporous, dense, defect-free UiO-66 does not achieve high volumetric gas storage capacity as a result of rapid onset micropore saturation. In the case of UiO-66_C – D, a finely tuned balance between SBET, rb and Vtot for optimised gas uptake was achieved. This novel capability to beneficially tune/control mesoporosity into an otherwise purely microporous MOF has not been explored before. These results suggest that industrially promising high- stability/low-cost MOFs, which have previously been deemed incapable of reaching gas storage targets, should be re-explored. Through synthetic inclusion of mesoporous cavities with careful tuning of their mesopore:micropore ratio, the volumetric gas storage capacity of such MOFs may be elevated to far beyond their theoretical limits. G as u pt ak e cm 3 (S TP ) c m -3 0.2 0.4 0.6 0.8 1.0 1.2 0 50 100 150 200 250 300 350 Micro/Mesoporosity ratio 216 1.4| Gas Adsorption Kinetics Figure 8| CH4 adsorption-desorption isotherms for UiO-66 monoliths. Absolute volumetric (cm3 (STP) cm–3) CH4 adsorption (filled markers) and desorption (hollow markers) isotherms for a, UiO-66_A (blue triangles); b, UiO-66_B (red diamonds); c, UiO-66_C (purple squares) and d, UiO-66_D (green circles) represented as absolute uptake between 0 – 100 bar (298 K). The kinetics of gas adsorption are of paramount industrial importance as they determine a materials ability to equilibrate to a target pressure. Hence, for efficient fuel accessibility as well as time-efficient re-fuelling, both of which are required for industrially viable gas storage, it is essential that the kinetics of gas adsorption/desorption are rapid. Firstly, unlike their N2 adsorption isotherms (Chapter II, Section 1.0), all adsorption/desorption CO2 and CH4 isotherms for monolithUiO-66 showed an absence of hysteresis (Figure 8 and Appendix, Supplementary Figures 19 – 22). This demonstrates the ready accessibility of the adsorbed fuel i.e. the materials pore geometry does not trap the adsorbed species (as is the case for hysteretic materials). Furthermore, the semi-linear Type II shape of these desorption isotherms means that consistent decreases in pressure should result in steady releases of stored gaseous Volumetric 0 20 40 60 80 100 0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 0 20 40 60 80 100 120 140 0 20 40 60 80 100 0 50 100 150 200 250 300 0 20 40 60 80 100 0 50 100 150 200 250 300 Pressure (bar) Pressure (bar) Pressure (bar)Pressure (bar) CH 4 up ta ke (S TP ) v v- 1 CH 4 up ta ke (S TP ) v v- 1 CH 4 up ta ke (S TP ) v v- 1 CH 4 up ta ke (S TP ) v v- 1 Pressure (bar) Pressure (bar) re (bar) Pressure (bar) CH 4 up ta ke (S TP ) c m 3 cm -3 CH 4 up ta ke (S TP ) c m 3 cm -3 CH 4 up ta ke (S TP ) c m 3 cm -3 CH 4 up ta ke (S TP ) c m 3 cm -3 Volumetric a b c d a c 217 fuel. This contradicts more typical MOF Type I desorption isotherms, in which the majority of gas is rapidly released over a small pressure range i.e. near the minimum desorption pressure. These novel materials may therefore offer practical benefits in terms of industrial fuel storage. The kinetics of gas uptake for UiO-66_A – D were recorded by measuring the time taken to reach 95% equilibrium pressure after gas dosing (Figure 9). Rapid, industrially viable equilibration was reached for both CO2 and CH4 (less than 270 s for all monoliths) with no discernible trends related to their density or porosity being apparent. These results are comparable to the fast adsorption kinetics recently reported for monolithHKUST-1 (200 s for CH4).4 Data suggest that despite its high density and monolithic macrostructure, UiO-66_D (the best performing material in this study) offers high industrial viability for gas-fuel storage both in terms of rapid gas-adsorption together with accessibility of the adsorbed species for fuel- release. 218 Figure 9| Gas uptake kinetics for UiO-66 monoliths. Decay of pressure in UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles) for a, CH4 at ca. 100 bar (298 K) and b, CO2 at ca. 40 bar (298 K). Dashed black line (inset) indicates the pressure at which 95% uptake equilibrium was obtained. a b 0 50 100 150 200 98.2 98.4 98.6 0 100 200 300 98.2 98.4 98.6 98.8 0 50 100 150 200 98.4 98.6 98.8 0 100 200 300 400 98.4 98.6 98.8 Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Time (s) Time (s) Time (s)Time (s) Pr es su re (b ar ) 0 20 40 60 80 100 38.80 38.84 38.88 0 25 50 75 100 125 150 39.72 39.78 39.84 0 50 100 150 200 250 300 39.56 39.60 39.64 39.68 0 25 50 75 100 39.60 39.64 39.68 Time (s) Time (s) Time (s) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Time (s) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Ti e (s) Ti (s) i ) i (s) Ti (s) i e (s) Ti e (s) Ti e (s) 219 1.5| Gas Separation Un-refined NG contains a significant proportion of contaminant species e.g. over 70% CO2, SO2 and H2O.20 Removal of the abundant CO2 from CH4 is an essential component of ‘NG sweetening’, the process of industrially purifying these gas streams for fuel applications. Membrane-based separation processes have been the subject of significant interest over recent years, with tuneable porous MOFs being of obvious appeal.21–23 Powdered MOFs are, however, unsuitable for gas phase filtration due to unavoidable powder densification during application – the result of which is pipe blockages and pressure drops. Highly porous monolithUiO-66 presents obvious benefits for this application. The materials high water stability coupled with both its large monolith particle size and good mechanical strength make it an ideal molecular sieve for industrial gas separation. Figure 10| Equimolar binary isotherms for UiO-66 monoliths. Binary gravimetric isotherms of CO2 (filled marker) and CH4 (hollow marker) uptake in monoliths (0 – 40 bar, 298 K); a, UiO-66_A (blue triangles), b, UiO-66_B (red diamonds), c, UiO-66_C (purple squares) and d, UiO-66_D (green circles). 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 Lo ad in g (m m ol g -1 ) Lo ad in g (m m ol g -1 ) Lo ad in g (m m ol g -1 ) Lo ad in g (m m ol g -1 ) Pressure (bar) Pressure (bar) Pressure (bar) Pressure (bar) a b c d 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 100 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 0 10 20 30 40 0 2 4 6 8 10 220 Equimolar binary CO2 and CH4 adsorption isotherms for the different UiO-66 monoliths were calculated from the experimental pure-component adsorption isotherms of each gas at 298 K (Figure 10) using IAST.24,25 The experimental isotherms were previously fitted to an analytical model: the BET model for CO2 isotherms and quadratic model for CH4 isotherms. From these data it is apparent that for each of the monoliths high preferential adsorption of CO2 is favoured across the entire pressure range. This is consistent with previous reports of the high affinity for CO2 of UiO-66.8 With the aim of better understanding the differences in gas adsorption selectivity between the different monoliths, selectivity data for a 50:50 gas mixture was calculated from the equimolar binary isotherms (Figure 11). Data showed similar selectivity in all cases; the average CO2:CH4 selectivity amongst the materials was 4.8 ± 0.3 at 10 bar. The small error associated with this average is an indication of the negligible differences in selectivity between the monoliths. Furthermore, each monolith exhibited an exponential reduction in selectivity at increased pressure. This is consistent with previous observations of reduced membrane selectivity in polymeric and mixed-membrane matrices at elevated pressures.26 These data indicate that the robust monolithic UiO-66 MOFs may be applicable, not only to dense NG fuel storage but also to selective gas separation for NG purification. Figure 11| Gas mixture adsorption selectivity for UiO-66 monoliths. CO2:CH4 uptake selectivity for monoliths UiO-66_A (blue triangles), UiO-66_B (red diamonds), UiO-66_C (purple squares) and UiO-66_D (green circles) for a 50:50 gas mixture (0 – 40 bar, 298 K). 0 10 20 30 40 0 1 2 3 4 5 6 7 Se le ct iv ity C O 2:C H 4 Pressure (bar) 221 1.6| Conclusions The industrial potential of monoliths UiO-66_A – D, synthesised in Chapter II, Section 1.0, towards a number of gas phase applications was experimentally and computationally studied. Firstly, each material’s potential towards ambient temperature NG storage was determined by measurement of its CH4 adsorption capacity up to 100 bar. While similar gravimetric capacities were obtained amongst the different materials, volumetric capacities varied as a function of their significantly differing rb. The highest volumetric uptake capacity was exhibited by the densest monolith, UiO-66_D, which represented a ca. 200% improvement over traditional storage technology i.e. a pressurised tank. Furthermore, each of the tested monoliths demonstrated a Type II CH4 adsorption isotherm which differed from the Type I isotherms reported, both experimentally for the powdered MOF, as well as computationally simulated in the defect-free material. As a result of its unusual isotherm shape, the volumetric uptake capacity of UiO-66_D at 100 bar exceeded both the theoretical maximum in the defect-free MOF as well as that of benchmark monolithHKUST-1. Computational simulations of mixed micro-/mesoporous UiO-66 were used to verify that the origin of this enhanced gas uptake capacity lies in the inclusion of non-crystalline mesoporosity in the monolithic materials. This serves to enhance gas uptake capacity at high pressure via condensation in the wide mesopores after micropore saturation. Yet, excessive mesoporosity reduces rb and diminishes volumetric capacity; as was the case for UiO-66_A and UiO-66_B. The data clearly indicated that an enhanced volumetric gas adsorption capacity, exceeding the theoretical maximum in the defect-free MOF, may be achieved by tuning the materials micro- /mesoporous ratio. This first report of enhanced gas storage via synthetically tuning, non- crystalline, mesoporosity into a traditionally microporous MOF suggests that high stability/low cost MOFs, which cannot reach gas storage targets in their typical microporous state, should be re-explored for potential industrial applications. For example, citing these results for monolithUiO-66 as motivation,27 Huang et al. subsequently reported analogous enhancements in the CO2, CH4, C2H2 and C2H4 uptake capacity of MIL-121; a thermally triggered decarboxylation procedure was used to post-synthetically introduce hierarchical mesoporous defects into the powdered MOF. As a result, CH4 uptake capacity was increased e.g. to 132.3 cm3 g–1 (80 bar, 298 K) from a previously negligible capacity at the same temperature and pressure in the defect-free material.28 222 Calculation of the working capacity of UiO-66_D over the U.S. DOE target pressure range (5 – 65 bar) demonstrated that it was ca. 10% lower than that of record-setting yet chemically unstable monolithHKUST-1. However, further increasing the maximum storage pressure to, still industrially viable, 100 bar resulted in a working capacity of 261 cm3 (STP) cm–3 (5 – 100 bar), exceeding that of monolithHKUST-1 by ca. 10%. This is the highest volumetric working capacity recorded in a conformed MOF over this pressure range. The practical viability of high stability monolithic UiO-66_D for potential gas storage applications was also supported by its rapid gas adsorption kinetics (comparable to monolithHKUST-1) as well as the absence of hysteresis in its desorption isotherm (which suggests ready access to the stored fuel). The CO2 adsorption capacity of monolithic UiO-66 was further studied, with Type II isotherms again observed in all cases. An outstanding 500% improvement in volumetric CO2 adsorption capacity was observed in monolithUiO-66_D, relative to a pressurised storage tank. Not only does this suggest that the material may be applicable to industrial CO2 storage but calculation of CO2:CH4 selectivity demonstrated significantly favoured adsorption of CO2, pointing towards potential applications in NG refinery. 223 2.0| H2 Storage in monolithHKUST-1 and Pd@monolithHKUST-1 In Chapter I, the numerous environmental benefits of H2 gas-fuel over traditional fossil fuels were introduced. These benefits include cleaner combustion (with water being the only by- product), absence of spilling/pooling (rapid gas dissipation occurs in the event of fuel escape), better energy security as well as its potential to be renewably generated by a number of environmental methods e.g. biological generation using cyanobacteria, water electrolysis powered by wind turbines or water photo-splitting catalysed by semi-conductors.29 If our current fossil-fuel based economy is reasonably to be replaced by a H2 based economy, the essential foundations of its infrastructure must be economically achieved: fuel-generation, transportation, storage, conversion and usage.30 H2 storage in particular presents a substantial challenge; energy dense storage systems (both mobile and stationary) are required if H2 is to compete with established fossil-fuels. While its gravimetric energy density is three-times greater than petrol, the volumetric energy density of H2 is four-times lower.29 Despite cryogenic and high pressure storage systems having been extensively applied to this fuel, they are both dangerous, impractical and costly.31 Furthermore, these storage means are incapable of achieving the U.S. DOE’s highly ambitious on-board storage targets (6.5 wt% and 50 g L–1).32 The capacity of MOFs to efficiently adsorb gaseous species in their abundant microporosity offers a safe and economic means of achieving dense H2 storage. While low temperatures (below the U.S. DOE targets) are often applied to H2 storage testing,33 promising results at ambient temperatures have also been reported. For example; Long et al. reported benchmark results for near ambient H2 storage in Ni2(m-dobdc) of 11.9 g L–1 at 298 K and 100 bar.34 However, this powdered material was not pelletised, making it practically and technologically unsuitable for industrial gas storage applications. In contrast, the desirable combination of high rb and Vtot provided by the robust monolithic MOFs studied herein may present a more industrially viable means of achieving dense H2 storage. While monolithMOFs have already reported outstanding results towards both CH4 storage and CO2 capture,4,27 their potential towards more challenging H2 storage has not been explored thus far. 224 2.1| Aims and Objectives As discussed (Chapter III, Section 3.0), the benchmark CH4 storage material, monolithHKUST- 1, reported by Tian et al. was used as a host for Pd NPs with the aim of enhancing its H2 storage capacity. For the first time, NP immobilisation inside this monolith was demonstrated with high loadings of 5% and 10% (by weight) being achieved. Significantly, the crystallinity, porosity, SBET and rb of the monolith were maintained reasonably well despite this substantial doping. Furthermore, XPS analysis of the materials indicated an electronic interaction between the Cu- MOF and Pd NPs. The aim of this section is therefore to determine the extent of both the Pd- doped and undoped monolithHKUST-1 material’s industrially potential towards H2 storage. Firstly, the low pressure (0 – 1 bar) H2 storage capacity of each monolithic material will be explored. Since the formation of PdH is highly energetically favourable, this is the pressure range where the influence of the dopant is expected to be most apparent. The results will be benchmarked against the literature report of powdered Pd@HKUST-135 as well as the current industry standard technology i.e. a pressurised gas cylinder, over the same pressure range. Furthermore, the mechanism of gas uptake into the composite’s components (i.e. Pd and MOF) will be explored by comparing variable temperature adsorption-desorption isotherms in the different monolithic materials as well as through in situ H2 adsorption XRD studies. Secondly, the high pressure, ambient temperature H2 adsorption capacity of each material will be recorded up to 100 bar, with the aim of determining their viability towards gas storage at higher pressures. The influence of doping on both the total uptake capacity and the accessible working capacity will be compared to the undoped material. Crucially, the obtained isotherms will be analysed for signs of enhanced uptake due to the electronic interaction of the two components, rather than total uptake simply being the sum of the two.36 H2 storage results will be benchmarked against the current record-setting MOF at ambient temperature34 as well as standard pressurised storage technology. 225 2.2| Low Pressure H2 Storage The excess H2 storage capacities of monolithHKUST-1, 5% and 10% doped Pd@monolithHKUST- 1 were recorded under ambient conditions (298 K) up to a maximum pressure of 1 bar (Figure 12) and converted to absolute uptake (as discussed in Chapter IV, Section 1.0). The Pd doping level of the materials significantly influenced their gravimetric H2 capacity (Figure 12a). A strong enhancement in this low pressure adsorption was observed to correlate with Pd doping level whereby, over the entire pressure range, H2 adsorption (wt%) for monolithHKUST-1 < 5% Pd@monolithHKUST-1 < 10% Pd@monolithHKUST-1. Data suggest that at low pressure, the loss of SBET incurred by the monolith doping (Chapter III, Section 3.0) did not significantly impact its overall capacity to store H2. Instead, the dopant was the most important factor in determining low pressure H2 uptake, as evidenced by the isotherm shapes. While monolithHKUST-1 exhibits a near linear uptake of H2 with increased pressure, both of the doped monoliths show rapid uptake below 0.1 bar. This sudden and immense uptake corresponds to chemisorption of H2 by the Pd NPs via the energetically favourable formation of PdH. When the Pd is saturated, above ca. 0.1 bar, the gas can no longer be chemisorbed and is instead physisorbed by the surrounding MOF, comparably to the undoped material. Hence, the greater the monolith doping level, the higher this initial spike in H2 chemisorption at low pressure due to the formation of PdH. Comparison of the different materials semi-logarithmic adsorption-desorption isotherms provides further evidence for this H2 storage mechanism (Figure 12c). Hysteresis at low pressure (< 0.1 bar) was observed only in the doped materials. This delayed H2 desorption is likely incurred by the energetic unfavorability of PdH returning to Pd under vacuum conditions.37 Each material’s gravimetric (wt%) isotherm was further converted to volumetric (g L–1) using their experimentally determined rb (Chapter III, Section 3.0). Due to their near identical densities, the trends observed in the gravimetric uptake were the same as those for volumetric; Pd doping of the monolith incurred significant enhancements in both volumetric and gravimetric capacity. Crucially, each of the materials showed a significant improvement in volumetric uptake capacity compared to a pressurised storage tank; the current standard technology (Figure 12b). In particular, the results of 10% Pd@monolithHKUST-1 are promising, showing not only a ca. 250% improvement over the undoped material but crucially a ca. 1000% improvement in uptake compared to a traditional storage tank. 226 Figure 12| Low pressure H2 isotherms. a, b, Gravimetric (wt%) and volumetric (g L–1), respectively, linear H2 adsorption isotherms and c, semi-logarithmic adsorption-desorption isotherms (volumetric). Data correspond to monolithHKUST-1 (green circles), 5% Pd@monolithHKUST-1 (blue squares), 10% Pd@monolithHKUST-1 (purple diamonds) and a traditional pressurised storage tank (black line). All isotherms were recorded at 298 K up to a maximum pressure of 1 bar. These preliminary H2 storage studies demonstrate that NP doping can significantly enhance the low-pressure gas storage capacity of the host MOF to achieve results significantly better than the current technology. Furthermore, this suggests that additional increases in doping level may increase the H2 storage capacity of the composite further. The results of these doped monoliths can be compared to that published by Kitagawa et al. where the low pressure H2 storage capacity of Pd@powderHKUST-1 was studied.35 Although those adsorption studies were reported only as a function of H atoms per Pd atom in the composite (Pd/H), the data was digitised and converted to gravimetric uptake using their reported MOF doping level, 24% Pd (Figure 13). Since the experimental rb of this powdered material was not published (rb is difficult to accurately quantify for powdered materials), the volumetric results could not be obtained. However, comparison of the gravimetric results (mmol g–1) supports the current work (Figure 13). The magnitude of the low pressure H2 adsorption in the 24% doped powder material is significantly higher than that of the analogous monoliths and correlates well with the observed trend in doping level i.e. H2 uptake (1 bar) 24% Pd@powderHKUST-1 (i.e. 0.98 mmol g–1) > 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.001 0.01 0.1 1 Pressure (bar) H 2 up ta ke (g L -1 ) 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0 0.2 0.4 0.6 0.8 1 H 2 up ta ke (g L -1 ) Pressure (bar)Pressure (bar) H 2 up ta ke (w t % ) a b c 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.2 0.4 0.6 0.8 1 227 10% Pd@monolithHKUST-1 (0.42 mmol g–1) > 5% Pd@monolithHKUST-1 (0.25 mmol g–1) > monolithHKUST-1 (0.11 mmol g–1). However, such high doping levels (i.e. 24%) could not be achieved in the monolithic materials, where attempts to do so were qualitatively observed to incur a significant reduction in mechanical properties; obtained materials became malleable, visually dull and non-monolithic. By extrapolation of these data, which demonstrate that elevating the composites Pd content increases H2 capacity, the best low-pressure H2 storage material would appear to be pure Pd NPs (in the absence of MOF). Yet despite their outstanding H2 storage capacity, without immobilising on a host material, toxic, expensive and impractical Pd NPs are non-viable in terms of both large-scale industrial usage and recycling. The fact that the high H2 adsorption capabilities of these NPs can likewise be achieved when immobilised in a densified monolith makes the materials very industrially promising, particularly for proposed gas storage applications. Based on these results, the mechanism of H2 storage in the doped composites was further studied. Figure 13| Comparison of powder and monolith H2 isotherms. Gravimetric (mmol g–1) linear adsorption isotherms for monolithHKUST-1 (green circles), 5% Pd@monolithHKUST-1 (blue squares), 10% Pd@monolithHKUST-1 (purple diamonds) and 24% Pd@powderHKUST-1 (red circles, digitised from Reference 35). All isotherms were recorded at 298 K up to a maximum pressure of 1 bar. Pressure (bar) H 2 up ta ke (m m ol g -1 ) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.2 0.4 0.6 0.8 1 228 2.3| Mechanistic Studies 2.3.1| Temperature Dependence of H2 Adsorption A practical and cost-effective ambient temperature of 298 K was used for the above adsorption studies. However, the U.S. DOE’s H2 targets (6.5 wt% and 50 g L–1) do not specify a fuel storage temperature, merely an acceptable temperature range for fuel delivery to the FCV (233 – 358 K).33 This offers some scope for temperature-swing conditions, i.e. different storage and release temperatures. This swing technique has previously achieved outstanding enhancements in the deliverable H2 capacity of numerous MOFs including NU-1000, UiO-67 and HKUST- 1.33 Hence, the effect of temperature on the low pressure H2 adsorption capacity was further studied by similarly recording the 0 – 1 bar adsorption-desorption H2 isotherms of the monoliths at the lower temperatures of 283 K and 273 K (Figure 14). It is immediately apparent that across all materials, hysteresis in the semi-logarithmic plots is more pronounced at reduced temperatures. This is consistent with the accepted origin of the hysteresis in Pd – an energy barrier to the hydride phase transformation that results from lattice strain.37 Figure 14| Variable temperature H2 isotherms. Semi-logarithmic, gravimetric (mmol g–1) H2 adsorption (filled marker) – desorption (hollow marker) isotherms for monolithHKUST-1 (green circles), 5% Pd@monolithHKUST-1 (blue squares) and 10% Pd@monolithHKUST-1 (purple diamonds) at a, 273 K, b, 283 K and c, 298 K up to a maximum pressure of 1 bar. 0 0.1 0.2 0.3 0.4 0.001 0.01 0.1 1 0 0.1 0.2 0.3 0.4 0.5 0.001 0.01 0.1 1 0 0.1 0.2 0.3 0.4 0.5 0.001 0.01 0.1 1 Pressure (bar) Pressure (bar) Pressure (bar) H 2 up ta ke (m m ol g –1 ) H 2 up ta ke (m m ol g –1 ) H 2 up ta ke (m m ol g –1 ) a b c 273 K 283 K 298 K 229 Improvements in the total uptake capacities of the doped materials were observed by reducing the storage temperature; for example, a 38% enhancement in storage capacity was achieved by decreasing the temperature from 298 K to 283 K in 10% Pd@monolithHKUST-1. This is consistent with numerous literature studies regarding the temperature dependence of H2 storage in pure Pd NPs; higher temperatures are thought to stabilise the metal phase.38 However, no further improvement was achieved through temperature reduction to 273 K. This suggests that, while low temperature storage and high temperature desorption may be used to enhance H2 storage capacity, the studied near-ambient temperature range appears too low to see meaningful trends. Further detailed testing under more substantial temperature/pressure swing conditions could be used to fully elucidate each materials potential over the entire deliverable temperature range proposed by the U.S. DOE (233 – 358 K). 230 2.3.2| Relative Contribution of Pd and Cu to Total H2 Uptake Kitagawa et al. previously reported that, for Pd@powderHKUST-1, the contribution of the host MOF towards the composite’s H2 storage capacity at low pressure was negligible relative to the vast H2 uptake of Pd.35 This assumption was based on their calculation of the number of H atoms per Pd atom and per Cu3(btc)2 unit in the separate components: they observed H2/Cu3(btc)2 < 0.02 (1 bar 303 K) in the pure MOF whereas H/Pd ca. 0.5 in the pure NPs (1 bar, 303 K) (Figure 15d). Hence, for the purposes of studying the H2 uptake in the composite material, they attributed the entirety of the recorded H2 storage to the Pd NPs. With the aim of better comparing the current Pd@monoltihHKUST-1 data to the literature results for both Pd@powderHKUST-1 and Pd NPs, the experimentally obtained isotherms (Figure 12) were likewise plotted as a function of H/Pd (Figure 15a – d), under the same assumptions as Kitagawa. In all cases, the same distinctive H2 isotherm shape was observed (Figure 15e). This is accounted for by the well-studied mechanism of H2 chemisorption by Pd. Firstly, in Stage 1 (Figure 15e), despite sharp increases in pressure, only a relatively small uptake of H2 into the Pd was recorded, as indicated by low values of H/Pd. This corresponds to the primary formation of the dilute a-hydride phase (PdHx) where x £ 0.03.39 The conduction band of Pd comprises dense 4d and broad 5s bands. In the presence of H2, a lower energy state is formed below the 4d band by the transfer of 1s H electrons to the 4d holes. This band formation means that the energy of the hydride system is lower than that of a solid solution.40 Hence, this primary stage is followed by a rapid plateau (Stage 2, Figure 15e) whereby, for a very small increased in pressure, a sudden increase in H/Pd is observed. This is the materials phase transition to the denser b-hydride (PdHx) phase (x ³ 0.58) which occurs when the chemical potential of the a and b phases is the same.40 Ultimately, Pd’s capacity form high density b-hydride is limited by the availability of 4d band holes.41 The plateau pressure at which this a-hydride to b-hydride transition occurs has been extensively studied and is known to correlate with Pd NP size, shape and temperature.38,42,43 For the monolithic composite materials synthesised herein, the transition occurs rapidly between 0.02 – 0.03 bar. However, Kitagawa et al. observed a more gradual b-hydride transition at the higher pressures of 0.05 – 0.1 bar. This may correlate with the slightly higher temperature used for their study (303 K vs. 298 K). Urban et al. previously explored the size and temperature 231 dependence of H2 storage in Pd NPs. Through combined experimental and computational studies, it was demonstrated that at elevated temperatures the metal phase is stabilised, delaying the hydride phase transition.42 Figure 15| H/Pd adsorption-desorption isotherms. Comparison of H2 adsorption-desorption semi-logarithmic isotherms (0 – 1 bar) for a, b, 5% Pd@monolithHKUST-1 (blue squares) and 10% Pd@monolithHKUST-1 (purple diamonds), respectively (at 298 K), to c, d, 24% Pd@powderHKUST-1 (red circles, digitised from Reference 35) and pure Pd NPs (black squares, digitised from Reference 35), respectively (303 K), as a function of H atoms per Pd atom. e, Graphical representation of the distinct stages of H2 interaction with Pd under applied pressure. 0.0001 0.001 0.01 0.1 1 0 0.2 0.4 0.6 0.8 0.0001 0.001 0.01 0.1 1 0 0.2 0.4 0.6 0.8 0.0001 0.001 0.01 0.1 1 0 0.2 0.4 0.6 0.8 0.0001 0.001 0.01 0.1 1 0 0.2 0.4 0.6 0.8 a b c d H/Pd H/Pd H/Pd H/Pd Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) 0.0001 0.001 0.01 0.1 1 0 0.2 0.4 0.6 0.8 H/Pd Pr es su re (b ar ) Stage 1 Stage 2 Stage 3 e 232 The transition stage discussed above is followed by a post-plateau region where a gradual increase in Pd/H is observed for large increases in pressure (Stage 3, Figure 15e). This, correlates with the formation of a solid solution in the metal b-hydride.44 For each of the Pd- doped MOF materials (both monoliths and the powder reported by Kitagawa et al.), a final H/Pd in excess of 0.8 was achieved at 1 bar H2 pressure. This is considerably higher than the uptake capacity of the pure NPs i.e. not encapsulated in the MOF, where only ca. 0.5 H/Pd was reported (Figure 15d). In a follow-up paper, Kitagawa reported using high energy synchrotron X-ray spectroscopy techniques to probe the electronic structure of the composite Pd@powderHKUST-1.41 Data suggested that charge transfer from the Pd 4d bands to hybridised Cu–O bands in HKUST-1 occurs at the NP-MOF interface. They further postulated that this serves to increase the density of 4d holes in the NPs which ultimately enhances hydride capacity. Hence, the comparable results obtained herein, in addition to the XPS data (Chapter III, Section 3.0) would provisionally appear to corroborate the results by Kitagawa: the MOF enhances the H2 storage capacity of the NPs at low pressure by an electron transfer interaction. As discussed earlier, the conclusions drawn by Kitagawa et al. rely on the assumption that the MOFs adsorption capacity is negligible. Hence, any improvements in uptake capacity were attributed to enhanced adsorption in the dopant as a result of its interaction with the host MOF, rather than to concurrent adsorption by said MOF. By their consideration of H2 uptake capacity in terms of relative molar uptake (i.e. H/Pd and H2/Cu3(btc)2), it would indeed appear that uptake by the MOF is very poor relative to the Pd. Yet, it must be conceded that the MOF comprises a significant proportion of the composite (90 and 95% for the monolithic materials and 76% for the powdered material). Hence, completely negating its contribution to the uptake as a result of its low H2/Cu3(btc)2 capacity may in fact not be correct since the total molar quantity of MOF far exceeds that of the dopant. A clear feature that suggests a contribution from the MOF towards H2 uptake is that conversion of the monolithHKUST-1 isotherm to H2 molecules per unit of Cu3(btc)2 demonstrates that at 1 bar, H2/Cu3(btc)2 reaches ca. 0.05 (Figure 16a). This is significantly higher that the < 0.02 H2/Cu3(btc)2 reported by Kitagawa for the powdered composite (Figure 16b) and further indicates that negation of the MOFs contribution may not be correct. Hence the adsorption data were re-explored under the alternative assumption that the MOFs adsorption capacity is non- negligible. By this method, the H/Pd isotherms were re-plotted by subtracting the pure monolithMOFs contribution from the total adsorption of each doped composite. 233 Figure 16| H2/Cu3(btc)2 adsorption-desorption isotherms. Comparison of H2 adsorption- desorption semi-logarithmic isotherms for a, monolithHKUST-1 (green circles) and b, powderHKUST-1 (blue triangles, digitised from Reference 35) as a function of H2 molecules per unit of MOF (Cu3(btc)2). The significant changes which occurred in the H/Pd isotherms (Figure 17a, b) as a result of these calculations indicate that the MOF contribution to the total adsorption capacity should not easily be dismissed. In fact, for each of the doped monoliths, the re-calculated isotherms much more closely resemble that of the pure NPs (Figure 17c) i.e. for 10% Pd@monolithHKUST-1, H/Pd at 1 bar is now 34% higher than that of the pure Pd NPs. This contrasts significantly with the 68% improvement calculated in this same material when negating the monolithMOFs contributions, as well as with the 74% improvement calculated by Kitagawa in the case of the powder material. Furthermore, it is clear that the a-hydride ® b-hydride transition (Stage 2, Figure 15e) for the MOF-immobilised NPs occurs over a much narrower pressure range than the same phase change in the pure NPs. These observations, under the alternatively proposed calculations, continue to support the previously discussed high energy synchrotron analysis of Pd@powderHKUST-1, which suggested there to be an electronic interaction between the dopant NPs and the host-MOF which serves to enhance b-hydride formation.41 However, data suggest that this effect is less pronounced that first supposed, with the chemisorption behaviour of 0.0001 0.001 0.01 0.1 1 0 0.02 0.04 0.06 0.08 H2 / Cu3(btc)2 Pr es su re (b ar ) 0.0001 0.001 0.01 0.1 1 0 0.02 0.04 0.06 0.08 H2 / Cu3(btc)2 Pr es su re (b ar ) a b 234 dopant Pd NPs (Figure 17a, b) being more comparable to pure Pd NPs (Figure 17c) after subtraction of the MOFs contribution to the adsorption (Figure 16a). Obviously, this method of modelling the gas adsorption must also make approximations; for example, it relies on the assumption that the MOFs H2 adsorption capacity is both significant and taking place concurrently to the NPs. Further, it assumes that the pure MOFs uptake remains relatively unchanged despite its proximity to/interaction with the immobilized NPs. While this is only an approximation to the true situation, it offers an alternative way to view the same data. Based on the meaningful effect that this alternative assumption makes on the results, as well as the known non-zero H2 uptake of HKUST-1 below 1 bar45 which comprises by far the major component of each composite, consideration of the MOFs contribution to the total uptake appears to be vital in understanding the mechanism of gas storage. Figure 17| Recalculated H/Pd adsorption-desorption isotherms. Comparison of H2 adsorption-desorption logarithmic isotherms in a, 5% Pd@monolithHKUST-1 (blue squares) and b, 10% Pd@monolithHKUST-1 (purple diamonds) at 298 K to c, pure Pd NPs (black squares, digitised from Reference 35) at 303 K, respectively, as a function of H/Pd. The H2 desorption from the composites under reduced pressure was also recorded (Figure 17). In all cases a hysteresis was associated with the a ® b phase transition, matching the results of both Pd@powderHKUST-1 and Pd NPs.35 This hysteresis has been attributed to the lattice strain induced by the transformation. As previously mentioned, when fcc Pd chemisorbs H2 to form 0.0001 0.001 0.01 0.1 1 0 0.2 0.4 0.6 0.8 0.0001 0.001 0.01 0.1 1 0 0.2 0.4 0.6 0.8 Pr es su re (b ar ) Pr es su re (b ar ) Pr es su re (b ar ) H/Pd H/Pd H/Pd a b c 0.0001 0.001 0.01 0.1 1 0 0.2 0.4 0.6 0.8 235 the hydride, its crystalline structure does not change, but the lattice undergoes a significant 3.5% expansion. The strain induced by returning to the relatively contracted Pd(0) state results in an energy barrier to the transition and hence H2 desorption under reduced pressure is inhibited, resulting in hysteresis.37 236 2.3.3| In Situ H2 Adsorption XRD Studies The lattice expansion that occurs during PdH formation, as was commented on above, accounts for observations made during in situ H2 adsorption XRD experiments. By this method XRD patterns were compared for each of the two doped monolithic composites under H2 atmospheres that varied between 0 – 1 bar (Figure 18 and Figure 19). In both cases, all diffraction maxima (hkl = [111], [200], [220], [311], and [222]) originating from the crystalline Pd NPs shifted to lower diffraction angles, i.e. greater lattice spacings, in the presence of H2. In this XRD experiment, at the lowest H2 pressures used (P < 0.02 bar), little lattice expansion was observed. However, at P = 0.2 bar, a sudden and pronounced peak shift was recorded and not further moved by subsequently elevated H2 pressures. These results demonstrate the slow formation of a-hydride at low pressures, followed by rapid expansion to the b-hydride phase at the critical plateau pressure, above which no further meaningful expansion takes place. No shift in the MOFs diffraction peaks was seen in the presence of H2 as HKUST-1 only weakly physisorbs the gas within its porosity, with no quantifiable lattice changes as a result. 237 Figure 18| XRD in situ H2 adsorption for 5% Pd@monolithHKUST-1. a, XRD patterns comparing peak positions under vacuum conditions to those under in situ exposure to H2 (0.01, 0.2 and 1.1 bar). Highlighted in pink are the Pd reflections; [111], [200], [220], [311] and [222]. The XRD pattern for defect-free HKUST-1 (simulated from its ideal crystal structure) is provided for comparison (black). b, Semi-logarithmic H2 adsorption-desorption isotherm (0 – 1 bar) where markers (red circles, inset) show the points in the isotherm at which XRD patterns were collected (a). c, Comparison of diffraction angles for each Pd reflection under exposure to different pressures of H2 (a). Data values are provided as both absolute (2θ) and relative (Δd (%), compared to the reflections under vacuum) angles. 0 20 40 60 80 100 Angle (2θ) In te n si ty ( a. u .) 1.1 bar H 2 0.2 bar H 2 0.02 bar H 2 Vacuum H KU ST-1 1.1 bar H2 0.2 bar H2 0.01 bar H2 Vacuum HKUST-1 0 20 40 60 80 100 In te ns ity (a .u .) Angle (2!) a 111 200 220 311 222 0.001 0.01 0.1 1 0 0.1 0.2 0.3 Pr es su re (b ar ) H2 uptake (mmol g-1) b hkl [111] [200] [220] [311] [222] 2θ 39.9 46.5 68.1 82.1 86.6 Δd (%) 0.0 0 .0 0.0 0.0 0.0 hkl [111] [200] [220] [311] [222] 2θ 39.9 46.5 67.9 82.0 86.4 Δd (%) 0.0 0.0 +0.3 +0.1 +0.2 hkl [111] [200] [220] [311] [222] 2θ 38.7 44.9 65.3 78.6 82.6 Δd (%) +3.0 +3.4 +3.8 +3.7 +3.9 hkl [111] [200] [220] [311] [222] 2θ 38.4 44.8 65.1 78.4 82.5 Δd (%) +3.8 +3.6 4.1 +3.9 +4.0 1.1 bar H 2 0.2 bar H 2 0.01 bar H 2 Vacuum c 238 Figure 19| XRD in situ H2 adsorption for 10% Pd@monolithHKUST-1. a, XRD patterns comparing peak positions under vacuum conditions to those under in situ exposure to H2 (0.01, 0.2 and 1.1 bar). Highlighted in blue are the Pd reflections; [111], [200], [220], [311] and [222]. The XRD pattern for defect-free HKUST-1 (simulated from its ideal crystal structure) is provided for comparison (black). b, Semi-logarithmic H2 adsorption-desorption isotherm (0 – 1 bar) where markers (red circles, inset) show the points in the isotherm at which XRD patterns were collected (a). c, Comparison of diffraction angles for each Pd reflection under exposure to different pressures of H2 (a). Data values are provided as both absolute (2θ) and relative (Δd (%), compared to the reflections under vacuum) angles. 0 20 40 60 80 100 In te n si ty ( a. u .) Angle (2θ) Vacu u m 0.01 b ar H 2 0.2 b ar H 2 1.1 b ar H 2 H KU ST-1 hkl [111] [200] [220] [311] [222] 2θ 40.0 46.6 68.1 82.2 86.7 Δd (%) 0.0 0 .0 0.0 0.0 0.0 hkl [111] [200] [220] [311] [222] 2θ 40.0 46.6 68.1 82.2 86.6 Δd (%) 0.0 0.0 0.0 0.0 +0.1 hkl [111] [200] [220] [311] [222] 2θ 38.7 44.9 65.4 78.7 83.0 Δd (%) + 3.2 + 3.6 + 3.6 + 3.6 + 3.6 hkl [111] [200] [220] [311] [222] 2θ 38.7 44.9 65.3 78.5 82.9 Δd (%) + 3.2 + 3.6 + 3.8 + 3.9 + 3.7 1.1 bar H 2 0.2 bar H 2 0.01 bar H 2 Vacuum 1.1 bar H2 0.2 bar H2 0.01 bar H2 Vacuum HKUST-1 0 20 40 60 80 100 In te ns ity (a .u .) Angle (2!) a b c 111 200 220 311 222 0.001 0.01 0.1 1 0 0.2 0.4 H2 uptake (mmol g-1) Pr es su re (b ar ) 239 2.4| High Pressure H2 Storage Currently, no industrially viable physisorptative material is capable of approaching the U.S. DOE’s highly ambitious H2 storage targets of 6.5 wt% and 50 g L–1. In fact, as a result of their weak physisorptative mechanism of gas storage, MOFs may never be capable of achieving such dense storage of non-polar H2 at very low pressures. Indeed, non-polar, light H2 gas (molecular mass, 2.02 g mol–1) is particularly difficult to store via physisorption as a consequence of its exclusively weak Van der Waals interactions with adsorbents.46 While low pressure H2 storage is both safe and cost-effective, it does not represent the only industrially viable pressure range; the Toyota Mirai FCV relies on ambient temperature H2 fuel storage at an immense 700 bar.18 Such high pressures are arduous to achieve, requiring energetically demanding and expensive multi-stage gas compressors. Yet, this commercial vehicle serves to demonstrate that substantially elevated pressures can be practically and economically viable. Hence monolithic MOF materials, while unlikely to achieve H2 storage targets, may still provide a means of advancing the current state of the art. As a consequence of the promising results achieved in the preliminary low pressure studies (above), each material’s H2 storage capacity was subsequently analysed at further elevated pressures. The ambient temperature H2 storage capacities of both pure monolithHKUST-1 and each of the Pd-doped composites were tested up to 100 bar (Figure 20). The excess values were converted to absolute and studied in terms of both their gravimetric (wt%) and volumetric (g L–1) capacity. As a result of their near identical rb (Chapter III, Section 3), the trends in gas storage capacity amongst the different monolithic materials were the same for both gravimetric and volumetric capacity. The obtained isotherms were not dissimilar, with close overlap between each of the monolithic materials. However, closer inspection revealed subtle differences (Table 3). At the lowest pressures studied, both the volumetric and gravimetric uptake capacities varied with Pd doping level – H2 uptake (5 bar) 10% Pd@monolithHKUST-1 > 5% Pd@monolithHKUST-1 > monolithHKUST-1. This resembles the sub-1 bar isotherms, as discussed earlier, where the NP contribution towards the total H2 storage was observed to outweigh that of the host MOF. While this trend appeared to initially continue for these higher pressure isotherms, by the time 40 bar was reached, the isotherms were observed to cross paths. Following this, the trend in gas uptake capacity was reversed so that H2 uptake (100 bar) 10% Pd@monolithHKUST-1 < 5% Pd@monolithHKUST-1 < monolithHKUST-1. 240 Figure 20| High pressure H2 adsorption isotherms. a, Gravimetric (wt%) and b, volumetric (g L–1) H2 adsorption isotherms (0 – 100 bar, 298 K). Data correspond to monolithHKUST-1 (green circles), 5% Pd@monolithHKUST-1 (blue squares), 10% Pd@monolithHKUST-1 (purple diamonds) and a standard pressurised storage tank (black line). These observations are highly revealing in terms of determining the H2 storage capabilities of these monolithMOF and NP@monolithMOF composites. At low pressures, the Pd NPs rapidly chemisorb H2 while the MOF concurrently physisorbs it to a lesser extent. Hence, NP doping of the MOF nominally increases a composite’s H2 storage capacity at low pressure. While electronic interaction of the composite’s two components appears to ultimately enhance the PdH (b-hydride) formation capacity to some extent, the material is nevertheless saturated above a critical pressure. Hence, at further elevated pressures the host-MOF overtakes the NPs to become the composite’s dominant storage medium. From this perspective, the presence of fully hydrogenated/saturated NPs subsequently serves to diminish the MOFs storage capacity. As expected, a slight negative trend in gravimetric SBET with NP doping was previously observed (Chapter III, Section 3.0) whereby SBET monolithHKUST-1 (1557 m2 g–1) > 5% Pd@ monolithHKUST-1 (1468 m2 g–1) > 10% monolithHKUST-1 (1408 m2 g–1). Yet, at higher pressures the MOFs surface area and porosity are key in determining its gas storage capacity; this ultimately determines the availability of accessible surface area capable of H2 physisorption. For example, doped monolith 10% Pd@monolithHKUST-1 showed a 10% reduction in gravimetric porosity (Vtot = 0.58 cm3 g–1) relative to monolithHKUST-1 (Vtot = 0.64 cm3 g–1) which Pressure (bar)Pressure (bar) H 2 up ta ke (w t % ) H 2 up ta ke (g L -1 ) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0 20 40 60 80 100 a b 0 2 4 6 8 10 12 0 20 40 60 80 100 241 correlates reasonably well with the observation of its 12% reduction in H2 uptake (11.0 g L–1, 100 bar, 298 K) relative to that of monolithHKUST-1 (12.5 g L–1, 100 bar, 298 K). The above relationship between monolith doping level and total H2 storage capacity is further highlighted by comparison of each materials working capacity, i.e. the accessible volume of gas under practical conditions. This is calculated by subtracting the volume of gas stored at the minimum release pressure (typically 5 bar) from the volume stored at the maximum storage pressure (in this case 100 bar). The heightened uptake capacities of the doped materials at low pressure combined with their diminished storage capacities at higher pressures inherently reduces working capacity. This is reflected by Table 3, in which the 5 – 100 bar working capacity of 10% Pd@monolithHKUST-1 is 18% lower than that of the pure monolithic MOF. Table 3| Comparison of total volumetric H2 storage capacity (g L–1) in monolithHKUST-1, 5% Pd@monolithHKUST-1 and 10% Pd@monolithHKUST-1 to that of a pressurised storage tank at 5, 40 and 100 bar H2 pressure (298 K) as well as that of the corresponding working capacities (5 – 100 bar). Volumetric H2 capacity (g L–1) 5 bar 40 bar 100 bar Working Capacity (5 – 100 bar) monolithHKUST-1 0.9 5.2 12.5 11.6 5% Pd@monolithHKUST-1 1.2 5.0 11.6 10.4 10% Pd@monolithHKUST-1 1.6 5.1 11.0 9.4 Pressurised storage tank 0.4 3.2 7.7 7.3 It is clear that while NP@monolithMOF doping may be used to obtain practical, densified materials with enhanced H2 storage at low pressures i.e. 1 bar, Pd doping of monolithHKUST-1 still cannot approach the U.S. DOE storage targets at 100 bar. Moreover, further elevated pressures find the NP doping to be detrimental to H2 storage through the loss of MOF surface area. The pure monolith of HKUST-1 is by far the best H2 storage material tested in this study 242 and offers an excellent 59% enhancement in volumetric capacity over that of an equally pressurised storage tank. As the best performing material in this study, monolithHKUST-1 was benchmarked against the record setting, ambient temperature, H2 storage MOF, Ni2(m-dobdc).34 At 298 K, the gravimetric uptake capacity (Figure 21a) of monolithHKUST-1 slightly exceeds the benchmark (i.e. it is 17% higher at 100 bar) while the volumetric capacities (Figure 21b) of the two are near identical. With a volumetric working capacity of 10.9 g L–1 (5 – 100 bar, 298 K), the accessible quantity of H2 stored in Ni2(m-dobdc) is 6% lower than that of monolithHKUST-1 (11.6 g L–1, 5 – 100 bar, 298 K). The high adsorption capacities of both Ni2(m-dobdc) and monolithHKUST-1 may be attributed to the high charge density provided by their coordinatively unsaturated cationic metal sites (Ni2+ and Cu2+ respectively), both of which are capable of polarising H2 to some extent.34 This benchmark result is especially significant considering that monolithHKUST-1 is densified to a robust and practical pellet with real volumetric capacities calculated from its experimentally recorded rb. Contrarily, the benchmark results of Ni2(m- dobdc) were performed on the non-industrially viable powdered material, which Long reported to have an experimental rb of only 0.37 g cm–3. Here, Long instead calculated the materials volumetric H2 capacity using its theoretical crystal density which, although not reported explicitly, was back-calculated from the reported gravimetric and volumetric results to be 1.22 g cm–3. As a result of the rb of non-densified powderNi2(m-dobdc) being 70% lower than its theoretical crystal density, a practical volumetric capacity ca. 70% lower than the reported value may be expected. Under traditional densification of this loosely packed powder to a pelletised material, which nominally increases rb, compressive loss of SBET is typically expected to result in a ca. 25 – 50% loss in uptake capacity (Figure 21b). Hence, monolithHKUST-1 is found to be significantly superior, in terms of practical H2 storage capacity than the previous benchmark material (both powder and theoretically pelletised). Stability of the adsorbent is a key consideration in the practical design of new adsorbent materials. The thermal stability of HKUST-1 is reasonably high (Chapter III, Section 3.0) and comparable to that of the benchmark Ni2(m-dobdc),47 both of which display gradual thermal decomposition between 300 – 400 °C. However, of note is the low chemical stability of hygroscopic HKUST-1, which rapidly becomes amorphous under exposure to ambient moisture (the chemical origin of which is discussed in Chapter I). While the water stability of Ni2(m-dobdc) has not been extensively explored, chemically analogous Ni2(dobdc) (which 243 contains the para-disubstituted, 2,5-dioxido-1,4-benzenedicarboxylate, rather than meta- disubstituted, 4,6-dioxido-1,3-benzenedicarboxylate, linker) has been reported. This analogous MOF displays significantly greater chemical resistance to decomposition under exposure to water vapour than that of HKUST-1,48 and furthermore its resistance even to high temperature steam exposure has been demonstrated.49 While high volumetric H2 storage capacity of monolithHKUST-1 is a significant result, its chemical instability relative to other MOFs remains a major concern in terms of practical implementation and long term usage. Figure 21| Benchmark H2 adsorption isotherms. a, Gravimetric (wt%) H2 uptake capacity in monolithHKUST-1 (green circles) compared to the benchmark powdered material Ni2(m-dobdc) (yellow triangles, digitised from Reference 34). b, Volumetric (g L–1) uptake capacity of the two materials, calculated from the experimental rb of monolithHKUST-1 and the theoretical crystal density of Ni2(m-dobdc). The expected loss in volumetric uptake capacity (25 – 50%) by pelletisation under pressure of powdered Ni2(m-dobdc) is indicated (yellow, inset in b). HKUST-1 is an extensively researched MOF with the H2 storage capacity of its powdered morphology having been comprehensively studied experimentally. It has yielded wide ranging results that both support and contradict the results presented herein.50,51 It is well known that there is a persistent issue of irreproducibility amongst reported H2 storage measurements; for example, Broom and Hirscher organised a ‘round-robin’ H2 adsorption study of the same doped MgH2 sample at different laboratories across Europe, the U.S., China and Japan yielding vastly 0 1 2 3 4 5 6 7 8 9 10 11 12 0 20 40 60 80 100 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 0 20 40 60 80 100 Pressure (bar) Pressure (bar) H 2 up ta ke (w t% ) H 2 up ta ke (g L- 1 ) -25 % -50 % pelletised Ni2(m-dobdc) a b 244 differing results.52 Across twelve different research groups the H2 storage capacity of the same MgH2 material was recorded to range between 5 – 6.5 wt% (6 bar, 593 K) i.e. ± 15%. In the particular case of MOFs, Hirscher identified the materials handling and activation (i.e. evacuation of the porosity) to be key considerations for reproducible testing. Furthermore, the low density of nanoporous materials was highlighted as a key source of error in the measurement of high-pressure gas adsorption isotherms. However, this source of error was discussed assuming the MOFs to have traditional, low-density powdered morphology. In the current case, the densified nature of the monolithic materials may serve to reduce such error to some extent. In terms of methodology, the standard volumetric measurements which were used in the current case are further susceptible to error i.e. internal volume calibration, gas leaks, stability of the temperature control, accuracy of the pressure measurement and error accumulation of successive measurements.52 While these errors may be minimised, they remain systemic to the technique and should not be overlooked. Hence, as per good practice recommendations, the high pressure H2 adsorption isotherm of seemingly benchmark monolithHKUST-1, originally tested at the U.S. Sandia National Laboratory, California, was re-tested by the U.S. DOE National Renewable Energy Laboratory, Colorado. Comparison of the isotherms obtained for these materials, prepared under the same experimental conditions, demonstrates essentially very similar behaviour, with the gravimetric capacity experimentally obtained by the U.S. DOE being 14% lower than the results from Sandia (Figure 22). Despite this error margin, the results remain comparable to those of the benchmark, Ni2(m-dobdc). Considering the wide range of literature values reported for both powderHKUST-1 and other round-robin H2 adsorption studies i.e. ± 15%, as discussed above for the case of MgH2,52 this ca. 14% difference in uptake capacity appears to be within a reasonable error margin and may be attributed to a number of experimental factors. In particular, handling and preparation of highly air-sensitive HKUST-1 is likely to be a source of significant error in the testing of this material. Additionally, the U.S. National Renewable Energy Laboratory collected the adsorption isotherm at 303 K whereas the Sandia National Laboratory, who reported a higher storage capacity, performed their studies at the lower temperature of 298 K. This is an obvious difference in the experimental data collection, and correlates well with the known enhancement in H2 storage capacity of MOFs at reduced temperatures. Finally, the H2 storage capacity of monolithHKUST-1 was further supported by comparison of the experimental 245 isotherms with the computationally simulated isotherm modelled on the MOFs crystal structure (Figure 22). Again, this shows an excellent match to both of the experimentally obtained results and supports the validity of this benchmark result. Figure 22| High pressure H2 adsorption comparison. Comparison between high pressure gravimetric H2 storage capacity for monolithHKUST-1 collected at Sandia National Laboratory California (filled green circles, 298 K) and the U.S. DOE National Renewable Energy Laboratory, Colorado (hollow green circles, 303 K). The theoretical isotherm of defect-free HKUST-1 obtained by GCMC simulation (black circles, 298 K) and experimentally reported Ni2(m-dobdc) (yellow triangles, 298 K, digitised from Reference 34) are also provided. 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 Pressure (bar) H 2 up ta ke (w t% ) 246 2.5| Conclusions In Chapter II, Section 3.0 the capability of monolithicHKUST-1 to host Pd NPs was explored with the aim of enhancing the MOFs H2 storage capacity. Based on the success of the synthesis, which yielded 5% and 10% Pd doped (by weight) monoliths with high crystallinity, porosity and rb, the H2 storage capacity of each material (monolithHKUST-1, 5% Pd@monolithHKUST-1 and 10% Pd@monolithHKUST-1) was studied. For adsorption isotherms up to 1 bar H2, a strong correlation between the Pd doping level and the H2 uptake capacity (both gravimetric and volumetric) was observed. Furthermore, the recorded uptake capacities were proportional to those of 24% Pd@powderHKUST-1 reported by Kitagawa et al.35 Moreover, the collection of 0 – 1 bar H2 isotherms at variable temperatures (273 – 298 K) demonstrated the enhancement in H2 uptake capacity which occurs at reduced temperatures. This preliminary study suggests that these doped monolithic materials may be suitable for temperature swing conditions though further analysis is required to determine the magnitude of each material’s H2 capacity over a wider temperature range. The low-pressure isotherms were also studied as a function of H/Pd and H2/Cu3(btc)2. Following the assumptions proposed by Kitagawa, namely that the contribution of the MOF to the total uptake capacity is negligible relative to that of the dopant,35 comparable conclusions were drawn from the present results; H2 uptake capacity of Pd NPs immobilised in the MOF appeared to be substantially enhanced relative to that of the pure NPs. However, contrarily to powderHKUST-1, monolithHKUST-1 was found to exhibit a non-negligible H2 uptake at 1 bar. Re- evaluation of the data making the alternative assumption that the MOF concurrently adsorbs H2 into its microporosity at sub-1 bar pressure, suggested a significant reduction in the calculated uptake capacity (H/Pd) of the dopant NPs within the composite. In the case of 10% Pd@monolithHKUST-1, instead of the 68% enhancement in H/Pd (relative to the pure Pd NPs) calculated using the previous assumption, only a 34% enhancement was now calculated. While this still supports the general conclusions drawn by Kitagawa, being that the Pd uptake capacity is enhanced by its electronic interaction with the host MOF, data suggest that the enhancement is to a lesser extent than first supposed and that the MOF itself contributes significantly to the H2 uptake of these composites. High pressure isotherms, up to 100 bar, lend further support to the above conclusion. Indeed, the gravimetric and volumetric H2 capacity of monolithHKUST-1 was found to be far from 247 negligible, rivalling the benchmark results of powdered MOF, Ni2(m-dobdc).34 This result was verified through a combined approach of repeat testing at different research facilities as well as computational simulation. Contradicting the above trends in the low-pressure isotherms, the presence of Pd in the monolith was found to be detrimental to the total gas uptake capacity at pressures in excess of 40 bar. This is suggested to be a consequence of SBET loss in the doped materials which serves to hinder the MOFs H2 storage capacity after saturation of the Pd. Correspondingly, the H2 uptake capacity of pure monolithHKUST-1 was found to be the highest amongst the monolithic materials at 100 bar. Furthermore, the fact that these benchmark results were obtained in a densified material, already pelletised for industrial application is very significant. Previous benchmark results for Ni2(m-dobdc) were recorded for the low density, powdered version of the MOF and can be expected to decrease by 25 – 50% under normal pelletisation procedures. Although the 5 – 100 bar H2 working capacity of monolithHKUST-1 at ambient temperature appears to be amongst the highest reported for any MOF, at 11.6 g L–1 it remains well below the U.S. DOE’s ultimate storage target of 50 g L–1. High density H2 storage is a substantial challenge and considering the exclusively weak Van der Waals interactions that H2 assumes with adsorbents, physisorbative MOFs may ultimately prove incapable of reaching such ambitious targets. For example, Siegel et al. computationally screened ca. 500, 000 MOFs for H2 storage capacity concluding that even at 77 K, a ‘ceiling’ in volumetric working capacity exists around 40 g L–1.53 The best performing MOF in Siegel’s study, NU-100, showed a volumetric working capacity of only 35.5 g L–1 (5 – 100 bar, 77 K). In spite of the aforementioned issues pertaining to high pressure H2 storage by MOFs, the current results do represent a significant 59% improvement in volumetric H2 storage capacity relative to that of a traditional pressurised storage tank over the same pressure and temperature range. Hence, though not reaching the ultimate storage targets, these data may instead offer a low-cost and practical means of improving the current technology. 248 3.0| References 1. Wu, H., Yildirim, T. & Zhou, W. Exceptional Mechanical Stability of Highly Porous Zirconium Metal-Organic Framework UiO-66 and Its Important Implications. J. Phys. Chem. Lett. 4, 925–930 (2013). 2. Saha, D., Grappe, H. A., Chakraborty, A. & Orkoulas, G. Postextraction Separation, On- Board Storage, and Catalytic Conversion of Methane in Natural Gas: A Review. Chem. Rev. 116, 11436–11499 (2016). 3. Deniz, C. & Zincir, B. Environmental and Economical Assessment of Alternative Marine Fuels. J. Clean. Prod. 113, 438–449 (2016). 4. Tian, T. et al. A Sol–Gel Monolithic Metal–Organic Framework with Enhanced Methane Uptake. Nat. Mater. 17, 174–179 (2018). 5. Álvarez, J. R. et al. Structure Stability of HKUST-1 Towards Water and Ethanol and Their Effect on its CO2 Capture Properties. Dalton Trans. 46, 9192–9200 (2017). 6. Kumar, K. V., Preuss, K., Titirici, M. M. & Rodríguez-Reinoso, F. Nanoporous Materials for the Onboard Storage of Natural Gas. Chem. Rev. 117, 1796–1825 (2017). 7. Cmarik, G. E., Kim, M., Cohen, S. M. & Walton, K. S. Tuning the Adsorption Properties of UiO-66 via Ligand Functionalization. Langmuir 28, 15606–15613 (2012). 8. Biswas, S. & Van Der Voort, P. A General Strategy For The Synthesis Of Functionalised UiO-66 Frameworks: Characterisation, Stability and CO2 Adsorption Properties. Eur. J. Inorg. Chem. 12, 2154–2160 (2013). 9. Sreedhar, I., Nahar, T., Venugopal, A. & Srinivas, B. Carbon Capture by Absorption – Path Covered and Ahead. Renew. Sustain. Energy Rev. 76, 1080–1107 (2017). 10. Nandasiri, M. I., Jambovane, S. R., McGrail, B. P., Schaef, H. T. & Nune, S. K. Adsorption, Separation, and Catalytic Properties of Densified Metal-Organic Frameworks. Coord. Chem. Rev. 311, 38–52 (2016). 11. Mason, J. A., Veenstra, M. & Long, J. R. Evaluating Metal–Organic Frameworks for Natural Gas Storage. Chem. Sci. 5, 32–51 (2014). 12. Wu, H. et al. Unusual and Highly Tunable Missing-Linker Defects in Zirconium Metal- Organic Framework UiO-66 and Their Important Effects on Gas Adsorption. J. Am. Chem. Soc. 135, 10525–10532 (2013). 249 13. Cui, X., Bustin, R. M. & Dipple, G. Selective Transport of CO2, CH4, and N2 in Coals: Insights From Modeling of Experimental Gas Adsorption Data. Fuel 83, 293–303 (2004). 14. Lemmon, E. W., McLinden, M. O. & Friend, D. G. ‘Thermophysical Properties of Fluid Systems’. in NIST Chemistry WebBook, NIST Standard Reference Database Number 69 (National Institute of Standards and Technology, Gaithersburg MD, 20899, 2018). 15. Peng, Y. et al. Methane Storage in Metal − Organic Frameworks: Current Records, Surprise Findings, and Challenges. J. Am. Chem. Soc. 135, 11887–11894 (2013). 16. Jiang, J., Furukawa, H., Zhang, Y. B. & Yaghi, O. M. High Methane Storage Working Capacity in Metal-Organic Frameworks with Acrylate Links. J. Am. Chem. Soc. 138, 10244–10251 (2016). 17. Li, B. et al. A Porous Metal − Organic Framework with Dynamic Pyrimidine Groups Exhibiting Record High Methane Storage Working Capacity. J. Am. Chem. Soc 136, 6207–6210 (2014). 18. Toyota Mirai Brochure, 2019. (2019). 19. Millward, A. R. & Yaghi, O. M. Metal-Organic Frameworks with Exceptionally High Capacity for Storage of Carbon Dioxide at Room Temperature. J. Am. Chem. Soc. 127, 17998–17999 (2005). 20. Jusoh, N., Yeong, Y. F., Chew, T. L., Lau, K. K. & Shariff, A. M. Current Development and Challenges of Mixed Matrix Membranes for CO2/CH4 Separation. Sep. Purif. Rev. 45, 321–344 (2016). 21. Zamidi, M., Navarro, M., Lhotka, M., Zornoza, B. & Téllez, C. Enhancement of CO2/CH4 Separation Performances of 6FDA-Based Co-Polyimides Mixed Matrix Membranes Embedded with UiO-66 Nanoparticles. Sep. Purif. Technol. 192, 465–474 (2018). 22. Mozafari, M., Abedini, R. & Rahimpour, A. Zr-MOFs-Incorporated Thin Film Nanocomposite Pebax 1657 Membranes Dip-Coated on Polymethylpentyne Layer for Efficient Separation of CO2/CH4. J. Mater. Chem. A 6, 12380–12392 (2018). 23. Guo, A., Ban, Y., Yang, K. & Yang, W. Metal-Organic Framework-Based Mixed Matrix Membranes: Synergetic Effect of Adsorption and Diffusion for CO2/CH4 Separation. J. Memb. Sci. 562, 76–84 (2018). 250 24. Myers, A. L. & Prausnitz, J. M. Thermodynamics of Mixed-Gas Adsorption. AIChE J. 11, 121–127 (1965). 25. Walton, K. S. & Sholl, D. S. Predicting Multicomponent Adsorption: 50 Years of the Ideal Adsorbed Solution Theory. AIChE J. 61, 2757–2762 (2015). 26. Chen, K. et al. Enhanced CO2/CH4 Separation Performance of Mixed-Matrix Membranes Through Dispersion of Sorption-Selective MOF Nanocrystals. J. Membr. Sci. 563, 360–370 (2018). 27. Connolly, B. M. et al. Tuning Porosity in Macroscopic Monolithic Metal-Organic Frameworks for Exceptional Natural Gas Storage. Nat. Commun. 10, 2345 (2019). 28. Chen, S. et al. Cleaving Carboxyls: Understanding Thermally Triggered Hierarchical Pores in the Metal-Organic Framework MIL-121. J. Am. Chem. Soc. 141, 14257–14271 (2019). 29. Niaz, S., Manzoor, T. & Pandith, A. H. Hydrogen storage: Materials, Methods and Perspectives. Renew. Sustain. Energy Rev. 50, 457–469 (2015). 30. Hanley, E. S., Deane, J. P. & Gallachóir, B. P. Ó. The Role of Hydrogen in Low Carbon Energy Futures–A Review of Existing Perspectives. Renew. Sustain. Energy Rev. 82, 3027–3045 (2018). 31. Zhang, F., Zhao, P., Niu, M. & Maddy, J. The Survey of Key Technologies in Hydrogen Energy Storage. Int. J. Hydrogen Energy 41, 14535–14552 (2016). 32. Abe, J. O., Popoola, A. P. I., Ajenifuja, E. & Popoola, O. M. Hydrogen Energy, Economy and Storage: Review and Recommendation. Int. J. Hydrogen Energy 44, 15072–15086 (2019). 33. García-Holley, P. et al. Benchmark Study of Hydrogen Storage in Metal-Organic Frameworks under Temperature and Pressure Swing Conditions. ACS Energy Lett. 3, 748–754 (2018). 34. Kapelewski, M. T. et al. Record High Hydrogen Storage Capacity in the Metal-Organic Framework Ni2(m-dobdc) at Near-Ambient Temperatures. Chem. Mater. 30, 8179–8189 (2018). 35. Li, G. et al. Hydrogen Storage in Pd Nanocrystals Covered with a Metal–Organic Framework. Nat. Mater. 13, 802–806 (2014). 36. Prins, R. Hydrogen Spillover. Facts and Fiction. Chem. Rev. 112, 2714–2738 (2012). 251 37. Syrenova, S. et al. Hydride Formation Thermodynamics and Hysteresis in Individual Pd Nanocrystals With Different Size and Shape. Nat. Mater. 14, 1236–1244 (2015). 38. Griessen, R., Strohfeldt, N. & Giessen, H. Thermodynamics of the Hybrid Interaction of Hydrogen with Palladium Nanoparticles. Nat. Mater. 15, 311–317 (2016). 39. Bugaev, A. L. et al. Core-Shell Structure of Palladium Hydride Nanoparticles Revealed by Combined X-ray Absorption Spectroscopy and X-ray Diffraction. J. Phys. Chem. C 121, 18202–18213 (2017). 40. Yamauchi, M., Ikeda, R., Kitagawa, H. & Takata, M. Nanosize Effects on Hydrogen Storage in Palladium. J. Phys. Chem. C 112, 3294–3299 (2008). 41. Chen, Y. et al. Electronic Origin of Hydrogen Storage in MOF-Covered Palladium Nanocubes Investigated by Synchrotron X-rays. Commun. Chem. 1, 61 (2018). 42. Bardhan, R. et al. Uncovering the Intrinsic Size Dependence of Hydriding Phase Transformations in Nanocrystals. Nat. Mater. 12, 905–912 (2013). 43. Li, G. et al. Shape-Dependent Hydrogen-Storage Properties in Pd nanocrystals: Which Does Hydrogen Prefer, Octahedron (111) or Cube (100)? J. Am. Chem. Soc. 136, 10222– 10225 (2014). 44. Bugaev, A. L. et al. Time-Resolved Operando Studies of Carbon Supported Pd Nanoparticles Under Hydrogenation Reactions by X-ray Diffraction and Absorption. Faraday Discuss. 208, 187–205 (2018). 45. Prestipino, C. et al. Local Structure of Framework Cu(II) in HKUST-1 Metal-Organic Framework: Spectroscopic Characterization Upon Activation and Interaction with Adsorbates. Chem. Mater. 18, 1337–1346 (2006). 46. Ding, F. & Yakobson, B. I. Challenges in Hydrogen Adsorptions: From Physisorption to Chemisorption. Front. Phys. 6, 142–150 (2011). 47. Kapelewski, M. T. et al. M2(m-dobdc) (M = Mg, Mn, Fe, Co, Ni) Metal-Organic Frameworks Exhibiting Increased Charge Density and Enhanced H2 Binding at the Open Metal Sites. J. Am. Chem. Soc. 136, 12119–12129 (2014). 48. Liu, J. et al. CO2/H2O Adsorption Equilibrium and Rates on Metal-Organic Frameworks: HKUST-1 and Ni/DOBDC. Langmuir 26, 14301–14307 (2010). 49. Liu, J. et al. Stability Effects on CO2 Adsorption for the DOBDC Series of Metal- Organic Frameworks. Langmuir 27, 11451–11456 (2011). 252 50. Lin, K. S., Adhikari, A. K., Ku, C. N., Chiang, C. L. & Kuo, H. Synthesis and Characterization of Porous HKUST-1 Metal Organic Frameworks for Hydrogen Storage. Int. J. Hydrogen Energy 37, 13865–13871 (2012). 51. Moellmer, J., Moeller, A., Dreisbach, F., Glaeser, R. & Staudt, R. High Pressure Adsorption of Hydrogen, Nitrogen, Carbon Dioxide and Methane on the Metal-Organic Framework HKUST-1. Microporous Mesoporous Mater. 138, 140–148 (2011). 52. Broom, D. P. & Hirscher, M. Irreproducibility in Hydrogen Storage Material Research. Energy Environ. Sci. 9, 3368–3380 (2016). 53. Ahmed, A. et al. Exceptional Hydrogen Storage Achieved by Screening Nearly Half a Million Metal-Organic Frameworks. Nat. Commun. 10, 1568 (2019). Chapter V Outlook 254 255 1.0| Conclusions The ultimate aim of Chapter II was to develop a synthetic procedure for the preparation of macroscopic, monolithic Zr-MOFs with physical properties comparable to those of their theoretical crystal structure. A preparation procedure for monolithUiO-66 was developed without using chemical binders or applied pressures. This is the first report of a monolithic Zr-MOF with porosity, density, thermal and mechanical stability comparable to the single crystal material.1 This finding has significant implications for the synthesis of industrially viable MOF materials with potential towards a number of gas storage applications. The developed procedure first required the synthesis of ca. 10 nm crystalline primary MOF particles as a gel.2 The experimental drying conditions of these primary particles were extensively explored, and optimised drying procedures were developed. It was observed that extended drying (i.e. at room temperature) via the evaporation of the reaction solvent enabled the synthesis of monoliths whose physical properties closely resembled those of the single crystal material. Using FLIM, chemically distinct micron-scale aggregates were observed in the monoliths. This strongly suggests that the monoliths may be obtained via an extended epitaxial polymerisation reaction between neighbouring primary MOF NPs in the drying gel. This explains why use of a solvent which enables the reaction permits dense monolith formation. Furthermore, through this insight into the monolith formation mechanism, the capacity to tune the material’s physical properties with altered drying conditions can be accounted for. For the first time, it was demonstrated that tuneable volumes of non-crystalline mesoporosity can be built into macroscopic MOF structures. Since the volume of this porosity was observed to change by drying the same primary particles with altered drying temperatures, washing solvents and centrifugation times, it was postulated that this mesoporosity arose from alterations in MOF primary NP packing rather than crystalline defects within the individual NPs. The ability to tune pore size has significant implications for gas storage. For example, pore-width is well known to affect the onset pressure at which gas-condensation occurs in an adsorbent.3 Furthermore, the inclusion of this tuneable mesoporosity allows alteration of the monoliths density, a key factor in determining the volumetric gas storage capacity. The generality of the synthetic procedure was demonstrated by extending the synthesis for monolithZr-MOF to a range of other structures. Firstly, the simple monofunctionalised analogue UiO-66-NH2 was synthesised by a facile ligand swap in the primary particle reaction. 256 Analogously to monolithUiO-66, the obtained monolithUiO-66-NH2 displayed bulk physical properties comparable to those of the theoretical MOF and the novel capacity to tune mesoporosity into the macrostructure was again demonstrated. The developed synthetic procedure for monolithZr-MOF was further extended to UiO-66-ndc, which contains a larger and more complex naphthalene functionality. In this case, simple equimolar ligand exchange in the primary particle reaction was unsuccessful at synthesising either a MOF gel or a monolithic material when dried. Through TEM, the primary particle size was determined to have increased five-fold, a consequence of altered MOF primary particle nucleation by the different chemical properties of the naphthalene organic linker. Yet through a modulation study, it was determined that the primary particle size could be optimised via a 63% and 73% reduction in AA and HCl concentration, respectively, in the reaction mixture. This permitted the formation of robust and crystalline monolithicUiO-66-ndc akin to the previous monoliths of UiO-66 and UiO-66-NH2. This optically transparent material has highly potential towards the fluorescence sensing applications for which this MOF is well known. Although preliminary testing has been started by co-workers at Sandia National Laboratories, it falls outside the time frame of this project. Additionally, FLIM analysis of this novel material may further aid in characterising its fluorescent properties, though this too falls outside this project’s timeframe. The above results serve to demonstrate that altered chemical properties of the organic linker exert a significant influence on MOF crystallisation. Yet, through reaction tuning the developed monolith procedure may be successfully applied to even UiO-66 derivatives with substantially differing functionality from that of the parent MOF. The synthesis of a non-isostructural Zr-MOF, NU-1000, as a monolith was next explored. This material has a large tetratopic linker, tbapyH4, which imbues much wider porosity and higher surface area than UiO-66-derived structures. It was found that while the MOF could be synthesised as a gelatinous suspension of sub-10 nm primary particles and subsequently dried to yield a robust and semi-optically transparent monolith, the crystallinity and porosity were significantly lower than those of the theoretical crystal MOF structure. This highlights a key weakness of the developed monolithic Zr-MOF procedure, which fundamentally requires very small MOF NPs. MOFs with long organic linkers inherently cannot be synthesised as sub-10 nm particles while maintaining long range crystalline order. This is a consequence of their increased unit cell dimensions. This prevents extensive unit cell repeats within corresponding NPs, which are thus rendered amorphous. 257 Considering the above results as a whole it is apparent that the ultimate aim, being the development of a synthetic procedure for novel monolithZr-MOF with optimised chemical and physical properties, has been achieved. Furthermore, the capacity to extend this procedure to a wider range of Zr-MOFs has been successfully demonstrated and the procedure’s limitations towards more complex, wide pore structures have also been explored. The developed materials possess numerous industrially valued physical properties i.e. porosity, density, crystallinity and optical transparency that lend them towards several promising research avenues, including high-density gas storage,4 selective gas adsorption5 and fluorescent sensing.6 In Chapter III, the proof of concept work by Mehta et al. regarding the immobilisation of photocatalytic SnO2 NPs in monolithZIF-87 was used as a foundation from which to explore the synthesis of novel NP@monolithMOF. Firstly, the effects of including organic and inorganic surface capping agents in the SnO2 nanocube synthesis was studied. While NaCl facilitated the asymmetric growth of nanorods, no other capping agent was found to significantly influence particle size or morphology. However, ICP-OES was used to demonstrate the dependence of monolith doping level on the surface functionality of the dopant. In all cases, the use of an organic capping agent during NP synthesis increased the functionalised NPs capacity to dope the MOF. This may be attributed to the particles reduced agglomeration as a result of surface capping and hence an increased capacity to form a stable colloid, around which the MOF is capable of assembling. This was further supported by the 11-fold increase in doping achieved by using undried SnO2 NPs, which are also expected to exhibit reduced agglomeration relative to their dried counterparts. Yet, despite these significant enhancements in MOF doping, proportional improvements in photocatalytic capacity towards the degradation of the toxic dye MB were not observed. Although statistically significant improvements in photocatalytic dye degradation were achieved by several of the modified composites, the results did not correlate with changes in doping level. This was postulated to arise from the NPs immobilisation in the microporous MOF – while water molecules (the catalysts substrate) may permeate the MOF to reach the active NPs surface, the rate of catalytic conversion to active radicals appeared to be diffusion limited. The result described above was further explored through the synthesis of a wider range of catalytic NP@monolithZIF-8 for alternative potential applications. The generality of the monolithic MOF to host compositionally and morphologically diverse catalysts (mono-, bi- and trimetallic NPs) was explored by the synthesis of Au@monolithZIF-8, PdO/TiO2@monolithZIF-8 258 and ((Au@PdO)/TiO2)@monolithZIF-8 with the aim of increasing the industrial viability and economic potential of these expensive noble metal catalysts. In each case, NPs with different sizes, compositions and morphologies were successfully immobilised in monolithic ZIF-8 without compromising the host’s monolithicity or crystallinity. Furthermore, in each case the NPs demonstrated a high doping level and reasonable dispersion throughout the host. This is a significant result, suggesting the monolith’s capacity to act as an industrially viable means of achieving economic and environmental catalysis for a wide range of reactions. Yet, each of the three composite materials (Au@monolithZIF-8, PdO/TiO2@monolithZIF-8 and Au@Pd/TiO2@monolithZIF-8) displayed an inability to achieve any measurable catalytic activity towards CO oxidation, 2-propen-1-ol hydrogenation and CH4 oxidation, respectively, despite all of these reagents theoretically being small enough to penetrate the hosts porosity. This further supports the previously discussed limitations of the composite systems; the rate of substrate diffusion through the MOFs micropores (< 2 nm) appears to be the limiting factor in the catalytic reaction. In the current case, the synthesised NPs appeared to have been completely passivated by their immobilisation. These data demonstrate that although a wide range of composites may be synthesised, the practical functionality of monolithZIF-8-derived composites is suited for very specific applications, the identification of which requires careful consideration. Rather than focusing on catalytic reactions, which are highly rate-dependant, the synthesis of new NP@monolithMOF composite materials for enhanced gas storage was instead explored. The capability of monolithic HKUST-1 towards metallic Pd NP immobilisation was studied with the aim of increasing the H2 storage capacity of the host. It was determined that, comparably to monolithZIF-8,7 monolithHKUST-18 was also capable of hosting NPs with a high doping level (5% and 10% by weight) while retaining its industrially valuable physical properties (rb, microporosity, stability and a monolithic macrostructure). This monolithic composite material offers significant potential for the pressurised storage of H2 fuel where the kinetics of gas diffusion are less significant than is the case for the catalytic reactions previously discussed. Finally, in Chapter IV, the potential of the monolithic materials synthesised in Chapter II (monolithUiO-66_A – D) and Chapter III (monolithHKUST-1 and Pd@monolithHKUST-1) towards several industrial gas-phase applications was explored through a combined experimental and computational approach. Firstly, the CH4 storage capacity of the mixed micro-/mesoporous monolithic UiO-66 materials was tested. While each of the monoliths (UiO-66_A – D) were 259 found to display near comparable gravimetric (g g–1) CH4 storage capacity, a consequence of their comparable surface areas, their volumetric capacities (cm3 (STP) cm–3) were found to vary as a function of their significantly differing rb. In all cases, the volumetric capacities of monolithUiO-66 were recorded to exceed that of the current technology (i.e. a standard pressurised tank). The densest material, monolithUiO-66_D, displayed the highest volumetric capacity and yielded a significant ca. 200% improvement over the current industry standard technology. At 296 cm3 (STP) cm–3 (100 bar, 298 K) the volumetric CH4 capacity of UiO-66_D exceeds the theoretical maximum for the defect-free MOF UiO-66 (200 cm3 (STP) cm–3, 100 bar, 298 K) by 50%. This result was attributed to the unexpected Type II adsorption isotherms observed in each of the monolithUiO-66 samples. The traditionally microporous crystal structure of UiO- 66 normally results in a Type I isotherm, a consequence of rapid onset CH4 uptake at low pressures followed by an adsorption plateau after micropore saturation. However, the presence of non-crystalline mesoporosity in each of the monoliths was postulated to induce their altered isotherm shapes via further CH4 condensation in the wider mesoporous cavities at higher pressures. This hypothesis was supported through novel computational simulation of CH4 adsorption in mixed micro-/mesoporous UiO-66. It was determined that for maximum CH4 storage capacity, an optimal PSD exists. Excess mesoporosity reduces the materials’ density, diminishing volumetric capacity. However, purely microporous MOFs rapidly saturate at low pressure and cannot store gas at further elevated pressures. Hence, an optimised micro- /mesopore ratio can be used to maximise CH4 storage capacity, as was the case for UiO-66_D. As a result of the material’s maximised CH4 uptake capacity, UiO-66_D was found to rival the total CH4 uptake capacity of benchmark densified adsorbent monolithHKUST-1.8 Furthermore, due to the weaker interaction of CH4 with UiO-66 at low pressure, relative to that of HKUST- 1, the working capacity of monolithUiO-66_D (261 cm3 (STP) cm–3, 5 – 100 bar, 298 K) was found to be 10% higher than that of monolithHKUST-1 over the same pressure range. This is the highest volumetric CH4 capacity recorded in a densified MOF. Although this maximum working pressure (100 bar) is higher than the U.S. DOE target pressure (65 bar), it may still be regarded as safe, economic and industrially viable. UiO-66_D is the first material with a working capacity reaching the U.S. DOE target (263 cm3 (STP) cm–3, 298 K) over this pressure range. 260 The UiO-66_A – D monolith’s capacity to store CO2 were also explored. Again, Type II isotherms were observed in all cases, the result of which was high volumetric CO2 storage capacities with comparable trends to those observed for CH4 storage in the same materials. The storage capacity of UiO-66_D was not only 500% higher than a comparably pressurised storage tank at 40 bar but also 50% higher than commercial adsorbent zeolite 13X.9 Additionally, the CH4:CO2 selectivity of the materials was calculated from these experimental isotherms demonstrating each material’s comparable capacity to selectively adsorb CO2 from an equimolar CO2/CH4 mixture. These data suggest that monolithUiO-66 may potentially be applied, not only to high density CH4 storage for fuel delivery applications, but to further stages of the fuel generation pipeline including NG refinery and waste CO2 capture.10 Continuing the promising theme of using monolithic materials for gas fuel storage, the H2 storage capacity of monolithHKUST-1 and its Pd-doped derivatives (5% and 10% doping) was studied. At low pressures (0 – 1 bar), the gravimetric and volumetric storage capacities of each material were found to substantially exceed that of a traditional storage tank and the results were further correlated with Pd doping level. Hence, at 1 bar the highest H2 storage capacity was observed in 10% Pd@monolithHKUST-1. This is consistent with the enhancement in H2 storage capacity of composite Pd@powderHKUST-1 relative to its pure components as reported by Kitagawa et al.11 Mechanistic studies (collection of the isotherms at variable temperatures, in situ H2 XRD studies and analysis of the isotherms in terms of relative uptake by the Pd and MOF components) demonstrated the Pd NPs to contribute significantly to the monolithic composite’s uptake capacity via the energetically favourable formation of PdH. Kitagawa et al. previously postulated the MOFs adsorption capacity to be negligible relative to that of the Pd NPs, and hence any enhancement in total uptake capacity of their composite was attributed to the electronic interaction of the components. An enhanced 4d band hole density in Pd was suggested via the electron withdrawal effect of the Cu-MOF,12 the result of which was enhanced hydride formation capacity (74% improvement). However, in the current case the monolithMOFs contribution to the total uptake capacity was determined to be non-negligible over this pressure range. Hence, subtraction of the MOFs concurrent H2 adsorption capacity indicated only a 34% enhancement of PdH formation for Pd@monolithHKUST-1 relative to the pure NPs. The H2 storage capacity of monolithHKUST-1 and Pd@ monolithHKUST-1 was further explored up to a maximum pressure of 100 bar. At these elevated (but nonetheless practical) pressures, the 261 presence of Pd NPs (which become saturated with H at low pressure) was found to be detrimental to the composite material’s total storage capacity. The presence of non-porous Pd NPs diminishes the monoliths SBET, limiting adsorption by the MOF. Consequently, at 100 bar pure monolithHKUST-1 was found to be the highest performing material in terms of both volumetric and gravimetric H2 storage capacity. At 11.6 g L–1 (100 bar, 298 K) this material is far from close to achieving the U.S. DOE’s ultimate H2 storage target (50 g L–1). However, it does represent a benchmark result for MOFs, with this volumetric capacity being near identical to that of industrially non-viable benchmark powdered MOF Ni2(m-dobdc).13 As monolithHKUST-1 is pelletised into a practical material, it may offer a means of improving the current technology, even if it is not capable of reaching the ultimate storage target. 262 2.0| Future Work Perhaps the most promising results in this research project were achieved by monolithic UiO- 66.1 This industrially viable, high stability monolith demonstrates benchmark results for high density NG storage and approaches the U.S. DOE’s ultimate target. Correspondingly, the practicalities of its pilot plant production have been discussed with chemical engineers at Immaterial Labs Ltd. The conversion of laboratory-scale reactions to kilogram-scale reactions for industrial production is obviously non-trivial. In the current case, while the primary MOF particle synthesis is not particularly challenging, it is the subsequent centrifugation of small quantities of this MOF gel that represents a practical challenge outside a research laboratory setting. Furthermore, as extensively discussed, the rate of drying is critical in controlling the material’s bulk physical properties and its corresponding gas adsorption capacity. Not only is scale-up of the procedure likely to influence drying rate, but ultimately, the slow ca. 5 days drying period required to synthesise the densest monoliths may present a production bottleneck which must be overcome. Finally, while all of the reagents required for synthesis are relatively simple and commercially available, a substantial benefit from an industrial perspective, the use of DMF as the reaction and washing solvent is less than ideal. This toxic and expensive solvent is well-known to be the best solvent for UiO-66 synthesis with only a handful of hydrothermal syntheses having been reported.14 A greener, water-based monolith synthesis would be preferential, requiring substantial further work. The mass production of monolithHKUST-1,8 which exhibits outstanding CH4 and H2 storage capabilities, is likewise being explored by Immaterial. While synthesis of this material also requires practically challenging centrifugation and the slow evaporation of flammable ethanol,8 it is the material’s substantial air sensitivity which represents the biggest barrier to its practical usage. This instability is also reflected in the derivative Pd@monolithHKUST-1 composite. Furthermore, additional stability problems are introduced by the presence of the NPs; Pd undergoes lattice expansion under H2 chemisorption. If the long-term viability of using Pd@monolithHKUST-1 for hypothetical low pressure H2 storage applications is to be explored, the effects of dopant NP expansion on the mechanical stability of the surrounding monolith needs to be verified. This could be achieved through H2 loading-unloading cycling experiments followed by repeated nanoindentation testing. 263 Considering the results of this project as a whole, the synthesis of a wide range of MOFs as monoliths appears to be possible. Over 75,600 unique MOF structures are currently known,15 each one with diverse and distinct chemical/physical properties suitable for different conceivable applications. While the earlier discussed catalytic testing results suggest that applications involving organic species may not be ideal, due to the poor diffusion kinetics of these large molecules through the MOFs typically small micropores, potential applications involving smaller inorganic ions may prove more promising. For example, the selective adsorption of small inorganic species for i.e. water purification is of significant industrial interest. MOFs capabilities towards the adsorption of e.g. mercury, fluoride, arsenic and uranium have been extensively reported.16 In particular, water desalination is generating continued attention,17,18 and water stable monolithMOF could provide a practical and low cost means of achieving this. From a purely synthetic perspective, the results of this project further our understanding of different monolith’s capabilities towards hosting NPs, even if meaningful catalytic activities in these composites were not achieved. The use of a MOF with larger pores may permit faster diffusion of substrates through its porosity, enhancing catalytic activity. Hence, a promising area for future research lies in the synthesis of monolithic MOFs with wider pores for the targeted immobilisation of catalytic NPs, so that recyclable and industrially viable composite materials can be obtained without loss of activity. However, when the results of Chapter II are considered, this target appears challenging. It was observed that the synthesis of crystalline monolithic MOFs with wide pores is more difficult than that of the corresponding small pore MOFs due to the inherent inability to synthesise sufficiently small primary NPs for monolith formation while maintaining long range crystalline order. Hence, monolithNU-1000 (pore diameter ca. 31 Å) was amorphous whereas comparably synthesised UiO-66 (pore diameters, 11 and 8 Å) displayed high crystallinity. UiO-67, which is isoreticular to parent UiO-66, may represent a promising research avenue.19 The length of this MOFs biphenyl linker results in porosity (12 and 16 Å diameter) between that of UiO-66 and NU-1000. Thus, it may yet prove possible to synthesise this MOF as a crystalline monolith with improved diffusion kinetics relative to that of small pore ZIF-8 (11.7 Å pores connected by small 3.4 Å windows).20 The capacity to tune mesoporosity in monolithZr-MOF samples presents a further means of nominally increasing pore size and consequently improving rate of diffusion of foreign species through the MOFs porosity. In this vein, the preliminary synthesis of NP@monolithUiO-66 was 264 explored. Anecdotally, NPs could not be immobilised in the MOF during primary particle synthesis and were instead observed to sediment in the reaction vessel below the seemingly undoped MOF gel. Alternatively, mixing the NPs into the pre-synthesised gel prior to drying appeared to yield a doped monolith. Therefore, the NPs were not immobilised within MOF primary NPs, but instead trapped between them. The synthesis of NP@monolithZr-MOF with variable PSD requires substantiation and represents a future research avenue. Finally, for the more than 75,600 known MOF structures, each with distinctive chemical and physical properties, numerous applications have been proposed. This raises the question; how do we choose the ideal material for a target application? In this respect, the considered combination of computational and experimental chemistry will likely yield the most promising results.21 Recent advances in data mining and computational simulation of known MOF structures have already provided unrivalled insights into the design and selection of materials for specific target applications.22 For example, Moghadam et al. reported a full cycle of novel material development for targeted O2 capture and storage.23 Starting from the high throughput computational screening of 2,932 structures using combined GCMC simulations and density functional theory calculations, structure-property relationships were elucidated. Subsequently, the computationally identified top-performing MOF UMCM-15224 was synthesised as a powder, which yielded an outstanding 22.5% improvement in O2 storage capacity over the previous record-setting material. This demonstrates not only the unrivalled potential of MOFs to revolutionise environmental gas storage/capture but further highlights the significance of fully utilising computational chemistry to both rationalise and unlock this potential experimentally.5 265 3.0| References 1. Connolly, B. M. et al. Tuning Porosity in Macroscopic Monolithic Metal-Organic Frameworks for Exceptional Natural Gas Storage. Nat. Comm. 10, 2345 (2019). 2. Bueken, B. et al. Gel-Based Morphological Design of Zirconium Metal-organic Frameworks. Chem. Sci. 8, 3939–3948 (2017). 3. Cychosz, K. A. & Thommes, M. Progress in the Physisorption Characterization of Nanoporous Gas Storage Materials. Engineering 4, 559–566 (2018). 4. DeSantis, D. et al. Techno-economic Analysis of Metal–Organic Frameworks for Hydrogen and Natural Gas Storage. Energy & Fuels 31, 2024–2032 (2017). 5. Petit, C. Present and Future of MOF Research in the Field of Adsorption and Molecular Separation. Curr. Opin. Chem. Eng. 20, 132–142 (2018). 6. Miller, S. E., Teplensky, M. H., Moghadam, P. Z. & Fairen-Jimenez, D. Metal-Organic Frameworks as Biosensors for Luminescence-Based Detection and Imaging. Interface Focus 6, 20160027 (2016). 7. Mehta, J. P. et al. Sol-Gel Synthesis of Robust Metal-Organic Frameworks for Nanoparticle Encapsulation. Adv. Funct. Mater. 28, 1705588 (2018). 8. Tian, T. et al. A Sol–Gel Monolithic Metal–Organic Framework with Enhanced Methane Uptake. Nat. Mater. 17, 174–179 (2018). 9. Adil, K. et al. Valuing Metal–Organic Frameworks for Postcombustion Carbon Capture: A Benchmark Study for Evaluating Physical Adsorbents. Adv. Mater. 29, 1702953 (2017). 10. Tanh Jeazet, H. B. et al. Increased Selectivity in CO2/CH4 Separation with Mixed-Matrix Membranes of Polysulfone and Mixed-MOFs MIL-101(Cr) and ZIF-8. Eur. J. Inorg. Chem. 2016, 4363–4367 (2016). 11. Li, G. et al. Hydrogen Storage in Pd Nanocrystals Covered with a Metal–Organic Framework. Nat. Mater. 13, 802–806 (2014). 12. Chen, Y. et al. Electronic Origin of Hydrogen Storage in MOF-Covered Palladium Nanocubes Investigated by Synchrotron X-rays. Commun. Chem. 1, 61 (2018). 13. Kapelewski, M. T. et al. Record High Hydrogen Storage Capacity in the Metal-Organic Framework Ni2(m-dobdc) at Near-Ambient Temperatures. Chem. Mater. 30, 8179–8189 266 (2018). 14. Hu, Z., Peng, Y., Kang, Z., Qian, Y. & Zhao, D. A Modulated Hydrothermal (MHT) Approach for the Facile Synthesis of UiO-66-Type MOFs. Inorg. Chem. 54, 4862–4868 (2015). 15. Moghadam, P. Z. et al. Development of a Cambridge Structural Database Subset: A Collection of Metal-Organic Frameworks for Past, Present, and Future. Chem. Mater. 29, 2618–2625 (2017). 16. Connolly, B. M., Mehta, J. P., Moghadam, P. Z., Wheatley, A. E. H. & Fairen-Jimenez, D. From Synthesis to Applications : Metal – Organic Frameworks for an Environmentally Sustainable Future. Curr. Opin. Green Sustain. Chem. 12, 47–56 (2018). 17. Wang, W. et al. Trade-Off in Membrane Distillation with Monolithic Omniphobic Membranes. Nat. Comm. 10, 3220 (2019). 18. Lu, X. et al. Tuning the Permselectivity of Polymeric Desalination Membranes via Control of Polymer Crystallite Size. Nat. Comm. 10, 2347 (2019). 19. Katz, M. J. et al. A Facile Synthesis of UiO-66, UiO-67 and Their Derivatives. Chem. Commun. 49, 9449–9451 (2013). 20. Fairen-Jimenez, D. et al. Flexibility and Swing Effect on the Adsorption of Energy- Related Gases on ZIF-8: Combined Experimental and Simulation Study. Dalton Trans. 41, 10752–10762 (2012). 21. Howarth, A. J. Experimentalists and Theorists Need to Talk. Nature 551, 433–434 (2017). 22. Coudert, F. X. & Fuchs, A. H. Computational Characterization and Prediction of Metal- Organic Framework Properties. Coord. Chem. Rev. 307, 211–236 (2016). 23. Moghadam, P. Z. et al. Computer-Aided Discovery of a Metal-Organic Framework with Superior Oxygen Uptake. Nat. Comm. 9, 1378 (2018). 24. Schnobrich, J. K. et al. Linker-Directed Vertex Desymmetrization for the Production of Coordination Polymers with High Porosity. J. Am. Chem. Soc. 132, 13941–13948 (2010). Appendix ii iii I| Experimental UiO-66 powder Modified from the literature procedure reported by Farha et al.1 A solution of zirconium propoxide (450 μL, 1 mmol, 70% in propanol), DMF (44 mL) and glacial AA (25 mL) was heated to 130 °C for 2 hours to produce a bright yellow solution. The solution was cooled to room temperature and benzene-1,4-dicarboxylic acid (0.474 g, 2.86 mmol) was added. The solution was sonicated for 30 s before being left to stir at room temperature for a further 18 hours. The white/yellow suspension was centrifuged in two portions, each of which was washed in ethanol (2 x 35 mL) and methanol (1 x 35 mL) before being left to dry overnight at room temperature. This yielded UiO-66 as a white powder. UiO-66 gel Modified from the literature procedure reported by Bennett et al.2 Benzene-1,4-dicarboxylic acid (1.20 g, 7.25 mmol) and zirconium(IV) oxychloride octahydrate (1.61 g, 5.0 mmol) were dissolved in DMF (30 mL, 99%). Concentrated hydrochloric acid (1.5 mL, 37%) and glacial AA (2.0 mL) were added under vigorous stirring. The resulting solution was sealed in a 100 mL Pyrex Schott bottle and heated to 100 °C for 2 hours. This yielded UiO-66 as a thick white gel. monolithUiO-66_A – D DMF (50 mL) was added to as synthesised UiO-66 gel and vigorously mixed. The diluted UiO- 66 suspension (7.5 mL per tube) was centrifuged (3 min, 5500 rpm) and the supernatant decanted. The gel was washed, centrifuged (5500 rpm) and dried to produce a range of monoliths (Supplementary Table 1). The obtained monoliths were soaked in acetone (3 x 5 mL, 24 hours) and methanol (3 x 5 mL, 24 hours) and then dried at room temperature overnight. The monoliths were activated by heating to 110 °C under vacuum for 8 hours. iv Supplementary Table 1| Experimental conditions (washing solvent, centrifugation time and drying temperature) for monolithUiO-66 synthesis from UiO-66 gel. Washing procedure Centrifugation procedure Drying temperature UiO-66_A Ethanol (3 × 30 mL) 3 × 10 min† 200 °C UiO-66_B Ethanol (3 × 30 mL) 3 × 10 min† 30 °C UiO-66_C DMF (1 × 30 mL) 1 × 10 min† 30 °C UiO-66_D DMF (1 × 30 mL) 1 x 10 min† + 1 x 180 min‡ 30 °C †Centrifugation (5500 rpm) performed after each wash (gel was shaken in the washing solvent) to re-obtain MOF gel as sediment. ‡Additional 180 min (5500 rpm) centrifugation performed on densified MOF gel after washing in DMF and decanting the supernatant. After the final extended centrifugation, any further obtained supernatant was carefully extracted with a micro-pipette without disturbing the gel. UiO-66-NH2 gel Modified from the literature procedure for UiO-66 gel reported by Bennett et al.2 2-aminobenzene-1,4-dicarboxylic acid (1.31 g, 7.23 mmol) and zirconium(IV) oxychloride octahydrate (1.61 g, 5.0 mmol) were dissolved in DMF (30 mL, 99%). Concentrated hydrochloric acid (1.5 mL, 37%) and glacial AA (2.0 mL) were added under vigorous stirring. The resulting solution was sealed in a 100 mL Pyrex Schott bottle and heated to 100 °C for 2 hours. This yielded UiO-66-NH2 as a viscous yellow gel. monolithUiO-66-NH2_A – C DMF (50 mL, 99%) was added to the UiO-66-NH2 gel, as synthesised above, and vigorously mixed. The diluted UiO-66-NH2 suspension (7.5 mL per tube) was centrifuged (3 min, 5500 rpm) and the supernatant decanted. The gel was washed, centrifuged (5500 rpm) and dried to produce a range of monoliths (Supplementary Table 2). The obtained monoliths were soaked in acetone (3 ´ 5 mL, 24 hours) and methanol (3 ´ 5 mL, 24 hours) and then dried at room temperature overnight. Monoliths were activated by heating to 110 °C under vacuum for 8 hours. v Supplementary Table 2| Experimental conditions (washing solvent, centrifugation time and drying temperature) for monolithUiO-66-NH2 synthesis from UiO-66-NH2 gel. Washing procedure Centrifugation Drying temperature UiO-66-NH2_A Ethanol (3 × 30 mL) 3 × 10 min† 30 °C UiO-66-NH2_B DMF (1 × 30 mL) 1 × 10 min† 30 °C UiO-66-NH2_C DMF (1 × 30 mL) 1 × 10 min† + 1 × 180 min‡ 30 °C †Centrifugation (5500 rpm) performed after each wash (gel was shaken in the washing solvent) to re-obtain MOF gel as sediment. ‡Additional 180 min (5500 rpm) centrifugation performed on densified MOF gel after washing in DMF and decanting the supernatant. After the final extended centrifugation, any further obtained supernatant was carefully extracted with a micro-pipette without disturbing the gel. UiO-66-ndc gel Naphthalene-1,4-dicarboxylic acid (1.57 g, 7.25 mmol) and zirconium(IV) oxychloride octahydrate (1.61 g, 5.0 mmol) were dissolved in DMF (30 mL, 99%). Concentrated hydrochloric acid (37%) and glacial AA (Supplementary Table 3) were added under vigorous stirring. The resulting solution was sealed in a 100 mL Pyrex Schott bottle and heated to 100 °C for 2 hours. This yielded UiO-66-ndc as sediment or gel. vi Supplementary Table 3| Experimental conditions for UiO-66-ndc primary particle synthesis showing volumes of modulator used and the visible appearance of the obtained product. UiO-66-ndc_1 – 15 The monolithUiO-66_C drying procedure (Supplementary Table 1) was applied to each UiO-66- ndc gel (Supplementary Table 3). DMF (50 mL, 99%) was added to the UiO-66-ndc suspension or gel and vigorously mixed. The diluted UiO-66-ndc gel (7.5 mL per tube) was centrifuged (3 min, 5500 rpm) and the supernatant decanted. The gel was washed, centrifuged (5500 rpm) and dried. The obtained materials were soaked in acetone (3 ´ 5 mL, 24 hours) and methanol (3 ´ 5 mL, 24 hours) and then dried at room temperature overnight. Materials were activated by heating to 110 °C under vacuum for 8 hours. Modulator Volume (mL) Monolithic MOF Primary particle product AA HCl UiO-66-ndc_1 Sediment 2 1.5 UiO-66-ndc_2 Gel 1 1.5 UiO-66-ndc_3 Gel 0.75 1.5 UiO-66-ndc_4 Gel 1 1 UiO-66-ndc_5 Gel 0.75 1 UiO-66-ndc_6 Gel 0.5 1 UiO-66-ndc_7 Gel 0.25 1 UiO-66-ndc_8 Gel 0.25 0.5 UiO-66-ndc_9 Gel 0.5 0.5 UiO-66-ndc_10 Gel 0.75 0.75 UiO-66-ndc_11 Gel 0.75 0.6 UiO-66-ndc_12 Gel 0.75 0.5 UiO-66-ndc_13 Gel 0.75 0.4 UiO-66-ndc_14 Gel 0.75 0.2 UiO-66-ndc_15 Gel 0.75 0 vii NU-1000 gel The NU-1000 gel and subsequent monolith synthesis was performed by J. Mehta. Zirconium(IV) oxychloride octahydrate (121 mg, 0.375 mmol), 4-aminobenzoic acid (154 mg, 1.12 mmol) and trifluroacetic acid (200 µl) were dissolved in DMF (10 mL). 1,3,6,8-tetrakis(p- benzoate)pyrene (25 mg) was dissolved in DMF (10 mL). The solutions were combined at 140 °C and maintained at this temperature for 1 hour yielding NU-1000 as a viscous yellow gel. monolithNU-1000 The yellow MOF gel (synthesised above) was collected by centrifugation (60 min, 5500 rpm) and washed (2 x acetone (30 mL), 2 x DMF (30 mL)). One additional centrifugation (60 min, 5500 rpm) was performed on the gel after the final wash. The material was dried at 30 °C yielding monolithNU-1000, which was activated under vacuum at 110 °C for 24 hours. monolithZIF-8 Synthesised according to the literature procedure reported by Tian et al.3 2-methyl imidazole (0.65 g, 7.92 mmol) was dissolved in absolute ethanol (20 mL) under sonication. Zn(NO3)2.6H2O (0.3 g, 1.00 mmol) was dissolved in absolute ethanol (20 mL) under sonication. With rapid stirring the solutions were combined, forming a white suspension. The suspension was mixed at room temperature for a further 15 minutes before the precipitate was collected under centrifugation (5500 rpm, 10 min). The white solid was washed in ethanol (2 x 40 mL) and methanol (1 x 40 mL) before being dried at room temperature overnight. The resulting material was activated under vacuum at 110 °C for 8 hours, yielding monolithic ZIF-8 as a transparent, glassy monolith. SnO2 NPs Synthesised according to the literature procedure reported by Mehta et al.4 Sodium hydroxide (0.7 g, 17.5 mmol) was dissolved in a mixture of milli-Q water (20 mL) and absolute ethanol (20 mL). The resulting solution was added dropwise to a solution of SnCl4.5H2O (0.787 g, 2.24 mmol) in milli-Q water (15 mL), until a pH of 12 was achieved. The resulting cloudy white solution was heated in a sealed autoclave at 200 °C for 24 hours. The solid was collected by centrifugation (5500 rpm, 5 minutes), washed six times using an alternating sequence of ethanol (40 mL x 3) and a 3:1 ethanol:water mixture (40 mL x 3) and dried under vacuum to yield SnO2 NPs as a white solid. viii Surface functionalised SnO2 NPs Modified from the literature procedure reported by Mehta et al.4 Sodium hydroxide (0.7 g, 17.5 mmol) was dissolved in a solution of milli-Q water (20 mL) and absolute ethanol (20 mL). The resulting solution was added dropwise to a solution of SnCl4.5H2O (0.787 g, 2.24 mmol) in milli-Q water (15 mL), until a pH of 12 was achieved. Surface capping agents were added (Supplementary Table 4) and the pH was re-adjusted to 12. The resulting cloudy white solution was heated in a sealed autoclave for 24 hours.* The resulting solid product was collected by centrifugation (5500 rpm, 5 minutes), washed six times under sonication using an alternating sequence of ethanol (40 mL x 3) and a 3:1 ethanol:water mixture (40 mL x 3) and dried under vacuum to yield SnO2 NPs as a white solid in each case.† Supplementary Table 4| Surface capping agents added to the synthesis of functionalised SnO2 NPs, SnO2-1 – 12. Sample name Additive‡ SnO2-1 No additive SnO2-2† No additive SnO2-3 OA (0.703 cm3, 0.629 g, 2.23 mmol) SnO2-4 CTAB (0.816 g, 2.24 mmol) SnO2-5 CA (0.430 g, 2.24 mmol) SnO2-6 NaOl (0.682 g, 2.24 mmol) SnO2-7 PVP10k (0.75 g, 2.24 mmol) SnO2-8 NaCl – 1 equiv. (0.980 g, 16.7 mmol) SnO2-9 NaCl – 2 equiv. (2.00 g, 34.2 mmol) SnO2-10 NaCl – 5 equiv. (5.00 g, 85.6 mmol) SnO2-11* NaCl – 2 equiv. (2.00 g, 34.2 mmol) SnO2-12* No additive *All samples were synthesised at 200 °C except SnO2-11 and SnO2-12 (150 °C). †All samples were dried under vacuum except SnO2-2 which was re-dispersed to a white suspension in absolute ethanol (40 mL) after washing. ‡For all organic additives, a 1:1 molar ratio to Sn was utilised. ix SnO2@monolithZIF-8 Modified from the literature procedure reported by Mehta et al.4 SnO2 NPs (SnO2-1 – 12, 30 mg) were added to a solution of 2-methyl imidazole (0.65 g, 7.92 mmol) in absolute ethanol (20 mL) and sonicated for 15 minutes to form a white suspension. Under rapid stirring, a solution of Zn(NO3)2.6H2O (0.3 g, 1.00 mmol) in absolute ethanol (20 mL) was added and mixed for a further 15 minutes at room temperature. The white suspension was collected by centrifugation (5500 rpm, 10 min) and washed in absolute ethanol (2 x 40 mL) and methanol (1 x 40 mL) before being dried at room temperature overnight. The resulting white monolith was activated by heating under vacuum at 110 °C for 8 hours. Preliminary Au NPs (4.9 ± 1.1 nm diameter) Modified from the literature procedure reported by Murphy et al.5 HAuCl4.3H2O (7.88 mg, 0.02 mmol) and trisodium citrate (5.88 mg, 0.02 mmol) were dissolved in milli-Q water (80 mL). Under rapid stirring, an ice-cold solution of NaBH4 (9.08 mg, 0.24 mmol) in milli-Q water (2.4 mL) was injected, causing the previously yellow solution to immediately turn bright red/purple. The solution was mixed for 30 seconds at room temperature before PVP10 K (0.5 g) was added and it was mixed for a further 30 min. Under reduced pressure, the aqueous colloid was concentrated to a volume of 2 mL before being re-diluted with distilled ethanol (18 mL). Au@monolithZIF-8 (6.0 ± 1.6 nm diameter NPs) Modified from the literature procedure for SnO2@monolithZIF-8 reported by Mehta et al.5 Au NPs synthesised according to the preliminary procedure were used (above). 2-methyl imidazole (0.65 g, 7.92 mmol) was added to the ethanolic suspension of Au NPs (20 mL) and sonicated for 30 min. Under rapid stirring, a solution of Zn(NO3)2.6H2O (0.3 g, 1.00 mmol) in absolute ethanol (20 mL) was added and mixed for a further 15 min at room temperature. The resulting purple suspension was centrifuged (5500 rpm, 10 min) and washed in ethanol (2 x 40 mL) and methanol (1 x 40 mL) before being dried at room temperature overnight. The resulting dark purple/black, transparent monolith was activated under vacuum (110 °C) for 8 hours. Au@monolithZIF-8 (2.6 ± 1.0 nm diameter Au NPs) Derived from the literature procedure reported by Fischer et al.6 HAuCl4.3H2O (8.5 mg, 0.02 mmol) and PVP10K (555 mg) were dissolved in milli-Q water (80 mL) at room temperature. Under rapid mixing, an ice-cold solution of NaBH4 (11 mg) in milli-Q water (2.4 mL) was quickly injected and the solution was mixed for a further 20 seconds. The NP growth was x quenched by pouring into ice cold acetone (500 mL) and the pink solution was stirred, in ice, for a further 30 minutes. The Au NP seeds were collected under centrifugation (4 minutes, 5500 rpm) before immediately being re-dispersed in an ice-cold solution of 2-methyl imidazole (0.65 g, 7.92 mmol) in absolute ethanol (40 mL). A portion of the ethanolic Au colloid (20 mL) was added to an ice-cold solution of Zn(NO3)2.6H2O (0.3 g, 1.00 mmol) in absolute ethanol (20 mL) and rapidly mixed for 20 minutes. The red precipitate was collected under centrifugation (5500 rpm, 10 min) and washed in ethanol (2 x 40 mL) and methanol (1 x 40 mL) before being dried at room temperature overnight. The resulting dark purple/black, monolith was activated under vacuum (110 °C) for 8 hours. PdO/TiO2 NPs Modified from the synthetic procedure for Pt/TiO2 reported by Mehta.7 Stearic acid (0.84 g, 2.95 mmol) was dissolved in a solution of absolute ethanol (80 mL) and milli-Q water (20 mL) at 85 °C under rapid stirring. Pd(acac)2 (0.03 g, 0.10 mmol) was added and the resulting yellow solution was heated under reflux for 4 hours yielding a black solution. A solution of Ti(OBu)4 (0.4 mL, 1.17 mmol) in absolute ethanol (2 mL) was added dropwise under rapid stirring. Reflux was maintained for a further 2 hours before being cooled to room temperature. The solid was collected by centrifugation (5500 rpm, 10 min) and washed under sonication in alternating water/ethanol (1:7, 2 x 40 mL) and ethanol (2 x 40 mL) solutions. The resulting NPs were dried under vacuum overnight before being heated in a furnace at 500 °C for 3 hours yielding a brown solid. PdO/TiO2@monolithZIF-8 Modified from the literature procedure for SnO2@monolithZIF-8 reported by Mehta et al.5 Pd/TiO2 NPs (30 mg) were added to a solution of 2-methyl imidazole (0.65 g, 7.92 mmol) in absolute ethanol (20 mL) and dispersed to form a grey suspension under sonication for 2 hours. With rapid stirring, a solution of Zn(NO3)2.6H2O (0.3 g, 1.00 mmol) in absolute ethanol (20 mL) was added and mixed for a further 15 minutes at room temperature. The resulting white/grey suspension was centrifuged (5500 rpm, 10 min) and washed in ethanol (2 x 40 mL) and methanol (1 x 40 mL) before being dried at room temperature overnight. The resulting dark grey solid was activated in a vacuum oven at 110 °C for 8 hours. xi Au@PdO NPs Modified from the procedure reported by Yamuchi et al.8 HAuCl4.3H2O (0.02 g, 0.05 mmol) was dissolved in milli-Q water (2.5 mL) under sonication. Na2PdCl4 (0.015 g, 0.05 mmol) was dissolved in milli-Q water (2.5 mL) under sonication. The two solutions were combined and PVP40 K (0.067 g) was added. Under rapid stirring, a solution of ascorbic acid (0.07 g, 0.04 mmol) in milli-Q water (0.67 mL) was rapidly injected, causing a colour change from bright orange to dark brown. The solution was sonicated for 5 minutes before being vigorously mixed for a further 6 hours at room temperature. The solution was centrifuged (5500 rpm, 20 min) to obtain a black/brown solid which was washed with milli-Q water (3 x 40 mL). The resulting NPs were re-suspended in milli-Q water (20 mL) and sonicated for 30 min to minimise aggregation. (Au@PdO)/TiO2 NPs To an aqueous Au@PdO suspension (20 mL, as synthesised above), a solution of stearic acid (0.84 g, 2.95 mmol) in absolute ethanol (80 mL) was added. Under rapid stirring at room temperature, a solution of Ti(OBu)4 (0.4 mL, 1.17 mmol) in ethanol (2 mL) was added dropwise. The resulting grey mixture was heated to 115 °C and maintained at this temperature for 2 hours. The solution was cooled to room temperature and washed in alternating water/ethanol (1:7, 2 x 40 mL) and ethanol (2 x 40 mL) solutions before being dried under vacuum. The resulting grey powder was heated in a furnace at 500 °C for 3 hours. (Au@PdO)/TiO2@monolithZIF-8 Modified from the procedure for SnO2@monolithZIF-8 by Mehta et al.5 (Au@PdO)/TiO2 NPs (30 mg) were added to a solution of 2-methyl imidazole (0.65 g, 7.92 mmol) in ethanol (20 mL) and sonicated for 4 hours to form a grey suspension. Under rapid stirring, a solution of Zn(NO3)2.6H2O (0.3 g, 1.00 mmol) in absolute ethanol (20 mL) was added and mixed for a further 15 minutes at room temperature. The resulting white/grey suspension was centrifuged (5500 rpm, 10 min) and washed in absolute ethanol (2 x 40 mL) and methanol (1 x 40 mL) before being dried at room temperature overnight. The dark grey solid was activated by heating in a vacuum oven at 110 °C for 8 hours. monolithHKUST-1 Synthesised according to the literature procedure reported by Tian et al.9 Benzene-1,3,5- tricarboxylic acid (0.13 g, 0.62 mmol) was dissolved in absolute ethanol (10 mL) under xii sonication. Cu(NO3)2 . 2.5H2O (0.15 g, 0.64 mmol) was dissolved in absolute ethanol (10 mL) under sonication. With rapid stirring the solutions were combined, rapidly forming a blue suspension. The suspension was mixed at room temperature for a further 10 minutes before the precipitate was collected under centrifuged (5500 rpm, 10 min). The blue solid was washed in ethanol (3 x 40 mL) before being dried at room temperature overnight. The resulting material was activated under vacuum at 150 °C for 8 hours, yielding monolithic HKUST-1 as a transparent, blue monolith. Pd NPs Synthesised according to the procedure reported by Kitagawa et al.10 Na2PdCl4 (285 mg, 9.7 mmol), L-ascorbic acid (300 mg, 1.7 mmol), KBr (1.5 g, 12.6 mmol) and PVP10K (0.54 g) were dissolved in milli-Q water (55 mL). The red solution was heated under reflux in an oil bath at 80 °C for 3 hours, causing a colour change to black. The cooled black solution was flocculated in excess acetone and the solid collected under centrifugation (10 minutes, 5500 rpm). The obtained black solid was washed (x 2) by dissolving in the minimum volume of water (< 1 mL) and sonicating, before again precipitating with acetone and collecting by centrifugation (10 min, 5500 rpm). The washed Pd NPs were re-suspended in absolute ethanol (30 mL). Pd@monolithHKUST-1 The concentration of the ethanolic Pd NP suspension (above) was determined by drying a known volume of the suspension and recording the remaining dry mass. The volume of NP suspension required to achieve the desired MOF loading was thus determined and this was diluted to a total volume of 10 mL with absolute ethanol. Benzene-1,3,5-tricarboxylic acid (0.13 g, 0.62 mmol) was added and the black suspension was sonicated for 1 hour. Cu(NO3)2 . 2.5H2O (0.15 g, 0.64 mmol) dissolved in absolute ethanol (10 mL) was added under rapid stirring. The stirring was continued at room temperature for a further 10 minutes before the precipitate was collected under centrifugation (5500 rpm, 10 min). The black solid was washed in ethanol (3 x 40 mL) before being dried at room temperature overnight. The resulting material was activated under vacuum at 150 °C for 8 hours, yielding monolithic Pd@HKUST-1 as a glassy, black monolith. xiii II| Supplementary Figures Supplementary Figure 1| N2 adsorption isotherms and BET analysis of monolithUiO-66_A. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. d 0.0 0.2 0.4 0.6 0.8 1.0 0 200 400 600 800 1,000 1,200 1E-7 1E-6 1E-5 1E-4 0.001 0 50 100 150 200 0.00 0.05 0.10 0.15 0 50 100 150 200 250 300 0.00 0.02 0.04 0.06 0.08 0.0E+00 1.0E-04 2.0E-04 3.0E-04 N 2 u pt ak e (c m 3 g- 1 ) P/Po P/Po N 2 u pt ak e (c m 3 g- 1 ) V( 1- P/ P o ) P/Po Maximum P/Po P/Po P/ V( P o -P ) y = 0.00369799x + 0.00000340 R2 = 0.9996 UiO-66_A a b c dP/Po / o y = .0036 799x + .00 00340 R2 = 0.9996 / o V( 1- P/ P o ) P/ V( P o -P ) N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 Maximum xiv Supplementary Figure 2| N2 adsorption isotherms and BET analysis of monolithUiO-66_B. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. d 0.0 0.2 0.4 0.6 0.8 1.0 0 200 400 600 800 1,000 1E-6 1E-5 1E-4 0.001 0 50 100 150 0.00 0.05 0.10 0.15 0 50 100 150 200 250 0.00 0.02 0.04 0.06 0.08 0.0E+00 1.0E-04 2.0E-04 3.0E-04 4.0E-04 N 2 u pt ak e (c m 3 g- 1 ) P/Po N 2 u pt ak e (c m 3 g- 1 ) P/Po Maximum P/Po P/Po V( 1- P/ P o ) y = 0.00437737x + 0.00000384 R2 = 0.9995 P/Po P/ V( P o -P ) UiO-66_B a b c d P/Po P/ o P/Po V( 1- P/ P o ) P/ V( P o -P ) N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 y = 0.00437737x + 0.00000384 Maximum R2 = 0.9995 xv Supplementary Figure 3| N2 adsorption isotherms and BET analysis of monolithUiO-66_C. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. d 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 500 600 1E-6 1E-5 1E-4 0.001 0 50 100 150 0.00 0.05 0.10 0.15 0 50 100 150 200 250 300 0.00 0.02 0.04 0.06 0.08 0.0E+00 1.0E-04 2.0E-04 3.0E-04 N 2 u pt ak e (c m 3 g- 1 ) P/Po N 2 u pt ak e (c m 3 g- 1 ) P/Po Maximum P/Po V( 1- P/ P o ) P/Po y = 0.00408414x + 0.00000361 R2 = 0.9995 P/ V( P o -P ) P/Po UiO-66_C a b c d P/Po P/P o / o V( 1- P/ P o ) P/ V( P o -P ) N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 y = 0.00408414x + 0.00000361 R2 = 0. 95 Maximum xvi Supplementary Figure 4| N2 adsorption isotherms and BET analysis of monolithUiO-66_D. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. d UiO-66_D 0.0 0.2 0.4 0.6 0.8 1.0 0 50 100 150 200 250 300 350 400 1E-7 1E-6 1E-5 1E-4 1E-3 0 50 100 150 0.00 0.05 0.10 0.15 0 50 100 150 200 250 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.0 5.0E-5 1.0E-4 1.5E-4 2.0E-4 2.5E-4 3.0E-4 3.5E-4 N 2 u pt ak e (c m 3 g- 1 ) P/Po N 2 u pt ak e (c m 3 g- 1 ) P/Po V( 1- P/ P o ) P/Po Maximum P/Po y = 0.00443303x + 0.00000386 R2 = 0.9994 P/ V( P o -P ) P/Po a b c d o P/P P/PoP/ V( 1- P/ P o ) P/ V( P o -P ) N 2 up ta ke (S TP ) c m 3 g-1 N 2 up ta ke (S TP ) c m 3 g-1 y = 0.00443303x + 0.00000386 R2 = 0. 994 Maximum xvii Supplementary Figure 5| FLIM studies of monolithUiO-66_A. (left) FLIM images (a – d), of monolithUiO-66_A showing the aggregated MOF primary particles which comprise the monolith. White dashed boxes (inset) indicate the area selected for high magnification imaging (ai – di). The colours correspond to the fluorescence lifetime (see colour bar, below). (right) A 2D histogram phasor plot generated from the averaged FLIM images. The colours correspond to the frequency of occurrences (see colour bar, right). 20 µm Lifetime Low (3.7 ns)High (5.9 ns) 20 µm 20 µm 20 µm a aib c d bi ci di 6 µm 6 µm 6 µm6 µm xviii Supplementary Figure 6| FLIM studies of monolithUiO-66_B. (left) FLIM images (a – d), of monolithUiO-66_B showing the aggregated MOF primary particles which comprise the monolith. White dashed boxes (inset) indicate the area selected for high magnification imaging (ai – di). The colours correspond to the fluorescence lifetime (see colour bar, below). (right) A 2D histogram phasor plot generated from the averaged FLIM images. The colours correspond to the frequency of occurrences (see colour bar, right). 20 µm 20 µm 20 µm 20 µm Lifetime Low (2.6 ns)High (5.3 ns) a aib c d bi ci di 6 µm 6 µm 6 µm 6 µm xix Supplementary Figure 7| FLIM studies of monolithUiO-66_C. (left) FLIM images (a – d), of monolithUiO-66_C showing the aggregated MOF primary particles which comprise the monolith. White dashed boxes (inset) indicate the area selected for high magnification imaging (ai – di). The colours correspond to the fluorescence lifetime (see colour bar, below). (right) A 2D histogram phasor plot generated from the averaged FLIM images. The colours correspond to the frequency of occurrences (see colour bar, right). Lifetime Low (2.5 ns)High (5.3 ns) 20 µm 20 µm 20 µm 20 µm 6 µm 6 µm 6 µm 6 µm a aib c d bi ci di xx Supplementary Figure 8| FLIM studies of monolithUiO-66_D. (left) FLIM images (a – d), of monolithUiO-66_D showing the aggregated MOF primary particles which comprise the monolith. White dashed boxes (inset) indicate the area selected for high magnification imaging (ai – di). The colours correspond to the fluorescence lifetime (see colour bar, below). (right) A 2D histogram phasor plot generated from the averaged FLIM images. The colours correspond to the frequency of occurrences (see colour bar, right). Lifetime Low (2.6 ns)High (5.2 ns) 20 µm 20 µm 20 µm20 µm 6 µm 6 µm 6 µm 6 µm a aib c d bi ci di xxi Supplementary Figure 9| N2 adsorption isotherms and BET analysis of monolithUiO-66- NH2_A. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. 0.00 0.02 0.04 0.06 0.08 0.10 0.0E+00 1.0E-04 2.0E-04 3.0E-04 4.0E-04 5.0E-04 P/ V( P o -P ) P/Po y = 0.00517322x + 0.00000843 R2 = 0.99996 0.0 0.1 0.2 0.3 0 50 100 150 200 250 P/Po V( 1- P/ P o ) Maximum P/Po 0.0 0.2 0.4 0.6 0.8 1.0 0 200 400 600 800 P/Po N 2 u pt ak e, cm 3 ( ST P) g -1 1E-06 1E-05 1E-04 1E-03 0 50 100 150 P/Po N 2 u pt ak e, cm 3 ( ST P) g -1 UiO-66-NH2_A a b c dP/Po P/Po P/PoP/Po Maximum N 2 up ta ke (S TP ) c m 3 g- 1 N 2 up ta ke (S TP ) c m 3 g- 1 V( 1- P/ P o ) P/ V( P o -P ) xxii Supplementary Figure 10| N2 adsorption isotherms and BET analysis of monolithUiO-66- NH2_B. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 N 2 u pt ak e, cm 3 ( ST P) g -1 P/Po 0.0 0.1 0.2 0.3 0 50 100 150 200 P/Po V( 1- P/ P o ) Maximum P/Po 0.00 0.02 0.04 0.06 0.08 0.10 0.0E+00 1.0E-04 2.0E-04 3.0E-04 4.0E-04 5.0E-04 P/ V( P o -P ) P/Po y = 0.00529054 + 0.00000874 R2 = 0.99997473 1E-06 1E-05 1E-04 1E-03 0 50 100 150 N 2 u pt ak e, cm 3 ( ST P) g -1 P/Po a c b d UiO-66-NH2_B P/Po P/Po P/PoP/Po N 2 up ta ke (S TP ) c m 3 g- 1 N 2 up ta ke (S TP ) c m 3 g- 1 V( 1- P/ P o ) P/ V( P o -P ) Maximum xxiii Supplementary Figure 11| N2 adsorption isotherms and BET analysis of monolithUiO-66- NH2_C. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. 0.0 0.2 0.4 0.6 0.8 1.0 0 50 100 150 200 250 300 N 2 u pt ak e, cm 3 ( ST P) g -1 P/Po 0.0 0.1 0.2 0.3 0 40 80 120 160 P/Po V( 1- P/ P o ) Maximum P/Po 0.00 0.02 0.04 0.06 0.08 0.10 0.0E+00 1.0E-04 2.0E-04 3.0E-04 4.0E-04 5.0E-04 6.0E-04 7.0E-04 P/ V( P o -P ) P/Po y = 0.00654108x + 0.00001081 R2 = 0.99998452 1E-06 1E-05 1E-04 1E-03 0 25 50 75 100 N 2 u pt ak e, cm 3 ( ST P) g -1 P/Po UiO-66-NH2_C a c d b N 2 up ta ke (S TP ) c m 3 g- 1 N 2 up ta ke (S TP ) c m 3 g- 1 V( 1- P/ P o ) P/ V( P o -P ) P/Po P/Po P/PoP/Po Maximum xxiv Supplementary Figure 12| N2 adsorption isotherms and BET analysis of monolithUiO-66- ndc. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. 0.0E+00 2.0E+06 4.0E+06 6.0E+06 8.0E+06 1.0E+07 1.2E+07 1.4E+07 1.6E+07 1.8E+07 0.00 0.10 0.20 0.30 0.40 0.0E+00 1.0E-04 2.0E-04 3.0E-04 4.0E-04 5.0E-04 6.0E-04 7.0E-04 0.00 0.05 0.10 y = 0.006054x + 0.000010 R2 = 0.9999 0 20 40 60 80 100 120 140 1.0E-06 1.0E-05 1.0E-04 1.0E-03 1.0E-02 0 50 100 150 200 250 300 350 400 450 0.0 0.2 0.4 0.6 0.8 1.0 P/Po P/Po P/Po P/Po N 2 up ta ke (S TP ) c m 3 g- 1 N 2 up ta ke (S TP ) c m 3 g- 1 V( 1- P/ P o ) P/ V( P o -P )) a b c d Maximum xxv Supplementary Figure 13| N2 adsorption isotherms and BET analysis of monolithNU-1000. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. 0.04 0.08 0.12 0.16 0.20 0.24 0.28 2.0E-04 4.0E-04 6.0E-04 8.0E-04 1.0E-03 1.2E-03 P/ V( P o -P ) P/Po y = 0.00404855x + 0.00005028 R2 = 0.9998 1E-05 1E-04 1E-03 1E-02 0 100 200 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po 0.0 0.2 0.4 0.6 0.8 1.0 0 200 400 600 800 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po 0.0 0.1 0.2 0.3 0.4 0.5 0 50 100 150 200 250 V( 1- P/ P o ) P/Po Maximum P/Po 0 200 400 600 800 0.0 0.2 0.4 0.6 0.8 1.0 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po a b c e d N 2 up ta ke (S TP ) c m 3 g- 1 N 2 up ta ke (S TP ) c m 3 g- 1 N 2 up ta ke (S TP ) c m 3 g- 1 V( 1- P/ P o ) P/ V( P o -P ) P/Po P/Po P/PoP/Po P/Po Maximum SBET = 1063 cm3 g-1 Wo = 0.38 cm3 g-1 Vtot= 1.11 cm3 g-1 xxvi Supplementary Figure 14| TEM images of SnO2 NPs. TEM images of SnO2 NPs with modified synthetic procedures corresponding to SnO2-3 (a – b), SnO2-4 (c – d), SnO2-5 (e – f), SnO2-6 (g – h), SnO2-7 (i – j), SnO2-8 (k – l), SnO2-9 (m – n), SnO2-10 (o – p), SnO2-11 (q – r) and SnO2-12 (s – t). TEM images for SnO2-1/SnO2-2 previously published.12 k l i j m nc d e f g h o p a b q r s t xxvii Supplementary Figure 15| Low magnification STEM elemental mapping of Pd@monolithHKUST-1. a, Electron image of material with dashed white box (inset) indicating the area selected for elemental analysis (b). c – g, Elemental maps of area b corresponding to carbon, copper (L), oxygen, palladium and copper (K) respectively. a b c d e f C K Cu KPd LO K g 1 µm 1 µm 1 µm 1 µm 1 µm 1 µm xxviii Supplementary Figure 16| N2 adsorption isotherms and BET analysis of monolithHKUST-1. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 500 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po 0.0 0.2 0.4 0 100 200 300 400 V( 1- P/ P o ) P/Po maximum 0.000 0.005 0.010 0.015 0.020 0E+00 1E-05 2E-05 3E-05 4E-05 5E-05 6E-05 P/Po P/ V( P o -P ) y = 0.00279754x + 0.00000009 R2 = 0.99999887 1E-07 1E-06 1E-05 1E-04 1E-03 0 100 200 300 400 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po a dc b xxix Supplementary Figure 17| N2 adsorption isotherms and BET analysis of 5% Pd@monolithHKUST-1. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. 0.0 0.2 0.4 0 50 100 150 200 250 300 350 V( 1- P/ P o ) P/Po maximum 0.000 0.005 0.010 0.015 0.020 0E+00 1E-05 2E-05 3E-05 4E-05 5E-05 6E-05 P/ V( P- P o ) P/Po y = 0.00296581x + 0.00000008 R2 = 0.99999913 1E-07 1E-06 1E-05 1E-04 1E-03 0 50 100 150 200 250 300 350 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 500 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po a c b d xxx Supplementary Figure 18| N2 adsorption isotherms and BET analysis of 10% Pd@monolithHKUST-1. N2 gas adsorption isotherms were collected between 0 – 1 bar at 77 K. a, b, N2 adsorption isotherms represented as linear and semi-log plots respectively. c, Determination of maximum P/Po following Rouquerol’s consistency criteria11 and d, BET representation of N2 isotherm showing the linear range utilised in SBET calculations for each monolith and giving also the equation and R2 value that show the linearity of data. 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 N 2 u pt ak e cm 3 ( ST P) g -1 P/Po 0.0 0.2 0.4 0 50 100 150 200 250 300 350 V( 1- P/ P o ) P/Po maximum 0.000 0.005 0.010 0.015 0.020 0E+00 1E-05 2E-05 3E-05 4E-05 5E-05 6E-05 7E-05 P/ V (P -P o) P/Po y = 0.00309319x + 0.0000001 R2 = 0.99999898 1E-07 1E-06 1E-05 1E-04 1E-03 0 100 200 300 N 2 u pt ak e cm 3 ( ST P) g -1 P/Pod b c a xxxi Supplementary Figure 19| Gravimetric CH4 adsorption-desorption isotherms for UiO-66 monoliths. Absolute gravimetric (g g–1) adsorption (filled markers) and desorption (hollow markers) isotherms for a, UiO-66_A (blue triangles); b, UiO-66_B (red diamonds); c, UiO- 66_C (purple squares) and d, UiO-66_D (green circles) represented as absolute uptake between 0 – 100 bar CH4 pressure (298 K). Gravimetric 0 20 40 60 80 100 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0 20 40 60 80 100 0.00 0.05 0.10 0.15 0.20 0.25 0 20 40 60 80 100 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0 20 40 60 80 100 0.00 0.05 0.10 0.15 0.20 0.25 Pressure (bar) Pressure (bar) Pressure (bar)Pressure (bar) CH 4 up ta ke (S TP ) g g -1 CH 4 up ta ke (S TP ) g g -1 CH 4 up ta ke (S TP ) g g -1 CH 4 up ta ke (S TP ) g g -1 CH 4 up ta ke (g g- 1 ) CH 4 up ta ke (g g- 1 ) CH 4 up ta ke (g g -1 ) CH 4 up ta ke (g g -1 ) Pres ure (bar) Pressure (bar) Press r ( r)Pressure (bar) xxxii Supplementary Figure 20| Volumetric CH4 adsorption-desorption isotherms for UiO-66 monoliths. Absolute volumetric (cm3 (STP) cm–3) adsorption (filled markers) and desorption (hollow markers) isotherms for a, UiO-66_A (blue triangles); b, UiO-66_B (red diamonds); c, UiO-66_C (purple squares) and d, UiO-66_D (green circles) represented as absolute uptake between 0 – 100 bar CH4 pressure (298 K). Volumetric 0 20 40 60 80 100 0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 0 20 40 60 80 100 120 140 0 20 40 60 80 100 0 50 100 150 200 250 300 0 20 40 60 80 100 0 50 100 150 200 250 300 Pressure (bar) Pressure (bar) Pressure (bar)Pressure (bar) CH 4 up ta ke (S TP ) v v- 1 CH 4 up ta ke (S TP ) v v- 1 CH 4 up ta ke (S TP ) v v- 1 CH 4 up ta ke (S TP ) v v- 1 Pressure (bar) Pressure (bar) re (bar) Pressure (bar) CH 4 up ta ke (S TP ) c m 3 cm -3 CH 4 up ta ke (S TP ) c m 3 cm -3 CH 4 up ta ke (S TP ) c m 3 cm -3 CH 4 up ta ke (S TP ) c m 3 cm -3 Volumetric a b c d xxxiii Supplementary Figure 21| Gravimetric CO2 adsorption-desorption isotherms for UiO-66 monoliths. Absolute gravimetric (g g–1) adsorption (filled markers) and desorption (hollow markers) isotherms for a, UiO-66_A (blue triangles); b, UiO-66_B (red diamonds); c, UiO- 66_C (purple squares) and d, UiO-66_D (green circles) represented as absolute uptake between 0 – 100 bar CH4 pressure (298 K). 0 10 20 30 40 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 10 20 30 40 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 10 20 30 40 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 10 20 30 40 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Pressure (bar) Pressure (bar) Pressure (bar) CO 2 up ta ke (S TP ) g g -1 CO 2 u pt ak e (S TP ) g g -1 CO 2 u pt ak e (S TP ) g g -1 CO 2 u pt ak e (S TP ) g g -1 Pressure (bar) Pressure (bar) Pressure (bar) Pressure (bar) r ss r ( r) CO 2 up ta ke ( g g-1 ) CO 2 up ta ke (g g -1 ) CO 2 up ta ke (g g- 1 ) CO 2 up ta ke (g g -1 ) xxxiv Supplementary Figure 22| Volumetric CO2 adsorption-desorption isotherms for UiO-66 monoliths. Absolute volumetric (cm3 (STP) cm–3) adsorption (filled markers) and desorption (hollow markers) isotherms for a, UiO-66_A (blue triangles); b, UiO-66_B (red diamonds); c, UiO-66_C (purple squares) and d, UiO-66_D (green circles) represented as absolute uptake between 0 – 100 bar CH4 pressure (298 K). 0 10 20 30 40 0 20 40 60 80 100 120 140 160 0 10 20 30 40 0 20 40 60 80 100 120 140 160 0 10 20 30 40 0 50 100 150 200 250 300 0 50 100 150 200 250 300 0 10 20 30 40 Pressure (bar) Pressure (bar) Pressure (bar) CO 2 u pt ak e (S TP ) v v- 1 CO 2 u pt ak e (S TP ) v v- 1 CO 2 u pt ak e (S TP ) v v- 1 CO 2 up ta ke (S TP ) v v- 1 Pressure (bar) Volumetric CO 2 up ta ke (S TP ) c m 3 cm 3 CO 2 up ta ke (S TP ) c m 3 cm 3 CO 2 up ta ke (S TP ) c m 3 cm 3 CO 2 up ta ke (S TP ) c m 3 cm 3 Pressure (bar) ress re ( ar) Pressure (bar) Pressure (bar) a b c d xxxv III| Supplementary References 1. DeStefano, M. R., Islamoglu, T., Garibay, S. J., Hupp, J. T. & Farha, O. K. Room- Temperature Synthesis of UiO-66 and Thermal Modulation of Densities of Defect Sites. Chem. Mater. 29, 1357–1361 (2017). 2. Bueken, B. et al. Gel-Based Morphological Design of Zirconium Metal-organic Frameworks. Chem. Sci. 8, 3939–3948 (2017). 3. Tian, T., Velazquez-Garcia, J., Bennett, T. D. & Fairen-Jimenez, D. Mechanically and Chemically Robust ZIF-8 Monoliths With High Volumetric Adsorption Capacity. J. Mater. Chem. A 3, 2999–3005 (2015). 4. Mehta, J. P. et al. Sol-Gel Synthesis of Robust Metal-Organic Frameworks for Nanoparticle Encapsulation. Adv. Funct. Mater. 28, 1705588 (2018). 5. Jana, N. R., Gearheart, L. & Murphy, C. J. Seeding Growth for Size Control of 5-40 nm Diameter Gold Nanoparticles. Langmuir 17, 6782–6786 (2001). 6. Rösler, C. et al. Encapsulation of Bimetallic Nanoparticles into a Metal-Organic Framework: Preparation and Microstructure Characterization of Pd/Au@ZIF-8. Eur. J. Inorg. Chem. 2014, 5514–5521 (2014). 7. Mehta, J. P. Incorporation of Metal Oxide Nanoparticles in a Monolithic Metal-Organic Framework for Applications in Photocatalysis. (The University of Cambridge, 2016). 8. Wang, L. & Yamauchi, Y. Strategic Synthesis of Trimetallic Au@Pd@Pt Core-Shell Nanoparticles from Poly(vinylpyrrolidone)-Based Aqueous Solution toward Highly Active Electrocatalysts. Chem. Mater. 23, 2457–2465 (2011). 9. Tian, T. et al. A Sol–Gel Monolithic Metal–Organic Framework with Enhanced Methane Uptake. Nat. Mater. 17, 174–179 (2018). 10. Li, G. et al. Hydrogen Storage in Pd Nanocrystals Covered with a Metal–Organic Framework. Nat. Mater. 13, 802–806 (2014). 11. Gómez-Gualdrón, D. A., Moghadam, P. Z., Hupp, J. T., Farha, O. K. & Snurr, R. Q. Application of Consistency Criteria to Calculate BET Areas of Micro- and Mesoporous Metal-Organic Frameworks. J. Am. Chem. Soc. 138, 215–224 (2016).