On Molybdenum Sulfides and Other Active Materials for Sustainable Energy Systems

A thesis with publication submitted to fulfil requirements for the degree of

Doctor of Philosophy

Jyah Strachan

Laboratory of Advanced Catalysis for Sustainability

Faculty of Science

The University of Sydney

2020

Preface

Preface Abstract To respond to the approaching climate crisis, the current energy landscape must shift towards sustainable, decarbonised systems. This shift will require the development of inexpensive and active energy conversion materials. The work within this thesis reports the investigation of several candidate materials, i.e. molybdenum sulfides and carbides, for use as catalysts and electrodes in energy conversion processes. These materials were chosen for their natural abundance, controllable morphologies, and varied chemistries.

Chapter 1 focuses on the scope and potential of Chevrel phases (MxMo6S8) as catalytic materials. It includes a critical literature review that emphasises the unique features of Chevrel Phase catalysts and summarises the published catalytic reports. This survey highlights the underutilisation of Chevrel Phases as catalysts, laying the foundation for further investigations into the synthesis and tuning of highly active Chevrel Phase morphologies. These studies yielded nanoparticulate catalysts that exhibited excellent performance for the hydrogen evolution reaction (a current of 10mAcm-2 at 0.23 V and a Tafel slope of 68 mVdec-1).

Chapter 2 focuses on the various morphologies of MoS2 and incorporates two reviews submitted for publication: the first review summarises all known reports of 3R-MoS2. The contradictory findings and nomenclature inconsistencies within the literature are clarified. The second review rectifies the errors in the literature on hydrothermally produced 1T-MoS2 and provides best practice analysis instructions for researchers to prevent future mistakes. The final section of Chapter 2 introduces new results of an investigation into the effect of crystal disorder on the electrochemical performance of MoS2 in hybrid batteries in order to aid the optimisation of MoS2 electrodes.

Chapter 3 is comprised of a study that elucidates the structure of a highly active lignin valorisation catalyst,

Mo2CxNx-1/TiN. A wide range of characterisation techniques was employed, including X-ray absorption spectroscopy, alongside microscopic and diffraction-based techniques, to show that the titanium nitride and molybdenum carbide phases are intimately mixed and that the catalyst incorporates terminal Mo≡N motifs.

Chapter 4 includes two published studies on the silver catalysed reduction of 4-nitrophenol; a ubiquitous model reaction. The first involves a detailed kinetic analysis of the reaction mechanism to elucidate the role of oxygen during the induction period of the reaction. The research-level chemistry and protocols were then adapted to be used as a pedagogical tool to teach nanoscience and kinetics to undergraduate students.

i Preface

Statement of Originality The content of this thesis is my own work. This thesis has not been submitted for any degree or other purposes. I certify that the intellectual content of this thesis is the product of my own work and that all the assistance received in preparing this thesis and relevant sources have been acknowledged. In addition to the statements above, in cases where I am not the corresponding author of a published item, permission to include the published material has been granted by the corresponding author.

Jyah Strachan, 20th December 2020

The work and results presented in this thesis are primarily my own, with the input of collaborators as below. Additionally, this thesis contains work which has been accepted or submitted for publication as follows:

Chapter 1:

All the work within this chapter was designed, conducted, and written by me with supervision from Anthony F Masters (AFM) and Thomas Maschmeyer (TM). Section 1.1 (The Catalytic Nature of Chevrel Phases in Review) has been published in the Materials Research Bulletin. Section 1.2 (Chevrel Phase Nanoparticles as Electrocatalysts for Hydrogen Evolution) has been published in ACS Applied Energy Materials.

Chapter 2:

All the work within this chapter was designed, conducted, and written by me with supervision from AFM and

TM. Section 2.1 (3R-MoS2 in Review: History, Status, and Outlook) has been accepted for publication in ACS

Applied Energy Materials. Section 2.2 (Critical Review: Hydrothermal Synthesis of 1T-MoS2 – an Important Route to a Promising Material) has been published in the Journal of Materials Chemistry A. These two sections (2.1 and 2.2) contain reproduced figures, used with permission, from articles as referenced. Section 2.3

(Influence of Crystal Disorder in MoS2 Cathodes for Secondary Hybrid Mg-Li Batteries) was designed, conducted, and written by Lu Chen and me with supervision from AFM and TM. Chapter 2.3 is in the final stages of preparation for submission to the Australian Journal of Chemistry.

Chapter 3:

This work was the continuation of a study by Alexander Yuen (AY) and co-workers. The EXAFS experiment was designed by all authors, conducted by Stuart A. Bartlett (SAB), AY, Bernt Johannessen, and me, and the data analysed by SAB. With the exception of the microscopy (for which Hongwei Liu aided in sample analysis), the remaining experiments were collaboratively designed by SAB, AY, and me, then conducted by me. The resulting manuscript was drafted by all authors. This work has been submitted to Physical Chemistry Chemical Physics and is currently under peer review.

Chapter 4:

The work in this chapter was designed, conducted, and written by Christopher D Barnett and me with supervision and input from AFM and TM. Section 4.1 has been published in ACS Catalysis.

ii Preface

Section 4.2 was designed, conducted, and written by Christopher D Barnett, Alice Motion and me, with supervision and input from AY, AFM, and TM. It has been published in the Journal of Chemical Education.

Appendix:

The appendix contains supplementary information to the chapters above.

As supervisor for the candidature upon which this thesis is based, I can confirm that the authorship attribution statements above are correct.

Professor Anthony Masters, 1st January 2021

Professor Thomas Maschmeyer, 20th December 2020

iii Preface

Acknowledgements Tony and Thomas, thank you for the mountain of support that you have provided to guide me through the wild journey that the last few years have been. Thanks for smiling while I described the latest equipment malfunction or failed experiment. The hour at which you responded to my anxious emails never ceased to amaze me. You are both fantastic role models and I consider my time with you both to have been invaluable.

Alex and Chris, thanks for dealing with my nonsense day in, day out. You’re both brilliant, and great friends. Thanks for fielding my infinite questions and putting in time to make sure this thesis looked as good as possible.

To the lab members (Alfonso, Ellen, Eddie, the Lisas, Grace, Jake, Kevin, Sid, Gelion and others), thank you for the support, encouragement, and mischief. This PhD wouldn’t have been nearly as fun without you. To the GOST lab (Alvise, Maurizio, Manu, Ale, Robe, Carlotta and others) and Casa Primus (Alvise, Alix, Amanda), thank you for welcoming me into your families with open arms.

SURMC, thank you for putting your faith in me and for being such a supportive community of misfits. Special mention to Sean, Tess, Matt, Mimes, Alex, and Jose for the friendship and adventures, and to Dec and Jess for your love and aggressive hospitality.

To my friends at home: Jono, Jess, Alex and crew, thank you for letting me disappear for months on end and welcoming me back every time. I always appreciate your patience and love.

To Rhiannon, thank you for your unwavering support through such a turbulent time in your life. I appreciate all that you gave and continue to give. This adventure truly couldn’t have been possible without you. To all the Griens, thank you for bringing me into your family – I put in the time; I know you’ll be proud!

Finally, thank you to Mum, Dad, Indi, Shay, Eva, and all the Suttons and Strachans. Without two and a half decades of your support, I never would have achieved anything close to this. Thank you for always believing in me and facilitating this endless education of mine. I probably might stop collecting degrees now. I love you all.

I acknowledge the Gadi of the Eora Nation, the traditional custodians of the land upon which this work took place. I pay respect to those who have cared and continue to care for Country.

iv Preface

Table of Contents

Preface ...... i Introduction ...... 1 Chapter 1.1 ...... 4 The Catalytic Nature of Chevrel Phases in Review Chapter 1.2 ...... 18 Chevrel Phase Nanoparticles as Electrocatalysts for Hydrogen Evolution Chapter 2.1 ...... 29

3R-MoS2 in Review: History, Status, and Outlook Chapter 2.2 ...... 49

Critical Review: Hydrothermal Synthesis of 1T-MoS2 – an Important Route to a Promising Material Chapter 2.3 ...... 64

Influence of Crystal Disorder in MoS2 Cathodes for Secondary Hybrid Mg-Li Batteries Chapter 3 ...... 74 Elucidation of Structure and Support Interactions for a Highly Active Molybdenum Carbo- nitride@Titanium Nitride Hydrodeoxygenation Catalyst Chapter 4.1 ...... 89 4-Nitrophenol Reduction: Probing the Putative Mechanism of the Model Reaction Chapter 4.2 ...... 99 Nanoparticles for Undergraduates: Creation, Characterization, and Catalysis Conclusions ...... 108 Appendix to Chapter 1.2 ...... 113 Appendix to Chapter 2.3 ...... 127 Appendix to Chapter 3 ...... 130 Appendix to Chapter 4.1 ...... 141 Appendix to Chapter 4.2 ...... 160

v

Introduction The earth is facing a climate crisis due to the energy demands of a growing, developing global population.1 To mitigate the effects of climate change, a paradigm shift towards sustainable systems is necessary.1 One such shift is the decarbonisation of the energy sector, which involves a transition from fossil fuels to renewable energy systems. The most significant scientific challenge of this goal is the development of suitable energy vector systems.2 While electricity may be produced from renewable sources, energy conversion and storage are necessary if the electrical energy is not used where and when it is produced.2 Therefore, the work within this thesis is united by the overarching aim to further the development of renewable energy systems. The following introduction serves to broadly introduce these systems; further detail is given in the respective chapters to both avoid redundancy and maintain the integrity of the work included within.

Hydrogen and Catalysis The first energy vector discussed is hydrogen, which may be produced renewably by the electrolysis of water – a process that converts electrical energy to chemical potential energy in the form of hydrogen gas.2 If the electricity source is renewable, then the process can be considered both sustainable and carbon neutral.3 The reaction efficiency of the energy conversion process (i.e. the ratio of the input electrical energy to stored chemical potential energy) may be enhanced by employing a catalyst.4 The development of these catalysts presents a substantial roadblock to the realisation of a viable hydrogen economy.2

Several categories of catalyst exist, however the stability and ease of recovery make heterogeneous catalysts the most suited to industrial applications.5 Ideally the catalytic material will be active, inexpensive, stable, and produced safely with minimal environmental and social impact.6 It is by these metrics that molybdenum sulfides were chosen to be studied. The catalytic behaviour of Chevrel Phases (MxMo6S8) is reviewed in Chapter 1.1, followed by an investigation into the synthesis and performance of unsupported, nanoparticulate Chevrel Phase morphologies used as hydrogen evolution reaction catalysts, Chapter 1.2. Finally, the catalytic behaviour of exotic MoS2 polytypes is reviewed in Chapter 2.

Electrochemical Energy Storage The second form of energy discussed in this thesis is that which is stored in electrochemical devices; the simplest of which are the cells that constitute batteries. These cells convert chemical potential energy to electrical energy by harnessing the energy released as a chemical system moves towards equilibrium.7 In practice, they consist of an electrolyte that conducts charge between two electrodes; which for reversible (i.e. rechargeable) systems are often typically composed of materials able to accommodate cyclic intercalation–deintercalation of charged ions.7 The factors that describe an optimal electrode material are similar to that of a catalyst: expense; durability; and capacity are of primary importance, with kinetics and gravimetric & volumetric density being of concern in certain applications.8 Molybdenum sulfides are promising materials due to their ability to intercalate ions and their low cost.9-10 One of these materials, specifically MoS2, is tested as intercalation cathodes in Li+ batteries in Chapter 2.3.

Renewable Hydrocarbon Production The final energy vector discussed in this thesis is represented by hydrocarbons, specifically those produced from renewable sources. One such source is biomass-derived lignin: a waste product of the paper industry and the most underutilised fraction of lignocellulose.11 Lignin is a complex macromolecule comprised of irregular polyphenolic units.12 If the long, oxygenated chains of lignin are both cracked and deoxygenated, a product stream similar to that of conventional fossil fuels can be fed directly into existing petrochemical infrastructure.13

Introduction

To selectively produce the desired, short-chain, deoxygenated products, an appropriate catalyst must be chosen.11 Molybdenum carbides show exceptional activity and selectivity for the valorisation of lignin into the desired hydrocarbons, however a structure–activity relationship has not yet been established for these catalysts. Chapter 3 examines the structure of the most active lignin valorisation catalyst studied to date: Mo2C1−xNx@TiN.

Catalytic Model Reactions As exemplified in the foregoing, catalysts are a critical component on the path towards efficient use and interconversion of energy vectors as well as of chemicals. To accelerate the research and development of new catalysts, model reactions can be used which simplify the assessment of a catalyst in a suitably analogous reaction. Model reactions may be designed to occur in a faster, safer, and more easily evaluated manner than the corresponding real-world application. They also facilitate the consistent benchmarking of catalysts across the literature.

The most widely used model reaction for hydrogenation in aqueous solutions is the reduction of 4-nitrophenol to the corresponding aminophenol.14 This reaction lacks a robust kinetic mechanism and is solely employed as a phenomenological model. Chapter 4.1 describes the investigation into the kinetic mechanism of this model in order to more fully establish the reaction as a reliable and accepted model in contemporary catalysis. Chapter 4.2 reports the design of an undergraduate laboratory practical based on the above chemistry and intended to excite and fascinate talented first year undergraduates in their introduction to tertiary study.

Thesis Organisation The information within this thesis is conceptually organised in themes, whereby the framework of sustainability guided research into the individual energy conversion systems. These systems in turn are investigated using multiple materials (MoS2, Mo6S8, molybdenum carbo-nitrides, etc). The above introduction has been structured in this manner. However, to improve the logical and narrative flow within the thesis body, the following chapters are organised by material. This is illustrated diagrammatically in Figure 1. Supplementary information to the chapters is contained in the appendix.

Figure 1 – Thesis structure, as organised by material.

2 Introduction

References 1. IPCC Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Geneva, Switzerland, 2014; p 151. 2. Staffell, I.; Scamman, D.; Abad, A. V.; Balcombe, P.; Dodds, P. E.; Ekins, P.; Shah, N.; Ward, K. R., The role of hydrogen and fuel cells in the global energy system. Energy & Environmental Science 2019, 12 (2), 463-491. 3. Turner, J. A., Sustainable hydrogen production. Sci 2004, 305 (5686), 972-974. 4. Zou, X. X.; Zhang, Y., Noble metal-free hydrogen evolution catalysts for water splitting. Chem. Soc. Rev. 2015, 44 (15), 5148- 5180. 5. Schlögl, R., Heterogeneous Catalysis. Angew. Chem. Int. Ed. 2015, 54 (11), 3465-3520. 6. Anastas, P. T.; Kirchhoff, M. M.; Williamson, T. C., Catalysis as a foundational pillar of green chemistry. Applied Catalysis A: General 2001, 221 (1), 3-13. 7. Bard, A. J.; Faulkner, L. R., Fundamentals and applications. Electrochemical Methods 2001, 2 (482), 580-632. 8. Larcher, D.; Tarascon, J.-M., Towards greener and more sustainable batteries for electrical energy storage. Nature Chem. 2015, 7 (1), 19-29. 9. Aurbach, D.; Lu, Z.; Schechter, A.; Gofer, Y.; Gizbar, H.; Turgeman, R.; Cohen, Y.; Moshkovich, M.; Levi, E., Prototype systems for rechargeable magnesium batteries. Natur 2000, 407, 724. 10. Nitta, N.; Wu, F.; Lee, J. T.; Yushin, G., Li-ion battery materials: present and future. Mater. Today 2015, 18 (5), 252-264. 11. Cattelan, L.; Yuen, A. K. L.; Lui, M. Y.; Masters, A. F.; Selva, M.; Perosa, A.; Maschmeyer, T., Renewable Aromatics from Kraft Lignin with Molybdenum-Based Catalysts. ChemCatChem 2017, 9 (14), 2717-2726. 12. Xu, C.; Arancon, R. A. D.; Labidi, J.; Luque, R., Lignin depolymerisation strategies: towards valuable chemicals and fuels. Chem. Soc. Rev. 2014, 43 (22), 7485-7500. 13. Ragauskas, A. J.; Beckham, G. T.; Biddy, M. J.; Chandra, R.; Chen, F.; Davis, M. F.; Davison, B. H.; Dixon, R. A.; Gilna, P.; Keller, M., Lignin valorization: improving lignin processing in the biorefinery. Sci 2014, 344 (6185). 14. Aditya, T.; Pal, A.; Pal, T., Nitroarene reduction: a trusted model reaction to test nanoparticle catalysts. Chem. Commun. 2015, 51 (46), 9410-9431.

3 Chapter 1

Chapter 1.1 The Catalytic Nature of Chevrel Phases in Review

This chapter has been published in the Materials Research Bulletin as Strachan, J.; Masters, A. F.; Maschmeyer,

T., The Catalytic Nature of Chevrel Phases (MxMo6S8) in Review. Materials Research Bulletin 2021, 139, 111286.

Abstract

Chevrel Phases, MxMo6S8, are a class of identifies trends and gaps within the literature, and heterogenous catalysts comprised of abundant provides perspective on future directions for the materials that have been reported to out-perform catalytic application of Chevrel Phases. In doing so, conventional MoS2 catalysts in several reactions. we hope to bring these promising materials to the Historically, their lengthy and energy-intensive attention of a wider audience to facilitate syntheses have restricted the use of these materials improvements in energy & chemical conversion as catalysts, however recent developments in CP processes. synthesis techniques, driven by energy storage research, have lessened these constraints. The catalytic community has been slow to adopt these advances, yet Chevrel Phases remain excellent candidates for fundamental catalytic reactions due to their unique combination of low-coordinate molybdenum atoms; tuneable oxidation state; the presence of electron-reservoirs within the of Mo6- clusters; thermal stability, and dual Lewis acidity & basicity. This review provides a concise, first summary of five decades of catalytic reports,

History and Motivation In response to the current climate crisis, new energy conversion systems are being developed, and the efficiency of existing reactions improved. Green Chemistry advocates for the use of catalytic materials to ensure chemical transformations take place selectively and with minimal energy input.1 Noble metals are frequently the most suited for these tasks, however the scarcity and related expense of these metals provides challenges.2 Molybdenum, on the other hand, is a relatively earth-abundant and comparatively cheap transition metal.3

Molybdenum sulfides are among the most active non-noble metal catalysts.4 MoS2 catalysts are well established; however the catalytic behaviour of a related class of molybdenum sulfides, the Chevrel Phases, has not yet been fully investigated.

Chevrel Phases (CPs; MxMo6X8, where M = most transition metals and X = a chalcogenide), or ternary molybdenum chalcogenides, possess several interesting behaviours that distinguish them from other non-noble metal materials: facile metal intercalation enables the use of CPs as electrodes5 and the band structure can allow superconductivity.6-7 In the context of catalysis, intrinsic properties such as the low coordination of molybdenum ions,8 tuneable oxidation state,9 the reservoir of Mo6-cluster electrons,10 thermal stability,11 and the potential for multi-site (Mo and S) substrate adsorption12 mean that CPs are generally highly active materials.

4 Chapter 1

CPs were first discovered in 1971 by Chevrel and co-workers.13 They attracted substantial attention throughout the 1970s and ’80s due to the superconductivity of several derivatives, then again in the 1980s and ’90s due to their catalytic potential. Interest waned in the late-‘90s as their lengthy and energy-intensive syntheses limited potential commercial use. In 2000, Aurbach and co-workers showed that CPs are excellent candidates for energy storage as Mg2+ ion intercalation cathodes.5 Since then, substantial progress has been made in the quest to overcome limitations of the conventional synthesis procedure. This was primarily driven by research into CP-based batteries. These improvements, however, have not yet fully benefitted and led to advances in CP- based catalysts. This may be because the literature spans five decades, dozens of catalytic reactions, and several disparate advances in synthetic procedures, making it at times prohibitively difficult to navigate. Collection and organisation of existing reports might serve to accelerate the full realisation of CP-based catalysis.

This review aims to: highlight recent developments in the synthesis of CPs; survey the entire catalytic CP- literature to identify trends and provide a baseline for future studies; summarise key reports to facilitate efficient progress of the literature; apply these insights to identify knowledge-gaps within the literature, and suggest where further study might be directed. To that end, this review is structured into four sections: a general description of Chevrel Phases; a critical review of developments in synthetic procedures; an exhaustive list of catalytic reports; and a perspective on the future directions of the field.

Chevrel Phases Chevrel Phases consist of six-atom octahedral molybdenum clusters enclosed within distorted cubes of chalcogenides that form a 3D-channelled framework (Figure 2). The Mo6X8 motifs are tightly bound (the Mo– Mo bond length is close to that in metallic Mo; 2.65–2.80 and 2.72 A respectively)14 and well-separated, i.e. they can be considered weakly-coupled, discrete units.6 The chalcogenide ions are located at the face of each triangle of the octahedron such that the surface of these crystals is an unsaturated pseudo-square pyramidal molybdenum centre with four in-plane chalcogenide ligands and an uncoordinated site at the apex of the pyramid.15 The metal M can be one of 40+ transition metals intercalated within the channels of the Chevrel

Phase.6, 16 These metals donate electrons to the Mo6X8 cluster, which can be considered as 20–24 electrons per octahedron or as an average molybdenum oxidation state of +2.00 to +2.66.17

Figure 2 – Mo6X8 crystal structure, showing Mo6 octahedral clusters (grey polyhedra) and chalcogen atoms (white spheres) which form distorted cubes and channels.

5 Chapter 1

Electrons within the Mo6 cluster are delocalised and only weakly coupled to adjacent clusters due to the large inter-cluster distance. The density of states at the Fermi level is large and consists primarily of Mo 4d and S 3d orbitals.18 These states are narrow due to the weak inter-cluster coupling and are separated from the conduction band by a HOMO–LUMO gap of ~1 eV.19-21 Electrons from the intercalated metal are donated into S 3p orbitals, which stabilises the cluster and increases the reactivity of the S p-bands.15

While CPs were first discovered (and are conventionally understood) as containing Mo6 clusters, several elements may substitute for Mo-atoms within the clusters (a more general CP formula that includes these pseudobinaries is MxMo(6-y)M′yXz, where M′ = Re, Rh, or Ru, X = a chalcogenide, and y ≤ 6; however MxMo6S8 will be used here for simplicity, as is commonplace).22 The non-Mo metal determines the valence electron concentration of the clusters.23 Furthermore, the unit cell parameters may be altered by intercalating variously sized metal ions or by replacing the conventional sulfur with a larger chalcogen (such as selenium or tellurium).22, 24 Intercalation can be achieved by post-synthetic treatment, however substitution of Mo-atoms for Re, Rh, or Ru has only been achieved during CP formation.22

Syntheses The first Chevrel Phases were produced using a lengthy, high-temperature solid state synthesis that generates large, crystalline, low surface-area materials. These properties are undesirable if the activity of the catalyst is to be maximised. Improved synthetic procedures have been developed to address these issues such that nanoparticulate samples may be synthesised at moderate temperatures in < 24 h. A critical review and historical perspective of the strategies that have been employed to overcome this limitation follows.

Solid-State Synthesis In the solid-state synthetic procedure that was conventionally used prior to the late 1980s, a stoichiometric amount of each element is ground together, sealed in an evacuated quartz tube, then heated to ~1000 °C for several days.25 Alongside the molybdenum and sulfur sources, an extra metal (usually copper) must be added into the reactant mixture. This is because Mo6S8 is electron deficient, therefore metastable and unable to be synthesised directly without the donation of two additional electrons.15, 26-27 The copper is oxidised by the nascent Mo6S8 cluster and in the process, a Cu+ ion is intercalated within the channels between the clusters.

Studies using X-ray diffraction have reported that CP synthesis occurs from MoS2 & Cu0 intermediates.28-29 We note that several common precursors (CuS and S) are known to decompose and/or evaporate during heating, which presents an explosion hazard if sealed tubes are used.30-31 If the reaction is performed in an open system, then a sub-stoichiometric reactant ratio may result. Wakihara, et al. 32 have determined the single-phase region of CuxMo6S8 as a function of stoichiometry, which is summarised in Figure 3.33 If the ‘empty’ CP, Mo6S8, is desired, electro- or chemical oxidation must be performed to remove the copper.34 This process typically involves treating the sample with HCl and air to produce [CuCl4]2− + H2O + H+ or bulk electrolysis at ~400 mV (vs NHE).34-35 These topotactic reactions do not impact the structure of the CP, though slight alteration to the surface of the crystals has been observed.36 If large ions are oxidatively deintercalated, then strain may cause the CP crystals to fragment, resulting in a smaller average particle size.37

6 Chapter 1

Figure 3 – (left) CP phase diagram as a function of precursor stoichiometry, CuxMo6Sy. Solid lines indicate regions where product distributions have been empirically determined by Yamamoto, et al. 38, dashed lines are extrapolations. (right) High-resolution transmission electron micrograph of CP sheet; adapted with permission from Mao, et al. 30, copyright 2020 American Chemical Society.

Improved Procedures To decrease the temperature, duration, and energy input of the conventional solid state synthesis, several alternative methods have been developed. Lancry, et al. 31 reported a molten salt method whereby precursor powders are ground with an inexpensive salt that acts as a liquid flux as the salt melts. This method has the added advantage of enabling less reactive or volatile precursors, such as MoS2 to be used instead of sulfur. Gershinsky, et al. 39 synthesised CPs in under 20 min through the use of a self-propagating high-temperature synthesis where precursor powders were ground, loaded into a quartz tube under argon, then combusted. The heat released from the exothermic reaction of Mo + S + Cu sustained a sufficient temperature to rapidly yield CPs. Boursicot, et al. 40 were able to synthesise CP powders and films in 6 h by calcination of Mo and Cu salts followed by sulfurisation at 600 °C. Saha, et al. 41 were able to synthesise CPs in less than 30 min if the precursor powders were first subject to high energy mechanical milling. The intimate mixing of the reactants in these methods decreases the diffusion pathway, accelerating the reaction.42 Flukiger, et al. 43 found that by increasing the temperature and pressure of the synthesis to 1700 °C and 100 bar, CPs could be produced in minutes. Similarly, Pantou, et al. 44 and Seghir, et al. 45 were able to decrease the heating duration to 2 h by either hot or cold pressing pellets composed of elemental precursor mixtures.

With the goal of maximising the surface area, several methods have been applied to decrease particle size. The earliest development was the discovery of soluble precursors of CPs, which can be dissolved, mixed with a support, dried, then heated to form supported-CPs. Nanjundaswamy, et al. 46 first developed Mx(NH4)yMo3S9 (M = Cu, Pb, La, or Gd) precursors, then Rabiller-Baudry, et al. 47 later applied the wetness impregnation technique using alumina to produce supported-CPs. The soluble precursor does not need to include all three elements (e.g. Cu, Mo, and S), instead the cation (here, Cu) is commonly added as a separate salt, e.g. CuClx +

(NH4)2MoS4.29, 47 A pure, stoichiometric Cu2Mo6S8 product is obtained as excess sulfur is removed as H2S under a reducing atmosphere and the counter ions as neutral gases.29

An alternative to the soluble precursor route was developed by Schrader and co-workers48 whereby high surface area, amorphous Mo6S8 clusters are formed by reaction of Mo6Cl12 with NaSH and NaOBu in butanol. The samples, which the authors denote as MMoS (where M = Na, Pt, La etc), are strictly ligated salts (e.g.

Na2x(Mo6S8)Sx) that are themselves catalytically active and can be converted to crystalline CPs by heat

7 Chapter 1 treatment.49 Unfortunately, upon heating, the surface area decreases by an order of magnitude due to the pore collapse as the stabilising solvent molecules are removed from them.49

To synthesise thin films, vapour transport or sputtering techniques may be employed.7 Mao, et al. 30 used iodine vapor transport to synthesise sheets of CPs that were tens of nanometres thick, micrometres in width, and showed excellent electrochemical performance due to their dimensions (Figure 3, right). Schewemiller, et al. 42 synthesised CPs by radio frequency, direct current, and reactive sputtering using a range of substrates and targets (Pb, Mo, MoS2, sapphire etc).

Low Temperature Syntheses Several reports claim that CPs were produced at temperatures lower than ~200 °C.50-51 These reports often lack evidence to substantiate their claims. To our knowledge, ~450 °C is the lowest temperature of CP synthesis for which crystallographic (or similarly convincing) data is provided.52 However the CPs in the report by 52 Gershinsky, et al. co-nucleated with MoS2, resulting in an impure product. Perhaps the lowest temperature possible for a pure CP product is that reported by Boursicot and co-workers at ~600 °C.40 It appears moderately high temperatures (500–1000 °C) are required to overcome the activation energy of CP formation, and that the duration of this heat treatment may be decreased if the precursor elements are in intimate contact (as is also the case for molecular precursors).

In cases where the structure of the materials synthesised is unclear (e.g. if crystallites are too small or disordered to be detected by powder X-ray diffraction – as is common in low temperature syntheses of molybdenum sulfides), X-ray absorption spectroscopy is the analytical technique of choice for product analysis – as performed by Rabiller-Baudry, et al. 47 and others15, 26, 53.

Stability Chevrel Phases are reasonably stable in air, though slowly oxidise over the course of months. The oxidation involves incorporation of oxygen atoms around the intercalated metal atoms to form hydroxido ligands.11 Micron-sized CP particles have been shown to incorporate up to 5 wt.% oxygen,11 however the effect of oxidation on samples with very large surface areas has not been established. The material can be deoxygenated by heating under a stream of H2 at ~250 °C.11, 32 The stability of CPs under catalytic conditions is discussed below.

Catalytic Reports and Behaviour Chevrel Phases possess a several intrinsic properties that make them well-suited to a range of catalytic reactions. In the following section, common trends in catalytic reports are identified and discussed. Specific examples are given to highlight how the aforementioned characteristics enable the catalytic performance of CPs. In addition, a summary of notable catalytic results is given in Table 1.

There exists many more examples of catalysis performed by CPs, however these studies did not report the activity of the CP relative to a common standard, thus, were not compiled in Table 1. Many of these omitted reports were performed by the Kamiguchi group and are discussed in the Scope and Selectivity of Clusters section.

Hydrodesulfurisation Chevrel Phase catalysts have showed promising performance in a several reactions where noble metals typically excel. The first report of a catalytic reaction performed by CPs was hydrodesulfurisation (HDS) – a critical industrial process that removes organosulfur species in crude oil feeds, which minimises the amount of SOx

8 Chapter 1 pollution released during fuel combustion.54 With the tightening of air-quality regulations and the depletion of cleaner petroleum feeds, this is an increasingly difficult task.42

Conventionally ‘CoMo/Al2O3’, or cobalt-doped MoS2 on alumina, catalysts are employed, however the lower- and variable-valence of the Mo atoms in CPs improve their catalytic activity for organosulfur removal:53 CPs exhibit higher HDS activity, selectivity, and stability than CoMo catalysts.14, 55 As they are partly comprised of sulfur atoms, they are not deactivated on stream, unlike platinum group catalysts.9 Furthermore, the tuneable nature of CPs allows for reaction optimisation – the catalytic activity for HDS is proportional to both the Mo- oxidation state and identity of intercalated ions within CPs.9, 17 The CPs intercalated by larger ions (Ho, Pb, Sn, Ag, and In) exhibited higher activity and resistance to surface oxidation than did the small-ion CPs (Cu, Ni, Fe, and Co).9, 17 McCarty, et al. 9 propose that the mobility of the smaller ions allows them to retreat into the bulk structure, decreasing the stability of the surface sites.9 The presence of MoS2 in Raman spectra and a change to the surface M/Mo ratio in X-ray photoelectron spectra of used catalysts supported this hypothesis.9 Catalytic HDS activity increases with intercalant concentration (i.e. inversely with oxidation state) across a homometallic series.56-57 These results are consistent with model studies where lower oxidation state metal catalytic sites exhibited increased surface–thiophene interactions.57 Kareem and Miranda 56 note that HDS activity further increases with the lability of the CP chalcogen (i.e. S is most active, followed by Se, then Te) – the authors ascribed this behaviour to the increased concentration of coordinatively unsaturated sites caused by the more mobile chalcogens at the surface.

The isotropic and symmetric nature of CP crystals results in very few unique catalytic sites, thus selectivity (of

HDS over hydrogenation and isomerisation) is improved relative to that of CoMo58 or MoS259 (for which there are several binding modes).9, 14, 42 By immobilising CPs on alumina or titania supports, HDS activity can be further enhanced.55 It is unclear whether the enhancement is due to a synergistic CP–support effect or simply an increase in CP surface area.

Electrocatalysis Alonso-Vante and co-workers10 performed the electrocatalytic oxygen reduction (ORR) and hydrogen evolution reactions (HER) with CPs. The activity of CPs approached that of platinum in acidic electrolytes. The authors note that the ‘reservoir’ of cluster electrons allows for multi-electron transfer because the change to oxidation state during catalysis has little influence on the Fermi level.10 The potential window in which intercalated-CPs are stable depends on the intercalated metal and restricts their use as electrocatalysts at positive potentials.10, 60

The HER activity of CPs (as determined by the Tafel slope and exchange current density) is influenced by the identity of the intercalated metal and is greater for ‘empty’ CPs than for intercalated CPs.10 The activity relationships are more pronounced for ORR than for HER – Alonso-Vante, et al. 10 propose that this is because intercalated metals are active catalytic species and that there is less need for cooperative interactions between the two metal centres in HER.61 These authors observe that the ORR activity increases in pseudobinary CPs

(i.e. MxMo(6-y)M′yXz) because the number of electrons in the Mo6 cluster increased, the metal–metal bonds become polarised, and there are more distinct adsorption sites.10, 60, 62 Where the number of electrons is equal, the ORR activity correlated with the size of the intercalated ion – the origin of this relationship is likely to be the expansion of lattice parameters as a function of ion size, which in turn changes the sites upon which catalysis takes place.10 Interestingly, this relationship is less pronounced for HER. Alonso-Vante, et al. 10 propose that there must be less interaction between surface intermediates during the HER mechanism than there are during the ORR mechanism.

9 Chapter 1

Scope and Selectivity of Clusters The Kamiguchi group has extensively tested both CPs (which they call solid-state molybdenum sulfide clusters) and molecular Mo6S8 clusters in a multitude of catalytic reactions.11-12, 63 The reactions generally relate to de/hydrogenation, alkylation, and ring opening processes. Kamiguchi, et al. 64 determined that the Mo6 clusters are activated above ~400 °C (Figure 4). The activation process involves the desaturation of a surface Mo site (Figure 4). These unsaturated sites are Lewis acidic and are active for hydrogenation,65 cracking64 etc. The neighbouring sulfur site is basic, thus the CP surface acts as a bifunctional acid–base catalyst.66-67 The bifunctionality increases the specificity with which substrates bind to the surface, resulting in a high catalytic selectivity. In the case of the dehydrogenation of 1-butanol (Figure 4), the selectivity for butyraldehyde was

91% and dehydration to alkenes and ethers was very low (a 3–10 fold improvement in selectivity over MoS2 catalysts).63 A review of the molecular M6S8 clusters reported by Kamiguchi and co-workers has been published previously.68

Figure 4 – (a) Catalytic cycle of alcohol dehydrogenation with a CP catalyst.63, 66 (b) Schematic of the ligand effect, showing electron donation

from the intercalated ion to the Mo6-cluster. (c) Schematic of the ensemble effect – CO2 adsorption step of methanol synthesis on an unmodified67 and adatom-modified69 CP surface.

The amorphous, Mo6S8-cluster containing materials synthesised by Schrader and co-workers can be understood as an intermediary between CPs and molecular Mo6S8 clusters. These materials exhibit much higher surface area than crystalline CPs and are comprised of the same cluster motifs.49, 70 The materials may be stabilised by several transition metal counter ions; Sn, La, Ho, Pt, and Na have been reported.70-71 They have been tested as catalysts in HDS for which they show improved conversion of thiophene relative to CPs, though when normalised for surface area are out-performed by CPs – an explanation for this behaviour was not given.

In Silico Mechanism Investigations Finally, Liu and co-workers have extensively studied the catalytic activity of single transition metal atoms supported on Mo6S8 clusters using DFT.8, 20, 69, 72-73 Their calculations show that the catalytic behaviour of the cluster is significantly influenced by the transition metal modifier. This may be via an indirect reduction of the Mo centre (a ligand effect, Figure 4b) or by direct participation of the metal modifier in the catalytic reaction (an ensemble effect, Figure 4c). The contribution of these effects depends on the choice of transition metal: in less electronegative metals, such as potassium, the ligand effect is strongest, however in all cases the ensemble effect dominates. The modifier metals can also vary the reaction pathway, such as between the reverse water- gas shift or formate pathways to methanol synthesis.72, 74 For a comparison of the reaction pathways for methanol synthesis on transition metal-modified and bare Mo6S8 clusters, see Liu and Liu 69 and Liu, et al. 67,

10 Chapter 1 respectively. The ensemble effect has recently been experimentally observed by Perryman, et al. 15, corroborating the computational reports.

Future Directions Remarks on Established Reactivities and Morphologies To facilitate the efficient progress of research into Chevrel Phase catalysis, it is critical that studies be performed using best practice methodologies. Of the currently published catalytic studies, few employ a common, accessible benchmark as a comparison catalyst (i.e. only those listed in Table 1). We therefore encourage future studies to include an appropriate standard (this may be commercial MoS2 or Pt/C, etc, depending on the conventions of the reaction in focus).

A yet unexplored extension of the computational literature on catalytic CPs is the study of CP–support interactions and the implications for catalysis, particularly for common supports such as Al2O3 and carbon.

Similar studies using MoS2 have shown that catalyst–support interactions influence substrate binding energies (and thus the catalytic activity of the material).75 Given that molecular octahedral clusters may be immobilised on solid supports,76 mechanistic studies for HDS, HER, ORR, and hydrogenation using these materials are a priority. Furthermore, we note that the computational studies in the current literature solely involve adatom modifiers on the surface of a single Mo6S8 cluster – while acknowleding the computational expense, the computation of more realistic models, whereby the modifier metal is incorporated in the inter-cluster channels would be extremely valuable.

Several of the computational reports (vide supra) identify that CPs are suitable as supports for single atom catalysts (which are recently reviewed by Mitchell and Pérez-Ramírez 77). In similar systems, such as on a MoS2 support, single atom Fe has been used for N2 electroreduction;78 Rh for selective hydrogenation of unsaturated aldehydes to unsaturated alcohols;79 and Co for biomass hydrodeoxygenation.80 Using controlled partial reduction, or partial oxidative leaching, it is possible to prepare Chevrel Phases with substoichiometric amounts of intercalated metal ions. These metal ions should be present as single ions, which might be converted to atoms, or enter a catalytic cycle as ions. For example, Su, et al. 78 prepared MoS2-supported single Fe sites by immersing MoS2 in ethanolic FeCl3 at 60 °C, generating single site FenSxOy (0 < n < 3) moieties at ~ 5 wt% loading. A related electrochemical procedure might be used to produce FexMo6S8, with x chosen to ensure the generation of isolated single Fe sites. Thus intercalated CPs may be an efficient way to produce support single atom catalysts.

There are several exotic morphologies of Mo6S8 which have been synthesised, though not yet been tested for their catalytic activity. These include condensed cluster oligomers81 and one-dimensional wires;82 both of which maximise the number of catalytically available Mo-sites on the Mo6 octahedra relative to bulk samples. The incorporation of the Mo6S8 cluster into novel frameworks (akin to metal-organic frameworks) has not yet been realised. By using conjugated linkers we expect that catalytically active, conductive materials might be produced.

The synthesis of quantum dots and similar unsupported nanoparticles may be possible using molecular Mo6S8 precursors. However, these morphologies likely require the development of lower-temperature synthetic routes.

Although substantial progress has been made since the initial reports by Chevrel and coworkers13, the inflexibility of current syntheses continues to hinder the widespread use of CPs as catalysts. Control of particle size and shape as well as surface facets and defects will inevitably lead to significant improvements to the performance of CP catalysts. We therefore submit that the development of alternative synthetic methods remains the greatest goal in this field. Further improvements to the temperature requirements of syntheses

11 Chapter 1 would likely have beneficial flow-on effects in the adjacent fields of energy storage and photonics where morphological control is equally paramount.

Chevrel Phases and Trends in Catalysis There still exist several unexplored avenues for further fundamental, applied, and computational inquiry. In general, we note that the unique features of CP catalysts have not been widely exploited, thus the scope of their utilisation remains unknown. We expect that the dual acid/base sites may yet yield promising results as catalytic sites for one-pot multistep sequential reactions, especially those that rely on acid-base bifunctionalities (e.g. dehydrohalogenation followed by aromatic alkylation), as reviewed by Corma and co-workers.83 Furthermore, due to the thermal stability and sulfur tolerance of CPs, we expect them to perform well in high-temperature catalytic reactions, such as combustion reactions in gas turbine combustors.84

Within the broader context of modern catalysis, the evolution of the energy and chemical sectors is constantly creating new challenges. For example, the transition from traditional refinery- and petrochemical-based processes to biomass transformation and renewable energy conversion systems (e.g. HER) means that new catalysts must be designed to meet the needs of these processes.85 Additionally, the electrification of (traditionally) thermally driven processes requires that catalysts operate within a wide electrochemical window and, for pharma and fine chemicals especially, at near-ambient conditions. If the drawbacks of current syntheses can be overcome, we anticipate that CPs will be good candidates to replace ‘mature’ catalysts in several modern catalytic reactions.

The Haber-Bosch process accounts for 2% of global energy consumption and, after fossil fuel refining, generates the most CO2 of the chemical manufacturing industry.86-87 Electrification is one path to improving the energy efficiency of this process. Martín, et al. 86 report that the largest challenge to large-scale electrocatalytic nitrogen reduction is the development of a suitable electrocatalytic system. In this system again, the selectivity of the catalyst of key. Side reactions (e.g. HER) can be minimised by appropriate solvent selection, though hydrogen gas formation continues to reduce the efficiency of this reaction. MoS2 can outperform most reported non-noble metal catalysts for this reaction,88 though the activity of CPs is yet to be fully realised.

Lanzafame, et al. 85 identified catalyst selectivity and stability as the “main challenges” of catalytic materials in biomass conversion processes. These authors also note that catalysts should be developed that are tolerant of mixed biomass-petroleum feeds and able to efficiently perform fluid catalytic cracking, hydrocracking, and hydrotreating. Other emerging industrial processes, such as electrocatalytic CO2 reduction (e.g. to acetic acid), HER, and ORR, require catalysts that are stable under high temperatures and pressures.85 The selectivity- inducing acid-base bifunctional active sites of CPs, alongside their sulfur tolerance and thermal stability make them promising materials for these applications.

Environmental issues caused by the longevity of everyday plastic waste are an increasing concern. Low-carbon emission solutions to their decomposition are not yet mature and the energy efficiency of many of the more promising solutions (e.g. pyrolysis and solvolysis) must still be improved.89 Vollmer, et al. 89 have identified that the development of active, selective, and stable catalysts is a key challenge to the commercialisation of these processes. While CPs have not yet been tested as catalysts for these conversion processes, similarly robust catalysts such as MoS2 and Mo2C have shown great promise.90

Conclusions Chevrel Phases are intrinsically active, tailorable, selective catalysts for numerous fundamental reactions including HDS, HER, ORR, and C–C bond formation. This review shows that their unique octahedral Mo6S8

12 Chapter 1 cluster allows for efficient multielectron transfer and gives rise to a high concentration of bifunctional, Lewis acid/base active sites. We propose that the length and energy intensity of conventional syntheses has precluded widespread utilisation of this highly active material, although that recent developments have improved the synthesis and therefore the extrinsic catalytic performance of CPs. Still, the activity of these materials will continue to be outperformed by established catalysts (such as MoS2) until significant improvements to synthesis procedures are made. We identify this obstacle as the greatest barrier to the field and proposed several possible routes to improvement. Whether by systematic surveys of the myriad of tunable CP synthesis parameters or by step-wise improvements to synthesis strategies, there remains abundant space for further development of CP catalysts.

This review summarised the syntheses, notable characteristics, and catalytic reports of Chevrel Phase catalysts to provide a baseline from which future studies can build. From within this summary, knowledge gaps were identified, and past and future trends were highlighted. It is hoped that by collecting, organising, and analysing the scattered Literature reports of Chevrel Phase catalysts we stimulate further research into this promising material.

13 Chapter 1

References 1. Anastas, P. T.; Kirchhoff, M. M.; Williamson, T. C., Catalysis as a foundational pillar of green chemistry. Applied Catalysis A: General 2001, 221 (1), 3-13. 2. Jaffe, R.; Price, J.; Ceder, G.; Eggert, R.; Graedel, T.; Gschneidner, K.; Hitzman, M.; Houle, F.; Hurd, A.; Kelley, R., Energy critical elements: securing materials for emerging technologies. American Physical Society, Materials Research Society, Washington, DC 2011, 63 (8), 33-34. 3. Haynes, W. M., Abundance of elements in the earth’s crust and in the sea. CRC handbook of chemistry and physics. Taylor & Francis 2014, 14-18. 4. Jaramillo, T. F.; Jørgensen, K. P.; Bonde, J.; Nielsen, J. H.; Horch, S.; Chorkendorff, I., Identification of active edge sites for electrochemical H2 evolution from MoS2 nanocatalysts. Sci 2007, 317 (5834), 100-102. 5. Aurbach, D.; Lu, Z.; Schechter, A.; Gofer, Y.; Gizbar, H.; Turgeman, R.; Cohen, Y.; Moshkovich, M.; Levi, E., Prototype systems for rechargeable magnesium batteries. Natur 2000, 407, 724. 6. Fischer, Ø., Chevrel phases: superconducting and normal state properties. Appl. Phys. A: Mater. Sci. Process. 1978, 16 (1), 1-28. 7. Peña, O., Chevrel phases: Past, present and future. Physica C (Amsterdam, Neth.) 2015, 514, 95-112. 8. Zheng, X.; Guo, L.; Li, W.; Cao, Z.; Liu, N.; Zhang, Q.; Xing, M.; Shi, Y.; Guo, J., Insight into the Mechanism of Reverse Water-gas Shift Reaction and Ethanol Formation Catalyzed by Mo6S8-TM Clusters. Molecular Catalysis 2017, 439, 155-162. 9. McCarty, K. F.; Anderegg, J. W.; Schrader, G. L., Hydrodesulfurization catalysis by Chevrel phase compounds. J. Catal. 1985, 93 (2), 375-387. 10. Alonso-Vante, N.; Schubert, B.; Tributsch, H., Transition metal cluster materials for multi-electron transfer catalysis. Mater. Chem. Phys. 1989, 22 (3), 281-307. 11. Kamiguchi, S.; Takeda, K.; Kajio, R.; Okumura, K.; Nagashima, S.; Chihara, T., Application of Solid-State Molybdenum Sulfide Clusters with an Octahedral Metal Framework to Catalysis: Ring-Opening of Tetrahydrofuran to Butyraldehyde. J. Cluster Sci. 2013, 24 (2), 559-574. 12. Kamiguchi, S.; Arai, K.; Okumura, K.; Iida, H.; Nagashima, S.; Chihara, T., Solid-state molybdenum sulfide clusters with an octahedral metal framework as hydrogenation, dehydrogenation, and hydrogenolysis catalysts similar to the platinum group metals. Applied Catalysis A: General 2015, 505, 417-421. 13. Chevrel, R.; Sergent, M.; Prigent, J., Sur de nouvelles phases sulfurées ternaires du molybdène. J. Solid State Chem. 1971, 3 (4), 515-519. 14. McCarty, K. F.; Schrader, G. L., Hydrodesulfurization by reduced molybdenum sulfides: activity and selectivity of Chevrel phase catalysts. Industrial & Engineering Chemistry Product Research and Development 1984, 23 (4), 519-524. 15. Perryman, J. T.; Ortiz-Rodriguez, J. C.; Jude, J. W.; Hyler, F. P.; Davis, R. C.; Mehta, A.; Kulkarni, A. R.; Patridge, C. J.; Velazquez, J. M., Metal-promoted Mo6S8 clusters: a platform for probing ensemble effects on the electrochemical conversion of CO2 and CO to methanol. Materials Horizons 2020, 7 (1), 193-202. 16. Caillat, T.; Fleurial, J.-P.; Snyder, G., Potential of Chevrel phases for thermoelectric applications. Solid State Sci. 1999, 1 (7-8), 535-544. 17. Afanasiev, P.; Bezverkhyy, I., Ternary transition metals sulfides in hydrotreating catalysis. Applied Catalysis A: General 2007, 322, 129-141. 18. Naik, K. M.; Sampath, S., Cubic Mo6S8-Efficient Electrocatalyst Towards Hydrogen Evolution Over Wide pH Range. Electrochim. Acta 2017, 252, 408-415. 19. Chevrel, R.; Hirrien, M.; Sergent, M., Superconducting Chevrel Phases - Prospects and Perspectives. Polyhedron 1986, 5 (1-2), 87-94. 20. Cao, Z.; Guo, L.; Liu, N.; Zheng, X.; Li, W.; Shi, Y.; Guo, J.; Xi, Y., Theoretical study on the reaction mechanism of reverse water-gas shift reaction using a Rh-Mo6S8 cluster. RSC Advances 2016, 6 (110), 108270-108279. 21. Hamard, C.; Peña, O.; Le Floch, M., Solid state 95Mo NMR studies of the Chevrel-phase solid solution Mo6Se8−xTex. Solid State Commun. 2000, 113 (9), 489-494. 22. Perrin, A.; Chevrel, R.; Sergent, M.; Fischer, O., Synthesis and Electrical-Properties of New Chalcogenide Compounds Containing Mixed (Mo,Me)6 Octahedral Clusters (Me=Ru or Rh). J. Solid State Chem. 1980, 33 (1), 43-47. 23. Yvon, K.; Paoli, A., Charge transfer and valence electron concentration in Chevrel phases. Solid State Commun. 1977, 24 (1), 41- 45. 24. Levi, E.; Aurbach, D., Chevrel Phases, MxMo6T8 (M = Metals, T = S, Se, Te) as a Structural Chameleon: Changes in the Rhombohedral Framework and Triclinic Distortion. Chem. Mater. 2010, 22 (12), 3678-3692. 25. Perrin, A.; Perrin, C.; Chevrel, R., Chevrel Phases: Genesis and Developments. In Ligated Transition Metal Clusters in Solid-State Chemistry: The Legacy of Marcel Sergent, Halet, J. F., Ed. 2019; Vol. 180, pp 1-30. 26. Wan, L. F.; Wright, J.; Perdue, B. R.; Fister, T. T.; Kim, S.; Apblett, C. A.; Prendergast, D., Revealing electronic structure changes in Chevrel phase cathodes upon Mg insertion using X-ray absorption spectroscopy. PCCP 2016, 18 (26), 17326-17329. 27. Afanasiev and Bezverkhyy (App Cat A: General 322 (2007) 129–141) note that reduction of MoS2 under hydrogen at temperature results in metallic Mo, not Mo6S8. 28. Cheng, Y.; Parent, L. R.; Shao, Y.; Wang, C.; Sprenkle, V. L.; Li, G.; Liu, J., Facile synthesis of Chevrel phase nanocubes and their applications for multivalent energy storage. Chem. Mater. 2014, 26 (17), 4904-4907. 29. Rabiller-Baudry, M.; Sergent, M.; Chevrel, R., Convenient syntheses of chevrel phase compounds from soluble sulfide precursors under flowing hydrogen atmosphere. Mater. Res. Bull. 1991, 26 (6), 519-526.

14 Chapter 1

30. Mao, M. L.; Lin, Z. J.; Tong, Y. X.; Yue, J. M.; Zhao, C. L.; Lu, J. Z.; Zhang, Q. H.; Gu, L.; Suo, L. M.; Hu, Y. S.; Li, H.; Huang, X. J.; Chen, L. Q., Iodine Vapor Transport-Triggered Preferential Growth of Chevrel Mo6S8 Nanosheets for Advanced Multivalent Batteries. ACS Nano 2020, 14 (1), 1102-1110. 31. Lancry, E.; Levi, E.; Mitelman, A.; Malovany, S.; Aurbach, D., Molten salt synthesis (MSS) of Cu2Mo6S8—new way for large- scale production of Chevrel phases. J. Solid State Chem. 2006, 179 (6), 1879-1882. 32. Wakihara, M.; Hinode, H.; Inoue, C., Decomposition of NO Using Chevrel-Phase Sulfides. Solid State Ionics 1992, 53, 413-417. 33. Cheung, K. Y.; Steele, B. C. H., Homogeneity range of copper molybdenum sulphide CuxMo6S8−y. Mater. Res. Bull. 1980, 15 (12), 1717-1725. 34. Schöllhorn, R.; Kümpers, M.; Besenhard, J. O., Topotactic redox reactions of the channel type chalcogenides Mo3S4 and Mo3Se4. Mater. Res. Bull. 1977, 12 (8), 781-788. 35. Lancry, E.; Levi, E.; Gofer, Y.; Levi, M.; Salitra, G.; Aurbach, D., Leaching Chemistry and the Performance of the Mo6S8 Cathodes in Rechargeable Mg Batteries. Chem. Mater. 2004, 16 (14), 2832-2838. 36. Murgia, F.; Antitomaso, P.; Stievano, L.; Monconduit, L.; Berthelot, R., Express and low-cost microwave synthesis of the ternary Chevrel phase Cu2Mo6S8 for application in rechargeable magnesium batteries. J. Solid State Chem. 2016, 242, 151-154. 37. Rabiller, P.; Rabiller-Baudry, M.; Even-Boudjada, S.; Burel, L.; Chevrel, R.; Sergent, M.; Decroux, M.; Cors, J.; Maufras, J., Recent progress in chevrel phase syntheses: A new low temperature synthesis of the superconducting lead compound. Mater. Res. Bull. 1994, 29 (5), 567-574. 38. Yamamoto, S.; Matsui, K.; Wakihara, M.; Taniguchi, M., Stable range of copper molybdenum sulfide CuxMo6S8−y and measurement of the superconducting critical temperature Tc. Mater. Res. Bull. 1983, 18 (11), 1311-1316. 39. Gershinsky, G.; Haik, O.; Salitra, G.; Grinblat, J.; Levi, E.; Daniel Nessim, G.; Zinigrad, E.; Aurbach, D., Ultra fast elemental synthesis of high yield copper Chevrel phase with high electrochemical performance. J. Solid State Chem. 2012, 188 (Supplement C), 50-58. 40. Boursicot, S.; Bouquet, V.; Péron, I.; Guizouarn, T.; Potel, M.; Guilloux-Viry, M., Synthesis of Cu2Mo6S8 powders and thin films from intermediate oxides prepared by polymeric precursor method. Solid State Sci. 2012, 14 (6), 719-724. 41. Saha, P.; Jampani, P. H.; Datta, M. K.; Hong, D.; Gattu, B.; Patel, P.; Kadakia, K. S.; Manivannan, A.; Kumta, P. N., A rapid solid-state synthesis of electrochemically active Chevrel phases (Mo6T8; T= S, Se) for rechargeable magnesium batteries. Nano Research 2017, 1-21. 42. Schewemiller, I. M.; Koo, K. F.; Columbia, M.; Li, F.; Schrader, G. L., Synthesis and Characterization of Lead Chevrel-Phase Thin-Films for Hydrodesulfurization Catalysis. Chem. Mater. 1994, 6 (12), 2327-2335. 43. Flukiger, R.; Devantay, H.; Jorda, J.; Muller, J., Metallurgical and physical properties of ternary Molybdenum Sulfides (MxMo3S4) as synthetized in the bulk state. IEEE Trans. Magn. 1977, 13 (1), 818-820. 44. Pantou, R.; Guilloux-Viry, M.; Burel, L.; Jegaden, J. C.; Chevrel, R.; Perrin, A.; Merdrignac-Conanec, O.; Lecroc, M.; Guyader, J., Hot pressing sintered CuxMo6S8 targets for laser ablation thin films deposition. Solid State Sci. 1999, 1 (7), 647-656. 45. Seghir, S.; Stein, N.; Boulanger, C.; Lecuire, J. M., Electrochemical determination of the diffusion coefficient of cations into Chevrel phase-based electrochemical transfer junction by potential step chronoamperometry and impedance spectroscopy. Electrochim. Acta 2011, 56 (6), 2740-2747. 46. Nanjundaswamy, K.; Vasanthacharya, N.; Gopalakrishnan, J.; Rao, C., Convenient synthesis of the Chevrel phases metal molybdenum sulfide, MxMo6S8 (M= copper, lead, lanthanum or gadolinium). Inorg. Chem. 1987, 26 (25), 4286-4288. 47. Rabiller-Baudry, M.; Chevrel, R.; Sergent, M., Syntheses of bulk and supported Chevrel phases. J. Alloys Compd. 1992, 178 (1), 441-445. 48. Paskach, T. J.; Schrader, G. L.; McCarley, R. E., Synthesis of Methanethiol from Methanol over Reduced Molybdenum Sulfide Catalysts Based on the Mo6S8 Cluster. J. Catal. 2002, 211 (2), 285-295. 49. Thompson, R. K.; Hilsenbeck, S. J.; Paskach, T. J.; McCarley, R. E.; Schrader, G. L., Pretreatment of new reduced ternary molybdenum sulfide catalysts. J. Mol. Catal. A: Chem. 2000, 161 (1), 75-87. 50. Fan, H. F.; Huang, J.; Chen, G. L.; Chen, W.; Zhang, R.; Chu, S. J.; Wang, X. Q.; Li, C. R.; Ostrikov, K. K., Hollow Ni-V-Mo Chalcogenide Nanopetals as Bifunctional Electrocatalyst for Overall Water Splitting. ACS Sustainable Chemistry & Engineering 2019, 7 (1), 1622-1632. 51. Hong, X. D.; Liu, Q.; Iocozzi, J.; Gong, C.; Kong, L. Q.; Liu, X. Y.; Ye, M. D.; Lin, Z. Q., Needle-Leaf-Like Cu2Mo6S8 Films for Highly Efficient Visible-Light Photocatalysis. Particle & Particle Systems Characterization 2018, 35 (1). 52. Gershinsky, G.; Haik, O.; Salitra, G.; Grinblat, J.; Levi, E.; Nessim, G. D.; Zinigrad, E.; Aurbach, D., Ultra fast elemental synthesis of high yield copper Chevrel phase with high electrochemical performance. J. Solid State Chem. 2012, 188, 50-58. 53. Kochubey, D.; Rogov, V.; Babenko, V., Low-temperature synthesis of supported hydrodesulfurization catalysts based on chevrel phases. Kinet. Catal. 2009, 50 (2), 270-274. 54. Babich, I.; Moulijn, J., Science and technology of novel processes for deep desulfurization of oil refinery streams: a review☆. Fuel 2003, 82 (6), 607-631. 55. Ooi, S.; Zhang, H.; Hinode, H., The hydrodesulfurization activity and characterization of cobalt Chevrel phase sulfides. React. Kinet. Catal. Lett. 2004, 82 (1), 89-95. 56. Kareem, S. A.; Miranda, R., Hydrodesulfurization catalysis over ternary molybdenum chalcogenides. J. Mol. Catal. 1989, 53 (2), 275-283. 57. Ekman, M. E.; Anderegg, J. W.; Schrader, G. L., Reduced Molybdenum Formal Oxidation-States in Hydrodesulfurization Catalysis by Chevrel Phases. J. Catal. 1989, 117 (1), 246-257. 58. Šarić, M.; Rossmeisl, J.; Moses, P. G., Modeling the active sites of Co-promoted MoS2 particles by DFT. PCCP 2017, 19 (3), 2017-2024.

15 Chapter 1

59. Tsai, C.; Chan, K.; Abild-Pedersen, F.; Nørskov, J. K., Active edge sites in MoSe2 and WSe2 catalysts for the hydrogen evolution reaction: a density functional study. PCCP 2014, 16 (26), 13156-13164. 60. Fischer, C.; Alonso-Vante, N.; Fiechter, S.; Tributsch, H., Electrocatalytic properties of mixed transition metal tellurides (Chevrel-phases) for oxygen reduction. J. Appl. Electrochem. 1995, 25 (11), 1004-1008. 61. Schubert, B.; Gocke, E.; SchoÈllhorn, R.; Alonso-Vante, N.; Tributsch, H., In situ X-ray-electrochemical studies on the origin of H2O2 production during oxygen reduction at transition metal cluster materials. Electrochim. Acta 1996, 41 (9), 1471-1478. 62. Schubert, B.; Vante, N. A.; Gocke, E.; Tributsch, H., Oxygen Reduction Electrocatalysis by Chevrel Phase Sulfides Supported on Carbon Paste Electrodes. Berichte der Bunsengesellschaft für physikalische Chemie 1988, 92 (11), 1279-1283. 63. Kamiguchi, S.; Nagashima, S.; Chihara, T., Application of solid-state early-transition metal clusters as catalysts. Tetrahedron Lett. 2018, 59 (14), 1337-1342. 64. Kamiguchi, S.; Seki, Y.; Satake, A.; Okumura, K.; Nagashima, S.; Chihara, T., Catalytic Cracking of Methyl tert-Butyl Ether to Isobutene over Bronsted and Lewis Acid Sites on Solid-state Molybdenum Sulfide Clusters with an Octahedral Metal Framework. J. Cluster Sci. 2015, 26 (3), 653-660. 65. Kamiguchi, S.; Kajio, R.; Yamada, H.; Yuge, H.; Okumura, K.; Iida, H.; Nagashima, S.; Chihara, T., Thermal Activation of Solid- State Molybdenum Halide Clusters with an Octahedral Cluster Framework and Their Application to Catalytic Synthesis of 3- Methylpyridine from Piperidine and Methanol. Bull. Chem. Soc. Jpn. 2015, 88 (8), 1116-1122. 66. Kamiguchi, S.; Okumura, K.; Nagashima, S.; Chihara, T., Catalytic dehydrogenation of alcohol over solid-state molybdenum sulfide clusters with an octahedral metal framework. Mater. Res. Bull. 2015, 72, 188-190. 67. Liu, P.; Choi, Y.; Yang, Y.; White, M. G., Methanol Synthesis from H2 and CO2 on a Mo6S8 Cluster: A Density Functional Study. The Journal of Physical Chemistry A 2010, 114 (11), 3888-3895. 68. Nagashima, S.; Kamiguchi, S.; Chihara, T., Catalytic Reactions over Halide Cluster Complexes of Group 5-7 Metals. Metals 2014, 4 (2), 235-313. 69. Liu, C.; Liu, P., Mechanistic Study of Methanol Synthesis from CO2 and H-2 on a Modified Model Mo6S8 Cluster. Acs Catalysis 2015, 5 (2), 1004-1012. 70. Hilsenbeck, S. J.; McCarley, R. E.; Goldman, A. I.; Schrader, G. L., Hydrodesulfurization activity and EXAFS characterization of novel ternary tin and lanthanum molybdenum sulfide catalysts. Chem. Mater. 1998, 10 (1), 125-134. 71. Hilsenbeck, S. J.; McCarley, R. E.; Thompson, R. K.; Flanagan, L. C.; Schrader, G. L., Metal cluster hydrodesulfurization catalysts based on ternary lanthanum molybdenum sulfides. Journal of Molecular Catalysis a-Chemical 1997, 122 (1), 13-24. 72. Liu, C.; Liu, P., Mechanistic Study of Methanol Synthesis from CO2 and H2 on a Modified Model Mo6S8 Cluster. ACS Catalysis 2015, 5 (2), 1004-1012. 73. Liu, P.; Choi, Y.; Yang, Y. X.; White, M. G., Methanol Synthesis from H-2 and CO2 on a Mo6S8 Cluster: A Density Functional Study. J. Phys. Chem. A 2010, 114 (11), 3888-3895. 74. Zhang, H. T.; Liu, C.; Liu, P.; Hu, Y. H., Mo6S8-based single-metal-atom catalysts for direct methane to methanol conversion. J. Chem. Phys. 2019, 151 (2). 75. Tsai, C.; Abild-Pedersen, F.; Nørskov, J. K., Tuning the MoS2 Edge-Site Activity for Hydrogen Evolution via Support Interactions. Nano Lett. 2014, 14 (3), 1381-1387. 76. Kumar, S.; Khatri, O. P.; Cordier, S.; Boukherroub, R.; Jain, S. L., Graphene oxide supported molybdenum cluster: first heterogenized homogeneous catalyst for the synthesis of dimethylcarbonate from CO2 and methanol. Chemistry-A European Journal 2015, 21 (8), 3488-3494. 77. Mitchell, S.; Pérez-Ramírez, J., Single atom catalysis: a decade of stunning progress and the promise for a bright future. Nature Communications 2020, 11 (1), 1-3. 78. Su, H.; Chen, L.; Chen, Y.; Si, R.; Wu, Y.; Wu, X.; Geng, Z.; Zhang, W.; Zeng, J., Single Atoms of Iron on MoS2 Nanosheets for N2 Electroreduction into Ammonia. Angew. Chem. Int. Ed. 2020. 79. Lou, Y.; Zheng, Y.; Li, X.; Ta, N.; Xu, J.; Nie, Y.; Cho, K.; Liu, J., Pocketlike Active Site of Rh1/MoS2 Single-Atom Catalyst for Selective Crotonaldehyde Hydrogenation. J. Am. Chem. Soc. 2019, 141 (49), 19289-19295. 80. Liu, G.; Robertson, A. W.; Li, M. M.-J.; Kuo, W. C. H.; Darby, M. T.; Muhieddine, M. H.; Lin, Y.-C.; Suenaga, K.; Stamatakis, M.; Warner, J. H.; Tsang, S. C. E., MoS2 monolayer catalyst doped with isolated Co atoms for the hydrodeoxygenation reaction. Nature Chem. 2017, 9 (8), 810-816. 81. Chevrel, R.; Gougeon, P.; Potel, M.; Sergent, M., Ternary Molybdenum Chalcogenides - A Route to New Extended Clusters. J. Solid State Chem. 1985, 57 (1), 25-33. 82. Chevrel, R.; Potel, M.; Sergent, M.; Decroux, M.; Fischer, O., One-Dimensional Condensation of Mo6 Octahedral Clusters - A New Cluster, Mo12, and a New Building Block, Mo12S14, In M2Mo9S11. J. Solid State Chem. 1980, 34 (2), 247-251. 83. Climent, M. J.; Corma, A.; Iborra, S., Heterogeneous catalysts for the one-pot synthesis of chemicals and fine chemicals. Chem. Rev. 2011, 111 (2), 1072-1133. 84. Arai, H.; Machida, M., Thermal stabilization of catalyst supports and their application to high-temperature catalytic combustion. Applied Catalysis A: General 1996, 138 (2), 161-176. 85. Lanzafame, P.; Perathoner, S.; Centi, G.; Gross, S.; Hensen, E., Grand challenges for catalysis in the Science and Technology Roadmap on Catalysis for Europe: moving ahead for a sustainable future. Catalysis Science & Technology 2017, 7 (22), 5182-5194. 86. Martín, A. J.; Shinagawa, T.; Pérez-Ramírez, J., Electrocatalytic Reduction of Nitrogen: From Haber-Bosch to Ammonia Artificial Leaf. Chem 2019, 5 (2), 263-283. 87. Bicer, Y.; Dincer, I.; Zamfirescu, C.; Vezina, G.; Raso, F., Comparative life cycle assessment of various ammonia production methods. Journal of Cleaner Production 2016, 135, 1379-1395. 88. Zhang, L.; Ji, X.; Ren, X.; Ma, Y.; Shi, X.; Tian, Z.; Asiri, A. M.; Chen, L.; Tang, B.; Sun, X., Electrochemical Ammonia Synthesis via Nitrogen Reduction Reaction on a MoS2 Catalyst: Theoretical and Experimental Studies. Adv. Mater. 2018, 30 (28), 1800191.

16 Chapter 1

89. Vollmer, I.; Jenks, M. J.; Roelands, M. C.; White, R. J.; van Harmelen, T.; de Wild, P.; van der Laan, G. P.; Meirer, F.; Keurentjes, J. T.; Weckhuysen, B. M., Beyond Mechanical Recycling: Giving New Life to Plastic Waste. Angew. Chem. Int. Ed. 2020. 90. Cattelan, L.; Yuen, A. K. L.; Lui, M. Y.; Masters, A. F.; Selva, M.; Perosa, A.; Maschmeyer, T., Renewable Aromatics from Kraft Lignin with Molybdenum-Based Catalysts. ChemCatChem 2017, 9 (14), 2717-2726.

17 Chapter 1

Chapter 1.2 Chevrel Phase Nanoparticles as Electrocatalysts for Hydrogen Evolution

This chapter has been published in ACS Applied Nano Materials as Strachan, J.; Masters, A. F.; Maschmeyer, T., Chevrel Phase Nanoparticles as Electrocatalysts for Hydrogen Evolution. ACS Applied Nano Materials 2021, 4 (2), 2030-2036.

Abstract

Chevrel Phases (MxMo6S8) are a class of than an equivalent MoS2 phase with similar molybdenum chalcogenide materials that are nanostructure. Furthermore, the nanostructured attractive candidates for active non-noble metal Chevrel Phases prove to be easily modified by catalysts due to the relatively low coordination of electrochemical intercalation, which allows their molybdenum moieties. Conventionally, the performance fine-tuning; revealing a new family of lengthy and energy intensive syntheses of Chevrel versatile and tuneable catalysts. Phases produce highly crystalline, low surface area materials. In this work, a novel synthetic approach leading to a Chevrel Phase with unprecedented nanostructure is presented. The resultant material is fully characterised using a variety of spectroscopic, microscopic, and electrochemical techniques. In electrochemical testing aimed at catalysing the hydrogen evolution reaction, this nanostructured catalyst shows a substantially lower overpotential

Introduction Hydrogen is gaining increasing significance as a vector to transport and store the energy generated from renewable sources.1 It has both a high molar energy density and a benign combustion product.2 Its current, mostly fossil-based, origin can be viewed as a transition step towards the phasing in of its renewable generation,3 based on the hydrogen evolution reaction (HER): 2H(aq)+ + 2e− ⇌ H2(g).2 In the absence of a catalyst, the HER requires a prohibitively large overpotential for most applications.4 Although noble metal electrocatalysts can substantially reduce this overpotential, they are too expensive for commercialisation at the large scales necessary to displace fossil-based approaches. Therefore, non-noble metal electrocatalysts such as MoS2 have received intensive attention.4 Unsaturated Mo-sites of molybdenum sulfides (edge-sites in MoS2, MoS3, amorphous

MoSx, clusters, etc.) have a near-zero DFT-calculated Gibbs free energy of adsorbed hydrogen and can in principle be expected to display a HER activity close to that of noble metals.5

MoS2 is an attractive alternative to Pt as a catalyst for the HER, as it is approximately three orders of magnitude cheaper, and the bulk material has an HER activity only 57 times less than has platinum. The basal planes

(Figure 5, right) of the thermodynamically stable 2H-MoS2 are inactive as the phase is semiconducting and the electron transfer kinetics of the fully coordinatively saturated Mo atoms limited,6-7 leaving only edge-sites to participate in the catalysis.5 Notably, edge-sites make up only a small fraction of crystalline MoS2, as the layered morphology leads to the surface area being dominated by the dimensions of the basal plane.5 A number of strategies to increase the proportion of edge-sites by preparing non-laminated single layer or de-laminated few-

18 Chapter 1 layer, nanoparticulates have been reported in the literature, including our work on ionic-liquid-based nanostructuring.8-11 Our strategy in making delaminated nanoparticulate MoS2 has been to use ionic liquids to lightly stabilise the high energy surfaces of the growing nanoparticle, thus inhibiting aggregation/lamination, restricting the size to nanoparticulate dimensions and generating a high concentration of active sites.

Despite the success of this approach, further optimisation, producing variants of MoS2 (other than doped derivatives) is limited and the majority of molybdenum atoms are still coordinatively saturated. Accordingly, we sought to explore whether this approach might by applied to other molybdenum/sulfur materials. One obvious choice is the class of Chevrel Phases (CPs).

Chevrel Phases (CPs) are molybdenum sulfides that, importantly, do not form layered structures12,13-14 but, instead, exhibit coordinatively unsaturated Mo atoms (Figure 5) on every crystal face. In a sense, most molybdenum sites are ‘edge sites’ according to the prior discussion. Due to these structural differences at the atomic level, CPs might reasonably be expected to outperform MoS2 as HER catalysts.

The CP unit cell is comprised of an octahedral “cluster” of molybdenum atoms inside a distorted chalcogen cube (herein designated a ‘Mo6S8 unit’; Figure 5).15 Each unit is linked to six others by S–S bonds, with the consequence that empty, distorted, S8 cubes surround each face of a Mo6S8 unit. These units are arranged into a framework solid with channels running in 3-dimensions along the axes of the rhombohedral unit cell (Figure 5). The channels can be empty or occupied; most group I/II and transition metals – as well as some cations and solvent molecules – can be intercalated.16

Figure 5 – Structures of selected molybdenum sulfides, Mo shown in purple, S shown in yellow. Images highlighting the octahedron-in-cube

‘Mo6S8 unit’ (left), channelled Mo6S8 structure (centre; Mo atoms replaced with polyhedra for clarity) and the edge sites of three layers of

layered 2H-MoS2 (right).

Regrettably, widespread use of CPs has been hampered so far due to their lengthy and energy-intensive synthesis. Temperatures in excess of 1000 °C and dwell times of up to 7 days are common.13 Although incremental improvements have lessened the temperatures and times required, current high temperature syntheses produce highly crystalline solids with low surface areas; characteristics that are undesirable for optimal catalytic activity.17 This temperature constraint means that the smallest reported particles are of the micro- scale.18-21 New unsupported, nanostructured materials are a key catalyst target that has as yet not been reported in the literature.22

Given our previous success in producing unsupported, catalytic MoS2 by an ionic liquid mediated solvothermal method, we sought to adapt this strategy towards CP synthesis. Of additional interest was the influence of the necessary high temperature reaction step on the morphology of the particles. A support-free method was

19 Chapter 1 chosen to maximise the available surface area of the catalysts and allow for flexibility in choice of support if so desired.

Previously, nanoparticulate CPs have been isolated on a variety of supports.23 We thus, faced four questions. Can our ionic liquid synthetic route be adapted to generate nanopariculate CPs? Can the ionic liquid provide sufficient stabilisation that the CPs can be made unsupported (and so, e.g., allow a well-defined CP to be added to a variety of supports)? Can a variety of CPs be prepared by this route? Would such unsupported nanoparticulate CPs function as superior catalysts (using, in this case, as perhaps the simplest test reaction, the HER)?

Here, we provide the answers to these four questions and are able to report such a desirable novel nanostructured Chevrel Phase, including its synthesis as well as its physical characterisation and performance evaluation (HER electrocatalysis) and comparison with delaminated MoS2. Furthermore, we test our edge-site hypothesis through the intercalation of several metals in combination with an assessment of the resulting effect on catalytic performance relative to that of the ‘empty’ Mo6S8.

Results and Discussion Solvothermal synthesis of chalcogenide nanostructures is well precedented24 – especially the solvothermal synthesis of MoS2.25 Research in our laboratory indicates that ionic liquids may be used as both a reaction medium and a stabilising agent in such syntheses.26 We sought to adapt our established ionic liquid mediated

MoS2 synthesis to produce CPs (Figure 6). The use of an ionic liquid medium permits the use of soluble precursors, potentially extending the range of possible precursors and synthetic yield. However, the non- intercalated CP, Mo6S8, is metastable and has never been synthesised directly. Instead, an intercalated form must be made first (commonly CuxMo6S8),27 then subsequent deintercalation by chemical or electrochemical oxidation is performed to yield the desired product.16, 27 Thus, in the current study, copper has been introduced into the soluble precursor to synthesise CuxMo6S8. The solvothermal, ionic liquid mediated, synthesis involved dissolving Cu(NH4)MoS4 in 1-butyl-3-methylimidazolium trifluoromethanesulfonate and heating to 300 °C under a reducing atmosphere. The product, described herein as IL-CuxMoSy, was then cooled, washed at room temperature, and subsequently heat-treated at 700 °C under a reducing atmosphere to produce the copper containing nano-Chevrel Phase, n-CuCP (for nomenclature, see Figure 6). For electrochemical testing, this material was then suspended in a water-ethanol solution and Nafion added as a binder. The resulting suspension was sonicated and then dropcast onto glassy carbon electrodes. Electrochemical oxidation, as described below, gave nanoparticulate Mo6S8, n-CP. Two additional materials were synthesised for comparison: a Chevrel Phase sample synthesised by conventional solid state synthesis (bulk-CP, via bulk-CuCP)28-29 and the delaminated

MoS2 prepared in our previous reports (Figure 6; IL-MoS2; see SI for further details)26.

20 Chapter 1

Figure 6 – Schematic of catalyst syntheses. CuATTM = Cu(NH4)MoS4 and IL = BMIM OTF. Further details provided in the SI.

The broad reflections of the X-ray diffractogram of the n-CuCP precursor, IL-CuxMoSy (Figure 7, green trace), are consistent with disordered, expanded MoSx – a well-established product of solvothermal molybdenum sulfide synthesis (Figure S 9).26 After further heat treatment at 700 °C, the diffractogram of the material obtained

(Figure 7, blue trace) matches CuMo6S8 (ICSD: 158985, grey trace). Refinement data are shown in the SI (Figures S 7 and 8). The n-CuCP diffraction peaks are significantly broadened due to the small size of the crystallites. Rietveld analysis30 indicates that the crystallite size is 15 nm and 55 nm, for n-CuCP and bulk-CuCP (orange trace) respectively, when both strain and size are considered.

IL-CuxMoSy bulk-CuCP ¨ n-CuCP

ICSD CuMo6S8

¨ Intensity (a.u.) Intensity

10 20 30 40 50 60 2Theta (°)

Figure 7 – XRD diffractograms of Chevrel Phases with reference pattern (ICSD: 158985). Diamonds indicate reflections due to Mo metal (detailed further in the SI). Copper X-ray source (λ = 1.5406 Å).

Analysis by Transmission Electron Microscopy (TEM) reveals that IL-CuxMoSy, the product of the solvothermal reduction of Cu(NH4)MoS4, consists of aggregated spheres of a disordered, layered material, likely

CuxMoS2 or CuxMoS3 (Figure 8). The spheres vary in size from 50–100 nm. After heat-treatment at 700 °C for

60 h (to produce CuMo6S8), the spheres appear to maintain their size, although the crystallite morphology is changed drastically (Figure 9). The small crystallites of IL-CuxMoSy sintered to afford larger particles that now clearly exhibit the channelled structure of the nanostructured Chevrel Phase (n-CuCP), CuMo6S8. These larger crystallites are 15–30 nm in width and 50–100 nm in length, consistent with the size estimated from XRD.

Prior to electrocatalysis the copper-intercalated Chevrel Phase electrodes were deintercalated by electrochemical oxidation using cyclic voltammetry in HCl-acidified electrolyte (methods and characterisation by CV, XRD,

21 Chapter 1 and XPS are supplied in the SI). Cyclic voltammograms of the electrochemical deintercalation exhibited a

CuxMo6S8-oxidation current at ~0.3 V (vs Ag/AgCl), consistent with previous reports (Figure S 1),16 and X-ray photoelectron spectra of a sample removed from the electrode using carbon tape, indicated successful deintercalation (Figure S 13).

After the removal of the copper to produce nanoparticulate Mo6S8 (n-CP), Scanning Electron Micrographs of the electrode material (SEM, Figure S 4) reveal that this ‘empty’ n-CP consists of spheroidal, nano-scale crystallites. The Nafion binder appears to cover large swathes of the surface of the particles. The SEM images of samples of the bulk material, bulk-CP revealed that the bulk-CP was composed of large particles, usually greater than a micron in size (Figure 10, Figure S 5). These larger particles were typically aggregates of well- defined single crystals of Mo6S8, which featured long, unbroken channels through the material.

Figure 8 – Transmission electron micrograph of the product of solvothermal reduction of Cu(NH4)MoS4 (IL-CuxMoSy) at different magnifications

(scale left = 200 nm, right = 50 nm). The morphology is almost identical to that of IL-MoS2 prepared by solvothermal reduction of (NH4)2MoS4 (Figure S 2).

Figure 9 – Transmission electron micrographs of the IL-CuxMoSy post-heat treatment (at 700 °C) at different magnifications.

22 Chapter 1

Figure 10 – Transmission electron micrograph of the bulk Chevrel Phase, bulk-CP. Image showing crystal size (left) and well-defined pore structure (right).

Benchmarking hydrogen evolution catalysis was performed by measuring the overpotential (at a current density of 10 mA cm-2)31-32 and the Tafel slope. A conventional three-electrode apparatus consisting of a 1 M HCl electrolyte, graphite counter electrode, Ag/AgCl (3 M NaCl) reference electrode, and drop-cast active material on a glassy carbon electrode was used.

The HER activity of the n-CP was compared to bulk-CP and, as the published benchmark (Figure 11), delaminated MoS2, prepared by solvothermal delamination26 and referred to as IL-MoS2, see SI for full characterisation details.26 Polarisation curves for IL-CuxMoSy and other samples are given in Figure S 6, all experiments were performed in triplicate and are normalised by the electrochemical surface area of the electrodes. The increased surface coverage of low coordinate Mo sites12-14, 33 and potential for Mo6 clusters to act as electron “reservoirs”34 suggest that CPs might be expected to out-perform MoS2 for HER.5 Even so, the bulk-CP exhibited poorer HER activity than both commercial- and IL-MoS2 (Figure S 6). The large particle size of the bulk-CP is likely the explanation. Accordingly, we expected both the vastly increased number of electrochemically active sites of the n-CP and the improved electrical contact of the smaller particles on the glassy carbon surface to improve the activity of the CPs. Indeed, the overpotential decreases by ~0.5 V when comparing the bulk-CP to the n-CP (Figure 11) and the charge transfer resistance of the n-CPs is markedly lower than the bulk-CP (as determined by the electrical impedance spectroscopy data shown in the SI), consistent with our hypothesis. The overpotentials at 10 mA/cm2 for the traces in Figure 11 are: -0.04 (Pt), -

0.23 (n-CP), -0.3 (IL-MoS2), -0.67 (bulk-CP), and -0.86 (glassy carbon electrode) V vs RHE. Thus, our n-CP significantly outperforms other Chevrel Phase samples from the literature where a potential of 0.32 V at 10 mA/cm-2 has twice been reported.35-36 Additionally, the overpotential of the n-CP is ~0.07 V more positive than that of IL-MoS2, which indicates that the n-CP surface sites possess higher activity, assuming equal accessibility.

23 Chapter 1

0

n-CP

IL-MoS2 ) 2 -5 Bulk-CP Platinum Glassy Carbon

-10

Current (mA/cm -15

-20 -0.8 -0.6 -0.4 -0.2 0.0 Potential (vs RHE) Figure 11 – Linear sweep voltammograms of various electrocatalysts on glassy carbon electrodes performing HER. Electrolyte = 1 M HCl.

Tafel analyses for the HER polarisation curves of the n-CP catalyst taken under hydrodynamic conditions at pH 0 (1 M HCl) are shown in Figure 12 (Left, black trace). These data illustrate the rate of change of the HER current as a function of potential applied. Tafel slope analysis is then used to determine the rate limiting step of the overall HER for a given electrocatalyst.37 The Tafel slopes of the two CP electrocatalysts are nearly identical (within ± 2 mV/dec experimental error), both, to each other and to previously reported CP HER electrocatalysts (approximately 70 mV/decade),35-36 which indicates a mixed Tafel-Volmer rate limiting step, regardless of the particle size.38-39 This shallow slope, in conjunction with the low overpotential makes the nano- Chevrel Phase comparable in activity to several of the highest performing nanoparticulate noble metal-free catalysts in the literature (including supported/unsupported nanostructured MoS2, Mo2C, transition metals etc).4, 40-41 The durability of the n-CP catalyst was tested by repeated cycles from 0 to −0.6 V (vs Ag/AgCl). After 1000 cycles, there was no significant change in the polarisation curve of the n-CP sample. TEM and Raman characterisation of the electrode material after 1000 cycles also show no significant changes after the stability tests (Figures S16 and S17). Additionally, the Faradaic efficiency of the n-CP catalyst was measured to be ~100% (± 2%) over this period (Figure S19). This suggests that the sample possessed excellent structural stability under the conditions tested (Figure 12, Right).

24 Chapter 1

800 n-CP 72 mV/dec 0

IL-MoS2

600 168 mV/dec ) Bulk-CP 2 -5 1st Scan

vs RHE) 1000th Scan 2 68 mV/dec Platinum 400 37 mV/dec

-10 Current (mA/cm Current 200

-15 Potential (-mV/cm Potential

0 -20 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -0.4 -0.2 0.0 log(Current (mA/cm2)) Potential (vs RHE) Figure 12 – Tafel slopes for select samples of HER catalysts in 1 M HCl electrolyte with scan rate = 50 mV/s. (Right) Stability test for the IL- Chevrel Phase sample showing the first and one thousandth scan. Electrolyte = 1 M HCl, scan rate = 50 mV/s.

The intrinsic activity of a HER electrocatalyst is determined by the Gibbs free energy of hydrogen adsorption and desorption on the catalyst surface.42 For the HER, this is dominated by interaction of the metal d-orbitals and hydrogen s-orbitals.43 By tuning the electronics of the adsorption site, the HER activity of the catalyst can therefore be optimised.44 For CPs, the catalytically active surface Mo moiety has a Gibbs free energy of H+ adsorption of −0.1 eV.45 Reduction of the Mo oxidation state is expected to further stabilise the adsorbed hydrogen, thus, hindering the HER. Molybdenum reduction was achieved by electron transfer during the reductive intercalation of metal ions into the Mo6S8 framework (see the SI and Figure S 1).46 A comparison of the activity of the ‘empty’ Chevrel Phase (black trace) with the Cu- and Zn-intercalated derivatives (blue and yellow traces, respectively) is shown in Figure 13. The reactions were performed in neutral aqueous media to avoid deintercalation of Cu- and Zn-species. Relative to the ‘empty’ Chevrel Phase, the overpotential of the intercalated derivatives is increased by ~0.1 V.

0

n-CP ) 2 -2 n-ZnCP n-CuCP Glassy Carbon -4

-6 Current (mA/cm Current -8

-10 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 Potential (vs RHE) Figure 13 – Linear sweep voltammograms of various electrocatalysts on glassy carbon electrodes performing HER. Electrolyte = 1 M NaCl.

25 Chapter 1

The HER activity of both the Zn- and Cu-intercalated n-CP samples (n-ZnCP and n-CuCP, respectively) show a marked decrease in performance, consistent with our hypothesis. Morphological differences between the samples are expected to be negligible as deintercalation has been shown not to impact the structure of CPs.47 A similar trend was noted by McCarty and co-workers in the hydrodesulfurisation (HDS) activity of CPs.48

Significantly, the two reactions, HDS and HER, both involve the reversible binding & dissociation of H2.49 In other catalytic reactions involving the binding & dissociation of H2, such as the reverse water gas shift reaction and methanol synthesis, transition metal atoms bound to the external surface of CPs have been shown to significantly influence the lowest energy catalytic reaction pathway offering different optionalities as compared to bare CP.12, 33, 46, 50 In these bimolecular reactions each substrate is able to bind to a favourable metal centre (Mo or M; where M = K, Ti, Co, Rh, Ni, Cu, Pt, Ag etc), from which it can react with the neighbouring reagent (the ‘ensemble effect’).46, 51 In this way the CP-supported transition-metals can act as synergistic catalytic sites.51- 52 Thus, in our system, partially intercalated Zn or Cu at the catalyst surface may directly participate in the reaction – analogous to the surface-bound metal atoms in reported previously.51 However, here these surface- bound metal atoms inhibit the adsorption of H+ to the active S–Mo–S sites, decreasing the HER activity, which confirms our postulated mechanism of action.

Conclusions The four questions we posed have been answered in the affirmative. This study reported the synthesis of nanoparticulate Chevrel Phases by a novel, two-step solvothermal approach. The size of the phase-pure, unsupported particles was revealed to be < 100 nm by XRD (Figure 7) and TEM (Figure 8) – an unprecedented result for Chevrel Phases given the high temperatures and long dwell times of previous syntheses (Figure 10) as well as the support-free nature of our method. A variety of unsupported CPs (Mo6S8, MMo6S8, M = Cu, Zn) can be prepared by this route. Facile de/intercalation of transition metals by electrochemical control (Figure S 1) was confirmed by XPS (Figure S 13). These results indicated the tuneable nature of these catalysts, which enables them to act as a platform for various other reactions – methanol synthesis and the reverse water gas shift reaction appear to be promising targets for future studies. Our data indicate that the ‘empty’ nanoparticulate Chevrel Phase, n-CP, was the most active for HER from among the intercalated derivatives tested – intercalation of copper and zinc ions reduces the catalytic activity. These results are consistent with an optimal stabilisation of bound H+ on the catalyst surface – as investigated by manipulation of the oxidation state of the active Mo6 cluster.

In conclusion, unsupported nanoparticulate Chevrel Phases were synthesised by a novel, solvothermal, ionic- liquid mediated route and characterised by several spectroscopic, microscopic and electrochemical techniques. The increased surface area of the nano-Chevrel Phases (relative to the bulk) significantly improved their electrocatalytic activity for the hydrogen evolution – such that they outperformed nanoparticulate, disordered

MoS2, a well-established HER catalyst. These results reveal the latent potential of nano-Chevrel Phases as contemporary electrocatalysts; the electrochemistry of which warrants further investigation in future studies.

26 Chapter 1

References 1. Lewis, N. S.; Nocera, D. G., Powering the planet: Chemical challenges in solar energy utilization. Proc. Natl. Acad. Sci. U. S. A. 2006, 103 (43), 15729-15735. 2. Abe, J.; Ajenifuja, E.; Popoola, O., Hydrogen energy, economy and storage: review and recommendation. Int. J. Hydrogen Energy 2019, 44 (29), 15072-15086. 3. Schlapbach, L.; Züttel, A., Hydrogen-storage materials for mobile applications. In Materials for Sustainable Energy, 2011; pp 265-270. 4. Eftekhari, A., Electrocatalysts for hydrogen evolution reaction. Int. J. Hydrogen Energy 2017, 42 (16), 11053-11077. 5. Jaramillo, T. F.; Jørgensen, K. P.; Bonde, J.; Nielsen, J. H.; Horch, S.; Chorkendorff, I., Identification of active edge sites for electrochemical H2 evolution from MoS2 nanocatalysts. Sci 2007, 317 (5834), 100-102. 6. Luo, Z.; Ouyang, Y.; Zhang, H.; Xiao, M.; Ge, J.; Jiang, Z.; Wang, J.; Tang, D.; Cao, X.; Liu, C.; Xing, W., Chemically Activating MoS2 Via Spontaneous Atomic Palladium Interfacial Doping Towards Efficient Hydrogen Evolution. Nature Communications 2018, 9 (1), 2120. 7. Voiry, D.; Fullon, R.; Yang, J.; de Carvalho Castro e Silva, C.; Kappera, R.; Bozkurt, I.; Kaplan, D.; Lagos, M. J.; Batson, P. E.; Gupta, G.; Mohite, Aditya D.; Dong, L.; Er, D.; Shenoy, V. B.; Asefa, T.; Chhowalla, M., The Role of Electronic Coupling Between Substrate and 2D MoS2 Nanosheets in Electrocatalytic Production of Hydrogen. Nat. Mater. 2016, 15 (9), 1003-1009. 8. Lv, R.; Robinson, J. A.; Schaak, R. E.; Sun, D.; Sun, Y.; Mallouk, T. E.; Terrones, M., Transition metal dichalcogenides and beyond: synthesis, properties, and applications of single-and few-layer nanosheets. Acc. Chem. Res. 2015, 48 (1), 56-64. 9. Yan, Y.; Xia, B.; Xu, Z.; Wang, X., Recent development of molybdenum sulfides as advanced electrocatalysts for hydrogen evolution reaction. Acs Catalysis 2014, 4 (6), 1693-1705. 10. Zhang, G.; Liu, H.; Qu, J.; Li, J., Two-dimensional layered MoS 2: rational design, properties and electrochemical applications. Energy & Environmental Science 2016, 9 (4), 1190-1209. 11. Guo, Y.; Park, T.; Yi, J. W.; Henzie, J.; Kim, J.; Wang, Z.; Jiang, B.; Bando, Y.; Sugahara, Y.; Tang, J., Nanoarchitectonics for transition‐metal‐sulfide‐based electrocatalysts for water splitting. Adv. Mater. 2019, 31 (17), 1807134. 12. Cao, Z.; Guo, L.; Liu, N.; Zheng, X.; Li, W.; Shi, Y.; Guo, J.; Xi, Y., Theoretical study on the reaction mechanism of reverse water-gas shift reaction using a Rh-Mo6S8 cluster. RSC Advances 2016, 6 (110), 108270-108279. 13. Peña, O., Chevrel phases: Past, present and future. Physica C (Amsterdam, Neth.) 2015, 514, 95-112. 14. Fischer, Ø., Chevrel phases: superconducting and normal state properties. Appl. Phys. A: Mater. Sci. Process. 1978, 16 (1), 1- 28. 15. Chae, M. S.; Heo, J. W.; Lim, S.-C.; Hong, S.-T., Electrochemical zinc-ion intercalation properties and crystal structures of ZnMo6S8 and Zn2Mo6S8 chevrel phases in aqueous electrolytes. Inorg. Chem. 2016, 55 (7), 3294-3301. 16. Schöllhorn, R.; Kümpers, M.; Besenhard, J. O., Topotactic redox reactions of the channel type chalcogenides Mo3S4 and Mo3Se4. Mater. Res. Bull. 1977, 12 (8), 781-788. 17. Afanasiev, P.; Bezverkhyy, I., Ternary transition metals sulfides in hydrotreating catalysis. Applied Catalysis A: General 2007, 322, 129-141. 18. Gershinsky, G.; Haik, O.; Salitra, G.; Grinblat, J.; Levi, E.; Daniel Nessim, G.; Zinigrad, E.; Aurbach, D., Ultra fast elemental synthesis of high yield copper Chevrel phase with high electrochemical performance. J. Solid State Chem. 2012, 188 (Supplement C), 50-58. 19. Even-Boudjada, S.; Burel, L.; Chevrel, R.; Sergent, M., New Synthesis Route of PbMo6S8 Superconducting Chevrel Phase From Ultrafine Precursor Mixtures: II. PbS, Mo6S8, and Mo Powders. Mater. Res. Bull. 1998, 33 (3), 419-431. 20. Saha, P.; Jampani, P. H.; Datta, M. K.; Okoli, C. U.; Manivannan, A.; Kumta, P. N., A Convenient Approach to Mo6S8 Chevrel Phase Cathode for Rechargeable Magnesium Battery. J. Electrochem. Soc. 2014, 161 (4), A593-A598. 21. Mao, M.; Lin, Z.; Tong, Y.; Yue, J.; Zhao, C.; Lu, J.; Zhang, Q.; Gu, L.; Suo, L.; Hu, Y.-S.; Li, H.; Huang, X.; Chen, L., Iodine Vapor Transport-Triggered Preferential Growth of Chevrel Mo6S8 Nanosheets for Advanced Multivalent Batteries. ACS Nano 2020, 14 (1), 1102-1110. 22. Cheng, Y.; Parent, L. R.; Shao, Y.; Wang, C.; Sprenkle, V. L.; Li, G.; Liu, J., Facile synthesis of Chevrel phase nanocubes and their applications for multivalent energy storage. Chem. Mater. 2014, 26 (17), 4904-4907. 23. Cheng, Y.; Luo, L.; Zhong, L.; Chen, J.; Li, B.; Wang, W.; Mao, S. X.; Wang, C.; Sprenkle, V. L.; Li, G.; Liu, J., Highly Reversible Zinc-Ion Intercalation into Chevrel Phase Mo6S8 Nanocubes and Applications for Advanced Zinc-Ion Batteries. ACS Appl. Mater. Interfaces 2016, 8 (22), 13673-13677. 24. Byrappa, K.; Adschiri, T., Hydrothermal technology for nanotechnology. Prog. Cryst. Growth Charact. Mater. 2007, 53 (2), 117-166. 25. Li, Y.; Wang, H.; Xie, L.; Liang, Y.; Hong, G.; Dai, H., MoS2 nanoparticles grown on graphene: an advanced catalyst for the hydrogen evolution reaction. J. Am. Chem. Soc. 2011, 133 (19), 7296-7299. 26. Lau, V. W. h.; Masters, A. F.; Bond, A. M.; Maschmeyer, T., Promoting the Formation of Active Sites with Ionic Liquids: A Case Study of MoS2 as Hydrogen‐Evolution‐Reaction Electrocatalyst. ChemCatChem 2011, 3 (11), 1739-1742. 27. Lancry, E.; Levi, E.; Mitelman, A.; Malovany, S.; Aurbach, D., Molten salt synthesis (MSS) of Cu2Mo6S8—new way for large- scale production of Chevrel phases. J. Solid State Chem. 2006, 179 (6), 1879-1882. 28. Chevrel, R.; Sergent, M.; Prigent, J., Un nouveau sulfure de molybdene: Mo3S4 preparation, proprietes et structure cristalline. Mater. Res. Bull. 1974, 9 (11), 1487-1498. 29. Chevrel, R.; Sergent, M.; Prigent, J., Sur de nouvelles phases sulfurées ternaires du molybdène. J. Solid State Chem. 1971, 3 (4), 515-519. 30. Williamson, G.; Hall, W., X-ray line broadening from filed aluminium and wolfram. Acta Metall. 1953, 1 (1), 22-31.

27 Chapter 1

31. McCrory, C. C.; Jung, S.; Peters, J. C.; Jaramillo, T. F., Benchmarking heterogeneous electrocatalysts for the oxygen evolution reaction. J. Am. Chem. Soc. 2013, 135 (45), 16977-16987. 32. McCrory, C. C.; Jung, S.; Ferrer, I. M.; Chatman, S. M.; Peters, J. C.; Jaramillo, T. F., Benchmarking hydrogen evolving reaction and oxygen evolving reaction electrocatalysts for solar water splitting devices. J. Am. Chem. Soc. 2015, 137 (13), 4347-4357. 33. Zheng, X.; Guo, L.; Li, W.; Cao, Z.; Liu, N.; Zhang, Q.; Xing, M.; Shi, Y.; Guo, J., Insight into the Mechanism of Reverse Water-gas Shift Reaction and Ethanol Formation Catalyzed by Mo6S8-TM Clusters. Molecular Catalysis 2017, 439, 155-162. 34. Alonso-Vante, N.; Schubert, B.; Tributsch, H., Transition metal cluster materials for multi-electron transfer catalysis. Mater. Chem. Phys. 1989, 22 (3), 281-307. 35. Ortiz-Rodríguez, J. C.; Singstock, N. R.; Perryman, J. T.; Hyler, F. P.; Jones, S. J.; Holder, A. M.; Musgrave, C. B.; Velázquez, J. M., Stabilizing Hydrogen Adsorption through Theory-Guided Chalcogen Substitution in Chevrel-Phase Mo6X8 (X=S, Se, Te) Electrocatalysts. ACS Appl. Mater. Interfaces 2020, 12 (32), 35995-36003. 36. Naik, K. M.; Sampath, S., Cubic Mo6S8-Efficient Electrocatalyst Towards Hydrogen Evolution Over Wide pH Range. Electrochim. Acta 2017, 252, 408-415. 37. Li, D.; Lin, C.; Batchelor-McAuley, C.; Chen, L.; Compton, R. G., Tafel analysis in practice. J. Electroanal. Chem. 2018, 826, 117-124. 38. Shinagawa, T.; Garcia-Esparza, A. T.; Takanabe, K., Insight on Tafel slopes from a microkinetic analysis of aqueous electrocatalysis for energy conversion. Scientific Reports 2015, 5 (1), 13801. 39. Murthy, A. P.; Theerthagiri, J.; Madhavan, J., Insights on Tafel constant in the analysis of hydrogen evolution reaction. The Journal of Physical Chemistry C 2018, 122 (42), 23943-23949. 40. Jayabal, S.; Saranya, G.; Wu, J.; Liu, Y.; Geng, D.; Meng, X., Understanding the high-electrocatalytic performance of two- dimensional MoS2 nanosheets and their composite materials. Journal of Materials Chemistry A 2017, 5 (47), 24540-24563. 41. Zhuang, Z.; Huang, J.; Li, Y.; Zhou, L.; Mai, L., The Holy Grail in Platinum‐Free Electrocatalytic Hydrogen Evolution: Molybdenum‐Based Catalysts and Recent Advances. ChemElectroChem 2019, 6 (14), 3570-3589. 42. Nørskov, J. K.; Bligaard, T.; Logadottir, A.; Kitchin, J.; Chen, J. G.; Pandelov, S.; Stimming, U., Trends in the exchange current for hydrogen evolution. J. Electrochem. Soc. 2005, 152 (3), J23. 43. Li, F.; Han, G.-F.; Noh, H.-J.; Jeon, J.-P.; Ahmad, I.; Chen, S.; Yang, C.; Bu, Y.; Fu, Z.; Lu, Y., Balancing hydrogen adsorption/desorption by orbital modulation for efficient hydrogen evolution catalysis. Nature communications 2019, 10 (1), 1-7. 44. Hakala, M.; Kronberg, R.; Laasonen, K., Hydrogen adsorption on doped MoS 2 nanostructures. Scientific reports 2017, 7 (1), 1-13. 45. Yang, T.; Bao, Y.; Xiao, W.; Zhou, J.; Ding, J.; Feng, Y. P.; Loh, K. P.; Yang, M.; Wang, S. J., Hydrogen evolution catalyzed by a molybdenum sulfide two-dimensional structure with active basal planes. ACS Appl. Mater. Interfaces 2018, 10 (26), 22042- 22049. 46. Liu, C.; Liu, P., Mechanistic Study of Methanol Synthesis from CO2 and H2 on a Modified Model Mo6S8 Cluster. ACS Catalysis 2015, 5 (2), 1004-1012. 47. Murgia, F.; Antitomaso, P.; Stievano, L.; Monconduit, L.; Berthelot, R., Express and low-cost microwave synthesis of the ternary Chevrel phase Cu2Mo6S8 for application in rechargeable magnesium batteries. J. Solid State Chem. 2016, 242, 151-154. 48. McCarty, K. F.; Anderegg, J.; Schrader, G. L., Hydrodesulfurization catalysis by Chevrel phase compounds. J. Catal. 1985, 93 (2), 375-387. 49. Zou, X. X.; Zhang, Y., Noble metal-free hydrogen evolution catalysts for water splitting. Chem. Soc. Rev. 2015, 44 (15), 5148- 5180. 50. Zheng, X.; Guo, S.; Guo, L., Ethanol synthesis catalyzed by single Ni atom supported on Mo6S8 support. Applied Catalysis A: General 2018, 553, 52-64. 51. Perryman, J. T.; Ortiz-Rodriguez, J. C.; Jude, J. W.; Hyler, F. P.; Davis, R. C.; Mehta, A.; Kulkarni, A. R.; Patridge, C. J.; Velazquez, J. M., Metal-promoted Mo6S8 clusters: a platform for probing ensemble effects on the electrochemical conversion of CO2 and CO to methanol. Materials Horizons 2020, 7 (1), 193-202. 52. Kamiguchi, S.; Nagashima, S.; Chihara, T., Application of solid-state early-transition metal clusters as catalysts. Tetrahedron Lett. 2018, 59 (14), 1337-1342.

28 Chapter 2

Chapter 2.1

3R-MoS2 in Review: History, Status, and Outlook

This chapter has been accepted for publication in ACS Applied Energy Materials as “3R-MoS2 in Review: History, Status, and Outlook”.

Abstract Transition metal dichalcogenides of molybdenum catalysis. As the literature on this material is rapidly and tungsten (TMDs) are layered van der Waals expanding, this review seeks to summarise the materials that exhibit a rich array of polytypes. The history, known and predicted characteristics, different possible arrangements of the constituents syntheses, applications, as well as common of the ‘two-dimensional’ MX2 sheets (where M = misconceptions of, and surrounding, 3R-MoS2. group 4–10 elements, X = chalcogen) gives rise to Although the review is chemically focused, it a host of interesting and tuneable phenomena. includes suggested reading to cover a broader

Molybdenite, or molybdenum disulfide (MoS2) in scope. its most abundant and thermodynamically stable form, 2H-MoS2, is perhaps the most widely used TMD, though the potential applications of its metastable polytypes have been recognised only recently. From among the polytypes, the 3R-MoS2 (rhombohedral) phase has attracted the most interest because of its thermodynamic stability, ABC stacking (as opposed to the AA′ of the more common 2H-MoS2), and lack of inversion symmetry. These properties make it an excellent candidate for photonics, optoelectronics, and

Introduction

Transition metal dichalcogenides such as MoS2 have been indispensable throughout the 20th century and, are highly promising materials for a range of modern chemical transformations and optoelectronic applications, as exemplified by increasing research activity.1-3 MoS2, like other TMDs (eg. MoSe2, WS2 etc), exists as several polytypes, of which the 2H, 1T and 3R have been discussed most extensively. These TMDs are composed of

MoS2 layers, held together by (relatively) weak van der Waals forces. Each layer is composed of covalently bonded condensed MoS6 centres in which alternating sheets of sulfur atoms sandwich a sheet of molybdenum atoms (Figure 15). The structural differences between the most common polytypes are in the shape of the MoS6 centre (octahedral or trigonal prismatic) and the arrangement of the layers. Thus, 2H-MoS2, the most thermodynamically stable arrangement, can be described by a hexagonal unit cell, with trigonal prismatic MoS6 centres, but other metastable polytypes are possible.4 Of these, the rhombohedral (3R-MoS2) polytype is of primary interest as it occurs naturally. It is also the second most stable polytype of MoS2, lending itself to practical importance in the fields of catalysis,5 engineering (e.g. lubricants),6-7 and mineral processing.8

Additionally, 3R-MoS2 lacks an inversion centre, has a tuneable bandgap, and has a high piezoelectric constant. These features make it a promising material for optoelectronic and photonic devices.9-10 This review summarises the chemical literature for 3R-MoS2, expanding on Wilson and Yoffe’s 1979 TMD review.11 Our focus is on

29 Chapter 2

the chemistry of the 3R-MoS2 phase, therefore discussions beyond this scope will refer to respected reviews in the appropriate fields.

The current review is arranged as follows: a general background of the 3R-MoS2 polytype; discussion of reported syntheses with particular reference to growth mechanisms; critical analysis of characterisation data, highlighting common mischaracterisations; and finally, the applications and outlook of 3R-MoS2.

Background History The atomic composition of naturally occurring molybdenite was first recognised in 1782 by Peter Hjelm,12 following his discovery of molybdenum the previous year, and the earliest synthesis is attributed to JJ Berzelius (1826) who passed hydrogen sulfide through a molybdenum salt solution.13 The crystal structure of the most abundant form of natural molybdenite was determined in 1923 by Pauling and Dickinson based on the hexagonal P63mmc group.14 In 1957, Bell and Herfert synthesised MoS2 in a carbonate flux and observed that it had a different crystal structure to that of the natural samples.15 However, soon after this discovery naturally occurring samples were also identified in Canada.16 The structure was later independently resolved by Semiletov 17 and Wildervanck & Jellinek,18-19 who described its rhombohedral crystal structure. Naturally occurring samples of this material have since been discovered all over the globe.20 Interestingly, in Pauling’s 1923 account, the authors note that “triangular” crystals synthesised “by fusing together ammonium molybdate, sulfur and potassium carbonate” were reported by Guichard in 1901.21-22 Furthermore, De Schulten’s (1889) account also describes a carbonate flux synthesis of MoS2 and produced triangular plates of MoS2.23 Due to Guichard and De Schulten’s reported synthesis conditions and crystal shapes (vide infra), we believe these crystals were likely rhombohedral 3R-MoS2, which would predate the commonly ascribed ‘earliest’ synthesis by 68 years (see footnote24).

Nomenclature To simplify naming of the various TMD polytypes, an adapted version of Ramsdell’s25 convention was introduced in 196316 and has since been endorsed by the International Union of Crystallography (IUCr).26-27 It is used in this review. The two-layered hexagonal and three-layered rhombohedral forms are named 2H-MoS2 and 3R-MoS2, respectively. The number signifies the number of “MoS2” layers required to describe the unit cell and the letter signifies the lattice system. 2H-MoS2 consists of 2 layers of S/Mo/S sheets in a Bernal stacking arrangement, which can be notated as AbA BaB (where upper case = sulfur, lower case = molybdenum; Figure 15). In this case, there are two locations for the sulfur atoms (‘A’ and ‘B’) and two locations for the molybdenum atoms (‘a’ and ‘b’). The letters refer to a position along the b-axis of the unit cell. Thus, the Mo atoms in one layer align with the sulfur atoms in the adjacent layers and vice versa. Within a layer, the sulfur atoms in the two sheets occupy the same repeat positions along the b-axis and are offset relative to each other in adjacent layers.

2H-MoS2 is the lowest energy stable phase of MoS2 and occurs most commonly. 3R-MoS2 is the lowest energy metastable phase and is described by a rhombohedral unit cell containing AbA CaC BcB layers. It is common to describe and visualise the unit cell with hexagonal axes (both axes are shown in Figure 15). The layers in 3R-

MoS2 are not only offset relative to those of 2H-MoS2, but they are also oriented differently. In 2H-MoS2, each layer is oriented antiparallel to its neighbours (also conceptualised as a layer rotation of 60°, or the orientation of a trigonal prism in one layer at 60° to one in the next layer, Figure 15a, b, see footnote28). Given there are only two b-axis positions for all the atoms this arrangement is also called AA′, where the prime indicates an antiparallel layer. For 3R-MoS2, all layers are oriented parallel to each other (Figure 14c, d). Thus, this configuration can be described as ABC. The grey chevrons and lines in Figure 15 indicate layer ‘direction’ (for octahedral phases like 1T-MoS2, ‘direction’ can be interpreted as the sign (+/-) of the slope of the line). This

30 Chapter 2 nomenclature is most common in TMD literature where (nano)materials only a few layers thick are reported.29- 31

Figure 14 – (a) 2H-MoS2 showing MoS6 trigonal prismatic polyhedra in an alternating antiparallel arrangement. (b) 2H-MoS2 from the c-axis

perspective, note polyhedra (viewed here as triangles) rotate by 60° between each layer. (c) 3R-MoS2 showing parallel arrangement. (d) 3R-

MoS2 from the c-axis perspective, note polyhedra (viewed here as triangles) are oriented parallel in each layer. ICSD: 24000 (2H-MoS2) and

38401 (3R-MoS2).

Suggested Search Terms Because of the multitude of ways in which 3R-MoS2 can be described, the reader is encouraged to use the search terms ‘rhombohedral’ and ‘molybdenite’ in petrological databases, ‘van der Waals’, ‘MX2’ and ‘transition metal dichalcogenides’, and ‘AB (or ABC) stacking’ within the few-layer and optics literature. We recommend that authors include the search tags (e.g. publication titles, abstracts, or keywords) ‘3R’ or ‘ABC stacked’ and discourage the use of the archaic term ‘synthetic’ (MoS2) so that research communications may be clearer, enhancing the speed of progress.

Polytype Definition The IUCr defines “an element or compound as polytypic if it occurs in several different structural modifications, each of which may be regarded as built up by stacking layers of (nearly) identical structure and composition, and if the modifications differ only in their stacking sequence.” A ‘polytype’ is a specific case of ‘polymorphism’, which is a subset of the various ‘phases’ in which a compound may occur. Guinier, A. et al. Nomenclature of polytype structures. Acta Crystallographica Section A, 198426-27 General

3R-MoS2 (space group R3m, point group C53v, a=b=c=6.403 Å, α=28.627° or a=b=3.166 Å, c=18.41 Å with hexagonal axes) is one of the many polytypes of MoS2.32 It consists of layered sheets of covalently bonded condensed Mo–S trigonal prisms (often called ‘sandwich layers’).6, 33 These sandwich layers are more

‘compacted’ in MoS2 than other TMDs due to the relatively small size of the Mo atom.11, 34 Electron–electron repulsion from filled non-bonding dz2 orbitals, protruding into the gallery, is balanced by van der Waals forces holding the layers together. This combination of effects results in an interlayer Mo–Mo distance of 6.13 Å with a ‘gallery height’ of 2.98 Å.11, 31-32 The trigonal prisms in adjacent sandwich layers are oriented in parallel (Figure

31 Chapter 2

14) – i.e. they have the same S–Mo–S direction (marked with a chevron in Figure 15b) and are staggered at three locations within the unit cell. The polytypes can also be differentiated when viewed down the c-axis (Figure 15d, e, f) of the unit cell by the relative intensities of the near and distant molybdenum and sulfur atoms.

The interlayer electron–electron interactions significantly influence the physical properties and band structure of the MoS2.35 The valence and conduction electronic states are comprised of coupled S p- and Mo d-orbitals, therefore perturbation of the bands of the S-atoms (by adjacent layers, and therefore, polytype) influences the band gap.36-37 The Mo- and S-atoms have formal oxidation states of IV and −II, respectively.38 DFT calculations suggest that due to interlayer interactions, the 2H-MoS2 is 5 meV/formula unit more stable than is the 3R-

MoS2 phase and the AbA phases are favoured over AbC phases (Figure 15).31, 39-41 The 3R-MoS2 phase occurs naturally, but is metastable, although the phase change to 2H-MoS2 does not occur below ~500 °C.4, 42-43

Figure 15 – Crystal structures of the most common MoS2 polytypes: 2H (a, d: ISCD: 24000), 3R (b, e; ICSD: 38401), and 1T (c, f: ICSD: 254956). a, b, c) Crystals viewed along the a-axis, where the unit cell is shown as a grey box (3R rhombohedral unit cell shown in red), layer directions shown by grey chevrons or lines, and layer names given on the right. d, e, f) Bilayers viewed along the c-axis. g, h) Mo–S coordination shapes (trigonal prismatic for 2H & 3R, octahedral for 1T).

Polytypes

Only 2H-MoS2, 3R-MoS2, and amorphous MoS2 (called jordisite)1 have been found in natural deposits, although the number of potential polytypes is not limited to two.32 In fact, there are 112 distinct possible structures that can be described by fewer than 7 layers of trigonal prismatic S–Mo–S units.44 Most of these are unstable and have not been observed in the bulk45. Further polytypes are generated if one sulfur layer is translated in the a,b plane, distorting the S6 trigonal prism into an octahedron (Figure 15). The most well-known polytype with condensed S6 octahedra is the 1T-MoS2 (trigonal) phase, which attracts attention because of its relatively high stability and metallic character.46 There are also several other polytypes in different space groups (e.g. 1M and 2O)47-48 and the number of polytypes increases further when superlattice distortions are included,49-51 however these are out of the scope of the current review.

Syntheses of 3R-MoS2

32 Chapter 2

Growth Mechanism

3R-MoS2 syntheses often involve crystal growth in a flux, which is known to introduce screw dislocations into the nascent crystals.52-53 A crystal growing from a screw dislocation will consist of one continuous layer growing in an upward spiral or helix (Figure 16c).53 This mechanism can occur at a range of temperatures and pressures, and in various fluxes (vide infra). Agarwal and Joseph 54 showed that synthetic 3R-MoS2 does indeed grow from screw dislocations by using a sodium peroxide and potassium nitrate melt to etch crystals. From a central defect, a continuous layer will grow, containing parallel S-Mo-S motifs regardless of layer (Figure 15b).19, 42-43 A clear implication is that 2H-MoS2, which consists of antiparallel layers, cannot be grown by this mechanism. A more thorough discussion on the screw dislocation growth mechanism of MoS2 is given by Zhang, et al. 55 and a recent review by Nie, et al. 56 includes several models and visualisations.

Large Scale Syntheses The salt flux method developed by De Schulten 23 has been adapted by several researchers and appears to be the most reliable way to synthesise bulk 3R-MoS2.15, 57-59 A sulfur flux has also been used in combination with high pressures to produce 3R-MoS2 (whereby sulfur exchange facilitates the phase transformation), and although a 2H-MoS2 to 3R-MoS2 phase conversion can occur under extreme conditions (40–75 kbar, ~2000 °C), this procedure is not as practicable as the carbonate flux method.11, 43, 60-62 Under the similarly extreme conditions of self-propagating high temperature syntheses, impure 3R-MoS2 samples appear to have been prepared by several groups.63-64 Little is mentioned of the sample characteristics, but the particles appear to be micron sized platelets with very few macroscopic defects.

Figure 16 – Screw dislocation growth diagrams. a) Side on (a-axis) model of the MoS2 spiral pyramid. b) Single edge dislocation (raised blocks,

right) and subsequent spiral growth. c) top down (c-axis) model of 3R-MoS2 spiral pyramid

Chemical vapour transport (CVT) has been used to synthesise crystals that range in size from monolayered to bulk.65-67 Crystal growth from non-volatile precursors is achieved by the addition of a volatile transport agent that forms unstable gaseous compounds with the precursor to transport the precursor across a temperature gradient within a closed vessel, where the precursor then deposits on a seed crystal.68 For MoS2 synthesis, halides and MoCl5 are common transport agents.65, 69 Product control is achieved by altering the reaction parameters including the transport agent, zone temperature, substrate, and reaction length. Suzuki, et al. 10 showed that 3R-MoS2 and 2H-MoS2 phase crystals could be preferentially grown when Cl2 and I2 were used as the transport agents, respectively. Crystals grown by CVT are often used as precursors for preparing mechanically exfoliated MoS2 monolayers because of the ease of synthesis and the quality of the product. Sigiro and Nasruddin 70 showed that the phase of CVT-grown crystals can be controlled by choice of dopant. The authors tested Re, Nb, Fe, Co, and Ni, and observed that Re and Nb dopants lead to preferential 3R-MoS2

33 Chapter 2 crystal growth. The authors did not propose a mechanism, though we believe it is likely that the larger ions induce defects that cause screw dislocations (see ‘Natural Occurrence’ below).

Micro/Nano 3R-MoS2

Nano 3R-MoS2 can be prepared by sulfurisation of MoO3. Li, et al. 71 describe the preparation of MoS2 nanowires by electrodeposition of MoO3 on graphite followed by reaction with H2S above a temperature of

800 °C. Deepak, et al. 39 prepared hydrothermal MoO3 that was subsequently reacted with a H2S/H2 gas mixture at 900 °C. Alternatively, Leidich, et al. 72 used a non-hydrolytic sol–gel synthesis that involved preparing an amorphous precursor by reaction of MoCl5 and hexamethyldisilathiane then subsequent heat treatment at

~1000 °C. The authors found that a higher 3R-MoS2:2H-MoS2 ratio was obtained at higher temperatures.

Hydrothermal preparations of sub-micron 3R-MoS2 are also possible and will be discussed in-depth later in this review.

Few-Layered 3R-MoS2

Mono- and few-layered 3R-MoS2 can be synthesised by a variety of methods. Some, like mechanical exfoliation, are suitable only for laboratory studies,73 though others can prepare larger TMD samples and control for polytype and number of layers. The “scotch tape method” of mechanical exfoliation involves applying conventional adhesive tape to a single crystal, which, upon peeling off the tape, cleaves the layers by overcoming the weak interlayer van der Waals forces.73 This method was first used to produce MoS2 and later, graphene.74

Na, et al. 35 showed that mixed-phase and mixed-layer MoS2 can also be produced. Kumar, et al. 75 prepared similar samples by the scotch tape method and investigated the effect of heating temperature and rate on the material. Other mechanical methods, such as grinding, have been used to induce regions of 3R-MoS2 in a 2H-

MoS2 crystal.76-77 For instance, Macchione, et al. 78 created 3R-MoS2 domains by ball milling a single crystal of

2H-MoS2. Practical applications of these methods are limited as the phase composition of the product cannot be controlled reliably.

High-quality samples of 3R-MoS2, prepared for fundamental studies, have mostly been prepared by chemical vapor deposition (CVD) which allows large (i.e. micro–millimeter sized), relatively defect-free sheets to be produced. Several reviews have covered the synthesis of MoS2 by chemical vapor methods,33, 79-80 therefore only reports of 3R-MoS2 will be included here. CVD involves the activation and subsequent reaction of gaseous reactants followed by deposition of a solid product.81 In few layered MoS2 synthesis, this usually involves evaporating sulfur and molybdenum sources by heating in a furnace under an inert carrier gas. The product is then deposited on a substrate such as Si by lowering the reactor temperature.82-83 Zhou, et al. 84 prepared mixed- phase MoS2 by CVD and showed that the grain boundary between the 2H-MoS2 and 3R-MoS2 regions was atomically sharp and “meandered” between ring defects. Yan, et al. 85 synthesised mono-, bi- and tri-layer MoS2 to observe the different stacking arrangements of the product. Meier, et al. 86 were able to synthesise a range of

MoS2 polytypic morphologies by altering the heating ramp rate and post-synthesis annealing time of their CVD synthesis. The resulting materials exhibited variable CO2 photoreduction activity, dependent on the synthesis conditions applied.

Zhang, et al. 87 synthesised vertically oriented and spiral 3R-MoS2 by CVD. The authors suggest that both morphologies are the result of a screw dislocation growth mechanism, where the collision of expanding crystallites results in defects which cause upwards or vertical growth (Figure 16). In a similar report, Ly, et al. 88 present 3R-MoS2 triangular spiral pyramids grown by CVD on SiO2 defects. These pyramids reach tens of microns in size and, because they are composed of a single spiral of MoS2, are highly conducting along the vertical axis. The growth mechanism of the 3R-MoS2 spiral pyramids remains contentious.89-92 Kumar and

34 Chapter 2

Viswanath 93 note that a high sulfur evaporation rate provides a sufficiently high driving force for continuous (rather than layer-by-layer) growth to dominate, which is postulated to result in the spiral pyramids. Hao, et al. 94 instead ascribe the growth to low supersaturation resulting from the high temperature (1000 °C) of the heating zone. Dong, et al. 95 propose that step edges, created by the coalescence of two growth fronts, are the origin of the screw dislocations. Wu, et al. 90 note that these morphologies result in supergiant elasticity of the spiral

‘nanosprings’. Similar morphologies have been observed for other CVD grown TMDs (e.g. GaSe,96 WSe2,97

WS2,98 and MoSe299).

3R-MoS2 may also be produced using variations of CVD such as liquid salt transport synthesis. Cevallos, et al.

100 produced 3R-MoS2 in ~3% yield in a NaCl and CsCl eutectic salt flux. The authors note that no evidence of screw dislocation growth was observed, but rather demonstrate that stacking faults resulted in 3R-MoS2 intergrowths in the 2H-MoS2 crystals. Li, et al. 101 used Vapour-Liquid-Solid Growth to synthesise 3R-MoS2 monolayer micro-ribbons from MoO3, S, and NaCl, suitable for use in nanoelectronic devices.

Few-Layer Manipulation

Artificial monolayer manipulation can produce 3R-MoS2 by folding, twisting, or restacking 2H-MoS2 monolayers. Weston, et al. 102 were able to apply a ‘tear-and-stamp’ transfer technique to pick up and deposit one MoS2 monolayer onto another. To select for 2H-MoS2 or 3R-MoS2 stacking, the authors aligned the edges of the two sheets at the appropriate rotation and report that the electronic properties of the material change as a function of rotation angle. Paradisanos, et al. 103 employed a water-assisted lift-off process to stack MoS2 monolayers at controlled angles. The authors were able to tune the valence band splitting over large areas of high quality MoS2. In a less controlled fashion, Kumar, et al. 75 were able to transform MoS2 monolayers into nanodomains of 3R-MoS2 by heating the material. The authors propose that the phase transformation occurs to stabilise the voids left by volatilised sulfur atoms.

Inorganic Fullerenes

As well as several polytypes, MoS2 forms a wide variety of morphologies, including rods, ribbons, tubes, and fullerene-like particles at a range of scales. Most of these occur as the 2H-MoS2 polytype and have been presented in a comprehensive review by Enyashin, et al. 104, although the 3R-MoS2 phase has been identified in some of the samples. Remskar et al.105-106 synthesised MoS2 nanotubes and nanopods by sulfurisation of precursor Mo6S2I8 nanowires. The authors found that the 3R-MoS2 polytype was more prevalent in large- diameter nanotubes. The geometry of 3R-MoS2 nanotubes and resultant electronic properties have been analysed by Milosevic, et al. 107.

Natural Occurrence

Finally, 3R-MoS2 may be obtained by extraction from molybdenite ore. An exhaustive list of globally reported instances of 3R-MoS2 samples is too large for this review (>1000 at the time of writing); instead, a selection of metastudies are provided (also see footnote for suggested search terms).20, 44, 108-111 Natural occurrence of the

3R-MoS2 polytype has been shown to strongly correlate to the presence and concentration of impurities;110 primarily rhenium, but also tungsten, niobium, vanadium, iron, and titanium.112 It appears that these impurities induce defects in nascent MoS2, which make 3R-MoS2 growth by screw dislocation more probable (vide supra).42, 113 Given that rhenium is both valuable and scarce (with an average crustal abundance of 0.5 ppb),114 and that

MoS2 is the primary deposit in which rhenium is found (it can substitute up to tens of percent of Mo atoms),115 the 3R-MoS2 fraction of natural molybdenite is an important indicator of the value of the molybdenite ore. Additionally, as 187Re will undergo β decay to produce 187Os, the Re/Os ratio may be used as a chronometer, which makes 3R-MoS2 deposits important for dating strata.116-117 The 3R-MoS2 fraction of molybdenite also

35 Chapter 2

has implications for the processing and recovery of the ore. McClung 8 showed that the 3R-MoS2 polytype is more difficult to grind and is harder to separate by flotation as it less buoyant than is the 2H-MoS2 phase.118

Characterisation

This section discusses the structural, microscopic, spectroscopic, and physical characterisation of 3R-MoS2 and is intended to help researchers to effectively and efficiently compare their data to those in the literature and concludes with a note as how some of the common mistakes in this complex field might be avoided.

Structure

As noted above, the crystal structure of 3R-MoS2 was solved more than 60 years ago.32 More recent geometry optimisation by DFT routinely calculates the lattice parameters to within 1% of the observed values.4 It is of note that powder X-ray diffractograms for 3R-MoS2 appear quite similar to those of 2H-MoS2; though the regions at ~40° and ~50° 2θ can be used to differentiate the two polytypes (Figure 17). Houben, et al. 119 have noted that because the scattering intensities are dominated by the heavier Mo atoms, which only slightly change positions between different polytypes, small domains of polytypic impurities are “practically invisible”. Schönfeld, et al. 83 measured Debye-Waller factors for the mean squared displacement of Mo and S atoms parallel and perpendicular to the basal plane and found that the mean squared displacement is larger in the c- axis and for Mo atoms, it does not differ between 2H-MoS2 and 3R-MoS2. Electron backscatter diffraction has been used to identify areas of high 3R-MoS2 concentration in molybdenite ore.120 The authors combined the characterisation with elemental analysis to show the correlation of 3R-MoS2 with high rhenium concentration.

2H (P63mmc)

a=b=3.15 Å c=12.3 103 3R (R3m)

a=b=c=6.403 Å α=28.6°

104 015 107 105 018

40 45 50 X-ray Diffraction Intensity Diffraction X-ray 10 20 30 40 50 60 70 80 90 °2Theta (Cu)

Figure 17 – X-ray diffractograms of 2H (ICSD: 24000) and 3R-MoS2 (ICSD: 38401). Inset shows magnified region from 37.5° to 52.5° 2-theta.

Microscopy

Microscopy is a key technique for MoS2 polytype analysis. With the increased accessibility of high-resolution instruments, our understanding of MoS2 and other TMDs has improved. The polytype of few-layered MoS2 is easily discerned using HR-TEM by analysing the distinctive phase patterns along the c-axis (Figure 15).75, 78, 121- 123 Often a grey-value trace is measured (and sometimes compared to simulated data) to differentiate Mo from S and can also give an indication of atom depth within the unit cell, from which the polytype can be determined.122 Using this approach, Shiojiri, et al. 76 investigated the formation of natural and synthetic 3R-

MoS2 by analysing stacking faults. Zhou, et al. 124 observed that sulfur vacancies can migrate between layers and predominately lie in the (Mo–2S) column. Yan, et al. 85 observed that electron-beam damage125 occurs under standard electron-gun potentials (80 keV) in MoS2. The same authors observed several metastable CVD grown

36 Chapter 2 bi/trilayer phases such as AA and AAB by dark field imaging. A useful review on the identification of polytypes by electron microscopy is provided by Zhao, et al. 122

The polytypes can be further differentiated by their Kikuchi band patterns:11 3R-MoS2 gives a star with 3-fold symmetry, and 2H-MoS2 gives a star with 6-fold symmetry.126 These symmetries are also present in electron diffraction images, for which conventional,67 convergent beam,10 and zone axis patterns127 (i.e. real space crystallography) have been recorded. In a follow-up study zone axis patterns were later simulated by Tatlock 128

Electronic Structure and Optical Properties Band structure TMDs display a range of interesting electronic and optical properties that are dependent on the elements, polytypes, and number of layers, which make TMD materials ideal candidates for optic and optoelectronic devices.3 The filling of the d-bands near the Fermi level determines the properties of the material.3, 11 The d- band filling is determined by several factors: the internuclear distances, the crystal symmetry, the electronic configuration of the metal atom, and the electronegativity of the ligand.11 For 3R-MoS2, crystal field theory dictates that the trigonal prismatic geometry causes splitting of the d-orbitals into a filled dz2, degenerate empty dx2-y2 and dxy, and degenerate empty dyz and dxz orbitals (Figure 18c).104, 129 This configuration results in semiconducting behaviour.130 A more complete picture is given by band structure calculations, which determine the valence band maximum for 3R-MoS2 to be near the K point and the conduction band minimum occurring between the K and Γ points, resulting in an indirect band gap (Figure 18b).4 In monolayer MoS2, the K and K′ points are degenerate, though when layers are stacked the degeneracy is removed by an interlayer interaction that causes splitting of the valence band maximum.131 The band gap of 3R-MoS2 (2H-MoS2) has been calculated to be 0.80 (0.86) eV for bulk and 1.81 (1.66) eV for monolayer samples; the gap changes from indirect to direct as the number of layers is decreased.71, 132 The band gap has been observed to decrease ~0.1 eV with increasing dopant (rhenium) content.133

Absorption Spectra

Two strong absorbance maxima (A and B) are observed between 1.8 and 2.0 eV in few-layer MoS2.31 These absorptions have been assigned to d–d transition excitons split by spin–orbit coupling.10, 134-135 The energy difference of these excitons in 3R-MoS2 has been measured by optical transmission spectroscopy and calculated using DFT to be lower than that of 2H-MoS2 (~150 and ~190 meV, respectively).103, 135-136 Unlike 2H-MoS2, the energy difference of optical transitions of 3R-MoS2 does not change as a function of either layer number or pressure because the interlayer waves are decoupled.131 Using reflectivity spectroscopy, Akashi, et al. 136 observed weak features on the shoulders of the main excitons that corresponded to the s-orbital states of the electron–hole pair and which shifted significantly between 2H-MoS2 and 3R-MoS2 samples.

Density of States

The density of states of 3R-MoS2 and 2H-MoS2 are quite similar; the states near the Fermi Level are dominated by Mo d-electrons with some contributions from the S p-electrons. Both are characterised by maxima at ~−1.0 and ~2 eV relative to the Fermi level, though the 2H-MoS2 exhibits additional features beyond 5 eV that arise from Mo 4p and 5s orbitals.4, 137 He, et al. 132 performed band structure calculations for 3R-MoS2 containing S vacancies and found that the band gap increased as the S vacancy distance increased and that spin polarisation had a negligible influence on the calculations. Su, et al. 138 note that sulfur defects in MoS2 monolayers trap carriers in mid-gap states and show experimentally that defect suppression results in an 8-fold photoluminescence enhancement.132

Symmetry

37 Chapter 2

The same-direction layer stacking arrangement of 3R-MoS2 results in broken inversion symmetry within the unit cell. Consequently, wavefunctions at the K point are localised, thus the K point excitons cannot relax by means of interlayer scattering (unlike in 2H-MoS2) and must instead follow an intralayer valley relaxation path between the K and K′ points (Figure 18a). This restriction of interlayer hopping results in extended exciton lifetime, which makes 3R-MoS2 a promising material for information storage.136, 139 Akashi, et al. 136 have experimentally confirmed these theoretical predictions.103

Second-Order Harmonic Generation

As 3R-MoS2 is non-centrosymmetric, it exhibits strong second-order-harmonic (SH) generation. This process involves the combination of two photons to generate a single photon with double the frequency/energy.66, 87

This behaviour has been experimentally confirmed in bulk140 and monolayer/thin films of 3R-MoS2.66 At the A and B exciton resonance frequencies, the SHs are observed to red shift ~1.5 and ~10 meV/layer under the thin film limit (~10 layers).66 The intensity of the SH has been found to scale quadratically with the number of layers until re-absorption causes intensity saturation at ~10 layers.66, 87 3R-MoS2 generates SHs for every layer number, unlike 2H-MoS2, for which SHs occur only for odd numbered layers under the thin film limit.66

Figure 18 – a) Simplified band structure diagram of the MoS2 K exciton electron-hole pair showing allowed (green arrow) and disallowed (red

arrow) intralayer (i) and interlayer (ii) relaxation pathways from spin up (blue) to spin down (orange) bands for 3R and 2H-MoS2. b) Primitive

rhombohedral Brillouin zone marked with symmetry points. c) Simplified density of states diagram of trigonal prismatic (3R, 2H etc) MoS2 where occupied states are shaded; adapted with permission from Mattheiss 141, copyright 2020 by the American Physical Society.

Photoluminescence

3R-MoS2 exhibits strong photoluminescence from the A and B excitons.31 The energy of the 3R-MoS2 photoluminescent exciton A peak does not change appreciably as a function of layer number as it arises primarily from the Mo d-orbitals, which are screened from interlayer interactions.131 Nevertheless, it is known to slightly blue-shift as a result of quantum confinement effects131 and increase in intensity as the layer number is decreased.87, 140 The indirect gap transition of bilayer 2H-MoS2 is ~0.1 eV higher than that of 3R-MoS2.30 Circular dichroic photoluminescence experiments show that the intensity difference of the photoluminescence spectra does not change as a function of layer number for 3R-MoS2, but decreases rapidly for 2H-MoS2, which indicates preservation of the valley information in stacked 3R-MoS2 layers.139 Na, et al. 35 observe that the intralayer vibration at ~405 cm−1 is split when an excitation energy of 1.96 eV is used and rationalise this phenomenon as a consequence of interlayer interaction (Davydov splitting).35 The degree of splitting was shown to be polytype dependent.

Raman

Xia, et al. 30 show that principle component analysis can discriminate 2H-MoS2 from 3R-MoS2 bilayers using Raman spectra. The absorption at 395 cm−1, assigned to the LA′(M) + TA′(M) and LA(M) + TA(M) modes, was the primary difference between the two polytypes. The intensity ratio of the shear/breathing modes is smaller for 3R-MoS2 than for 2H-MoS2 69 and both modes have been observed to branch (split into non- equivalent peaks) as a function of the number of layers in both polytypes.35, 69 The breathing mode frequency

38 Chapter 2 has been calculated to blue-shift as the number of layers increases. Indeed, this absorption disappears for bulk samples as the displacement difference of the top and bottom layers (relative to the absolute distance between these layers) becomes negligible.142 Similarly, the E2g1–A1g frequency difference of ~25 cm−1 has been observed to decrease to 19 cm−1 in monolayers which allows for Raman based quantification of layer number.10, 87 These resonances have been observed to blue-shift as a function of strain, as measured on inorganic fullerenes of 3R-

MoS2.105 van Baren, et al. 69 used DFT to calculate the Raman spectra of 2H-MoS2 and 3R-MoS2 and found that the 3R-MoS2 shear peaks red-shift as a function of layer number, while the 2H-MoS2 absorptions blue- shift. Yan, et al. 143 experimentally showed that features within the ultralow frequency Raman region may discriminate different MoS2 stacking orders.

Table 1 – Raman modes for 2H-MoS2 and 3R-MoS2 Wavenumber (cm-1) Mode 3R- 2H-MoS2 Notes MoS2 E21g 383 383 Both blue shift.87 Refs.144-145 A1g 405 405 Davydov splitting under 1.96 eV laser.35 Both red shift.87 Refs.144-145 b band 420 ~429 Refs.87, 144 2LA(M) ~450 ~450 Refs.87, 146 A2u ~472 ~472 Forbidden; excited under ~2.7 eV laser.146 Shear modes 28–19 27–13 3R-MoS2 (red shift); 2H-MoS2 (blue shift).35, 69 Breathing modes 40–18 41–16 Similar thickness dependence69. Both red shift.35, 69, 146 The ranges tabulated under ‘Wavenumber’ and the behaviour under ‘Notes’ are relative to a layer number increase such that the high-end of the range corresponds to the fewest layers.

Miscellaneous Materials that can absorb or dissipate microwave radiation are important for shielding, telecommunications, and digital processor technology. High efficiency, light weight, and thin materials are considered optimal.147

Although 3R-MoS2 fulfils the latter two criteria, unfortunately the microwave absorptivity of 3R-MoS2 at 2.45 GHz is negligible. However, studies have shown that the imaginary permittivity and reflection loss are dependent on stacking order, and improve as defects are introduced into the material.148-149

MoS2 sheets are promising piezoelectric materials for use in sensors and energy converters owing to their crystallinity and strain tolerance.150 As 2H-MoS2 is centrosymmetric, in bulk it does not exhibit the piezoelectric effect, although at the few-layer scale, it does exhibit limited piezoelectric behaviour that decreases with increasing layer number. Tan, et al. 9 studied the piezoelectric constant of 3R-MoS2 by large-scale molecular dynamics simulation. They observe that the piezoelectric constant changes non-linearly as a function of layer number, reaching a maximum value of 0.457 Cm−2 at 5 layers. The non-linearity is a result of complicated surface effects and electronic interactions.

Mechanical Properties

McClung 8 describes natural 3R-MoS2 as occurring often as elliptical or ball shaped particles with a low aspect ratio, oxide-rich faces, hydrophilic, being slow to float, malleable and frequently containing a high rhenium content. These properties make it difficult to grind and separate by flotation, which results in a lower yield relative to 2H-MoS2 from MoS2-containing ore. The 3R-MoS2 phase has a marginally higher density (5.009 gcm−3) than does the 2H-MoS2 phase (4.995 gcm−3), although this is unlikely to affect processing.60 Bulk synthetic 3R-MoS2 crystals are often triangular platelet shaped and larger than 2H-MoS2 crystals grown under

39 Chapter 2

comparable conditions.15, 21, 23 Wildervanck and Jellinek 19 noted for MoS2 prepared by a carbonate flux that although 3R-MoS2 crystals comprised only 10% of the yield, that 50% of the largest crystals were 3R-MoS2.

Table 2 – Mechanical properties of MoS2 Characteristic 3R 2H Density (gcm−3) 5.009 4.995 Elastic Constants (GPa) 223 (C11), 54 (C12), 12 (C13), 223 (C11), 53 (C12), 10 (C13), 43–48b (C33), 16–17b (C44) 49-55b (C33), 15-18b (C44) Bulk Modulus (GPa)b 55 56 Shear Modulus (GPa)b 39 39 Young’s Modulus (GPa)b 95 95 Poisson’s Ratiob 0.23 0.23 Microhardness Parameter (GPa)b 7.25 7.25 Shear mode frequency, ωS, (cm−1)c 31.85 32.85 Shear force constant, Kx/y, (× 1019 Nm−3)c 2.76 2.94 Breathing mode frequency, ωB, (cm−1)c 53.85 57.40 Compressive force constant, Kz, (× 1019 7.89 8.98 Nm−3)c Data from reference 7 unless otherwise noted,b values estimated from plots in Yengejeh, et al. 7c reference.69

The elastic properties of 2H-MoS2 and 3R-MoS2 are near identical because of their similar stacking (Table 2). In general, the elastic constants are greater than those of other polytypes because of the more favourable interlayer interaction and thermodynamic stability (see General, above).7 Coutinho, et al. 4 calculated the thermodynamic properties using DFT and observed that 3R-MoS2 has a larger heat capacity than does 2H-

MoS2 for all temperatures. Khalil, et al. 145 performed a similar study using dispersion corrected DFT leading to very similar results. They also noted that the free energy of formation of 2H-MoS2 is lower than that of 3R- MoS2 for all temperatures, indicating a higher stability for the 2H-MoS2 phase.

Electrical conductivity through the basal layers of MoS2 is several orders of magnitude less than across the sheets as electrons must tunnel between layers to travel.88, 133 Ly, et al. 88 synthesised spiral pyramids of 3R-

MoS2 and found that, because the structure was composed of a single, helical sheet, a current of 60 nA could be obtained with a 1 V bias over a height of 25 nm (~40 layers) in conductance-atomic force microscopy experiments. 2H-MoS2, which cannot grow as a spiral-wound single sheet cannot exhibit this behaviour.

Mischaracterisations of MoS2

Mischaracterisation of MoS2 as the 3R-MoS2 polytype is often the result of misinterpreted XRD data. This error is most common in samples prepared by hydrothermal syntheses,151-154 but also occurs for samples obtained by a variety of procedures where disordered MoS2 is synthesised.155-156 Mischaracterisation occurs when broadening of the 2H-MoS2 10퓁 (퓁=1, 3 and 5) reflections are mistaken for the convolution of 104, 105, 107, and 108 reflections from crystalline 3R-MoS2 onto the 2H-MoS2 pattern.60 While this certainly could be the 119 157 case (as in the 50:50 2H-MoS2:3R-MoS2 mixtures simulated by Houben, et al. and Goloveshkin, et al. ), considering that the synthesis is known to produce disordered, intercalated materials,158 it is much more likely that Warren’s criterion is being satisfied where random layer (a, b plane) translation and/or rotation (Figure 19b) reduces the observed reflections to 00퓁 and hk0.159 This phenomenon was further developed by Wilson

160 and a more MoS2 focused discussion has been presented by Moser and Lévy 161. Most materials that exhibit broadening of the 00퓁 lines would be better described as randomly stacked/oriented layers of MoS2 where an intercalated compound prevents interlayer van der Waals forces from orienting the layers into the thermodynamically stable 2H-MoS2 polytype.

40 Chapter 2

Figure 19 – Diagrams of possible MoS2 layer arrangements and the effect on the x-ray diffraction reflections. Adapted with permission from Wildervanck and Jellinek 19, copyright John Wiley and Sons.

Given the rhombohedral (trigonal) crystal system of the 3R-MoS2 phase, the shape that minimises surface energy (thus the most thermodynamically stable shape) is a particle with 3-fold symmetry. Similarly, for the 2H-

MoS2 phase, the hexagonal symmetry should produce hexagonal platelets.162 However, several studies report the CVD synthesis of triangular MoS2 particles for which the morphology was determined to be 2H-MoS2. This morphology has been shown to grow in a range of shapes, i.e. triangular platelets, octahedra, wires, disks, etc. under kinetic control and hexagonal platelets under thermodynamic control.162 A kinetic regime may be induced in CVD synthesis if non-stoichiometric Mo:S ratios are used163-164 or the atoms in the support sufficiently bind to the growing MoS2 (which can even result in phase changes as a function of particle size165 and substrate lattice parameters166).167 These criteria have been demonstrated experimentally168-169 and are supported by DFT calculations.170-171 Thus, our recommendation is that particle shape not be used as an indicator of polytype for CVD grown particles.

Applications TMDs are foreseen to find use in most categories of optical modulation.172 Their semiconducting behaviour allows for use in optoelectronic devices and switches,173 complementing graphene (a conductor) for applications that require 2D materials.174 These general applications have been abundantly covered in several recent reviews.3, 172, 174 For 3R-MoS2 specifically, the breaking of inversion symmetry inhibits interlayer exciton relaxation, which generates an additional degree of freedom in the band valleys (vide supra). By using this valley degree of freedom as an information carrier, valleytronic devices might be created.10, 175 Park, et al. 176 have designed a 2×2 states per cell device with a doubled information density by harnessing the valley-index degree of freedom.

At present, there are very few reports on chemical applications of 3R-MoS2. There are some reported catalyses, such as that by Chua, et al. 177, who compared the electrocatalytic activity of several commercial MoS2 samples for the hydrogen evolution reaction (HER). Although some samples contained up to 40% 3R-MoS2, the authors reported no significant difference in activity between the 2H-MoS2 and 3R-MoS2 containing samples. Zhao, et al. 5 calculated the free energy of oxygen reduction on 2H-MoS2 and 3R-MoS2 They observed that the free energies of each intermediate, over-potentials, and reaction pathways were almost identical for each polytype.

Saber, et al. 178 report the hydrothermal synthesis, and subsequent use of 3R-MoS2 for the catalytic degradation of methyl orange. While the catalysts might be better described as disordered MoS2, the authors note that the activity of 2H-MoS2 and 3R-MoS2 is largely the same. They also perform DFT calculations on defect-rich monolayers and note that the catalytic activity for the HER is similar for the two polytypes. Toh, et al. 179 reports the synthesis of 3R-MoS2 samples in a carbonate flux for use as HER electrocatalysts. However, as their materials are likely a disordered, rather than mixed phase MoS2, the reported activity trend follows an inverse

41 Chapter 2 synthesis temperature (and by implication, crystallinity and particle size) relationship. Furthermore, the activities were not normalised for electrochemical surface area and, as a consequence, the results may well be a function of edge-site density, rather than intrinsic phase activity. In all studies reported so far, the phase of the MoS2 is not confirmed after catalysis. It is not currently known whether catalytic conditions (e.g. heat, applied voltage, photoirradiation, redox cycles, etc.) cause 3R-MoS2 to convert into the more stable 2H-MoS2 in-situ. Therefore, the catalyst comparison studies above should be interpreted with caution until the stability of 3R-MoS2 is established.

There are several properties of 3R-MoS2 that have not yet been studied for chemical reactions. The effect of layer rotation (relative to the 2H-MoS2 polytype) on the edge characteristics of 3R-MoS2 samples is not yet understood. It seems likely that the unique layer terminations will lead to tuneable reaction free energies for a range of catalytic transformations.180 The difference in band gap and optical properties of 3R-MoS2, relative to

2H-MoS2, have not yet been exploited for photocatalysis, though these appear to be promising avenues for further study (especially in few layered samples, where the layer dependent indirect band gap is expected to increase the number of charge carriers).181-182 Additionally, the unique morphologies of 3R-MoS2 crystals are yet to be used in applications that demand highly crystalline, edge-site rich materials. For example, the surface of screw-dislocation grown triangular pyramids is predominately composed of edge-sites; such edge-sites in

MoS2 more generally have been shown to be the active site for the electrocatalytic HER.183 Given that these materials are highly crystalline and highly conducting along the vertical axis, they are also promising intercalation materials for electrochemical applications such as batteries.

Outlook and Conclusions

Compared to the 2H-MoS2 morphology, 3R-MoS2 research is in its infancy. In addition to the promising optoelectronic applications, there is ample room for studies on the catalytic and electrochemical behaviour of

3R-MoS2. The areas for which MoS2 shows the most promise (i.e. as a non-noble metal catalyst and intercalation cathode) are of principle importance for investigation. To rival the activity of the highly active, nano- architectured 2H-MoS2 in the current literature, fine control over the morphology and crystallinity of 3R-MoS2 must be realised.174 TMD derivatives, such as polytypic and polyatomic heterostructures should be prepared in an attempt to optimise the HER and intercalation performance of MoS2. The reliable, high yield, high purity synthesis of bulk 3R-MoS2 remains perhaps the most significant challenge that must be overcome before the possibilities of the use of this phase can be realised at scale. To this end, bottom up syntheses such as solvothermal and CVD seem to be the most promising methods to control polytype due to the myriad of synthetic variables available. Additionally, post-synthesis treatment by defect introduction or intercalation may also provide polytypic control.

This review has summarised the history, synthetic methodologies, characteristics, and applications of 3R-MoS2; a once overlooked material that has prompted renewed excitement over its unique structural and electronic characteristics. While 3R-MoS2 may indeed prove to be revolutionary for optoelectronics, further improvements to polytype and morphology control are required before the full potential for chemical transformations can be realised.

42 Chapter 2

References 1. King, B., Minerals explained 39: molybdenite. Geology Today 2004, 20 (1), 34-37. 2. Lv, R.; Robinson, J. A.; Schaak, R. E.; Sun, D.; Sun, Y.; Mallouk, T. E.; Terrones, M., Transition metal dichalcogenides and beyond: synthesis, properties, and applications of single-and few-layer nanosheets. Acc. Chem. Res. 2015, 48 (1), 56-64. 3. Mak, K. F.; Shan, J., Photonics and optoelectronics of 2D semiconductor transition metal dichalcogenides. Nat. Photonics 2016, 10 (4), 216-226. 4. Coutinho, S. S.; Tavares, M. S.; Barboza, C. A.; Frazao, N. F.; Moreira, E.; Azevedo, D. L., 3R and 2H polytypes of MoS2: DFT and DFPT calculations of structural, optoelectronic, vibrational and thermodynamic properties. J. Phys. Chem. Solids 2017, 111, 25-33. 5. Zhao, S. Y.; Wang, K.; Zou, X. L.; Gan, L.; Du, H. D.; Xu, C. J.; Kang, F. Y.; Duan, W. H.; Li, J., Group VB transition metal dichalcogenides for oxygen reduction reaction and strain-enhanced activity governed by p-orbital electrons of chalcogen. Nano Research 2019, 12 (4), 925-930. 6. Lv, R.; Terrones, H.; Elias, A. L.; Perea-Lopez, N.; Gutierrez, H. R.; Cruz-Silva, E.; Rajukumar, L. P.; Dresselhaus, M. S.; Terrones, M., Two-dimensional transition metal dichalcogenides: Clusters, ribbons, sheets and more. Nano Today 2015, 10 (5), 559-592. 7. Yengejeh, S. I.; Liu, J. X.; Kazemi, S. A.; Wen, W.; Wang, Y., Effect of Structural Phases on Mechanical Properties of Molybdenum Disulfide. Acs Omega 2020, 5 (11), 5994-6002. 8. McClung, C. R., Molybdenite polytypism and its implications for processing and recovery: A geometallurgical-based case study from Bingham Canyon Mine, Utah. Miner. Metall. Process. 2016, 33 (3), 149-154. 9. Tan, D.; Willatzen, M.; Wang, Z. L., Prediction of strong piezoelectricity in 3R-MoS2 multilayer structures. Nano Energy 2019, 56, 512-515. 10. Suzuki, R.; Sakano, M.; Zhang, Y. J.; Akashi, R.; Morikawa, D.; Harasawa, A.; Yaji, K.; Kuroda, K.; Miyamoto, K.; Okuda, T.; Ishizaka, K.; Arita, R.; Iwasa, Y., Valley-dependent spin polarization in bulk MoS2 with broken inversion symmetry. Nat. Nanotechnol. 2014, 9 (8), 611-617. 11. Wilson, J. A.; Yoffe, A., The transition metal dichalcogenides discussion and interpretation of the observed optical, electrical and structural properties. AdPhy 1969, 18 (73), 193-335. 12. CRC handbook of chemistry and physics (Online). CRC handbook of chemistry and physics (Online) 1978. 13. Berzelius, J. J., Ueber die Schwefelsalze. AnP 1826, 82 (4), 425-458. 14. Dickinson, R. G.; Pauling, L., The Crystal Structure of Molybdenite. J. Am. Chem. Soc. 1923, 45 (6), 1466-1471. 15. Bell, R. E.; Herfert, R. E., Preparation and characterization of a new crystalline form of molybdenum disulfide. J. Am. Chem. Soc. 1957, 79 (13), 3351-3354. 16. Traill, R., A rhombohedral polytype of molybdenite. The Canadian Mineralogist 1963, 7 (3), 524-526. 17. Semiletov, S., Crystal structure of the rhombohedral molybdenum (IV) sulfide. Kristallografiya 1961, 6 (6), 536-540. 18. Jellinek, F.; Brauer, G.; Müller, H., Molybdenum and Niobium Sulphides. Natur 1960, 185 (4710), 376-377. 19. Wildervanck, J. C.; Jellinek, F., Preparation and Crystallinity of Molybdenum and Tungsten Sulfides. Z. Anorg. Allg. Chem. 1964, 328 (5‐6), 309-318. 20. Mandarino, J.; Gait, R., Molybdenite polytypes in the Royal Ontario Museum. The Canadian Mineralogist 1970, 10 (4), 723-729. 21. Guichard, M., Annales de Chimie et de Physique 1901, 7 (23), 504. 22. Hintze, C. A. F.; Linck, G. E.; Chudoba, K., Handbuch der Mineralogie: von Carl Hintze. Veit: 1904; Vol. 1. 23. De Schulten, A., Reproduction de la molybdénite. Bulletin de Minéralogie 1889, 12 (9), 545-546. 24. It should be noted that Pauling proposed three explanations for the shape discrepancy: the crystal structure was not hexagonal, the crystals were twinned, or that 3-fold symmetry dominated at scales larger than the unit cell. The first explanation was correct (if indeed the samples were 3R-MoS2), though he was not able to obtain satisfactory data to confirm this. 25. Ramsdell, L. S., Studies on silicon carbide. American Mineralogist: Journal of Earth and Planetary Materials 1947, 32 (1-2), 64- 82. 26. Guinier, A.; Bokij, G.; Boll-Dornberger, K.; Cowley, J.; Ďurovič, S.; Jagodzinski, H.; Krishna, P.; De Wolff, P.; Zvyagin, B.; Cox, D., Nomenclature of polytype structures. Report of the International Union of Crystallography Ad hoc Committee on the nomenclature of disordered, modulated and polytype structures. Acta Crystallogr. Sect. A: Found. Crystallogr. 1984, 40 (4), 399- 404. 27. Bailey, S.; Frank-Kamenetskii, V.; Goldsztaub, S.; Kato, A.; Pabst, A.; Schulz, H.; Taylor, H.; Fleischer, M.; Wilson, A., Report of the International Mineralogical Association (IMA)–International Union of Crystallography (IUCr) Joint Committee on Nomenclature. Acta Crystallographica Section A: Crystal Physics, Diffraction, Theoretical and General Crystallography 1977, 33 (4), 681-684. 28. The symmetry of the hexagonal unit cell requires that layers only need to rotate 60°, not 180°, to become antiparallel. 29. Peng, T.; Huai-Hong, G.; Teng, Y.; Zhi-Dong, Z., Stacking stability of MoS2 bilayer: an ab initio study. Chinese Physics B 2014, 23 (10), 106801. 30. Xia, M.; Li, B.; Yin, K.; Capellini, G.; Niu, G.; Gong, Y.; Zhou, W.; Ajayan, P. M.; Xie, Y.-H., Spectroscopic signatures of AA′ and AB stacking of chemical vapor deposited bilayer MoS2. ACS Nano 2015, 9 (12), 12246-12254. 31. He, J. G.; Hummer, K.; Franchini, C., Stacking effects on the electronic and optical properties of bilayer transition metal dichalcogenides MoS2, MoSe2, WS2, and WSe2. PhRvB 2014, 89 (7). 32. Takeuchi, Y.; Nowacki, W., Detailed crystal structure of rhombohedral MoS2 and systematic deduction of possible polytypes of molybdenite. Schweizer Mineralogische und Petrographische Mitteilungen 1964, 44 (1), 105-120.

43 Chapter 2

33. Chhowalla, M.; Shin, H. S.; Eda, G.; Li, L. J.; Loh, K. P.; Zhang, H., The chemistry of two-dimensional layered transition metal dichalcogenide nanosheets. Nature Chem. 2013, 5 (4), 263-275. 34. Kertesz, M.; Hoffmann, R., Octahedral vs. trigonal-prismatic coordination and clustering in transition-metal dichalcogenides. J. Am. Chem. Soc. 1984, 106 (12), 3453-3460. 35. Na, W.; Kim, K.; Lee, J. U.; Cheong, H., Davydov splitting and polytypism in few-layer MoS2. 2d Materials 2019, 6 (1). 36. Chang, C.-H.; Fan, X.; Lin, S.-H.; Kuo, J.-L., Orbital analysis of electronic structure and phonon dispersion in MoS 2, MoSe 2, WS 2, and WSe 2 monolayers under strain. PhRvB 2013, 88 (19), 195420. 37. Fan, X. F.; Zheng, W. T.; Kuo, J. L.; Singh, D. J.; Sun, C. Q.; Zhu, W., Modulation of electronic properties from stacking orders and spin-orbit coupling for 3R-type MoS2. Scientific Reports 2016, 6. 38. Voiry, D.; Mohite, A.; Chhowalla, M., Phase engineering of transition metal dichalcogenides. Chem. Soc. Rev. 2015, 44 (9), 2702-2712. 39. Deepak, F. L.; Mayoral, A.; Steveson, A. J.; Mejía-Rosales, S.; Blom, D. A.; José-Yacamán, M., Insights into the capping and structure of MoS2 nanotubes as revealed by aberration-corrected STEM. Nanoscale 2010, 2 (10), 2286-2293. 40. Weiss, K.; Phillips, J. M., Calculated specific surface energy of molybdenite (MoS 2). PhRvB 1976, 14 (12), 5392. 41. Rydberg, H.; Dion, M.; Jacobson, N.; Schröder, E.; Hyldgaard, P.; Simak, S. I.; Langreth, D. C.; Lundqvist, B. I., Van der Waals Density Functional for Layered Structures. Phys. Rev. Lett. 2003, 91 (12), 126402. 42. Newberry, R. J., Polytypism in molybdenite (I): a nonequilibrium impurity-induced phenomenon. Am. Mineral. 1979, 64, 758- 767. 43. Wang, S.; Zhang, J.; He, D.; Zhang, Y.; Wang, L.; Xu, H.; Wen, X.; Ge, H.; Zhao, Y., Sulfur-catalyzed phase transition in MoS2 under high pressure and temperature. J. Phys. Chem. Solids 2014, 75 (1), 100-104. 44. Wickman, F. E.; Smith, D. K., Molybdenite polytypes in theory and occurrence. I. Theoretical considerations of polytypism in molybdenite. American Mineralogist: Journal of Earth and Planetary Materials 1970, 55 (11-12), 1843-1856. 45. In this review, "bulk" refers to samples above the "thin film limit" or "few-layer limit" (i.e. ~20 layers; where the phenomena observed at the few-layer scale cease). . 46. Wypych, F.; Schöllhorn, R., 1T-MoS2, a new metallic modification of molybdenum disulfide. J. Chem. Soc., Chem. Commun. 1992, (19), 1386-1388. 47. Caputo, R., Polytypism of MoS2. Jacobs Journal of Inorganic Chemistry 2016, 1 (1), 004. 48. Lévy, F. A., Intercalated layered materials. 1 ed.; Springer Science & Business Media: 2012; Vol. 6. 49. Zhao, W.; Pan, J.; Fang, Y.; Che, X.; Wang, D.; Bu, K.; Huang, F., Metastable MoS2: crystal structure, electronic band structure, synthetic approach and intriguing physical properties. Chemistry–A European Journal 2018, 24 (60), 15942-15954. 50. Huang, Q.; Li, X.; Sun, M.; Zhang, L.; Song, C.; Zhu, L.; Chen, P.; Xu, Z.; Wang, W.; Bai, X., The Mechanistic Insights into the 2H-1T Phase Transition of MoS2 upon Alkali Metal Intercalation: From the Study of Dynamic Sodiation Processes of MoS2 Nanosheets. Advanced Materials Interfaces 2017, 4 (15), 1700171. 51. Manzeli, S.; Ovchinnikov, D.; Pasquier, D.; Yazyev, O. V.; Kis, A., 2D transition metal dichalcogenides. Nature Reviews Materials 2017, 2 (8). 52. Elwell, D.; Neate, B., Mechanisms of crystal growth from fluxed melts. J. Mater. Sci. 1971, 6 (12), 1499-1519. 53. Frank, F., The influence of dislocations on crystal growth. Discussions of the Faraday Society 1949, 5, 48-54. 54. Agarwal, M. K.; Joseph, B., Dislocation etching in molybdenite. J. Mater. Sci. 1974, 9 (8), 1262-1264. 55. Zhang, L.; Liu, K.; Wong, A. B.; Kim, J.; Hong, X.; Liu, C.; Cao, T.; Louie, S. G.; Wang, F.; Yang, P., Three-dimensional spirals of atomic layered MoS2. Nano Lett. 2014, 14 (11), 6418-6423. 56. Nie, Y. F.; Barton, A. T.; Addou, R.; Zheng, Y. P.; Walsh, L. A.; Eichfeld, S. M.; Yue, R. Y.; Cormier, C. R.; Zhang, C. X.; Wang, Q. X.; Bang, C. P.; Robinson, J. A.; Kim, M.; Vandenberghe, W.; Colombo, L.; Cha, P. R.; Wallace, R. M.; Hinkle, C. L.; Cho, K., Dislocation driven spiral and non-spiral growth in layered chalcogenides. Nanoscale 2018, 10 (31), 15023-15034. 57. Toh, R. J.; Sofer, Z.; Luxa, J.; Sedmidubský, D.; Pumera, M., 3R phase of MoS2 and WS2 outperforms the corresponding 2H phase for hydrogen evolution. Chem. Commun. 2017, 53 (21), 3054-3057. 58. Lieth, R., Preparation and crystal growth of materials with layered structures. Springer Science & Business Media: 2013; Vol. 1. 59. Milbauer, J., Über die Einwirkung des Sulfocyankaliums auf Metalloxyde bei höheren Temperaturen. Zeitschrift für anorganische Chemie 2013, 42 (1), 433-449. 60. Wang, S. M.; Zhang, J. Z.; He, D. W.; Zhang, Y.; Wang, L. P.; Xu, H. W.; Wen, X. D.; Ge, H.; Zhao, Y. S., Sulfur-catalyzed phase transition in MoS2 under high pressure and temperature. J. Phys. Chem. Solids 2014, 75 (1), 100-104. 61. Towle, L. C.; Oberbeck, V.; Brown†, B. E.; Stajdohar, R. E., Molybdenum Diselenide: Rhombohedral High Pressure-High Temperature Polymorph. Sci 1966, 154 (3751), 895-896. 62. Silverman, M. S., Ultrahigh pressure-high temperature synthesis of rhombohedral dichalcogenides of molybdenum and tungsten. Inorg. Chem. 1967, 6 (5), 1063-1064. 63. Takahashi, N.; Shiojiri, M., Stacking faults in hexagonal and rhombohedral MoS2 crystals produced by mechanical operation in relation to lubrication. Wear 1993, 167 (2), 163-171. 64. Bozheyev, F.; An, V.; Irtegov, Y., Properties of Copper and Molybdenum Sulfide Powders Produced by Self-propagating High- temperature Synthesis. In Nanomaterials for Structural, Functional and Biomedical Applications, Khasanov, O., Ed. 2014; Vol. 872, pp 191-196. 65. Shi, J.; Yu, P.; Liu, F.; He, P.; Wang, R.; Qin, L.; Zhou, J.; Li, X.; Zhou, J.; Sui, X., 3R MoS2 with broken inversion symmetry: a promising ultrathin nonlinear optical device. Adv. Mater. 2017, 29 (30), 1701486. 66. Zhao, M.; Ye, Z. L.; Suzuki, R.; Ye, Y.; Zhu, H. Y.; Xiao, J.; Wang, Y.; Iwasa, Y.; Zhang, X., Atomically phase-matched second- harmonic generation in a 2D crystal. Light-Science & Applications 2016, 5.

44 Chapter 2

67. Remškar, M.; Škraba, Z.; Ballif, C.; Sanjines, R.; Levy, F., Stabilization of the rhombohedral polytype in MoS2 and WS2 microtubes: TEM and AFM study. Surf. Sci. 1999, 433, 637-641. 68. Klosse, K.; Ullersma, P., Convection in a chemical vapor transport process. J. Cryst. Growth 1973, 18 (2), 167-174. 69. van Baren, J.; Ye, G. H.; Yan, J. A.; Ye, Z. P.; Rezaie, P.; Yu, P.; Liu, Z.; He, R.; Lui, C. H., Stacking-dependent interlayer phonons in 3R and 2H MoS2. 2d Materials 2019, 6 (2). 70. Sigiro, M.; Nasruddin, M. N., Growth and optical characterization of MoS2 single crystals with different dopants. Optik 2015, 126 (6), 666-670. 71. Li, Q.; Walter, E.; Van der Veer, W.; Murray, B.; Newberg, J.; Bohannan, E.; Switzer, J.; Hemminger, J.; Penner, R., Molybdenum disulfide nanowires and nanoribbons by electrochemical/chemical synthesis. The Journal of Physical Chemistry B 2005, 109 (8), 3169-3182. 72. Leidich, S.; Buechele, D.; Lauenstein, R.; Kluenker, M.; Lind, C., “Non-hydrolytic” sol–gel synthesis of molybdenum sulfides. J. Solid State Chem. 2016, 242, 175-181. 73. Novoselov, K. S.; Jiang, D.; Schedin, F.; Booth, T. J.; Khotkevich, V. V.; Morozov, S. V.; Geim, A. K., Two-dimensional atomic crystals. Proc. Natl. Acad. Sci. U. S. A. 2005, 102 (30), 10451. 74. Frindt, R.; Yoffe, A., Physical properties of layer structures: optical properties and photoconductivity of thin crystals of molybdenum disulphide. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences 1963, 273 (1352), 69-83. 75. Kumar, P.; Horwath, J. P.; Foucher, A. C.; Price, C. C.; Acero, N.; Shenoy, V. B.; Stach, E. A.; Jariwala, D., Direct visualization of out-of-equilibrium structural transformations in atomically thin chalcogenides. Npj 2d Materials and Applications 2020, 4 (1). 76. Shiojiri, M.; Isshiki, T.; Saijo, H.; Yabuuchi, Y.; Takahashi, N., Cross-Sectional Observations of Layer Structures and Stacking- Faults in Natural and Synthesized Molybdenum-Disulfide Crystals by High-Resolution Transmission Electron-Microscopy. J. Electron Microsc. 1993, 42 (2), 72-78. 77. Isshiki, T.; Nishio, K.; Saijo, H.; Shiojiri, M.; Yabuuchi, Y.; Takahashi, N., High‐resolution transmission electron microscopy of hexagonal and rhombohedral molybdenum disulfide crystals. Microsc. Res. Tech. 1993, 25 (4), 325-334. 78. Macchione, M. A.; Mendoza-Cruz, R.; Bazan-Diaz, L.; Velazquez-Salazar, J. J.; Santiago, U.; Arellano-Jimenez, M. J.; Perez, J. F.; Jose-Yacaman, M.; Samaniego-Benitez, J. E., Electron microscopy study of the carbon-induced 2H-3R-1T phase transition of MoS2. New J. Chem. 2020, 44 (4), 1190-1193. 79. Butler, S. Z.; Hollen, S. M.; Cao, L. Y.; Cui, Y.; Gupta, J. A.; Gutierrez, H. R.; Heinz, T. F.; Hong, S. S.; Huang, J. X.; Ismach, A. F.; Johnston-Halperin, E.; Kuno, M.; Plashnitsa, V. V.; Robinson, R. D.; Ruoff, R. S.; Salahuddin, S.; Shan, J.; Shi, L.; Spencer, M. G.; Terrones, M.; Windl, W.; Goldberger, J. E., Progress, Challenges, and Opportunities in Two-Dimensional Materials Beyond Graphene. ACS Nano 2013, 7 (4), 2898-2926. 80. Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M. S., Electronics and optoelectronics of two-dimensional transition metal dichalcogenides. Nat. Nanotechnol. 2012, 7 (11), 699-712. 81. Choy, K. L., Chemical vapour deposition of coatings. Prog. Mater Sci. 2003, 48 (2), 57-170. 82. Chen, J.; Zhao, X.; Grinblat, G.; Chen, Z.; Tan, S. J.; Fu, W.; Ding, Z.; Abdelwahab, I.; Li, Y.; Geng, D., Homoepitaxial Growth of Large‐Scale Highly Organized Transition Metal Dichalcogenide Patterns. Adv. Mater. 2018, 30 (4), 1704674. 83. Schönfeld, B.; Huang, J.; Moss, S., Anisotropic mean-square displacements (MSD) in single-crystals of 2H-and 3R-MoS2. Acta Crystallogr. Sect. B: Struct. Sci. 1983, 39 (4), 404-407. 84. Zhou, S.; Wang, S. S.; Shi, Z.; Sawada, H.; Kirkland, A. I.; Li, J.; Warner, J. H., Atomically sharp interlayer stacking shifts at anti- phase grain boundaries in overlapping MoS2 secondary layers. Nanoscale 2018, 10 (35), 16692-16702. 85. Yan, A. M.; Chen, W.; Ophus, C.; Ciston, J.; Lin, Y. Y.; Persson, K.; Zettl, A., Identifying different stacking sequences in few- layer CVD-grown MoS2 by low-energy atomic-resolution scanning transmission electron microscopy. PhRvB 2016, 93 (4). 86. Meier, A. J.; Garg, A.; Sutter, B.; Kuhn, J. N.; Bhethanabotla, V. R., MoS2 Nanoflowers as a Gateway for Solar-Driven CO2 Photoreduction. Acs Sustainable Chemistry & Engineering 2019, 7 (1), 265-275. 87. Zhang, J. L.; Ye, M. X.; Bhandari, S.; Muqri, A. K. M.; Long, F.; Bigham, S.; Yap, Y. K.; Suh, J. Y., Enhanced second and third harmonic generations of vertical and planar spiral MoS2 nanosheets. Nanot 2017, 28 (29). 88. Ly, T. H.; Zhao, J.; Kim, H.; Han, G. H.; Nam, H.; Lee, Y. H., Vertically Conductive MoS2 Spiral Pyramid. Adv. Mater. 2016, 28 (35), 7723-+. 89. Zhou, Q. W.; Su, S. Q.; Cheng, P. F.; Hu, X. B.; Gao, X. S.; Zhang, Z.; Liu, J. M., Vertically conductive MoS2 pyramids with a high density of active edge sites for efficient hydrogen evolution. Journal of Materials Chemistry C 2020, 8 (9), 3017-3022. 90. Wu, J. Y.; He, J. Y.; Ariza, P.; Ortiz, M.; Zhang, Z. L., Supergiant elasticity and fracture of 3D spirally wound MoS2. IJFr. 91. Samaniego-Benitez, J. E.; Mendoza-Cruz, R.; Bazan-Diaz, L.; Garcia-Garcia, A.; Arellano-Jimenez, M. J.; Perez-Robles, J. F.; Plascencia-Villa, G.; Velazquez-Salazar, J. J.; Ortega, E.; Favela-Camacho, S. E.; Jose-Yacaman, M., Synthesis and structural characterization of MoS2 micropyramids. J. Mater. Sci. 2020, 55 (26), 12203-12213. 92. Roy, A.; Ghosh, R.; Rai, A.; Sanne, A.; Kim, K.; Movva, H. C. P.; Dey, R.; Pramanik, T.; Chowdhury, S.; Tutuc, E.; Banerjee, S. K., Intra-domain periodic defects in monolayer MoS2. Appl. Phys. Lett. 2017, 110 (20). 93. Kumar, P.; Viswanath, B., Effect of sulfur evaporation rate on screw dislocation driven growth of MoS2 with high atomic step density. Cryst. Growth Des. 2016, 16 (12), 7145-7154. 94. Hao, S.; Yang, B. C.; Gao, Y. L., Controllable growth and characterizations of hybrid spiral-like atomically thin molybdenum disulfide. Physica E-Low-Dimensional Systems & Nanostructures 2016, 84, 378-383. 95. Dong, X.; Yan, C.; Tomer, D.; Li, C. H.; Li, L., Spiral growth of few-layer MoS2 by chemical vapor deposition. Appl. Phys. Lett. 2016, 109 (5).

45 Chapter 2

96. Diep, N. Q.; Liu, C.-W.; Wu, S.-K.; Chou, W.-C.; Huynh, S. H.; Chang, E. Y., Screw-Dislocation-Driven Growth Mode in Two Dimensional GaSe on GaAs (001) Substrates Grown by Molecular Beam Epitaxy. Scientific reports 2019, 9 (1), 1-8. 97. Chen, L.; Liu, B.; Abbas, A. N.; Ma, Y.; Fang, X.; Liu, Y.; Zhou, C., Screw-dislocation-driven growth of two-dimensional few- layer and pyramid-like WSe2 by sulfur-assisted chemical vapor deposition. ACS Nano 2014, 8 (11), 11543-11551. 98. Sarma, P. V.; Patil, P. D.; Barman, P. K.; Kini, R. N.; Shaijumon, M. M., Controllable growth of few-layer spiral WS2. RSC Advances 2016, 6 (1), 376-382. 99. Wang, X.; Yang, H.; Yang, R.; Wang, Q.; Zheng, J.; Qiao, L.; Peng, X.; Li, Y.; Chen, D.; Xiong, X., Weakened interlayer coupling in two-dimensional MoSe 2 flakes with screw dislocations. Nano Research 2019, 12 (8), 1900-1905. 100. Cevallos, F. A.; Guo, S.; Heo, H.; Scuri, G.; Zhou, Y.; Sung, J. O.; Taniguchi, T.; Watanabe, K.; Kim, P.; Park, H.; Cava, R. J., Liquid Salt Transport Growth of Single Crystals of the Layered Dichalcogenides MoS2 and WS2. Cryst. Growth Des. 2019, 19 (10), 5762-5767. 101. Li, S.; Lin, Y.-C.; Zhao, W.; Wu, J.; Wang, Z.; Hu, Z.; Shen, Y.; Tang, D.-M.; Wang, J.; Zhang, Q., Vapour–liquid–solid growth of monolayer MoS 2 nanoribbons. Nat. Mater. 2018, 17 (6), 535. 102. Weston, A.; Zou, Y. C.; Enaldiev, V.; Summerfield, A.; Clark, N.; Zolyomi, V.; Graham, A.; Yelgel, C.; Magorrian, S.; Zhou, M. W.; Zultak, J.; Hopkinson, D.; Barinov, A.; Bointon, T. H.; Kretinin, A.; Wilsons, N. R.; Beton, P. H.; Fal'ko, V. I.; Haigh, S. J.; Gorbachev, R., Atomic reconstruction in twisted bilayers of transition metal dichalcogenides. Nat. Nanotechnol. 103. Paradisanos, I.; Shree, S.; George, A.; Leisgang, N.; Robert, C.; Watanabe, K.; Taniguchi, T.; Warburton, R. J.; Turchanin, A.; Marie, X.; Gerber, I. C.; Urbaszek, B., Controlling interlayer excitons in MoS2 layers grown by chemical vapor deposition. Nature Communications 2020, 11 (1). 104. Enyashin, A.; Gemming, S.; Seifert, G., Nanosized allotropes of molybdenum disulfide. The European Physical Journal Special Topics 2007, 149 (1), 103-125. 105. Viršek, M.; Jesih, A.; Milošević, I.; Damnjanović, M.; Remškar, M., Raman scattering of the MoS2 and WS2 single nanotubes. Surf. Sci. 2007, 601 (13), 2868-2872. 106. Remškar, M.; Mrzel, A.; Viršek, M.; Jesih, A., Inorganic nanotubes as nanoreactors: the first MoS2 nanopods. Adv. Mater. 2007, 19 (23), 4276-4278. 107. Milosevic, I.; Vukovic, T.; Damnjanovic, M.; Nikolic, B., Symmetry based properties of the transition metal dichalcogenide nanotubes. EPJB 2000, 17 (4), 707-712. 108. Ayres, D., Distribution and occurrence of some naturally‐occurring polytypes of molybdenite in Australia and Papua New Guinea. Journal of the Geological Society of Australia 1974, 21 (3), 273-278. 109. Frondel, J. W.; Wickman, F. E., Molybdenite polytypes in theory and occurrence. II. Some naturally-occurring polytypes of molybdenite. American Mineralogist: Journal of Earth and Planetary Materials 1970, 55 (11-12), 1857-1875. 110. Barton, I. F.; Rathkopf, C. A.; Barton, M. D., Rhenium in Molybdenite: a Database Approach to Identifying Geochemical Controls on the Distribution of a Critical Element. Mining, Metallurgy & Exploration 2020, 37 (1), 21-37. 111. For further information, the reader is encouraged to use the search terms “Molybdenite” and “3R” in a database that contains petrographical, mineralogical, and geological entries. 112. Ciobanu, C.; Cook, N.; Kelson, C.; Guerin, R.; Kalleske, N.; Danyushevsky, L., Trace element heterogeneity in molybdenite fingerprints stages of mineralization. Chem. Geol. 2013, 347, 175-189. 113. Newberry, R., Polytypism in molybdenite (II); Relationships between polytypism, ore deposition/alteration stages and rhenium contents. Am. Mineral. 1979, 64 (7-8), 768-775. 114. Rathkopf, C.; Mazdab, F.; Barton, I.; Barton, M. D., Grain-scale and deposit-scale heterogeneity of Re distribution in molybdenite at the Bagdad porphyry Cu-Mo deposit, Arizona. J. Geochem. Explor. 2017, 178, 45-54. 115. John, D. A.; Taylor, R. D., By-products of porphyry copper and molybdenum deposits: Chapter 7. 116. Luck, J. M.; Allègre, C. J., The study of molybdenites through the187Re-187Os chronometer. Earth. Planet. Sci. Lett. 1982, 61 (2), 291-296. 117. Stein, H.; Markey, R.; Morgan, J.; Hannah, J.; Scherstén, A., The remarkable Re–Os chronometer in molybdenite: how and why it works. Terra Nova 2001, 13 (6), 479-486. 118. Li, H.; He, T. S.; Wang, Y. B.; Jin, J. P.; Yuan, H., XRD and SEM Analyses of Molybdenite with Different Particle Sizes and Its Floatability Difference. Spectroscopy and Spectral Analysis 2018, 38 (11), 3588-3592. 119. Houben, L.; Enyashin, A. N.; Feldman, Y.; Rosentsveig, R.; Stroppa, D. G.; Bar-Sadan, M., Diffraction from Disordered Stacking Sequences in MoS2 and WS2 Fullerenes and Nanotubes. J. Phys. Chem. C 2012, 116 (45), 24350-24357. 120. Plotinskaya, O. Y.; Shilovskikh, V. V.; Najorka, J.; Kovalchuk, E. V.; Seltmann, R.; Spratt, J., Grain-scale distribution of molybdenite polytypes versus rhenium contents: mu XRD and EBSD data. Mineral. Mag. 2019, 83 (5), 639-644. 121. Enyashin, A. N.; Bar-Sadan, M.; Houben, L.; Seifert, G., Line defects in molybdenum disulfide layers. The Journal of Physical Chemistry C 2013, 117 (20), 10842-10848. 122. Zhao, X. X.; Ning, S. C.; Fu, W.; Pennycook, S. J.; Loh, K. P., Differentiating Polymorphs in Molybdenum Disulfide via Electron Microscopy. Adv. Mater. 2018, 30 (47). 123. Isshiki, T.; Nishio, K.; Saijo, H.; Shiojiri, M.; Yabuuchi, Y.; Takahashi, N., High-Resolution Transmission Electron-Microscopy of Hexagonal and Rhombohedral Molybdenum-Disulfide Crystals. Microsc. Res. Tech. 1993, 25 (4), 325-334. 124. Zhou, S.; Wang, S. S.; Li, H. S.; Xu, W. S.; Gong, C. C.; Grossman, J. C.; Warner, J. H., Atomic Structure and Dynamics of Defects in 2D MoS2 Bilayers. ACS Omega 2017, 2 (7), 3315-3324. 125. Egerton, R., Control of radiation damage in the TEM. Ultramicroscopy 2013, 127, 100-108. 126. Uyeda, R.; Nonoyama, M., The Observation of Thick Specimens by High Voltage Electron Microscopy. II. Experiment with Molybdenite Films at 50-1200 kV. JaJAP 1968, 7 (3), 200.

46 Chapter 2

127. Tatlock, G.; Steeds, J., Real space crystallography in molybdenite. NPhS 1973, 246 (155), 126-128. 128. Tatlock, G., The simulation and interpretation of zone axis patterns in molybdenite. Philos. Mag. 1975, 32 (6), 1159-1170. 129. Song, I.; Park, C.; Choi, H. C., Synthesis and properties of molybdenum disulphide: from bulk to atomic layers. RSC Advances 2015, 5 (10), 7495-7514. 130. He, Z. L.; Que, W. X., Molybdenum disulfide nanomaterials: Structures, properties, synthesis and recent progress on hydrogen evolution reaction. Applied Materials Today 2016, 3, 23-56. 131. Ci, P. H.; Chen, Y. B.; Kang, J.; Suzuki, R.; Choe, H. S.; Suh, J.; Ko, C.; Park, T.; Shen, K.; Iwasa, Y.; Tongay, S.; Ager, J. W.; Wang, L. W.; Wu, J. Q., Quantifying van der Waals Interactions in Layered Transition Metal Dichalcogenides from Pressure- Enhanced Valence Band Splitting. Nano Lett. 2017, 17 (8), 4982-4988. 132. He, W.; Shi, J.; Zhao, H. K.; Wang, H.; Liu, X. F.; Shi, X. H., Bandgap engineering of few-layered MoS2 with low concentrations of S vacancies. RSC Advances 2020, 10 (27), 15702-15706. 133. Tiong, K. K.; Liao, P. C.; Ho, C. H.; Huang, Y. S., Growth and characterization of rhenium-doped MoS2 single crystals. J. Cryst. Growth 1999, 205 (4), 543-547. 134. Roxlo, C.; Chianelli, R.; Deckman, H.; Ruppert, A.; Wong, P., Bulk and surface optical absorption in molybdenum disulfide. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1987, 5 (4), 555-557. 135. Clark, A.; Williams, R., The optical absorption properties of synthetic MoS2. J. Phys. D: Appl. Phys. 1968, 1 (9), 1222. 136. Akashi, R.; Ochi, M.; Bordacs, S.; Suzuki, R.; Tokura, Y.; Iwasa, Y.; Arita, R., Two-Dimensional Valley Electrons and Excitons in Noncentrosymmetric 3R-MoS2. Physical Review Applied 2015, 4 (1). 137. Chen, X.; Chen, Z.; Li, J., Critical electronic structures controlling phase transitions induced by lithium ion intercalation in molybdenum disulphide. Chin. Sci. Bull. 2013, 58 (14), 1632-1641. 138. Su, W. T.; Jin, L.; Qu, X. D.; Huo, D. X.; Yang, L., Defect passivation induced strong photoluminescence enhancement of rhombic monolayer MoS2. PCCP 2016, 18 (20), 14001-14006. 139. Suzuki, R.; Sakano, M.; Zhang, Y.; Akashi, R.; Morikawa, D.; Harasawa, A.; Yaji, K.; Kuroda, K.; Miyamoto, K.; Okuda, T., Valley-dependent spin polarization in bulk MoS 2 with broken inversion symmetry. Nat. Nanotechnol. 2014, 9 (8), 611. 140. Wagoner, G. A.; Persans, P. D.; Van Wagenen, E. A.; Korenowski, G. M., Second-harmonic generation in molybdenum disulfide. Journal of the Optical Society of America B-Optical Physics 1998, 15 (3), 1017-1021. 141. Mattheiss, L., Band structures of transition-metal-dichalcogenide layer compounds. PhRvB 1973, 8 (8), 3719. 142. Liang, L.; Puretzky, A. A.; Sumpter, B. G.; Meunier, V., Interlayer bond polarizability model for stacking-dependent low- frequency Raman scattering in layered materials. Nanoscale 2017, 9 (40), 15340-15355. 143. Yan, J.; Xia, J.; Wang, X.; Liu, L.; Kuo, J.-L.; Tay, B. K.; Chen, S.; Zhou, W.; Liu, Z.; Shen, Z. X., Stacking-dependent interlayer coupling in trilayer MoS2 with broken inversion symmetry. Nano Lett. 2015, 15 (12), 8155-8161. 144. Livneh, T.; Spanier, J. E., A comprehensive multiphonon spectral analysis in MoS2. 2D Materials 2015, 2 (3), 035003. 145. Khalil, R. M. A.; Hussain, F.; Rana, A. M.; Imran, M.; Murtaza, G., Comparative study of polytype 2H-MoS2 and 3R-MoS2 systems by employing DFT. Physica E-Low-Dimensional Systems & Nanostructures 2019, 106, 338-345. 146. Lee, J.-U.; Kim, K.; Han, S.; Ryu, G. H.; Lee, Z.; Cheong, H., Raman signatures of polytypism in molybdenum disulfide. ACS Nano 2016, 10 (2), 1948-1953. 147. Liu, J.; Che, R.; Chen, H.; Zhang, F.; Xia, F.; Wu, Q.; Wang, M., Microwave Absorption Enhancement of Multifunctional Composite Microspheres with Spinel Fe3O4 Cores and Anatase TiO2 Shells. Small 2012, 8 (8), 1214-1221. 148. Ning, M.-Q.; Lu, M.-M.; Li, J.-B.; Chen, Z.; Dou, Y.-K.; Wang, C.-Z.; Rehman, F.; Cao, M.-S.; Jin, H.-B., Two-dimensional nanosheets of MoS 2: a promising material with high dielectric properties and microwave absorption performance. Nanoscale 2015, 7 (38), 15734-15740. 149. Hayashi, K.; Serikawa, D.; Chijimatsu, Y.; Shimakawa, M.; Kume, S.; Manabe, K.; Takahashi, T., Microwave absorption of mixed-layer transition metal dichalcogenides. J. Alloys Compd. 1997, 262, 325-330. 150. Wu, W.; Wang, L.; Li, Y.; Zhang, F.; Lin, L.; Niu, S.; Chenet, D.; Zhang, X.; Hao, Y.; Heinz, T. F.; Hone, J.; Wang, Z. L., Piezoelectricity of single-atomic-layer MoS2 for energy conversion and piezotronics. Natur 2014, 514 (7523), 470-474. 151. Swathi, S.; Rani, B. J.; Ravi, G.; Yuvakkumar, R., Dopant Influence on Phase and Electrochemical Performance of Molybdenum Sulfide Nanostructures. In Proceedings of the International Conference on Advanced Materials, Sadasivuni, K. K.; Kurian, J.; Damodaran, S. V.; Joseph, J.; Joseph, D.; Tom, E.; Thomas, D., Eds. 2019; Vol. 2162. 152. Rani, B. J.; Pradeepa, S. S.; Hasan, Z. M.; Ravi, G.; Yuvakkumar, R.; Hong, S. I., Supercapacitor and OER activity of transition metal (Mo, Co, Cu) sulphides. J. Phys. Chem. Solids 2020, 138. 153. Ali, B. A.; Omar, A. M. A.; Khalil, A. S. G.; Allam, N. K., Untapped Potential of Polymorph MoS2: Tuned Cationic Intercalation for High-Performance Symmetric Supercapacitors. ACS Appl. Mater. Interfaces 2019, 11 (37), 33955-33965. 154. Geioushy, R. A.; El-Sheikh, S. M.; Hegazy, I. M.; Shawky, A.; El-Sherbiny, S.; Kandil, A. H. T., Insights into two-dimensional MoS2 sheets for enhanced CO2 photoreduction to C-1 and C-2 hydrocarbon products. Mater. Res. Bull. 2019, 118. 155. Wang, H.; Skeldon, P.; Thompson, G.; Wood, G., Synthesis and characterization of molybdenum disulphide formed from ammonium tetrathiomolybdate. J. Mater. Sci. 1997, 32 (2), 497-502. 156. Moser, J.; Levy, F.; Bussy, F., Composition and growth mode of MoS x sputtered films. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1994, 12 (2), 494-500. 157. Goloveshkin, A. S.; Bushmarinov, I. S.; Lenenko, N. D.; Buzin, M. I.; Golub, A. S.; Antipin, M. Y., Structural Properties and Phase Transition of Exfoliated-Restacked Molybdenum Disulfide. The Journal of Physical Chemistry C 2013, 117 (16), 8509- 8515. 158. Geng, X.; Sun, W.; Wu, W.; Chen, B.; Al-Hilo, A.; Benamara, M.; Zhu, H.; Watanabe, F.; Cui, J.; Chen, T.-p., Pure and stable metallic phase molybdenum disulfide nanosheets for hydrogen evolution reaction. Nature communications 2016, 7, 10672. 159. Warren, B. E., X-Ray Diffraction in Random Layer Lattices. PhRv 1941, 59 (9), 693-698.

47 Chapter 2

160. Wilson, A., X-ray diffraction by random layers: ideal line profiles and determination of structure amplitudes from observed line profiles. Acta Crystallographica 1949, 2 (4), 245-251. 161. Moser, J.; Lévy, F., Random stacking in MoS2−x sputtered thin films. Thin Solid Films 1994, 240 (1-2), 56-59. 162. Xie, S.; Xu, M.; Liang, T.; Huang, G.; Wang, S.; Xue, G.; Meng, N.; Xu, Y.; Chen, H.; Ma, X., A high-quality round-shaped monolayer MoS 2 domain and its transformation. Nanoscale 2016, 8 (1), 219-225. 163. Mahyavanshi, R. D.; Kalita, G.; Sharma, K. P.; Kondo, M.; Dewa, T.; Kawahara, T.; Tanemura, M., Synthesis of MoS2 ribbons and their branched structures by chemical vapor deposition in sulfur-enriched environment. Appl. Surf. Sci. 2017, 409, 396- 402. 164. Yang, Y.; Pu, H. B.; Di, J. J.; Zhang, S.; Chen, C. L.; Zang, Y.; Wang, X., Morphology engineering of MoS2 nanostructures by controlling MoO3-x concentration using a quasi-closed crucible. Chem. Phys. 2018, 513, 78-82. 165. Bruix, A.; Lauritsen, J. V.; Hammer, B., Size-dependent phase transitions in MoS2 nanoparticles controlled by a metal substrate. arXiv:1805.01244 2018. 166. Duerloo, K.-A. N.; Li, Y.; Reed, E. J., Structural phase transitions in two-dimensional Mo-and W-dichalcogenide monolayers. Nature communications 2014, 5 (1), 1-9. 167. Wang, X.; Feng, H.; Wu, Y.; Jiao, L., Controlled synthesis of highly crystalline MoS2 flakes by chemical vapor deposition. J. Am. Chem. Soc. 2013, 135 (14), 5304-5307. 168. Chen, W.; Zhao, J.; Zhang, J.; Gu, L.; Yang, Z.; Li, X.; Yu, H.; Zhu, X.; Yang, R.; Shi, D.; Lin, X.; Guo, J.; Bai, X.; Zhang, G., Oxygen-Assisted Chemical Vapor Deposition Growth of Large Single-Crystal and High-Quality Monolayer MoS2. J. Am. Chem. Soc. 2015, 137 (50), 15632-15635. 169. Yang, S. Y.; Shim, G. W.; Seo, S.-B.; Choi, S.-Y., Effective shape-controlled growth of monolayer MoS 2 flakes by powder- based chemical vapor deposition. Nano Research 2017, 10 (1), 255-262. 170. Schweiger, H.; Raybaud, P.; Kresse, G.; Toulhoat, H., Shape and edge sites modifications of MoS2 catalytic nanoparticles induced by working conditions: a theoretical study. J. Catal. 2002, 207 (1), 76-87. 171. Joswig, J.-O.; Lorenz, T.; Wendumu, T. B.; Gemming, S.; Seifert, G., Optics, mechanics, and energetics of two-dimensional MoS2 nanostructures from a theoretical perspective. Acc. Chem. Res. 2015, 48 (1), 48-55. 172. Sun, Z. P.; Martinez, A.; Wang, F., Optical modulators with 2D layered materials. Nat. Photonics 2016, 10 (4), 227-238. 173. Park, J.; Yeu, I. W.; Han, G.; Hwang, C. S.; Choi, J.-H., Ferroelectric switching in bilayer 3R MoS 2 via interlayer shear mode driven by nonlinear phononics. Scientific Reports 2019, 9 (1), 1-9. 174. Li, X.; Zhu, H., Two-dimensional MoS2: Properties, preparation, and applications. Journal of Materiomics 2015, 1 (1), 33-44. 175. Xu, X.; Yao, W.; Xiao, D.; Heinz, T. F., Spin and pseudospins in layered transition metal dichalcogenides. Nat. Phys. 2014, 10 (5), 343-350. 176. Park, J.; Yeu, I. W.; Han, G.; Jang, C.; Kwak, J. Y.; Hwang, C. S.; Choi, J. H., Optical control of the layer degree of freedom through Wannier-Stark states in polar 3R MoS2. Journal of Physics-Condensed Matter 2019, 31 (31). 177. Chua, X. J.; Tan, S. M.; Chia, X. Y.; Sofer, Z.; Luxa, J.; Pumera, M., The Origin of MoS2 Significantly Influences Its Performance for the Hydrogen Evolution Reaction due to Differences in Phase Purity. Chemistry-a European Journal 2017, 23 (13), 3169- 3177. 178. Saber, M. R.; Khabiri, G.; Maarouf, A. A.; Ulbricht, M.; Khalil, A. S., A comparative study on the photocatalytic degradation of organic dyes using hybridized 1T/2H, 1T/3R and 2H MoS 2 nano-sheets. RSC Advances 2018, 8 (46), 26364-26370. 179. Toh, R. J.; Sofer, Z.; Luxa, J.; Sedmidubsky, D.; Pumera, M., 3R phase of MoS2 and WS2 outperforms the corresponding 2H phase for hydrogen evolution. Chem. Commun. 2017, 53 (21), 3054-3057. 180. Benck, J. D.; Hellstern, T. R.; Kibsgaard, J.; Chakthranont, P.; Jaramillo, T. F., Catalyzing the hydrogen evolution reaction (HER) with molybdenum sulfide nanomaterials. ACS Catalysis 2014, 4 (11), 3957-3971. 181. Li, Z.; Meng, X.; Zhang, Z., Recent development on MoS2-based photocatalysis: A review. Journal of Photochemistry and Photobiology C: Photochemistry Reviews 2018, 35, 39-55. 182. Han, B.; Hu, Y. H., MoS2 as a co-catalyst for photocatalytic hydrogen production from water. Energy Science & Engineering 2016, 4 (5), 285-304. 183. Jaramillo, T. F.; Jørgensen, K. P.; Bonde, J.; Nielsen, J. H.; Horch, S.; Chorkendorff, I., Identification of Active Edge Sites for Electrochemical H2 Evolution from MoS2 Nanocatalysts. Sci 2007, 317 (5834), 100.

48 Chapter 2

Chapter 2.2

Critical Review: Hydrothermal Synthesis of 1T-MoS2 – an Important Route to a Promising Material

This chapter has been published in the Journal of Materials Chemistry A as Strachan, J.; Masters, A. F.;

Maschmeyer, T., Critical Review: Hydrothermal Synthesis of 1T-MoS2 – an Important Route to a Promising Material. Journal of Materials Chemistry A 2021, 9 (15), 9451-9461.

Abstract

The unique anisotropy, polytypism, and abundance in the literature it is unclear whether MoS2 or an of molybdenum disulfide make it a singularly intercalated (Mn+)(1/n)MoIIIS2 variant is produced. versatile material for a range of catalytic, Given the high potential of this material for many electrochemical, and tribological applications. By cutting-edge applications, it is important to clarify employing a hydrothermal synthesis, a high surface the literature to help facilitate rapid progress. In this area, optionally-supported, gas-repellent metallic context the review sets out to provide clear and

MoS2 (1T-MoS2) has been reported to be produced unambiguous synthetic and analytical strategies. from benign reagents at scale. This hydrothermally produced material has been shown to exhibit outstanding performance as a capacitor, and as an electrocatalyst (e.g. for the hydrogen evolution reaction). However, synthetic ambiguity and sample mischaracterisations are extremely common within the literature of reports of hydrothermally produced 1T-MoS2, and these occur across a range of analysis techniques such as Raman spectroscopy, X-ray diffraction, magnetic susceptibility, and X- ray photoelectron spectroscopy. These oversights have led to significant inconsistencies in the prevalent understanding of 1T-MoS2. In most cases

Introduction Sustainable solutions to the climate crisis require highly active, abundant materials for the myriad challenges of the evolving energy sector. 1T-MoS2 is one such material; its noble metal-like edge-site activity, layered structure, and metallic properties make it an ideal candidate for catalysing the hydrogen evolution reaction 1 (HER) and storing energy as either an electrode or capacitor. It is one of several MoS2 polytypes, of which the thermodynamically most stable form is 2H-MoS2.

Previous reviews have comprehensively covered the various applications of 1T-MoS2. Additionally, research by

Geng, et al. 2 showed that 1T-MoS2 exhibits excellent activity for the HER (-175 mV overpotential @ 10 mAcm- 2 with a Tafel slope of 41 mVdec-1) and subsequent reviews provide comparison to other outstanding materials.3- 4 We direct newer researchers to the work of the Jaramillo group, who have published several papers on best practice electrocatalytic methods.5 Acerce, et al. 1 showed that 1T-MoS2 can achieve capacitance values up to

49 Chapter 2

700 F cm−3, which has been contextualised within the field by a later review by Chia, et al. 4 For energy storage applications, including an exhaustive list of reports of hydrothermal (HT) synthesised electrodes and a report of a HT 1T-MoS2 electrode reaching a reversible capacity of 1150 mAhg-1, see Xu, et al. 6 and Chhowalla, et al. 7, 8. Finally, applications in photovoltaics, including dye-sensitised and Schottky barrier solar cells have been reviewed by Wang, et al. 9

The focus of this review is the hydrothermal (HT) synthesis of 1T-MoS2. Due to the low cost, scalability, operational stability, and benign precursors/solvent of hydrothermal 1T-MoS2 (relative to other preparations), this material is expected to facilitate commercialisation of hydrogen production by water electrolysis which is “the greatest barrier” in the decarbonisation of the energy sector.10 Common misconceptions and errors in the literature that retard timely progress in the understanding and development of applications for this highly promising material are discussed.

1T-MoS2 is conventionally synthesised by the intercalation of alkali metals into MoS2, although the inherent reactivity of the alkali metal precursors (e.g., n-BuLi) makes this synthesis difficult due to risk mitigation, and expense, and therefore unattractive for commercialisation.10 In 2015, Song and co-workers reported the production of 1T-MoS2 by a hydrothermal method,11 enabling cost-efficient 1T-MoS2 to be produced at scale from benign precursors.12 Unfortunately, this study and many published subsequently contain sample mischaracterisation and data misinterpretation, which have resulted in an inconsistent and unnavigable body of literature. The material was described as “ammonium ion intercalated” 1T-MoS2, or “colloidal metallic 1T‐

MoS2 layers highly stabilised by intercalated ammonium ions”, but no counterion was identified. We posit that for hydrothermally produced 1T-MoS2:

- All proposed synthesis mechanisms contradict XPS and XAS data;

- An estimated 95% of Raman studies mistakenly present data for MoO3 in place of 1T-MoS2; - XRD and TEM are routinely misused in an attempt to identify polytypes; - Thermal (DSC/TGA) data are inconsistent across the literature; - Sources of magnetic susceptibility are left unaccounted for; - Distorted polytypes (1T′ and 1T″) are not recognised.

- Most importantly, the nomenclature “1T-MoS2” does not accurately describe hydrothermally

produced, 1T-MoS2-like samples.

The current literature perpetuates these errors. This review, therefore, summarises the characteristics of hydrothermal 1T-MoS2 from reliable, self-consistent data, identifies contradictions within the current literature, and provides direction for future studies.

Nomenclature and Description The International Union of Crystallography implemented the ‘1T’ nomenclature in 1984 to differentiate the polytypes of layered materials, including MoS2.13 The ‘1’ signifies that the unit cell contains atoms from only one layer of stoichiometry MoS2 and the ‘T’ describes the trigonal unit cell. Each layer is comprised of stacked planes of sulfur, molybdenum, and sulfur atoms (thus are often called ‘sandwich layers’ of S–Mo–S). The Mo atoms in 1T-MoS2 are octahedrally coordinated to S atoms, whereas those in the naturally occurring, thermodynamically stable, 2H polytype have a trigonal prismatic co-ordination (Figure 20).14 This geometry determines the crystal field splitting of the molybdenum valence d-orbitals; the valence states of the trigonal prismatic polytypes are comprised of a singular, filled dz2 orbital as opposed to degenerate dxy, dxz, and dyz

50 Chapter 2 orbitals in the octahedral polytypes (Figure 20a & b).15 Therefore, the density of states of the octahedral polytypes includes a wide singular valence band, which contains the Fermi level, whereas the Fermi level of the trigonal prismatic polytypes lies within a bandgap (Figure 20e & f).15 As a result, the octahedral polytypes, including 1T-MoS2, are metallic and paramagnetic and the trigonal prismatic polytypes, including 2H-MoS2, are diamagnetic and semiconducting. Although the trigonal prismatic 2H-MoS2 polytype is approximately 60 kJ/mol more stable than 1T-MoS2 for MoIVS2 configuration, the 1T-MoS2 octahedral arrangement is favoured for d3 transition metals, such as reduced MoIIIS2¯.16 This is a consequence of the overall lower energy of the three electrons in the triply degenerate d-orbitals of the octahedral configuration (Figure 1). An explanation of the nomenclature used in this review is given in Table 3. The phase transition from 1T- to 2H-MoS2 has an activation energy of about 40 kJ/mol and involves shear of a sulfur plane relative to the molybdenum and sulfur planes of that layer.17 1T-MoS2 samples are stable at ambient conditions if intercalated ions are present to stabilise the lattice. We note that several superlattice distortions of the 1T-MoS2 phase may be formed as a result of layer interactions with intercalated ions (1T′ and 1T″), however due to errors in phase analysis by Raman and TEM (i.e. the very techniques by which the distorted phases are most easily identified, vide infra) within the literature, it is difficult to identify reliable data for these phases within published articles that reportedly synthesise and characterise 1T-MoS2 prepared by HT methods. Thus, this review will refer to all three phases simply as 1T. For further reading see Voiry, et al. 18, Zhao, et al. 19 and references therein.

Figure 20 – Geometry, crystal field splitting, band structure, and density of states for 2H-MoS2 (a, c, e) and 1T-MoS2 (b, d, f). Additional electron responsible for inducing the Oh distortion has been added to (b). Adapted with permission from Enyashin, et al. 20. Copyright 2020 American Chemical Society. Adapted with permission from Zhao, et al. 15. Copyright 2020 John Wiley and Sons.

We note that the data in Figure 20 are of a calculated 1T-MoS2 structure, which, as we will stress below, may not be representative of samples prepared by a hydrothermal method. Critically, no counterion has been observed in single crystal samples prepared by solid state methods, whereas hydrothermally prepared samples

51 Chapter 2 routinely include an intercalated ammonium species. Similarly, the sample used for the single crystal structure determination of 1T-MoS2 was not prepared hydrothermally, but rather by hydrolysis, then oxidation, of LiMoS2.21 The hydrolysis was accompanied by gas evolution and the oxidation used an acetonitrile solution of iodine. The

LiMoS2 was prepared by heating an evacuated, sealed mixture of Li2S, Mo and S (1:2:3) at 850 °C for 20 h. This synthesis is similar to that reported in 1992, for which the development is described below.22 These considerations have implications for the nomenclature used to describe 1T-MoS2 polytypes and samples. To most accurately describe the variations in samples and convey the ambiguity of the oxidation states in many cases, refined descriptors, presented in Table 3, are used here forth.

Table 3 – 1T-MoS2 nomenclature used within this review Nomenclature Notes Literature This review

IV 2H-MoS2 2H-MoS2 2H-Mo S2 1T-MoS2 1T-[MoS2]n 0 ≥ n ≥ −1. No counter ion has yet been observed for n = −1 samples. 1T-MoS2 (from HT 1T- 0 ≥ n ≥ −1. Ammonium ions (commonly) are present hydrothermal syntheses) [MoS2]n between expanded sheets as a counterion when n = −1 (i.e. forming 1T-(NH4)xMo4-xS2 structures).

n History of HT 1T-[MoS2]

1T-[MoS2]n was likely first synthesised as an alkali-metal (Na+, Li+) intercalate in 1959 by Rüdorff and Sick 23, not by a hydrothermal route, but by reacting (preformed) MoS2 with a solution of alkali metal in liquid ammonia. This was part of continued investigations on the intercalation of layered materials. They characterised their product as, e.g., Na0.8(NH3)0.2MoS2. Rüdorff later published powder X-ray diffraction data for his alkali- intercalated MoS2 materials,24 though the data quality was insufficient to recognise the change in polytype. We note that several other procedures may also be used to synthesise intercalated MoS2.25 In 1973, Somoano, et al.

26 noted the metallic behaviour of alkali metal-intercalated MoS2. Although octahedral, metallic transition metal dichalcogenides of the 1T-[MoS2]n polytype were known at the time,27 a structural distortion leading to octahedral Mo centres was not proposed until 1983.28 Improvements to the quality of subsequent crystallographic data enabled Mulhern 29 to resolve the crystal structure and label the Li-intercalated MoS2 as a

1T-MoS2 polytype. This material, LixMoS2, was prepared by reacting a solid mixture of Li2S, Mo and S at 800

– 1000 °C. Sandoval, et al. 30 later used group theory analysis to assign the Raman spectrum of 1T-[MoS2]n samples and Wypych and Schöllhorn 22 were able to synthesise pure 1T-[MoS2]n crystals by oxidation of

KMo(III)S2, indisputably confirming the octahedral Mo geometry. Research then became focused on the various distortions of the 1T-[MoS2]n polytype, for which a history is given by Chou, et al. 31.

Contemporaneously with the above crystal structure controversy, Chianelli, et al. 32 synthesised MoS2 by a hydrothermal method in 1979. However, it was not until 2015 that Song and co-workers discovered 1T-[MoS2]n may be produced (as an ammonium intercalate) by hydrothermal synthesis and reported the material as “1T-

MoS2”.12 Since that report, the number of reports of hydrothermally produced 1T-[MoS2]n has grown at an ever-increasing rate.

n Hydrothermal 1T-[MoS2] Synthesis Background

1T-[MoS2]n may be produced by post-synthetic reduction of MoS2 or in-situ. Post-synthetic treatment involves

(electro)chemical reduction of the MoS2 lattice, accompanied by intercalation with positive ions; Li+ being the

52 Chapter 2 most common as its small size results in rapid intercalation kinetics.33 In-situ reduction may be realised by a considered choice of reagents and conditions: the hydrothermal method is one such approach and facilitates the reduction of Mo precursors with sulfur compounds. HT syntheses produce highly active, defect rich, high surface area MoS2 in a range of Green, recoverable, non-flammable solvents (if water is not used as the solvent, the method is referred to as solvothermal). The products are also hydrophilic, gas-repellent, and may be easily synthesised as hierarchical structures on a range of supports in a one-pot reaction.2, 9 These characteristics make

1T-[MoS2]n produced using the hydrothermal method (HT 1T-[MoS2]n) an ideal catalyst and energy storage material.

Procedure The HT method is straightforward. However, care must be taken because of the high pressures involved. A typical procedure involves loading a homogeneous solution containing Mo- and S-containing precursors into a suitable (usually Teflon-lined) pressure vessel that is sealed, brought to temperature and typically held at temperature for 12–24 h. The vessel is cooled, then the precipitates are removed and purified by repeated centrifuging and washing, and the solid is finally dried. It is critical that the expected pressure is calculated beforehand. This calculation should include the pressure contributed by the production of gases from reactant decomposition. Consulting an isochoric database,34 using a pressure gauge, and applying a significant safety margin and pressure release for the pressure limit of the vessel are recommended. Typical pressures for a reactor half-filled with a 0.5 mmol/mL aqueous solution of ammonium molybdate are 10 – 20 bar at 200 °C.

Synthesis Mechanism

The reaction mechanism of the HT production of 1T-[MoS2]n is understood as the reduction of a MoVI complex using a S2¯ species as a sulfur source and reductant. Given that the octahedral Mo-coordination is only favoured when Mo is reduced to MoIII (vide supra), we assert that all published synthesis mechanisms insufficiently explain the reaction because, to our knowledge, no mechanism has been proposed that incorporates reduced (i.e. d3)

MoIIIS2¯ as a product. Without the presence of a MoIIIS2¯ species, the intercalation of a charged species such as NH4+ (absent a charge compensating anion(s)) would lead to charge accumulation in the MoS2 crystal.

Despite this, several authors propose that the intercalation of NH4+ produces the 1T-[MoS2]n phase, yet there is no evidence of causation for this correlation. Given it is the reduction of a Mo(VI) precursor to a d3 electronic configuration that stabilises the 1T phase, NH4+ ions are likely only necessary for charge balance.12, 19, 28 The following discussion presents X-ray photoelectron spectra (XPS) and X-ray absorption spectra (XAS) that clearly point to a MoIII oxidation state and X-ray diffraction (XRD) and transmission electron microscopy

(TEM) analyses that identify the presence of an intercalated species (vide infra). Thus, HT 1T-[MoS2]n might be better conceptualised as [(NH4)MoS2]0.35 As a working hypothesis and while fully acknowledging the need to probe these reaction cascades more deeply with, e.g. labelling studies, we propose [(NH4)MoS2]0 to be the product of a complex set of competing reactions that occur at significant pressure in the HT vessel. These reactions are likely to involve: the hydrolysis of thioamides, the decomposition of amides to generate ammonia, the reduction of Mo cations by S2- derivatives (H2S, S(NH3)2 etc), and the oxidation of ammonia by MoIVS2.

Using the typical reagents MoO3 and thiourea in an acidic solution, as an archetypical example, these reactions may be written as:

H2NCSNH2 + H2O → H2NCONH2 + H2S H2NCONH2 + H2O → CO2 + 2NH3 MoVIO3 + 3H2S → MoIVS2 + 3H2O + S 2MoIVS2 + 2NH3 + H2S → 2[(NH4)MoIIIS2] + S

53 Chapter 2

n Scheme 1 - Reaction processes in HT 1T-[MoS2] synthesis from a thioamide.

If instead a combined sulfur-molybdenum source is used, such as the ammonium tetrathiomolybdate, which was employed by the Kang group,36-38 we propose that the MoIII product is formed by a different mechanism:

2(NH4)2MoVIS4 → 2[(NH4)MoIIIS2] + 2NH3 + H2S + 3S

n Scheme 2 – Synthesis of HT 1T-[MoS2] from ammonium tetrathiomolybdate

We note that the proposed by-product in both schemes, elemental sulfur, has not been reported in the literature. This would be consistent with a combination of causes: the sulfur will likely be nanoparticulate, well dispersed, and amorphous, and therefore go unnoticed by experimenters unless specifically looked for. Given the presence of water and the nascent catalyst, the sulfur may undergo disproportionation (e.g. by the reverse Claus reaction).39-40 Additionally, thermal decomposition studies of (NH4)2MoS4 indicate that excess sulfur in MoS2 is quite stable, even under forcing conditions (< 800 °C, inert atmosphere).41-43

Synthesis Parameters

The HT synthesis of 1T-[MoS2]n can be performed using a wide set of experimental parameter values. An extensive list of reports has already been summarised by Shi, et al. 44. The molybdenum source is typically an ammonium or sodium molybdate,44 although MoCl5 and phosphomolybdic acid have also been used.45

Thioamides such as thiourea or thioacetamide are the most common sulfur sources.44 To ensure 1T-[MoS2]n is the dominant phase, an intercalation agent must also be present in sufficient excess; a role typically assigned to

NH4+.12 Some syntheses use a separate reducing agent, such as oxalic acid, while others that employ even stronger reducing agents such as hydrazine have produced 2H-MoS2.2 Syntheses typically occur at between

160–200 °C. Above ~200 °C, by a mechanism not yet fully understood, 2H-MoS2 is formed instead (we do n however note that conversion of isolated 1T-[MoS2] samples occurs in this temperature range, vide infra). Dwell times typically range from 12–24 h, beyond which the 2H-MoS2 phase begins to dominate over the metastable

1T-[MoS2]n phase. In addition, the product becomes more crystalline.6 The synthesis time may be decreased to <10 min by employing a focused microwave reactor, which induces hotspots in a controlled manner.46

It is possible to synthesise 1T-[MoS2]n in a range of solvents. Water is commonly used as it is readily available, non-flammable, and Green, however ethanol, supercritical CO2, dimethylformamide, propionic acid, and ethylene glycol have also been employed. It has been reported that in some instances the solvent plays a key role in the stabilisation of the 1T polytype. Water, for example, may act as both a proton donor and an oxygen source in the conversion of thiourea to urea and oxidised dimethylformamide species may be intercalated to 12, 47 facilitate layer separation. The rational design of hydrothermally produced MoS2 (HT MoS2) catalysts would be aided by a systematic study on the effects of solvent parameters such as pH, polarity, and zeta potential.

It has been proposed that the geometry of the Mo source significantly influences the polytype of the product of the reaction. Geng, et al. 2 report that the 1T-[MoS2]n phase could only be obtained in pH ranges (pH <4) where the octahedral geometry of the MoO3 precursor was said to be maintained. However, we note that octahedral Mo complexes have been observed at higher pH48 and additionally, appropriate control experiments were not performed; it may be that the low pH facilitates the acid-catalysed decomposition of thiourea. He, et al. 49 used L-cysteine as a tridentate sulfur precursor to stabilise the octahedral geometry and found that the L- cysteine complexed strongly to the Mo centre prior to HT treatment, such that an absorbance assigned to a

54 Chapter 2

Mo5+ d–d transition was observed in the UV-vis spectrum. It was also found that the use of L-cysteine allowed the 1T-[MoS2]n polytype to be synthesised at a record high temperature of 240 °C.

1T-[MoS2]n formation may be optimised by the addition of transition metal dopants, intercalants and supports during the HT synthesis, as illustrated in the following paragraph.

As an example of doping, Nethravathi, et al. 50 generated Co-doped HT 1T-[MoS2]n by addition of cobalt acetate into the precursor solution. Metal ion doping of MoS2 is reported to enable tuning of its electronic and magnetic properties, in the case of Nethravathi, et al. 50, producing magnetic MoS2 sheets that outperformed non-doped samples in the reduction of nitroarenes. HT 1T-[MoS2]n quantum dots may be selectively synthesised by use of a diphenyl-functionalised disulfide source.51 There are no reports that propose a mechanism for the morphological control of the quantum dot, therefore we propose that, as the disulfide bonds break to reduce the Mo6+ source, the phenyl moieties are retained and act as capping agents to inhibit further growth.

The use of intercalants is exemplified by the work of Kang and coworkers38 who have reported that it is possible to intercalate a range of molecules (methylammonium, alkylated phenylenediamines, protonated melamine, porphyrins, and Fe-phthalocyanine) between the MoS2 layers, resulting in considerable expansion of the interlayer gallery (e.g. using melamine, an interlayer gallery distance of ~ 21 Å is observed). This intercalation stabilises the 1T-[MoS2]¯ phase and improves the catalytic performance of the catalysts for the HER. The catalytic enhancement is proportional to the concentration of the intercalated species up to ~10 mol%. Calculations indicate that the improved stability is due to charge transfer from nitrogen moieties on the intercalant to sulfur vacancies in the MoS2 sheets.36 Ding, et al. 52 were able to control the polytype of their HT

MoS2 by applying a magnetic field during HT synthesis (which involved heating (NH4)6Mo7O24·4H2O ~0.03M as Mo7) and thiourea (~ 32 equivalents) to 210 °C for 18 h). The authors report that phase pure “1T-MoS2” may be obtained by applying a 9 T field. The field induces a magnetic-free-energy term, for which the value is more negative for 1T- than it is for 2H-MoS2, resulting in a lower ground-state free energy of the 1T polytype.

A full list of supports upon which HT 1T-[MoS2]n has been grown has been reviewed by Chia, et al. 4 This review covers supported transitional metal dichalcogenides, whereas MoS2/graphene composites are reviewed by Wang, et al. 9 and supported quantum dots are discussed by Li, et al. 51

Despite these detailed investigations, several key questions remain, such as: by what mechanism is the MoS2 reduced (to a d3 Mo centre)? Is 1T-[MoS2]n not observed above ~200 °C because NH4+ ions deintercalate, or because they react further? Is an isothiouronium-type species present? Does pH solely stabilise the octahedral

Mo centre or does it play another role (e.g. produce NH4+ from thioamides)? Does water play a physical as well as chemical role (e.g. intercalation behaviour)? What effect do pressure and concentration have on the products?

Strong reductants typically produce 2H-, rather than 1T-[MoS2]n . Why?

Characterisation The following section is intended as both a reference to peer-reviewed high-quality data, and as a guide to the characterisation of HT 1T-[MoS2]n in terms of techniques used and their opportunities and limitations. The section is arranged by the mode of analysis; spectroscopy is covered first, followed by diffraction, microscopy, then physical and electrochemical techniques.

55 Chapter 2

Figure 21 – 1T/2H polytype comparison by (a) STEM-ADF, (b) UV-vis spectroscopy, (c) Raman spectroscopy, (d) grey value traces (across dotted yellow lines in (a)) of STEM-ADF, (e) X-ray photoelectron spectroscopy of Mo region, (f) Photoluminescence spectroscopy. Colours in legend n apply to all figures. a MoO3 was produced by oxidation of a HT 1T-[MoS2] sample by a 5 mW Raman laser. (a, d) adapted with permission from Zhao, et al. 19, (c) adapted from Fang, et al. 53, copyright 2020 John Wiley and Sons; (b, e) adapted from Geng, et al. 2 under Creative Commons; (f) Adapted from Friedman, et al. 54, with permission from The Royal Society of Chemistry.

Spectroscopy

The 1T polytype is most easily distinguished from 2H-MoS2 by spectroscopic techniques. Because the 2H phase has a band gap of 1.2 eV, it absorbs visible light at ~615 and 660 nm, giving solutions of 2H-MoS2 a green colour (Figure 21b).18 Smaller sheets of 2H-MoS2 may also absorb at ~440 nm due to a quantum confinement effect.44 Conversely, the metallic 1T-[MoS2]n samples will merely scatter visible light, giving them a grey colour.44 Thin films of 2H-MoS2 exhibit photoluminescence as a result of direct band gap excitons (Figure 21f).55 Two absorptions between 600–700 nm, labelled A and B, arise from spin-orbit coupled transitions at the K point of the Brillouin zone. It should be noted that the decomposition products of N-methylpyrrolidone, a solvent commonly used in the dispersal of MoS2, exhibit photoluminescence and Tyndall scattering – thus solution colour should not be relied upon as the sole indicator for distinguishing MoS2 polytype.56 Infrared spectroscopy as applied to HT 1T-[MoS2]n can only provide a limited amount of information such as confirmation of the presence of the intercalant or derivatisation of a known compound, such as for the alkylammonium-intercalated MoS2 produced by the Kang group.38

HT 1T-[MoS2]n is frequently mischaracterised by Raman spectroscopy. We estimate that 90% of the current reports in the literature present, in place of 1T-[MoS2]n, the spectrum of MoO3 (for which unambiguous data is readily available57). We strongly suspect that this mischaracterisation is often the result of sample oxidation under the Raman laser as 1T-[MoS2]n and MoO3 unfortunately exhibit somewhat similar Raman spectra (Figure 21c). It has been shown that the 1T phase will readily convert to the 2H phase under a laser power stronger than ~100 µW.17, 21 In our experience, if a sufficiently high laser power is used (~5 mW), the phase transition will be followed by the oxidation of the 2H-MoS2 to MoO3 due to local heating. Although bulk MoS2 begins to oxidise at 600 °C, HT 1T-[MoS2]n samples are typically composed of unstacked monolayers of MoS2, thus

56 Chapter 2 only a small energy input is required for these processes to occur. While single crystals of non-hydrothermally produced 1T-[MoS2]n may provide spectra with clear, well resolved Raman absorptions,21 the disordered nature of HT 1T-[MoS2]n samples results in broader, lower signal-to-noise Raman signals.2 To easily differentiate HT

1T-[MoS2]n from MoO3, we strongly recommend that data are collected (and presented) up to 900 wavenumbers, so as to test for the presence of the intense and diagnostic MoO3 absorptions at 603 and 876 cm-1 (Table 4).

Table 4. Raman modes of MoS2 and MoO3 1T-[MoS2]n a 2H-MoS2b MoO3c 116 B1g, m 128 B2g, m 147 151 J1 156 B1g, s 226 J2 228 B3u, w 286 E1g 286 B2g, m 333 J3 338 B1g, m 380 E2g 412 A1g 407 A1g 603 B1u, vs 876 B3u, vs Values in wavenumbers (cm-1). a includes absorptions from superlattice octahedral- distortions. Data and assignments from references 17, 21, 30. b & c data and assignments from 17 & 57 and references therein. Underlined numbers indicate diagnostic signals.

The polytypes also exhibit distinct XPS spectra due to the different electron densities of the Mo and S in 1T- (d3) and 2H- (d2) MoS2 (Figure 21e). XPS spectra of single crystal- and HT- 1T-[MoS2]n have been reported by Yu, et al. 21 and Geng, et al. 2, respectively. In both cases, the Mo 3d and S 2p absorptions were shifted to lower binding energies by ~0.9 eV relative to the 2H-MoS2 polytype, consistent with a MoIII oxidation state in the former (see Table 5).58-59 Signals attributable to intercalated compounds, such as N 1s peaks from ammonium derivatives, may also be observed in the XPS data.38 The high vacuum conditions of the XPS measurements slowly induce a 1T → 2H phase change,2 thus care should be taken when estimating the relative polytype ratio from XPS measurements.

Table 5. XPS values for 1T- and 2H-MoS2 Element and Spectral Line 1T 2H Mo 3d3/2 231.5 232.2 Mo 3d5/2 228.4 229.1 S 2p1/2 162.7 163.8 S 2p3/2 161.6 162.6 Values are for binding energy (in eV), charge referenced to adventitious carbon. Data from Yu, et al. 21 and Geng, et al. 2

The Kang group36, 38 has characterised several ammonium-derivative intercalated HT 1T-[MoS2]n samples by X-ray absorption spectroscopy (XAS) including both the X-ray absorption near edge (XANES) and extended X-ray absorption fine structure (EXAFS) regions. They note that the Mo XANES K-edge energy of HT 1T-

[MoS2]n samples is consistently lower than that of 2H-MoS2, which indicates the intercalated 1T samples are more metallic in character and contain lower-valent Mo than their 2H polytype – consistent with the XPS findings and therefore consistent with our interpretation of ‘1T-MoS2’ as being 1T-[Mo4+nS2]n.60 The Fourier- transformed EXAFS curves of HT 1T-[MoS2]n are characterised by strong and weak peaks at 2.40 and 2.76 Å associated with Mo–S and Mo–Mo pathways, respectively. These data are consistent with the 1T Mo–Mo being

57 Chapter 2 shortened by 0.4 Å relative to the 2H polytype and the Mo–S pathway being both shortened in octahedral complexes and split in distorted, 1T′, samples.

Diffraction

The MoS2 polytypes exhibit unique X-ray diffraction patterns16 – for 1T-[MoS2]n data, see Fang, et al. 53 and references therein. However, for the highly disordered, few-layer, nano-sized samples that the HT method produces, XRD analysis cannot differentiate between polytypes because of the broadening associated with small particle sizes and crystalline defects. The XRD patterns of these samples typically consist of broad reflections at ~9, ~33, and 57° 2θ (λ = 1.54 Å) and occasionally very broad basal plane reflections. The broadening of the basal plane reflections of HT MoS2 samples, i.e. the (2H) 10퓁 reflections (where 퓁 = 1, 3, 5), indicates both few-layer stacking in the c direction and rotational disorder of the planes.61 The shift of the 002 reflection towards a larger d-spacing is associated with interlayer ion intercalation, which effectively enlarges the c-axis of the unit cell. The 2θ value of this reflection depends on the identity of the intercalated species; for the commonly synthesised (NH4+)-intercalated MoS2, this reflection is positioned at 9.7° 2θ (0.91 nm).1

Microscopy

Transmission electron microscopy (TEM) can be used to discriminate the various MoS2 polytypes. It is most useful when analysing mono/bi/few-layered samples, for which the phases appear most dissimilar. Low resolution TEM does not generally contain enough information to differentiate between polytypes, but spherical aberration corrected scanning TEM annular dark field (STEM-ADF) may readily do so as the intensity along columns of atoms varies with atomic number (Figure 21a).19 Best practice phase identification involves iterative template matching and comparison with simulated samples (Figure 21a).19 Grey value traces across the columns of atoms may then be used to reveal the unique patterns of each polytype (Figure 21d).19

Thermogravimetric Analysis (TGA) and Differential Scanning Calorimetry (DSC)

The irreversible 1T → 2H phase change occurs with constant mass at ~96 °C in pure 1T-[MoS2]n samples.21-22

HT 1T-[MoS2]n samples invariably contain an intercalated species, which further stabilises against phase change.

Therefore, HT 1T-[MoS2]n samples undergo a phase change to 2H at a higher temperature (~200–250 °C) than non-intercalated analogues, with a corresponding weight decrease due to the loss of the intercalated species. The temperature and heat flow of the phase change are dependent on the concentration and identity of the intercalated species, and can be measured by DSC. The phase conversion in HT 1T-[MoS2]n samples typically coincides with a 10–40% weight loss often assigned to the deintercalation of NH4+ ions.38 There are no reports that propose a mechanism for this topotactic reaction; we propose that the [MoIIIS2]¯ is likely oxidised by the ammonium ions to produce MoIVS2, liberating H2 and NH3 gases:

2[(NH4)MoS2]0(s) → 2NH3(g) + H2(g) + 2MoS2(s)

We also note that the DSC data for HT 1T-[MoS2]n reported in the literature lack consistency with respect to the endo/exothermic nature of the phase change, which is consistently endothermic in pure/single crystal samples and DFT calculations.16, 21-22, 62 Additionally, multiple heat flow and weight loss features are often unassigned. There have been no reports that attempt to characterise the products of the phase change by TGA- mass spectrometry or other methods. An understanding of the volatile products would facilitate the rational design of stabilised HT 1T-[MoS2]n derivatives.

58 Chapter 2

Electrochemistry

The metallic band structure of 1T-[MoS2]n imparts high conductivity along the layers. The 1T phase increases conductivity by ~7 orders of magnitude relative to the semiconducting 2H polytype.1 Nyquist plots, obtained by electrochemical impedance spectroscopy, of electrodes containing MoS2 typically reveal charge transfer resistance values on the order of 1 and 50 Ω for 1T- and 2H-MoS2, respectively, depending on the electrolyte and electrode characteristics such as conducting and binding agents.2 Further discussion is provided by Shi, et al. 44. Because of its high conductivity, the 1T polytype is substantially more prone to oxidation than is the 2H analogue. Wang, et al. 63 show that oxidation of 1T-[MoS2]n can lead to degradation to the 2H polytype. They also propose that a two-site corrosion reaction, facilitated by 1T-[MoS2]n as a conductive medium, proceeds overall as:

2MoS2 + 6H2O + 9O2 → 2H2MoO4 + 4H2SO4 E° = 0.8 V.

n HT 1T-[MoS2] samples should therefore be kept under an inert atmosphere to avoid oxidation and subsequent phase change.

Magnetism

1T-[MoS2]n is paramagnetic due to the unpaired electrons in the triply degenerate valence orbitals (Figure 20), whereas the spin pairing of the electrons in the dz2 orbital of the d2 2H phase results in diamagnetism.64 We note that paramagnetism is expected for octahedral 1T-[MoS2]n, regardless of whether it is understood as Mo(III) or

Mo(IV). However, 2H-MoS2 has been observed to exhibit weak magnetism as a consequence of unpaired electrons along the unsaturated edges of MoS2 films. The effect also may arise from the presence of basal plane sulfur vacancies, adsorbed oxygen, and Mo5+ species (see Yan, et al. 64, Martinez, et al. 65 and references therein). Researchers are therefore encouraged to perform appropriate control experiments and to ensure that the oxidation state and polytype are firmly established if reporting magnetic susceptibility data.

Outlook

Although the HT method and MoS2 catalysts are well-established, the recent synthesis of HT 1T-[MoS2]n opens new pathways for further development. The full scope of the synthetic parameters is yet to be explored (vide supra) and only very few intercalation compounds have been reported. Although XPS and XAS data clearly identify a d3 MoIIIS2 species, the literature has not yet incorporated this information into mechanistic considerations. Until the formation mechanism is understood (for which a systematic study of the relevant parameters is likely required) the development and potential of HT 1T-[MoS2]n will be hindered. Morphological control is, at present, limited to the few studies on quantum dots. If catalytic activity for the HER and electrochemical performance as an electrode are to be optimised by understanding, rather than serendipity, rational morphological control must be achieved. Further elucidation of the support–intercalant interaction may prove useful in the development of more stable HT 1T-[MoS2]n. This rationale extends to the distorted, 1T′ and 1T″, phases as well. The support–intercalant DFT calculations by Kang and coworkers38 provide a promising start, but the rational selection of intercalation compounds to control polytype and distortion has not yet been realised and ambiguity remains about both the oxidation state of the MoS2 and the charge of intercalated species.

The catalytic activity of 1T-[MoS2]n is proportional to the interlayer distance, however the limit to this relationship is unknown.38 1T-[MoS2]n samples appear to outperform the 2H polytype in HER reactions;2 if the

1T polytype is reconceptualised as [MoIIIS2]− (as proposed here), then the mechanism of catalytic activity improvement should be determined. It may be that the reduced molybdenum provides Lewis basic sites to

59 Chapter 2 better co-ordinate H+ ions or the improved conductivity facilitates electron transfer. The performance of HT

1T-[MoS2]n may also be enhanced by the incorporation of dopants, for which only Co has been investigated.50 The similarities of the distorted (1T′ and 1T″) phases, together with the prevalent mischaracterisation of HT

MoS2 means that the true polytype, and therefore the true performance, of many HT 1T-[MoS2]n samples is still unknown.

There exists ample room for further development of HT 1T-[MoS2]n and derivatives. However, to achieve the latent potential of the material, common errors in the literature must be addressed. Only through careful, thorough analysis and critique is progress assured.

Conclusions

Metallic 1T-[MoS2]n is a promising material for a range of catalytic, electrochemical, and tribological applications. When synthesised using a hydrothermal method, the material exhibits improved (electro)catalytic performance due to the increased proportion of active sites and gas-repellent behaviour to the extent that it outperforms most non-noble metal HER catalysts. At present, extensive sample mischaracterisation exists within the HT 1T-[MoS2]n literature. Thus, in an effort to rectify these errors, this review summarised the available, reliable data and establishes clear guidelines for researchers regarding synthetic and analytical procedures. We have shown that all synthesis mechanisms proposed contradict XPS and XAS data, since the

XPS and XAS data of 1T-[MoS2]n samples are consistent with the presence of Mo in an oxidation state <4; most Raman studies mistake the spectrum of MoO3 for 1T-[MoS2]n because samples undergo oxidation under sufficient laser irradiation; the hydrothermal method produces particles too small and disordered to be analysed by PXRD; the assignment of heat flow and mass loss events in calorimetric and gravimetric analyses of HT 1T-

[MoS2]n is, at present, inconsistent and unsubstantiated and complicated by the desorption of intercalants; terminal sulfur moieties contribute to the magnetic susceptibility of HT 1T-[MoS2]n samples and the distorted polytypes (1T′ and 1T″) of 1T-[MoS2]n are not discussed with in the HT 1T-[MoS2]n literature – although they are very likely present in hydrothermally produced samples. The hydrothermal method opens new pathways for tuning the characteristics of 1T-[MoS2]n, and experiments within the large possible parameter space are only in their infancy. Given the potential of this synthesis procedure and material for many cutting-edge materials applications, the contradictions within the literature identified and rectified here may yet assist the realisation of commercial electrocatalysts based on the 1T-[MoS2]n polytype.

60 Chapter 2

References 1. Acerce, M.; Voiry, D.; Chhowalla, M., Metallic 1T phase MoS2 nanosheets as supercapacitor electrode materials. Nat. Nanotechnol. 2015, 10 (4), 313. 2. Geng, X.; Sun, W.; Wu, W.; Chen, B.; Al-Hilo, A.; Benamara, M.; Zhu, H.; Watanabe, F.; Cui, J.; Chen, T.-p., Pure and stable metallic phase molybdenum disulfide nanosheets for hydrogen evolution reaction. Nature communications 2016, 7, 10672. 3. Chacko, L.; Rastogi, P. K.; Aneesh, P. M., Phase Engineering from 2H to 1T-MoS2 for Efficient Ammonia PL Sensor and Electrocatalyst for Hydrogen Evolution Reaction. J. Electrochem. Soc. 2019, 166 (8), H263-H271. 4. Chia, X. Y.; Eng, A. Y. S.; Ambrosi, A.; Tan, S. M.; Pumera, M., Electrochemistry of Nanostructured Layered Transition-Metal Dichalcogenides. Chem. Rev. 2015, 115 (21), 11941-11966. 5. McCrory, C. C.; Jung, S.; Ferrer, I. M.; Chatman, S. M.; Peters, J. C.; Jaramillo, T. F., Benchmarking hydrogen evolving reaction and oxygen evolving reaction electrocatalysts for solar water splitting devices. J. Am. Chem. Soc. 2015, 137 (13), 4347-4357. 6. Xu, J.; Zhang, J. J.; Zhang, W. J.; Lee, C. S., Interlayer Nanoarchitectonics of Two-Dimensional Transition-Metal Dichalcogenides Nanosheets for Energy Storage and Conversion Applications. Advanced Energy Materials 2017, 7 (23). 7. Chhowalla, M.; Shin, H. S.; Eda, G.; Li, L. J.; Loh, K. P.; Zhang, H., The chemistry of two-dimensional layered transition metal dichalcogenide nanosheets. Nature Chem. 2013, 5 (4), 263-275. 8. Hu, X. L.; Zhang, W.; Liu, X. X.; Mei, Y. N.; Huang, Y., Nanostructured Mo-based electrode materials for electrochemical energy storage. Chem. Soc. Rev. 2015, 44 (8), 2376-2404. 9. Wang, Z. H.; Zhao, C. Q.; Gui, R. J.; Jin, H.; Xia, J. F.; Zhang, F. F.; Xia, Y. Z., Synthetic methods and potential applications of transition metal dichalcogenide/graphene nanocomposites. Coord. Chem. Rev. 2016, 326, 86-110. 10. Liu, Z. P.; Zhao, L.; Liu, Y. H.; Gao, Z. C.; Yuan, S. S.; Li, X. T.; Li, N.; Miao, S. D., Vertical nanosheet array of 1T phase MoS2 for efficient and stable hydrogen evolution. Applied Catalysis B-Environmental 2019, 246, 296-302. 11. This method involved heating a ~0.03 M (per Mo7) aqueous solution of (NH4)6Mo7O24·4H2O with 15 equivalents (per Mo7) of thiourea under autogeneous pressure at 200 °C for 24 h. 12. Liu, Q.; Li, X.; He, Q.; Khalil, A.; Liu, D.; Xiang, T.; Wu, X.; Song, L., Gram‐Scale Aqueous Synthesis of Stable Few‐Layered 1T‐MoS2: Applications for Visible‐Light‐Driven Photocatalytic Hydrogen Evolution. Small 2015, 11 (41), 5556-5564. 13. Guiner, A.; Bokij, G.; Boll-Dornberger, K.; Cowley, J.; Durovic, S.; Jagogzinski, H.; Krishna, P.; de Wolff, P.; Zvyagin, B.; Cox, D., Nomenclature of polytype structures. Acta Crystallogr. A 1984, 40, 399-404. 14. For comparison, the complexes [Mo(dithiolene)3]n¯, which have an MoS6 core, typically have a trigonal prismatic molybdenum centre, for complexes with n = 0 and distortions toward octahedrally co-ordinated molybdenum for n > 0 . . 15. Zhao, W.; Pan, J.; Fang, Y.; Che, X.; Wang, D.; Bu, K.; Huang, F., Metastable MoS2: crystal structure, electronic band structure, synthetic approach and intriguing physical properties. Chemistry–A European Journal 2018, 24 (60), 15942-15954. 16. Caputo, R., Polytypism of MoS2. Jacobs Journal of Inorganic Chemistry 2016, 1 (1), 004. 17. Guo, Y.; Sun, D.; Ouyang, B.; Raja, A.; Song, J.; Heinz, T. F.; Brus, L. E., Probing the Dynamics of the Metallic-to- Semiconducting Structural Phase Transformation in MoS2 Crystals. Nano Lett. 2015, 15 (8), 5081-5088. 18. Voiry, D.; Mohite, A.; Chhowalla, M., Phase engineering of transition metal dichalcogenides. Chem. Soc. Rev. 2015, 44 (9), 2702-2712. 19. Zhao, X. X.; Ning, S. C.; Fu, W.; Pennycook, S. J.; Loh, K. P., Differentiating Polymorphs in Molybdenum Disulfide via Electron Microscopy. Adv. Mater. 2018, 30 (47). 20. Enyashin, A. N.; Yadgarov, L.; Houben, L.; Popov, I.; Weidenbach, M.; Tenne, R.; Bar-Sadan, M.; Seifert, G., New Route for Stabilization of 1T-WS2 and MoS2 Phases. The Journal of Physical Chemistry C 2011, 115 (50), 24586-24591. 21. Yu, Y.; Nam, G.-H.; He, Q.; Wu, X.-J.; Zhang, K.; Yang, Z.; Chen, J.; Ma, Q.; Zhao, M.; Liu, Z., High phase-purity 1T′-MoS 2- and 1T′-MoSe 2-layered crystals. Nature Chem. 2018, 10 (6), 638. 22. Wypych, F.; Schöllhorn, R., 1T-MoS2, a new metallic modification of molybdenum disulfide. J. Chem. Soc., Chem. Commun. 1992, (19), 1386-1388. 23. Rüdorff, W.; Sick, H., Einlagerungsverbindungen von Alkali‐und Erdalkalimetallen in Molybdän‐und Wolframdisulfid. Angew. Chem. 1959, 71 (3), 127-127. 24. Rudorff, W., Inclusion of base metals in graphite and in metallic chalcogenides of the type MeX 2. Chimia 1965, 19 (9), 489- 499. 25. The synthesis of KMoS2 was reported in 1965. “K2MoS6” was heated at 550 °C for 24 h under H2, annealed at 650 °C, cooled under H2 and washed with water to give a black solid with a metallic lustre. An alternate preparation involves heating K2MoO4 with dry H2S at 320 °C for 12 h, then at 470 °C for 12 h, to give a product, which was reduced under 20/80 vol% H2/N2 for 72 h. This KMoS2 can be partially oxidised by washing copiously with water (Wypych, Schöllhorn, 1T-MoS2,. Chemical Communications 1992, (19), 1386-1388). The washed product has a Mo:S molar ratio of 1:2; no such data are reported for the KMoS2. The synthesis of AMoS2 (A = Na+, K+) has also been reported by electrolytic reduction in a KCl/NaCl melt at 700 °C (Gao, et al. PCCP 2014, 16 (36), 19514-19521). Heating a stoichiometric mixture of MoS2, A2S (A = Li, Na, K) and Mo at 800 °C for 10h gives crystalline AMoS2 (Guo, et. al., J of Mat Chem C 2017, 5 (24), 5977-5983). Interestingly, the authors report that LiMoS2 prepared by this route is more stable than LixMoS2 prepared by reaction of MoS2 with BuLi. No elemental analyses were reported. In summary, several authors report the synthesis of (incompletely characterised) AMoS2, by other than hydrothermal means. . 26. Somoano, R.; Hadek, V.; Rembaum, A., Alkali metal intercalates of molybdenum disulfide. The Journal of Chemical Physics 1973, 58 (2), 697-701. 27. Mattheiss, L. F., Band structures of transition-metal-dichalcogenide layer compounds. PhRvB 1973, 8 (8), 3719-3740.

61 Chapter 2

28. Py, M.; Haering, R., Structural destabilization induced by lithium intercalation in MoS2 and related compounds. Can. J. Phys. 1983, 61 (1), 76-84. 29. Mulhern, P. J., Lithium intercalation in crystalline Li x MoS2. Can. J. Phys. 1989, 67 (11), 1049-1052. 30. Sandoval, S. J.; Yang, D.; Frindt, R.; Irwin, J., Raman study and lattice dynamics of single molecular layers of MoS 2. PhRvB 1991, 44 (8), 3955. 31. Chou, S. S.; Sai, N.; Lu, P.; Coker, E. N.; Liu, S.; Artyushkova, K.; Luk, T. S.; Kaehr, B.; Brinker, C. J., Understanding catalysis in a multiphasic two-dimensional transition metal dichalcogenide. Nature communications 2015, 6 (1), 1-8. 32. Chianelli, R.; Prestridge, E.; Pecoraro, T.; Deneufville, J., Molybdenum disulfide in the poorly crystalline 'rag' structure. Sci 1979, 203, 1105-1107. 33. He, Z. L.; Que, W. X., Molybdenum disulfide nanomaterials: Structures, properties, synthesis and recent progress on hydrogen evolution reaction. Applied Materials Today 2016, 3, 23-56. 34. Eric W. Lemmon; Mark O. McLinden; Daniel G. Friend, Thermophysical Properties of Fluid Systems. National Institute of Standards and Technology, Gaithersburg MD, 20899: Vol. NIST Standard Reference Database Number 69. 35. Note: this is an ideal formula, and samples may contain substoichiometric [NH4] or Mo. 36. Kwak, I. H.; Kwon, I. S.; Debela, T. T.; Seo, J.; Ahn, J. P.; Yoo, S. J.; Kim, J. G.; Park, J.; Kang, H. S., Two-dimensional MoS2- melamine hybrid nanostructures for enhanced catalytic hydrogen evolution reaction. Journal of Materials Chemistry A 2019, 7 (39), 22571-22578. 37. Kwon, I. S.; Kwak, I. H.; Abbas, H. G.; Jung, G.; Lee, Y.; Park, J.; Yoo, S. J.; Kim, J. G.; Kang, H. S., Intercalation of aromatic amine for the 2H-1T phase transition of MoS2 by experiments and calculations. Nanoscale 2018, 10 (24), 11349-11356. 38. Kwak, I. H.; Kwon, I. S.; Abbas, H. G.; Jung, G.; Lee, Y.; Park, J.; Kang, H. S., Stable methylammonium-intercalated 1T '-MoS2 for efficient electrocatalytic hydrogen evolution. Journal of Materials Chemistry A 2018, 6 (14), 5613-5617. 39. Tsuchiya, N.; Suto, Y.; Kabuta, T.; Morikawa, S.; Yokoyama, S., Sustainable hydrogen production system with sulfur–water– organic materials by hydrothermal reaction. J. Mater. Sci. 2008, 43 (7), 2115-2122. 40. Laperdrix, E.; Sahibed-dine, A.; Costentin, G.; Bensitel, M.; Lavalley, J.-C., Evidence of the reverse Claus reaction on metal oxides: Influence of their acid–base properties. Applied Catalysis B: Environmental 2000, 27 (2), 137-142. 41. Wang, H. W.; Skeldon, P.; Thompson, G. E.; Wood, G. C., Synthesis and characterization of molybdenum disulphide formed from ammonium tetrathiomolybdate. J. Mater. Sci. 1997, 32 (2), 497-502. 42. Walton, R. I.; Dent, A. J.; Hibble, S. J., In situ investigation of the thermal decomposition of ammonium tetrathiomolybdate using combined time-resolved X-ray absorption spectroscopy and X-ray diffraction. Chem. Mater. 1998, 10 (11), 3737-3745. 43. Prasad, T.; Diemann, E.; Müller, A., Thermal decomposition of (NH4) 2MoO2S2,(NH4) 2MoS4,(NH4) 2WO2S2 and (NH4) 2WS4. J. Inorg. Nucl. Chem. 1973, 35 (6), 1895-1904. 44. Shi, S. L.; Sun, Z. X.; Hu, Y. H., Synthesis, stabilization and applications of 2-dimensional 1T metallic MoS2. Journal of Materials Chemistry A 2018, 6 (47), 23932-23977. 45. Zheng, X.; Wang, S.; Xiong, C. X.; Hu, G. H., In situ growth of 1T-MoS2 on liquid-exfoliated graphene: A unique graphene- like heterostructure for superior lithium storage. Carbon 2018, 133, 162-169. 46. Lee, Y. B.; Kim, S. K.; Ji, S.; Song, W.; Chung, H. S.; Choi, M. K.; Lee, M.; Myung, S.; Lim, J.; An, K. S.; Lee, S. S., Facile microwave assisted synthesis of vastly edge exposed 1T/2H-MoS2 with enhanced activity for hydrogen evolution catalysis. Journal of Materials Chemistry A 2019, 7 (8), 3563-3569. 47. Wu, M. H.; Zhan, J.; Wu, K.; Li, Z.; Wang, L.; Geng, B. J.; Wang, L. J.; Pan, D. Y., Metallic 1T MoS2 nanosheet arrays vertically grown on activated carbon fiber cloth for enhanced Li-ion storage performance. Journal of Materials Chemistry A 2017, 5 (27), 14061-14069. 48. Gheller, S.; Sidney, M.; Masters, A.; Brownlee, R.; O'Conner, M.; Wedd, A., Applications of molybdenum-95 NMR spectroscopy. X. Polyoxomolybdates. Aust. J. Chem. 1984, 37 (9), 1825-1832. 49. He, D. P.; Ooka, H.; Li, Y. M.; Jin, F. M.; Nakamura, R., Phase-selective Hydrothermal Synthesis of Metallic MoS2 at High Temperature. Chem. Lett. 2019, 48 (8), 828-831. 50. Nethravathi, C.; Prabhu, J.; Lakshmipriya, S.; Rajamathi, M., Magnetic Co-Doped MoS2 Nanosheets for Efficient Catalysis of Nitroarene Reduction. Acs Omega 2017, 2 (9), 5891-5897. 51. Li, X. T.; Lv, X. D.; Li, N.; Wu, J. J.; Zheng, Y. Z.; Tao, X., One-step hydrothermal synthesis of high-percentage 1T-phase MoS2 quantum dots for remarkably enhanced visible-light-driven photocatalytic H-2 evolution. Applied Catalysis B- Environmental 2019, 243, 76-85. 52. Ding, W.; Hu, L.; Dai, J. M.; Tang, X. W.; Wei, R. H.; Sheng, Z. G.; Liang, C. H.; Shao, D. F.; Song, W. H.; Liu, Q. N.; Chen, M. Z.; Zhu, X. G.; Chou, S. L.; Zhu, X. B.; Chen, Q. W.; Sun, Y. P.; Dou, S. X., Highly Ambient-Stable 1T-MoS2 and 1T-WS2 by Hydrothermal Synthesis under High Magnetic Fields. ACS Nano 2019, 13 (2), 1694-1702. 53. Fang, Y.; Pan, J.; He, J.; Luo, R.; Wang, D.; Che, X.; Bu, K.; Zhao, W.; Liu, P.; Mu, G., Structure Re‐determination and Superconductivity Observation of Bulk 1T MoS2. Angew. Chem. 2018, 130 (5), 1246-1249. 54. Friedman, A. L.; Perkins, F. K.; Hanbicki, A. T.; Culbertson, J. C.; Campbell, P. M., Dynamics of chemical vapor sensing with MoS 2 using 1T/2H phase contacts/channel. Nanoscale 2016, 8 (22), 11445-11453. 55. Eda, G.; Yamaguchi, H.; Voiry, D.; Fujita, T.; Chen, M.; Chhowalla, M., Photoluminescence from chemically exfoliated MoS2. Nano Lett. 2011, 11 (12), 5111-5116. 56. Ogilvie, S. P.; Large, M. J.; Fratta, G.; Meloni, M.; Canton-Vitoria, R.; Tagmatarchis, N.; Massuyeau, F.; Ewels, C. P.; King, A. A.; Dalton, A. B., Considerations for spectroscopy of liquid-exfoliated 2D materials: emerging photoluminescence of N-methyl- 2-pyrrolidone. Scientific reports 2017, 7 (1), 16706.

62 Chapter 2

57. Seguin, L.; Figlarz, M.; Cavagnat, R.; Lassègues, J. C., Infrared and Raman spectra of MoO3 molybdenum trioxides and MoO3 · xH2O molybdenum trioxide hydrates. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 1995, 51 (8), 1323-1344. 58. Choi, J. G.; Thompson, L. T., XPS study of as-prepared and reduced molybdenum oxides. Appl. Surf. Sci. 1996, 93 (2), 143- 149. 59. Baker, M. A.; Gilmore, R.; Lenardi, C.; Gissler, W., XPS investigation of preferential sputtering of S from MoS2 and determination of MoSx stoichiometry from Mo and S peak positions. Appl. Surf. Sci. 1999, 150 (1), 255-262. 60. George, G.; Cleland Jr, W.; Enemark, J.; Smith, B.; Kipke, C.; Roberts, S.; Cramer, S. P., L-Edge spectroscopy of molybdenum compounds and enzymes. J. Am. Chem. Soc. 1990, 112 (7), 2541-2548. 61. Moser, J.; Lévy, F., Random stacking in MoS2−x sputtered thin films. Thin Solid Films 1994, 240 (1-2), 56-59. 62. We note that the DFT calculations in Caputo, R. (2016) used Mo(IV) 63. Wang, Z. Y.; Zhang, Y. J.; Liu, M. C.; Peterson, A.; Hurt, R. H., Oxidation suppression during hydrothermal phase reversion allows synthesis of monolayer semiconducting MoS2 in stable aqueous suspension. Nanoscale 2017, 9 (17), 5398-5403. 64. Yan, S.; Qiao, W.; He, X.; Guo, X.; Xi, L.; Zhong, W.; Du, Y., Enhancement of magnetism by structural phase transition in MoS2. Appl. Phys. Lett. 2015, 106 (1), 012408. 65. Martinez, L. M.; Karthik, C.; Kongara, M.; Singamaneni, S. R., Paramagnetic defects in hydrothermally grown few-layered MoS 2 nanocrystals. J. Mater. Res. 2018, 33 (11), 1565-1572.

63 Chapter 2

Chapter 2.3

Influence of Crystal Disorder in MoS2 Cathodes for Secondary Hybrid Mg-Li Batteries

This chapter is in the final stages of preparation for submission to the Australian Journal of Chemistry as

“Investigating the Influence of Crystal Disorder in MoS2 Cathodes for Secondary Hybrid Mg-Li Batteries”

Abstract The full extent to which the electrochemical consistent with a decrease in disorder. These properties of MoS2 electrodes are influenced by findings aid the optimisation of MoS2 electrodes their morphological characteristics, such as which show promise in several battery crystalline disorder, remains unclear. Here, we technologies. report that disorder introduced by ball-milling decreases the Faradaic component of cell capacity and leads to increasingly pseudocapacitive behaviour. After high temperature annealing, a more battery-like character of the cell is restored,

Introduction Lithium ion batteries are unmatched in portable energy storage applications for their high power and energy density.1-3 A range of Li-intercalation cathodes have been studied, amongst which MoS2 excels in cost, capacity, and power.4 5-7 Like graphite, 2H-MoS2 consists of Bernal-stacked monolayers held together by weak van der Waals forces. This layer structure allows for the intercalation of guest ions, such as Li+.8-10 Historically, intercalation has been understood as a process whereby Li+ ions diffuse between the MoS2 layers and are stabilised in octahedral interstices between sulfur atoms.11-12 The rate of diffusion between the basal layers is rapid as the activation energy of Li+ ion migration between these layers is low.11 Li+ ion diffusion does, however, cause expansion of the c-axis by ~5%.11, 13 This expansion induces strain in the MoS2 layers, which leads to formation of dislocations.14 As the lithium concentration nears saturation, the resultant strain can cause particles to break into polycrystalline fragments, potentially creating alternative pathways for Li+ ion diffusion into the

MoS2 particles.15 Recently, Cui and co-workers have shown that intercalation can also occur through the basal planes.16 The authors propose that Li+ ions may additionally diffuse through “natural defects” in the MoS2 lattice, though this process occurs at a relatively lower rate. As a result, it might be expected that the kinetics of the cathodic reaction (Equation 2) are enhanced in MoS2 particles with a high defect concentration, i.e. substantial disorder.

Defect introduction in MoS2 can be achieved both during and post-synthesis.17 Although sulfur vacancies, grain boundaries and strain can be controlled by varying synthesis parameters in techniques such as chemical vapour deposition,17-19 in-situ approaches have limited scalability. Post-synthetic treatments are better suited to upscaling with lower cost bulk materials, as they allow MoS2 to be sourced rather than synthesised. In MoS2 crystals, sulfur vacancies may be induced with electron beam or argon plasma treatment,20-21 Mo-O moieties incorporated by oxygen plasma treatment and hydrothermal treatment22-23 and lattice ripples induced by SF6,

CF4, and CHF3 plasmas.24 Ozone treatment and ion, proton, and alpha particle bombardment also introduce

64 Chapter 2 defects, though the effects are less well understood.17 An alternative to these high energy processes is mechanical milling, which is a scalable method to manipulate crystal morphology by applying impact and shear stress.25 When applied to layered materials, ball-milling has been shown to cleave, exfoliate, and twist sheets,25 introduce vacancies26 and expose edge sites.27 The coordinatively unsaturated atoms along exposed edge sites have been shown to preferentially bind Li+.12, 28 Thus, mechanical milling is expected to improve the capacity of MoS2 cathodes in Li+ ion batteries by introducing defects, and increasing the number of edge sites. Additionally, ball-milling is expected to decrease the particle size, increasing surface area and reducing diffusion path length, all of which are known to further improve the capacity of electrode materials in general.11

Lithium anodes are prone to forming dendrites, which can short circuit the cell, initiating an exothermic reaction between the electrode and electrolyte that can cause a battery to ignite.29-31 Efforts to inhibit dendrite formation have focused on modifying the electrode,32-33 electrolyte,34-37 separator,38-39 and battery management system,30 although at high current, low temperature, or low overcharge, dendrites still present a significant problem.31 A solution might be to change the anode material whilst retaining Li+ as the charge carrier. Magnesium is a viable alternative as it is not prone to dendrite formation and has suitable electrochemical characteristics (Table 6). Although research into Mg batteries is active, they are not commercially available yet, in part due to a lack of suitable cathode materials;40-45 the known cathode materials typically possess low capacity and/or poor intercalation kinetics.45-46 An alternative approach is to use Mg/Li dual-ion hybrid batteries to exploit the benefits of both Mg and Li batteries by combining the dendrite-free Mg anode with the fast kinetics of a Li+ intercalation cathode.47-49

Table 6 – Comparison of Li-, Mg-, and hybrid-ion battery components using a variety of metrics Metric Li Mg Hybrid References -1 a b 50 Gravimetric Capacity (mAhg ) 3861 , 372 anode 2205 anode 50 Volumetric Capacity (mAhcm-3) 2066a, 837 b anode 3833 anode Dendrites Yes No No 51-53 Sustainability Scarce (20 ppmc) Earth Abundant (21000 ppmc) Mixed 50, 54-55 Notes: aLi Metal, bgraphite, ccrustal abundance.

Mg2+ + 2e-  Mg – 2.4 V vs SHE (1) xLi++ MoS2 + xe-  LixMoS2 – 0-2.0 to – 0.1 V vs SHE (depending on x) (2) xMg2+ + MoS2 + 2xe-  MgxMoS2 – 1.6 V vs SHE (3) Li+ + e-  Li – 3.1 V vs SHE (4)

As the rate of reversible Li+ ion intercalation is the kinetic bottleneck of the cathodic reaction, the modification of bulk MoS2 morphology by ball-milling and associated changes at the atomic level as outlined above will the electrochemical performance batteries. This study, therefore, aims to investigate the effect of changes in the cathode disorder/crystallinity on the electrochemical performance of Mg/Li hybrid cells.

Results and Discussion Structural Characterisation

Two commercial bulk MoS2 samples, herein designated 2 µm and 90 nm (from the average layer stack height in the c-axis direction), were chosen for baseline measurements. Adjusting the treatment duration of mechanical milling led to samples with varying particle sizes and degrees of crystal defects. The 90 nm MoS2

65 Chapter 2

was ball-milled for 4, 24, and 60 hours to produce BM4-, BM24-, BM60-MoS2, respectively. Characterisation by transmission electron microscopy (TEM), Powder X-ray diffraction (PXRD), and Raman spectroscopy are shown in Figure 22 and Figure 23 and summarised alongside N2-sorption analysis (fitted with the Brunner- Emmett-Teller analysis; BET) in Table 7. Scanning electron microscopy (SEM) are shown in Figure S 21. As the ball-milling duration was lengthened (4, 24 and 60 hours), the MoS2 flakes became increasingly more broken and smaller in size (Figure S 21), the surface area rose from 8 to 90 m2g-1 (Figure S 21), the XRD reflections broadened (Figure 23), the layers in the crystallites became distorted with the introduction of defects (Figure 22), and the defect-induced LA(M) Raman mode56-57 increased in intensity (Figure 23). These results are all consistent with the crystals of the samples breaking into smaller, increasingly disordered crystallites (by the introduction of defects and strain) as a result of the ball milling.

Figure 22 – Transmission electron micrographs of post-milling sample, BM60-MoS2

A

101

002

100

105 110 008 1g 103 1 E2g LA(M) BM60- BM60- 2TA(X) BM24- BM24- BM4- BM4-

90nm-

Intensity (a.u.) Intensity Intensity (a.u.) Intensity 90nm- 2µm-MoS2

2µm-MoS2 P63/mm MoS2

150 200 250 300 350 400 450 10 20 30 40 50 60 Wavenumber (cm-1) °2theta (Cu)

Figure 23 –Raman spectra and X-ray diffractograms of MoS2 samples before and after ball-milling for various durations. Raman assignments 58 relate to MoS2 with the exception of the 2TA(X) mode, which arises from the Si substrate.

Table 7 – Summary of physical analyses of the ball-milled samples

Sample Surface Area Williamson-Hall Crystallite Size LA(M)/E2g ratio (-MoS2) (BET; m2g-1) (XRD; nm)a (Raman; arb.) 2 µm 5 500 - 90 nm 8 440 - BM4 26 180 0.06

66 Chapter 2

BM24 58 40 0.28 BM60 88 20 0.51 a Williamson-Hall analysis performed on the 002 MoS2 reflection.

Cell Preparation

The MoS2 samples, 2µm-, 90nm-, BM4-, BM24-, and BM60-MoS2, were suspended in N-methylpyrrolidone with polyvinylidene difluoride and carbon black (8:1:1 active: binder: carbon mass ratio), then coated onto stainless steel and dried at 120 °C. The coated electrodes were transferred to an argon-filled glove box and assembled as part of conventional coin cells using a 0.8 M Mg + 1.0 M LiCl in tetrahydrofuran and a Mg metal anode. The excess anode material and electrolyte concentration, in addition to the kinetically favoured anodic reaction, ensured that the cathode was the kinetic and capacity limiting component of the cell (see SI for details).

Electrochemical Performance The gravimetric capacity density, reversibility and stability of the coin cells were assessed. At a charging rate of 250 mAg-1, the capacity density of the samples increased from 170 to 266 mAhg-1 as ball-milling duration of the electrode precursors increased (Table 8).

Table 8 – Summary of electrochemical performance of the MoS2 samples Coulombic Value of b Sample Capacity density (mAhg-1) Capacity Decay (500 cycles, %) Efficiency (%) (average, CV) 0.82 2 µm 170 21.1  100 0.86 90 nm 227 24.4  100 0.85 BM4 225 30.0  100  n/a BM24 243 63.6 100 n/a BM60 266 52.7  100 Analysis of the voltage plateau (the region where discharging voltage remains nearly constant across a given capacity range) was used to reveal the charge-storage mechanism of the system. For our samples, electrodes prepared with MoS2 samples that were not ball-milled exhibit voltage plateaus for approximately a third of the entire discharge capacity. As the ball-milling duration of the electrode precursors increases, the plateau begins to increase in slope, then disappears entirely for the BM24- and BM60-MoS2 samples. This behaviour is consistent with a change to the charge-storage mechanism from battery-like to pseudo-capacitor-like.59

Figure 24 – Voltage profiles, coulombic efficiency and capacity retention graphs for MoS2 samples before and after ball-milling for various durations.

67 Chapter 2

Given the voltage profile of the ball-milled samples indicates a change in the relative extent with which the various charge storage mechanisms operate in parallel, the samples were further analysed by cyclic voltammetry to quantify the ratio of battery-/pseudo-capacitor-like behaviour. The cyclic voltammograms inFigure 26 shows two sets of maxima (at 0.7/1.4 and 1.6/1.8 V vs Mg/Mg+ for the cathodic/anodic reactions respectively) which correspond to the two plateaus in the cycling plots (Figure 24). These current maxima progressively broaden and decrease in magnitude as the ball-milling duration of the electrode precursors is increased. The current was analysed as a function of scan rate to further investigate the charge storage mechanism according to Equation 5.60-62

i = avb (5)

Where i = current, a and b = adjustable parameters, and v = scan rate; b indicates the charge storage mechanism

Values of b of 0.5 or 1 reflects typical battery (Faradaic) or capacitor (non-Faradaic) behaviour, respectively, with possible solutions to Equation 5 existing on a continuum between these two values.63 The values for b, derived from fitting both current maxima, are given in Figure 25 and Table 8. Values for b for the BM24- and

BM60-MoS2 could not be obtained as the maxima could not be resolved.

Figure 25 – Cyclic voltammograms and peak current analysis of MoS2 samples before and after ball-milling for various durations.

The stability of an electrode is assessed by examining the capacity retention over a given number of charge/discharge cycles. For the electrodes prepared from crystalline samples, 2µm- and 90nm-MoS2, a capacity decay of ~20% is observed after 500 cycles and this value increases to 50% for the electrodes prepared from ball-milled samples. The Coulombic Efficiency describes the reversibility of the cathodic reaction (Equation 6), which, alongside the capacity retention reflects the lifespan of a battery.64 Values of the Coulombic Efficiency less than unity indicate irreversible Li+ intercalation, changes to the cathode, or the presence of side reactions.65

퐶1 (푑𝑖푠푐ℎ푎푟𝑔푒/𝑖푛푡푒푟푐푎푙푎푡𝑖표푛) 푎푚표푢푛푡 표푓 퐿푖 푖푛푡푒푟푐푎푙푎푡푒푑 h퐶 = = (6) 퐶2 (푐ℎ푎푟𝑔푒/푑푒𝑖푛푡푒푟푐푎푙푎푡𝑖표푛) 푎푚표푢푛푡 표푓 퐿푖 푑푒푖푛푡푒푟푐푎푙푎푡푒푑

For all samples, after conditioning (~50 cycles), the Coulombic Efficiency reaches around 100 %.

68 Chapter 2

Influence of defects

Mechanical milling had a significant effect on the morphology and electrochemical performance of the MoS2 electrodes in this Mg/Li hybrid battery. The morphology of the samples becomes increasingly disordered (Table 7) and the charge storage mechanism shifts from battery-like to pseudo capacitor-like as the ball-milling duration of the electrode precursors is increased (Table 8).

The effect of ball-milling on crystalline samples is established; sheet exfoliation, particle size decrease, and defect generation are well precedented.66 In the current study, the creation of defects increases the variety of

MoS2 sites that can accept a Li+ ion.67 These sites may be any combination of edges, dislocations, and vacancies both on the surface and within the layers of MoS2. Each unique site possesses a different thermodynamic potential for the cathodic reaction (Equation 2), which results in a slope in the voltage profile of the ball-milled samples (Figure 24).67 These results are contrasted by the voltage profiles of the non-ball-milled samples, where the vast majority of the sites of the cathodic reaction (i.e. intercalation) are equivalent because of the crystalline nature of the sample. This results in the voltage plateaus observed for 90nm- and 2µm-MoS2. Additionally, because of the creation of new Li+ accepting sites upon milling, the capacity density was increased. Similar behaviour has been previously observed in graphite67 and graphene.68

(a) (b)

2.0 2 m 2 um 200 90 nm 0.25 A g-1 BM4 0.25 A g-1 1.5 0.50 A g-1 150 1.0 A g-1

1.0 2.0 A g-1

100 3.0 A g-1 5.0 A g-1 Voltage(V) 0.5

50 Capacity(mAhg-1) 2.0 A g-1 0.5 A g-1 -1 -1 -1 0.0 5.0 A g-1 3.0 A g 1.0 A g 0.25 A g 0 0 25 50 75 100 125 150 0 10 20 30 40 50 60 70 -1 Capacity(mAh g ) Cycles(N)

Figure 26 (a) Cycling profile of the coin cell with 2µm-MoS2 as the cathode material at different charging discharging current. (b) summary of the cathode capacity density with 3 types of cathode material at different current density.

Since cells incorporating electrodes fabricated from 2m-, 90 nm-, and BM4-MoS2 show relatively good capacity retention upon cycling, they were assessed at different current densities to investigate the influence of the current density on the resulting capacity of the cathode material. The discharging capacity was recorded at increasing current densities until a maximum current density of 5 A.g-1. A final scan at the original current density of 0.25A.g-1 was then performed to check whether the capacity can recover. Figure 26b plots the discharge capacity of cells with 2 m-, 90 nm-, and BM4-MoS2 electrodes as a function of current density. All these three batteries exhibit similar features: the discharging capacity gradually decreases as the current density is increased and returns to the original value when the current density is lowered. At the lowest current density

(0.25 A.g-1), the cell with the electrode prepared from 2m-MoS2 exhibits the lowest capacity (140 mAh.g-1) of the three and that with the electrode prepared from BM4-MoS2 shows the highest specific capacity (170 mAh.g-1). At the highest current density of 5 A.g-1, the ranking of the discharging capacity reverses: the cell with the electrode prepared from 2m-MoS2 is least affected (with a capacity decay of 60%) and that with an electrode prepared from BM4-MoS2 is most affected (with a capacity decay of 80%). We conclude that this is

69 Chapter 2 due to the undamaged crystal structure, which enables intercalation to be maintained more efficiently even at the higher charging rates. The ball-milled samples are more disordered as well as defect-rich and simply have a lower inherent intercalation capacity, being dominated by more capacitive characteristics.16, 69

Thus, we propose that to optimise Li+-ion MoS2 electrodes, the crystallinity should be maximised and the particle size minimised. The crystallinity determines the capacity density of the intercalation region and the particle size determines the rate of Li+ intercalation. In order to test this hypothesis, we annealed the BM60-

MoS2 sample at 800 °C in a reducing atmosphere for 24 hours – the resultant sample was designated Ann’d-

MoS2. The physical characterisation data for the Ann’d-MoS2 is shown in Figure 27. In summary, the sample exhibited sharper XRD reflections (which corresponded to a crystallite size of 100 nm; up from 20 nm) and the LA(M) signal in the Raman spectrum decreased in intensity relative to the un-annealed sample (to an

LA(M):E2g ratio of 0.33; down from 0.51). These changes are consistent with improved particle crystallinity.56- 57, 70

Wavenumber (cm-1) 100 200 300 400 500 600 12 2.0

6 1.5

1.0 6 0

Voltage (V) Voltage 0.5

4

Counts (x10000) Counts

Intensity (arb. ×1000) (arb. Intensity 2 0.0 10 20 30 40 50 60 0 50 100 150 °2theta (Cu) Capacity (mAh g-1)

Figure 27 –XRD (black) and Raman spectroscopy (red) data for the Ann’d-MoS2 sample (i.e. BM60-MoS2 annealed for 24 hours) (right) and (left)

coin cell performance with thermal annealed MoS2 cathode.

By checking whether the cycling curve of the cell incorporating the Ann’d-MoS2 sample possesses a plateau, one can determine whether the battery-like behaviour of the cell has increased (or correspondingly, whether the capacitor-like behaviour has decreased). As seen in Figure 27, the departure of the curve from linearity indicates an increased contribution from battery-like intercalation processes and decreased contribution from the capacitor-like adsorption processes. Relative to the non-annealed, ball-milled samples, which did not exhibit a plateau, this is a significant improvement. The data are consistent with our proposed hypothesis – that defects introduced by ball-milling decrease the faradaic contribution of a cell’s capacity and lead to increasingly pseudo- capacitive behaviour. When the defects are removed by high temperature annealing, the battery-like character of the cell is restored, supporting our hypothesis above.

Conclusions

We have shown that the electrochemical performance of MoS2 cathodes in Mg/Li batteries is primarily determined by two factors: the particle size and crystallinity. Commercial MoS2 samples were ball milled to manipulate these two factors. To separate the effects of sample crystallinity from particle size, the ball milled samples were annealed and the electrochemistry of the two samples compared. The annealed samples exhibited a larger capacity density plateau, consistent with the homogenisation of the active sites, in turn, consistent with increased crystallinity. These samples also exhibited a greater capacity than the samples that had not undergone ball-milling which was attributed to the increased surface are of the ball milled particles.

70 Chapter 2

References 1. Nitta, N.; Wu, F.; Lee, J. T.; Yushin, G., Li-ion battery materials: present and future. Mater. Today 2015, 18 (5), 252-264. 2. Li, M.; Lu, J.; Chen, Z.; Amine, K., 30 years of lithium‐ion batteries. Adv. Mater. 2018, 30 (33), 1800561. 3. Scrosati, B.; Hassoun, J.; Sun, Y.-K., Lithium-ion batteries. A look into the future. Energy Environ. Sci. 2011, 4 (9), 3287-3295. 4. Hsu, C.-J.; Chou, C.-Y.; Yang, C.-H.; Lee, T.-C.; Chang, J.-K., MoS 2/graphene cathodes for reversibly storing Mg 2+ and Mg 2+/Li+ in rechargeable magnesium-anode batteries. Chem. Commun. 2016, 52 (8), 1701-1704. 5. Teng, Y.; Zhao, H.; Zhang, Z.; Li, Z.; Xia, Q.; Zhang, Y.; Zhao, L.; Du, X.; Du, Z.; Lv, P.; Świerczek, K., MoS2 Nanosheets Vertically Grown on Graphene Sheets for Lithium-Ion Battery Anodes. ACS Nano 2016, 10 (9), 8526-8535. 6. Azhagurajan, M.; Kajita, T.; Itoh, T.; Kim, Y.-G.; Itaya, K., In Situ Visualization of Lithium Ion Intercalation into MoS2 Single Crystals using Differential Optical Microscopy with Atomic Layer Resolution. J. Am. Chem. Soc. 2016, 138 (10), 3355-3361. 7. Xiong, F.; Wang, H.; Liu, X.; Sun, J.; Brongersma, M.; Pop, E.; Cui, Y., Li intercalation in MoS2: in situ observation of its dynamics and tuning optical and electrical properties. Nano Lett. 2015, 15 (10), 6777-6784. 8. Zhang, J.; Yang, A.; Wu, X.; van de Groep, J.; Tang, P.; Li, S.; Liu, B.; Shi, F.; Wan, J.; Li, Q.; Sun, Y.; Lu, Z.; Zheng, X.; Zhou, G.; Wu, C.-L.; Zhang, S.-C.; Brongersma, M. L.; Li, J.; Cui, Y., Reversible and selective ion intercalation through the top surface of few-layer MoS2. Nature Communications 2018, 9 (1), 5289. 9. Sun, D.; Ye, D.; Liu, P.; Tang, Y.; Guo, J.; Wang, L.; Wang, H., MoS2/Graphene Nanosheets from Commercial Bulky MoS2 and Graphite as Anode Materials for High Rate Sodium‐Ion Batteries. Advanced Energy Materials 2018, 8 (10), 1702383. 10. Spirko, J. A.; Neiman, M. L.; Oelker, A. M.; Klier, K., Electronic structure and reactivity of defect MoS2: I. Relative stabilities of clusters and edges, and electronic surface states. Surf. Sci. 2003, 542 (3), 192-204. 11. Stephenson, T.; Li, Z.; Olsen, B.; Mitlin, D., Lithium ion battery applications of molybdenum disulfide (MoS2) nanocomposites. Energy & Environmental Science 2014, 7 (1), 209-231. 12. Li, Y.; Wu, D.; Zhou, Z.; Cabrera, C. R.; Chen, Z., Enhanced Li adsorption and diffusion on MoS2 zigzag nanoribbons by edge effects: a computational study. The journal of physical chemistry letters 2012, 3 (16), 2221-2227. 13. Whittingham, M. S.; Gamble Jr, F. R., The lithium intercalates of the transition metal dichalcogenides. Mater. Res. Bull. 1975, 10 (5), 363-371. 14. Mulhern, P. J., Lithium intercalation in crystalline LixMoS2. Can. J. Phys. 1989, 67 (11), 1049-1052. 15. Chrissafis, K.; Zamani, M.; Kambas, K.; Stoemenos, J.; Economou, N. A.; Samaras, I.; Julien, C., Structural Studies of MoS2 Intercalated by Lithium. Materials Science and Engineering B-Solid State Materials for Advanced Technology 1989, 3 (1-2), 145-151. 16. Zhang, J.; Yang, A.; Wu, X.; van de Groep, J.; Tang, P.; Li, S.; Liu, B.; Shi, F.; Wan, J.; Li, Q., Reversible and selective ion intercalation through the top surface of few-layer MoS 2. Nature communications 2018, 9 (1), 1-9. 17. Lin, Z.; Carvalho, B. R.; Kahn, E.; Lv, R.; Rao, R.; Terrones, H.; Pimenta, M. A.; Terrones, M., Defect engineering of two- dimensional transition metal dichalcogenides. 2D Materials 2016, 3 (2), 022002. 18. Huang, C.; Wu, S.; Sanchez, A. M.; Peters, J. J.; Beanland, R.; Ross, J. S.; Rivera, P.; Yao, W.; Cobden, D. H.; Xu, X., Lateral heterojunctions within monolayer MoSe 2–WSe 2 semiconductors. Nat. Mater. 2014, 13 (12), 1096-1101. 19. Castellanos-Gomez, A.; Roldán, R.; Cappelluti, E.; Buscema, M.; Guinea, F.; van der Zant, H. S.; Steele, G. A., Local strain engineering in atomically thin MoS2. Nano Lett. 2013, 13 (11), 5361-5366. 20. Komsa, H.-P.; Kurasch, S.; Lehtinen, O.; Kaiser, U.; Krasheninnikov, A. V., From point to extended defects in two-dimensional MoS${}_{2}$: Evolution of atomic structure under electron irradiation. PhRvB 2013, 88 (3), 035301. 21. Chow, P. K.; Jacobs-Gedrim, R. B.; Gao, J.; Lu, T.-M.; Yu, B.; Terrones, H.; Koratkar, N., Defect-induced photoluminescence in monolayer semiconducting transition metal dichalcogenides. ACS Nano 2015, 9 (2), 1520-1527. 22. Xie, J.; Zhang, J.; Li, S.; Grote, F.; Zhang, X.; Zhang, H.; Wang, R.; Lei, Y.; Pan, B.; Xie, Y., Controllable disorder engineering in oxygen-incorporated MoS2 ultrathin nanosheets for efficient hydrogen evolution. J. Am. Chem. Soc. 2013, 135 (47), 17881- 17888. 23. Kang, N.; Paudel, H. P.; Leuenberger, M. N.; Tetard, L.; Khondaker, S. I., Photoluminescence quenching in single-layer MoS2 via oxygen plasma treatment. The Journal of Physical Chemistry C 2014, 118 (36), 21258-21263. 24. Chen, M.; Nam, H.; Wi, S.; Priessnitz, G.; Gunawan, I. M.; Liang, X., Multibit Data Storage States Formed in Plasma-Treated MoS2 Transistors. ACS Nano 2014, 8 (4), 4023-4032. 25. Zhang, S.; Cui, Y.; Wu, B.; Song, R.; Song, H.; Zhou, J.; Chen, X.; Liu, J.; Cao, L., Control of graphitization degree and defects of carbon blacks through ball-milling. RSC advances 2014, 4 (1), 505-509. 26. Zhang, B.; Lu, L.; Lai, M., Evolution of vacancy densities in powder particles during mechanical milling. Physica B: Condensed Matter 2003, 325, 120-129. 27. Wang, D.; Wang, Z.; Wang, C.; Zhou, P.; Wu, Z.; Liu, Z., Distorted MoS2 nanostructures: An efficient catalyst for the electrochemical hydrogen evolution reaction. Electrochem. Commun. 2013, 34, 219-222. 28. Wang, H.; Zhang, Q.; Yao, H.; Liang, Z.; Lee, H.-W.; Hsu, P.-C.; Zheng, G.; Cui, Y., High Electrochemical Selectivity of Edge versus Terrace Sites in Two-Dimensional Layered MoS2 Materials. Nano Lett. 2014, 14 (12), 7138-7144. 29. Scrosati, B.; Garche, J., Lithium batteries: Status, prospects and future. J. Power Sources 2010, 195 (9), 2419-2430. 30. Wen, J.; Yu, Y.; Chen, C., A review on lithium-ion batteries safety issues: existing problems and possible solutions. Materials express 2012, 2 (3), 197-212.

71 Chapter 2

31. Wu, H.; Zhuo, D.; Kong, D.; Cui, Y., Improving battery safety by early detection of internal shorting with a bifunctional separator. Nature communications 2014, 5, 5193. 32. Park, K.-S.; Im, D.; Benayad, A.; Dylla, A.; Stevenson, K. J.; Goodenough, J. B., LiFeO2-incorporated Li2MoO3 as a cathode additive for lithium-ion battery safety. Chem. Mater. 2012, 24 (14), 2673-2683. 33. Ding, F.; Xu, W.; Graff, G. L.; Zhang, J.; Sushko, M. L.; Chen, X.; Shao, Y.; Engelhard, M. H.; Nie, Z.; Xiao, J., Dendrite-free lithium deposition via self-healing electrostatic shield mechanism. J. Am. Chem. Soc. 2013, 135 (11), 4450-4456. 34. Zhang, S. S., A review on electrolyte additives for lithium-ion batteries. J. Power Sources 2006, 162 (2), 1379-1394. 35. Yang, C.; Chen, J.; Ji, X.; Pollard, T. P.; Lü, X.; Sun, C.-J.; Hou, S.; Liu, Q.; Liu, C.; Qing, T.; Wang, Y.; Borodin, O.; Ren, Y.; Xu, K.; Wang, C., Aqueous Li-ion battery enabled by halogen conversion–intercalation chemistry in graphite. Nature 2019, 569 (7755), 245-250. 36. Xu, K., Electrolytes and Interphases in Li-Ion Batteries and Beyond. Chem. Rev. 2014, 114 (23), 11503-11618. 37. Xu, K., Nonaqueous Liquid Electrolytes for Lithium-Based Rechargeable Batteries. Chem. Rev. 2004, 104 (10), 4303-4418. 38. Arora, P.; Zhang, Z., Battery Separators. Chem. Rev. 2004, 104 (10), 4419-4462. 39. Kang, S. M.; Ryou, M.-H.; Choi, J. W.; Lee, H., Mussel-and diatom-inspired silica coating on separators yields improved power and safety in Li-ion batteries. Chem. Mater. 2012, 24 (17), 3481-3485. 40. Zhang, R.; Ling, C., Status and challenge of Mg battery cathode. MRS Energy & Sustainability 2016, 3. 41. Mao, M.; Gao, T.; Hou, S.; Wang, C., A critical review of cathodes for rechargeable Mg batteries. Chem. Soc. Rev. 2018. 42. Muldoon, J.; Bucur, C. B.; Gregory, T., Fervent Hype behind Magnesium Batteries: An Open Call to Synthetic Chemists‐ Electrolytes and Cathodes Needed. Angew. Chem. 2017. 43. Canepa, P.; Sai Gautam, G.; Hannah, D. C.; Malik, R.; Liu, M.; Gallagher, K. G.; Persson, K. A.; Ceder, G., Odyssey of multivalent cathode materials: open questions and future challenges. Chem. Rev. 2017, 117 (5), 4287-4341. 44. Song, J.; Sahadeo, E.; Noked, M.; Lee, S. B., Mapping the Challenges of Magnesium Battery. The Journal of Physical Chemistry Letters 2016, 7 (9), 1736-1749. 45. Huie, M. M.; Bock, D. C.; Takeuchi, E. S.; Marschilok, A. C.; Takeuchi, K. J., Cathode materials for magnesium and magnesium- ion based batteries. Coord. Chem. Rev. 2015, 287, 15-27. 46. Liang, Y.; Yoo, H. D.; Li, Y.; Shuai, J.; Calderon, H. A.; Robles Hernandez, F. C.; Grabow, L. C.; Yao, Y., Interlayer-expanded molybdenum disulfide nanocomposites for electrochemical magnesium storage. Nano Lett. 2015, 15 (3), 2194-2202. 47. Yao, H.-R.; You, Y.; Yin, Y.-X.; Wan, L.-J.; Guo, Y.-G., Rechargeable dual-metal-ion batteries for advanced energy storage. Phys. Chem. Chem. Phys. 2016, 18 (14), 9326-9333. 48. Rashad, M.; Li, X.; Zhang, H., A Magnesium/Lithium ion hybrid battery with high reversibility by employing NaV3O8. 1.69 H2O nanobelts as positive electrode. ACS Appl. Mater. Interfaces 2018. 49. Li, H.; Okamoto, N. L.; Hatakeyama, T.; Kumagai, Y.; Oba, F.; Ichitsubo, T., Fast Diffusion of Multivalent Ions Facilitated by Concerted Interactions in Dual‐Ion Battery Systems. Advanced Energy Materials 2018, 8 (27), 1801475. 50. Cheng, Y.; Chang, H. J.; Dong, H.; Choi, D.; Sprenkle, V. L.; Liu, J.; Yao, Y.; Li, G., Rechargeable Mg-Li hybrid batteries: status and challenges. J. Mater. Res 2016, 31 (20), 3125-3141. 51. Chen, L.; Zhao, S.; Liu, Y.; Horne, M.; Bond, A. M.; Zhang, J., Room temperature electrodeposition of metallic magnesium from ethylmagnesium bromide in tetrahydrofuran and ionic liquid mixtures. J. Electrochem. Soc. 2016, 163 (4), H3043-H3051. 52. Matsui, M., Study on electrochemically deposited Mg metal. J. Power Sources 2011, 196 (16), 7048-7055. 53. Ling, C.; Banerjee, D.; Matsui, M., Study of the electrochemical deposition of Mg in the atomic level: Why it prefers the non- dendritic morphology. Electrochim. Acta 2012, 76, 270-274. 54. Larcher, D.; Tarascon, J.-M., Towards greener and more sustainable batteries for electrical energy storage. Nature Chem. 2015, 7 (1), 19-29. 55. Rumble, J. R.; Lide, D. R.; Bruno, T. J., Abundance of Elements in the Earth’s Crust and in the Sea - CRC Handbook of Chemistry and Physics. 2017. 56. Li, H.; Zhang, Q.; Yap, C. C. R.; Tay, B. K.; Edwin, T. H. T.; Olivier, A.; Baillargeat, D., From Bulk to Monolayer MoS2: Evolution of Raman Scattering. Adv. Funct. Mater. 2012, 22 (7), 1385-1390. 57. McDevitt, N.; Zabinski, J.; Donley, M.; Bultman, J., Disorder-induced low-frequency Raman band observed in deposited MoS2 films. Appl. Spectrosc. 1994, 48 (6), 733-736. 58. El Garah, M.; Bertolazzi, S.; Ippolito, S.; Eredia, M.; Janica, I.; Melinte, G.; Ersen, O.; Marletta, G.; Ciesielski, A.; Samorì, P., MoS2 nanosheets via electrochemical lithium-ion intercalation under ambient conditions. FlatChem 2018, 9, 33-39. 59. Yu, F.; Liu, Z.; Zhou, R.; Tan, D.; Wang, H.; Wang, F., Pseudocapacitance contribution in boron-doped graphite sheets for anion storage enables high-performance sodium-ion capacitors. Materials Horizons 2018, 5 (3), 529-535. 60. Jiang, Y.; Liu, J., Definitions of pseudocapacitive materials: a brief review. Energy & Environmental Materials 2019, 2 (1), 30- 37. 61. Wang, J.; Polleux, J.; Lim, J.; Dunn, B., Pseudocapacitive Contributions to Electrochemical Energy Storage in TiO2 (Anatase) Nanoparticles. The Journal of Physical Chemistry C 2007, 111 (40), 14925-14931. 62. Lindström, H.; Södergren, S.; Solbrand, A.; Rensmo, H.; Hjelm, J.; Hagfeldt, A.; Lindquist, S.-E., Li+ ion insertion in TiO2 (anatase). 2. Voltammetry on nanoporous films. The Journal of Physical Chemistry B 1997, 101 (39), 7717-7722. 63. Simon, P.; Gogotsi, Y.; Dunn, B., Where do batteries end and supercapacitors begin? Science 2014, 343 (6176), 1210-1211. 64. Dahn, J.; Burns, J.; Stevens, D., Importance of Coulombic Efficiency Measurements in R&D Efforts to Obtain Long-Lived Li- Ion Batteries. The Electrochemical Society Interface 2016, 25 (3), 75-78. 65. Dahn, J.; Burns, J.; Stevens, D., Importance of coulombic efficiency measurements in R&D efforts to obtain long-lived Li-ion batteries. Electrochem. Soc. Interface 2016, 25 (3), 75.

72 Chapter 2

66. Zhang, Y.; Tao, L.; Xie, C.; Wang, D.; Zou, Y.; Chen, R.; Wang, Y.; Jia, C.; Wang, S., Defect engineering on electrode materials for rechargeable batteries. Adv. Mater. 2020, 32 (7), 1905923. 67. Sivakkumar, S. R.; Milev, A. S.; Pandolfo, A. G., Effect of ball-milling on the rate and cycle-life performance of graphite as negative electrodes in lithium-ion capacitors. Electrochim. Acta 2011, 56 (27), 9700-9706. 68. Dong, Y.; Zhang, S.; Du, X.; Hong, S.; Zhao, S.; Chen, Y.; Chen, X.; Song, H., Boosting the electrical double‐layer capacitance of graphene by self‐doped defects through ball‐milling. Adv. Funct. Mater. 2019, 29 (24), 1901127. 69. Hwang, H.; Kim, H.; Cho, J., MoS2 nanoplates consisting of disordered graphene-like layers for high rate lithium battery anode materials. Nano Lett. 2011, 11 (11), 4826-4830. 70. Mittemeijer, E. J.; Welzel, U., The “state of the art” of the diffraction analysis of crystallite size and lattice strain. Zeitschrift für Kristallographie - Crystalline Materials 2008, 223 (9), 552-560.

73 Chapter 3

Chapter 3 Elucidation of Structure and Support Interactions for a Highly Active Molybdenum Carbo-nitride@Titanium Nitride Hydrodeoxygenation Catalyst

This work has been submitted to Physical Chemistry Chemical Physics as “Elucidation of Structure and Support Interactions for a Highly Active Molybdenum Carbo-nitride@Titanium Nitride Hydrodeoxygenation Catalyst” and is currently under peer review.

Abstract This study unveils the key structural features and titanium nitride support with the molybdenum compositional aspects of several molybdenum carbo-nitride phase. The titanium nitride and hemicarbide-based catalysts using a combination of molybdenum carbide phases are intimately mixed synchrotron techniques, X-ray diffraction and and, surprisingly, the result of the EXAFS analysis elemental mapping, in conjunction with electron indicates that the catalyst incorporates terminal microscopy. These molybdenum hemicarbides and Mo≡N motifs. a molybdenum carbo-nitride composite material were prepared by thermal decomposition of a molybdenum–melamine polymer in the presence of titanium nitride at 650 °C, under a reducing atmosphere. Although this composite was shown to be more effective than either Mo2C, or Mo2N for the hydrodeoxygenation of lignin to single-ring arenes in supercritical ethanol, its structure and mechanism of action were not well understood. Extended X-ray absorption fine structure experiments at 300 °C, under a stream of ethanol vapor to approximate in operando conditions, elucidate the importance of the interaction of the Introduction Almost 50 years ago, Boudart and Levy published their landmark study in which the “platinum-like” properties of molybdenum carbide were shown to be due to d-band broadening upon carbon insertion into the molybdenum lattice.1 Subsequently, work by Oyama and co-workers demonstrated the potential of early transition metal carbides and nitrides as petroleum hydrotreating catalysts.2 Despite these advances having driven interest in transition metal carbides as catalysts for reductive processes generally, it is only since 2012 that a significant body of literature has accrued in which molybdenum carbides, nitrides, and subsequently their composites, have been demonstrated to outperform traditional platinum-based catalysts: especially in the context of electrocatalytic water splitting and the Hydrogen Evolution Reaction (HER). Whilst it is not easy to predict the activity of a catalyst for hydrogenation or hydrogenolysis based on its HER performance, in some cases overlap of these catalytic abilities is to be expected on the basis of micro-reversible interactions of hydrogen with the catalyst surface.

From a sample of the recent literature in which high-performance, molybdenum carbide-based electrocatalysts were used for HER, two clear trends are apparent: (1) the nano-architecture of the material and support interactions play crucial roles in determining the working properties of the catalyst. Thus the reported materials include pomegranate-like N- & P-doped Mo2C@C nanospheres;3 holey reduced graphene oxide Mo2N–Mo2C heterojunctions;4 N-doped Mo2C nanoparticles;5 Mo2C@graphene–N-doped

74 Chapter 3 porous C-microspheres;6 phosphorus-doped molybdenum nitride;7 N-doped molybdenum carbide & phosphide;8 and well-dispersed [email protected] (2) The use of dopants and composite materials vastly improves the electrochemical properties of the molybdenum carbo-nitride catalysts compared with either the pure metal carbide or nitride phase.

Chemical valorisation of biomass is another field of research in which a key objective is to produce new, high-performance catalysts using elements that are abundant and inexpensive, yet outperform the current, more expensive ones. In-depth structural investigations are therefore crucial for providing information as the basis for further refinement of synthetic protocols leading to improved catalysts and for correlation with insights gained from mechanistic & in operando studies. The depolymerisation of lignin is an incredibly active field of current research. More than 20 reviews, covering a wide range of approaches and potential applications, have been published since 2010.10 Almost invariably reports of lignin depolymerisation focus on the production of oxygenated arenes which have niche applications, but are of limited commercial value. Conversely, we have recently demonstrated that molybdenum hemicarbide-based catalysts both depolymerise and deoxygenate lignin to commercially desirable benzene, toluene, ethylbenzene, xylene (BTEX) arenes under basic conditions in supercritical ethanol.11 We have continued to screen support materials for beneficial synergistic interactions with the carbide phase. As a result, we have discovered that a highly dispersed molybdenum carbide phase immobilised on TiN is even more active for lignin depolymerisation and hydrodeoxygenation and is more selective for BTEX arenes than catalysts previously reported by us (See Table 9)12 and others: bulk Mo2C,11 carbon-supported MoC1−x13 and Mo2N14 being inferior catalysts under comparable reaction conditions.

Table 10. Highlights of catalytic conversion of Kraft lignin to BTEX-type arenes under supercritical conditions.a, 12 Catalyst Liquefaction (%)b Oxygenates (mg/g)c Arenes (mg/g)c None 85 77 26 280 °C 330 °C 80 129 145 Mo2C 64 201 226 280 °C 330°C 82 27 514 Mo2C@Al2O3 68 250 166 280 °C 330 °C 100 24 579 Mo2C1−xNx@TiN 100 0 666d 330 °C a Reaction conditions: Kraft lignin (1.0 g); ethanol (100 mL); 330 °C; 1.3 bar N2 (initial pressure), 110–190 bar (final pressure); Catalyst (15 wt% loading of Mo w.r.t. lignin), 6 h. b The reaction produces gas, liquefied (EtOH- soluble) and solid fractions. The extent of liquefaction is used as a proxy for conversion. c The liquefied fraction includes single-ring aromatic and non-aromatic products. The aromatics were identified & quantified by GC- MS and their yield is expressed as mg of product per g of lignin starting material. d 70% of these fall within the BTEX group.

Our approach to lignin valorisation, directly addresses 4 of the 6 research goals recently identified as necessary for the production of aromatics from lignin:15 the need for more realistically complex model compounds; robust, impurity-tolerant, recyclable catalysts; improved technologies; multi- or hybrid catalysts to accommodate multiple sequential and/or cascade reactions. These issues are not limited to lignin, but rather, are key technological barriers to be overcome for the successful utilization of biomass as a raw material for chemicals and fuels in general and hence our focus in the current paper is to understand which structural features of our Mo2C1−xNx@TiN composite are important for its inherently high selectivity and robustness, and whether the activities of this class of catalytic materials can be further optimised.

75 Chapter 3

In the current study, several heterogeneous, molybdenum carbide-based catalysts were prepared and characterised. The overall aim was to correlate these materials’ previously observed catalytic properties with their structural features. We focused on the composition of the active phase; its interaction with the support material and how this interaction affected the catalyst morphology and nanoparticle distribution; and, in the case of our best candidates, how these features changed upon heating: in the absence or presence of ethanol (i.e. under approximate in operando conditions).

Results and Discussion

Unsupported Mo2C, Al2O3-supported Mo2C and the TiN-supported catalyst were prepared according to the procedures described in the Supporting Information.† The synthetic procedures reliably furnished several batches of each catalyst and variation between batches was not detected under our characterisation regime.

The samples were analysed by PXRD at room temperature and subsequent Rietveld analysis confirmed that the Mo2C sample was identical to pure Mo2C (P63/mmc, ICSD 43669) and the Mo2C1−xNx@TiN as containing Mo2C and TiN (Mo2C: P63/mmc, ICSD 43669; TiN: Fm−3m, ICSD 26947). For the

Mo2C@Al2O3 sample, Mo2C and Al2O3 with minor MoO2 impurities were identified. (Mo2C: P63/mmc,

ICSD 43669; MoO2: P21/c, ICSD 23722; Al2O3: R-3ch, ICSD 9770; Figure 30). The MoO2 domains on the

Al2O3 support are likely due to oxygen migration from the Al2O3 to the nascent Mo species during the thermal decomposition of the [Mo–melamine] precursor complex, as these phases are not present in the bulk Mo2C or Mo2C1−xNx@TiN samples prepared under identical conditions.

Bright field transmission electron microscopic analysis of the Mo2C1−xNx@TiN sample did not reveal any clear differences between the molybdenum carbo-nitride and TiN phases (Figure 28). All particles were approximately the same size, contrast and morphology. Analysis of the Mo2C@Al2O3 sample by TEM showed two distinct morphologies: those of Al2O3 and Mo2C (Figure 29).

Figure 28. Bright field TEM image of the Mo2C1−xNx@TiN sample. Scale bar (50 nm) shown in bottom right.

XAS Studies X-ray Absorption Spectroscopy (XAS) is a useful technique to probe the electronic and structural influence of different support materials on the active phase within heterogeneous catalysts. Hence, the Mo k-edge spectra (20 keV) of the molybdenum carbide phases within bulk Mo2C, Mo2C@Al2O3 and

Mo2C1−xNx@TiN were examined on the XAS beamline at the Australian Synchrotron. The X-ray data were collected at 25 °C and at 300 °C under an inert (N2) or reductive (EtOH/N2) atmospheres for all three samples.

76 Chapter 3

Figure 29. Bright field TEM image of the Mo2C@Al2O3 sample. Scale bar (200 nm) shown in bottom right.

The X-ray Absorption Near Edge Structure (XANES) regions of these three materials are shown in Figure 31a, comparing them at (a) 25 °C and at (b) 300 °C. Aside from the Mo foil, the room temperature bulk

Mo2C sample has the lowest energy absorption edge at 20.012 keV (assigned from the peak in the first derivative), whereas the supported samples have higher energy edge positions at 20.014 keV and 20.016 keV for the Mo2C@Al2O3 and Mo2C1−xNx@TiN samples, respectively. Given that a higher energy edge shift indicates a higher apparent oxidation state,16 all supported catalysts have less electron density at Mo relative to that in unsupported Mo2C.

Figure 30. XRD diffractograms of Mo2C, Mo2C@Al2O3 and Mo2C1−xNx@TiN at 25 °C and 500 °C with peaks from ICSD entries: Mo2C: 43669,

TiN: 26947, MoO2: 23722

Recent literature has suggested the relationship of Mo oxidation with edge position is linear (See Figure

S23), which would suggest in this case bulk Mo2C is close to Mo(0) (metallic) and the supported samples are on average approximately Mo(III).17-18 Previously reports of molybdenum XANES studies, where carbon nanotubes have been used as a support for Mo2C, state that a higher energy absorption edge relative 19 to bulk Mo2C demonstrates a charge-transfer from molybdenum to the support. We therefore interpret the significant variation in edge energies in our catalyst set to be indicative of differences in the charge-

77 Chapter 3 transfer from molybdenum to the various support materials. In a related study, Liu et al. suggested such changes in edge energies result in a downshift in the d-band centre of Mo, thus decreasing its hydrogen binding energy. As a consequence, enhanced hydrogen evolution due to relatively moderate Mo–H binding was expected on the basis of molecular orbital calculations of Mo2C nanoparticles.20 For our catalysts the presence of the supports is therefore similarly expected to have a significant influence on the binding energy of hydrogen to the active site during catalysis compared with the unsupported molybdenum carbide.

XANES data were also collected on these samples at 300 °C under N2 (Figure 31b). In comparison to the 25 °C data, a clear edge shift to lower energy is observed for the supported catalysts, suggesting a reduction in the charge-transfer from the Mo to the support. Despite this shift, the Mo absorptions remain at higher energies than that of Mo in the bulk Mo2C sample, which remains relatively unchanged at higher temperature. The largest shift in edge energy is observed for the Mo2C@Al2O3 sample and is especially noticeable in the same region as the pre-edge feature for the [(NH4)6Mo7O24] reference sample. This feature has been attributed to a MoO pre-edge transition and hence it may be that within the Mo2C@Al2O3 sample, some MoO has formed at the interface of the alumina support with the Mo2C phase.

Figure 31. Mo k-edge XANES spectra for Mo2C (blue), Mo2C@Al2O3 (green), Mo2C1−xNx@TiN (red) at (a) 25 °C and (b) 300 °C. Reference

spectra of Mo foil (black) and [(NH4)6Mo7O24] (purple) recorded at room temperature are shown in both sets of spectra.

The XANES analysis at 300 °C was conducted in the presence and absence of a stream of ethanol vapour in N2. Identical edge shifts were observed compared with the data collected at 25 °C indicating that the apparent reduction of Mo is not likely to be due to ethanol but more likely a change in the charge transfer of the support to the metal at high temperature. One possibility that would appear to be supported by EXAFS analysis (vide infra) is that the anchored nanoparticles that make up the molybdenum carbide phase within this material physically spread or disperse upon heating becoming more intimately embedded in the support.

Extended X-ray Absorption Fine Structure (EXAFS) spectroscopy of the 3 samples was also collected in order to investigate differences in structure between supported and unsupported catalysts. For each sample, data were collected at 25 °C and at 300 °C to investigate the changes as a result of the temperature increase (Figure 32).

78 Chapter 3

Comparing the Fourier transform (FT) EXAFS spectra of the three molybdenum samples at 25 °C, the

Mo2C and Mo2C@Al2O3 samples show very little difference in structure despite a significant difference in their absorption edge energies being observed by XANES. Both fit the Mo2C structure to a high statistical degree (Table 11, Figure 32).21 Taken together these data suggest that although Mo2C@Al2O3 is predominately bulk Mo2C (EXAFS), oxidation of the Mo at or near the support is more pronounced than for the unsupported material (XANES). The FT EXAFS of the Mo2C1−xNx@TiN however, is markedly different. The peak observed at 2.25 Å is much lower in amplitude than the corresponding ones in the spectra for the other two materials.

3 Figure 32. Mo k-edge Fourier transform EXAFS spectra for Mo2C (black), Mo2C@Al2O3 (red), Mo2C1−xNx@TiN (green) at (a) 25 °C [k - weighted] and (b) 300 °C [k2-weighted].

The fitting of the Mo2C data indicates this peak at 2.25 Å arises from the Mo–Mo bonding at ~2.95 Å and

~4.21 Å. Given there is a significantly smaller contribution of this pathway in the Mo2C1−xNx@TiN sample, it appears that the Mo2C on TiN has much less of a bulk Mo2C character, suggesting small or integrated

Mo2C particles on the support, resulting in the break-up of much of the Mo–Mo bonding which is so characteristic of the bulk Mo2C structure. According to our fitting analysis of Mo2C1−xNx@TiN (Table 11.

, Figure 32), the most likely explanation for the reorganisation of the bulk Mo2C structure is the presence of an Mo to N bond that is not observed in the other samples. This Mo to N bond is resolved at 1.65 Å with very high statistical confidence. Such a short distance points to the presence of a Mo to N triple bond and bonds of this type of very similar lengths have been observed in discrete organometallic MoN complexes.22-23 The presence of a nitride support appears to influence the thermal decomposition of the precursor complex, resulting in significant incorporation of nitrogen in the molybdenum phase and inhibition of the growth of larger molybdenum carbo-nitride nanoparticles. Nitrogen essentially acts as a ‘capping’ agent to give a supported catalyst with terminal Mo≡N motifs.

Fitting analysis of the EXAFS data required that 3 neighbouring carbons still be present in the model, evidence of retention of the core Mo2C structure. Hence this sample was designated Mo2C1−xNx@TiN, rather than simply Mo2C@TiN.

Higher disorder in the structure of the samples was evident in the EXAFS spectra recorded at 300 °C, which made the analysis less accurate for the determination of Mo–Mo coordination. A significant increase in the Debye–Waller factor (indication of high disorder) in the Mo–Mo coordination at ~4 Å for unsupported Mo2C was calculated (Table 11). No structural changes to Mo2C were readily discernible by variable temperature XRD analysis over this temperature range in air (vide infra). Therefore the change in the EXAFS spectra for these two temperatures must be a result of higher frequency vibration rather than any real structural change. The populations of low Z near neighbour atoms (Mo–C & Mo–N) remain relatively unchanged as a result of the increase in temperature. Significant loss of the Mo–Mo backscattering pathway is observed for all 3 samples.

79 Chapter 3

The EXAFS analysis indicates relatively little change in the structure of the Mo2C1−xNx@TiN comparing data at 25 °C and 300 °C, with only the weighting of the Mo–Mo coordination at ~4.27 Å decreasing which is most likely derived from an increase in the disorder and lower k-range data at 300 °C. In the Mo2C@Al2O3 sample however, a significant decrease of the peak at ~2.7 Å is observed when comparing the 300 °C to the room temperature data collection.

Table 11. Details of the Mo K-edge EXAFS fitting analysis of Mo2C and Mo2C1−xNx@TiN at 25 °C and 300 °C (full fitting parameters and fitting plots are provided in the SI, standard deviations in the reported variables are indicated in brackets). Sample Path Distance (Å) σ2 (×10−3Å2s) Mo2C 3 Mo–C 2.03(1), 2.09(1), 2.13(1) 0. 1 (25 °C) 12 Mo–Mo 2.91(1)–3.00(1) 4.1 6 Mo–Mo 4.15(1)–4.19(1) 1.3 8 Mo–C–Mo 4.20(1)–4.22(1) 1.2 Mo2C 3 Mo–C 2.03(1), 2.09(1), 2.12(1) 1.1 (300 °C) 12 Mo–Mo 2.91(6)–3.01(6) 9.1 6 Mo–Mo 4.1(7)–4.1(6) 20.2 8 Mo–C–Mo 4.20(6)–4.22(6) 10.3 Mo2C1−xNx 1 Mo–N 1.66(1) 1.1 @TiN 3 Mo–C 2.12(1) 2.3 (25 °C) 6 Mo–Mo 2.92(1)–3.02(1) 5.1 3 Mo–Mo 4.25(1), 4.27(1), 4.29(1) 3.1 8 Mo–C–Mo 4.28(2)–4.30(2) 9.2 Mo2C1−xNx 1 Mo–N 1.63(4) 2.3 @TiN 3 Mo–C 2.08(2) 1.1 (300 °C) 4 Mo–Mo 2.91(2)–3.00(2) 3.1 8 Mo–C–Mo 4.30(2) 16.3

Fitting of both the data from the unsupported Mo2C and the Mo2C@Al2O3 at 25 °C indicates that the peak at ~2.7 Å corresponds to the Mo–Mo coordination at ~2.9 Å. Given that unsupported Mo2C does not react or change oxidation state appreciably at 300 °C, the diminished amplitude of this peak was used as a reference for the corresponding amplitude decrease due to thermal disorder within the supported samples.

For Mo2C@Al2O3, the loss in amplitude of the corresponding peak is greater than for the unsupported

Mo2C, potentially indicating a structural change within this sample at 300 °C or a difference in disorder at higher temperatures between the unsupported and supported materials. Furthermore, the XAS data for the Mo2C@Al2O3 was not of sufficient quality to accurately resolve these structural changes.

Analyses of the EXAFS fits of the unsupported Mo2C and Mo2C1−xNx@TiN at 25 °C and 300 °C, (Table

11. ) indicate the most labile sites within the Mo2C structure are the two single scattering Mo–Mo paths at 2.92–3.02 Å and 4.25–4.29 Å, and a multi-scattering path Mo–C–Mo at 4.28–4.30 Å that can be fitted reliably. The difference in coordination between the unsupported and TiN-supported samples reflects the degree of dispersal of the molybdenum phase by the TiN support during the synthesis.

The shorter Mo–Mo scattering and the Mo–C–Mo multi-scattering pathways are derived from the Mo- octahedra containing a central carbon atom. The longer distance Mo–Mo pathway is attributed to coordination between opposing Mo-atoms within an octahedron as illustrated by the crystal structure (Figure 33). At 300 °C, the disorder (Debye–Waller factor, σ2 in Table 11. ), increases dramatically for the longer of the two Mo–Mo scattering pathways. In the case of the Mo2C1−xNx@TiN sample, this pathway is lost entirely: most likely lost due to increased thermal disorder rather than a chemical reaction occurring. Whereas, the shorter Mo–Mo and the Mo–C–Mo scattering pathways are still present in the 300 °C data: the increase in their disorder being comparatively less.

80 Chapter 3

The above interpretation of our data indicates that the Mo–Mo coordinations at opposing octahedral vertices are the most labile (shown in yellow inFigure 33). Therefore, it is likely that upon heating, these pathways in Mo2C1−xNx@TiN reorganise and interact with the support material forming new MoN sites, and the remaining Mo2C octahedra become further integrated with the support and isolated from each other.

Figure 33. Detail of the P63/mmc crystal structure of Mo2C (ICSD 43669) used in the analysis of the EXAFS data (C sites are half occupied) for all 3 samples showing representative labile pathways in red and yellow (Mo-atoms teal; C-atoms grey).

Several other models were tested, however they are not presented here as they did not result in statistically significant analyses. Fitting a Mo≡N shell for Mo2C and Mo2C@Al2O3 was one such analysis. Although the presence of the Mo≡N motif in these samples cannot be completely discounted, as Figure 32(a) does show some residual peaks at the same position, these weren’t able to be resolved as they are not distinct. By contrast, the Mo2C1−xNx@TiN material has a well- resolved peak at 1.8 Å, relating to a bond distance known for complexes containing a Mo≡N triple bond and the rest of the spectrum is very distinct. These results indicate that the presence of TiN during synthesis increases nitrogen incorporation in the Mo2C1-xNx, specifically, as Mo≡N. The hypothesis that MoO is present in the sample could not be eliminated because the inclusion of C, N and O in the fitting leads to an excessive number of parameters for the degrees of freedom, given the available k and R range. The results of such an analysis contain a high degree of undesired correlation. The data that are presented here are the highest statistically verifiable fits for each data set and are consistent with the following XRD and XPS data.

XRD Studies To investigate the influence of the supports on the molybdenum carbide phases further, the samples were analysed by variable temperature powder X-ray diffraction (VT-XRD). The VT-XRD studies were performed in air to probe the temperature at which the samples oxidised.

Despite the evidence of nitrogen incorporation into the molybdenum-containing phase according to our

EXAFS analysis, there was no indication of a crystalline MoxNy phase in the Mo2C1−xNx@TiN sample (Figure 34). This result is to be expected as being consistent with our interpretation that the nitrogen is present in the form of Mo≡N motifs forming an amorphous or disordered interface between the TiN support and the molybdenum-containing phase.

81 Chapter 3

Figure 34. Variable Temperature XRD of Mo2C1−xNx@TiN. The red–blue colour palette represents high–low diffraction intensity. Diamonds

indicate reflections due to MoO2.

For each catalyst, the samples were loaded into 0.5 mm capillaries and heated using a gas blower. The capillaries were fixed to a spinning sample stage with wax, and so were effectively sealed at both ends after being handled in air. The Mo2C1−xNx@TiN VT-XRD (Figure 34) shows that the molybdenum carbo-nitride phase began to oxidise to MoO2 above 255 °C. The rise of peaks due to MoO2 is accompanied by a simultaneous decrease of approximately 10% in the intensity of peaks associated with Mo2C, indicating the surface of the Mo2C comprises a large fraction of the material.

For unsupported Mo2C, no change in the diffractogram was observed upon increasing temperature from 25 °C to 500 °C in air (Figure S36), confirming any apparent loss in coordination of Mo–Mo bonding in the EXAFS analysis is due to increasing disorder rather than a significant structural change. The sample’s resistance to oxidation can likely be attributed to the relatively large size of the Mo2C particles, compared to the more dispersed, supported catalysts in which the surface area to volume ratio of the molybdenum containing phases is greater.

Somewhat surprisingly, no change to the Mo2C@Al2O3 VT-XRD was observed upon heating (Figure S37), also likely due, in part, to the large particle size of the Mo2C in this sample. The presence of MoO2 from the outset suggests that it formed during catalyst synthesis, as the precursor decomposed on an oxidic support, but once formed, the catalyst is overall very stable to oxidation in air. This result is in apparent contrast with the changes observed in this sample by XAS. Given that the sample is more oxidised to begin with than the others, it seems the structural changes observed by EXAFS most likely arise due to the presence of both MoOx and Mo2C in the sample. At 300 °C, the loss of signal from Mo–Mo interactions due to thermal disorder means that the Mo–O bonding becomes more apparent even though the oxidation of molybdenum does not occur (in fact according to the XANES data, the electron density around Mo increases). Conversely for the unsupported Mo2C, very little oxidised Mo is present to begin with. Hence at 300 °C, these differences seem more apparent from fitting the EXAFS data, than by direct inspection of the PXRD data.

We reiterate that the XRD studies were performed under different conditions to those used for XAS and that the formation of oxides at high temperature seen by XRD does not necessarily indicate that these oxides are present under the reductive conditions used during XAS analysis.

82 Chapter 3

Taken together, the variable temperature X-ray diffraction studies of the sample series indicate that the choice of support drastically influences the physical nature of the Mo2C. When TiN is present, the molybdenum-containing phase forms as finely dispersed crystallites of which the surface comprises a large fraction of the sample volume and nitrogen is incorporated into the Mo-rich phase. Whereas in the presence of Al2O3, bulkier Mo2C particles result with more complete surface oxidation, of these particles.

Figure 35. XPS spectra of C, Mo, N and Ti (Counts per second vs Binding Energy) showing peak deconvolution and envelope residual (axes right; note scale change). Raw data (crosses), envelope (navy), peak deconvolution as indicated by the respective legends.

XPS Characterisation

Having performed the above analyses, it was apparent that, structurally, the unsupported Mo2C and

Mo2C@Al2O3 presented as expected. The Mo2C1−xNx@TiN sample on the other hand, required further scrutiny and in order to test our interpretation of the of the EXAFS data, this sample was further analysed by X-ray photoelectron spectroscopy (XPS) to look for the presence of a Mo–N species. Despite etching the sample with an Argon-ion beam prior to analysis, some surface oxides were still observed to be present, most likely because the nano-particulate nature of the sample resulting in oxides being found throughout the material (vide infra). The XPS spectra for the C 1s, Mo 3d, N 1s and Ti 2p regions are shown in Figure 35. The binding energies and peak information, fitted by using Thermo Scientific’s Avantage Software, are listed in Table S5 All spectra were referenced to the 1s peak of adventitious carbon, set at 284.8 eV24 and fitted using the software’s ‘Smart’ background (a modified Shirley background model). The C 1s shows typical surface carbon species (aliphatic/alcoholic/acidic carbons). The model used to fit this region was highly constrained, with the surface alcohol, aldehyde and acid contributions fixed at +1.5 (±0.1), +3 (±0.1) and +4 (±0.1) eV, respectively, relative to the aliphatic carbon.25 The full-width at half maximum (FWHM) of each peak was fixed to be the same as the aliphatic carbon main peak (1.78 eV). The Mo2C contribution was unconstrained (FWHM = 1.7 eV) and its position was consistent with previously reported analyses.26

The Mo 3d region shows the presence of multiple overlapping (3d5/2/3d3/2) doublets. These are best deconvoluted as contributions from Mo2C, MoN, MoO2 and MoO3 (Table S5). The oxides are attributed to surface contamination (as they are absent in the XRD studies). In each case, the spin–orbit splitting was fixed at Δ=3.15 eV and the height ratio of the 3d5/2/3d3/2 doublets fixed at 1:0.69.27 The FWHM of the doublets were constrained to be equal. All FWHM are less than 2 and all peaks (except the Mo(0) in Mo2C)

83 Chapter 3

were fixed to be symmetric (Mo2C tail height = 0.39% and tail exponent = 0.0821) as recommended by

Castañeda et al.28. The MoN and Mo2C are strongly overlapped as previously noted by Jauberteau et al. 29.

The N 1s region shows the presence of peaks from TiN and MoN, and is overlapped by Mo 3p3/2 contributions from the various Mo species.30 The contributions from TiN and MoN are difficult to deconvolute with certainty as the N 1s binding energies are very close.29 The differences in FWHM of the two peaks (MoN and TiN) are therefore irrelevant. The FWHM of the MoOx 3p is large (3.5 eV), indicating that multiple contributions (likely MoO2, MoO3) are being fitted by the single peak. The Ti 2p region exhibits a complex series of peaks due to TiN and TiO2 (surface impurities) as well as the plasmons and shake-up features associated with these species (Figure 35).31 Spin–orbit splitting for the 2p5/2/2p3/2 features were fixed at Δ=5.7 eV and Δ=6 eV for TiO2 and TiN respectively.32 The TiN satellite features were fixed at +2 eV higher than their respective 2p features.32 The fit is consistent with previous analyses in the literature.33 Further detail can be found in Table S5.

The XPS analysis is consistent with both the EXAFS and VT-XRD interpretations. The presence of oxide impurities adds complexity to the spectra, however all assignments are consistent, both internally and with prior reported studies of the individual components. Importantly for this study, the Mo and N regions are most accurately fitted by the inclusion of a discrete MoN species and not simply a material containing a physical mixture of Mo2C and TiN.

Microscopy and Microanalysis Energy Dispersive X-ray Spectroscopy (EDS) and Electron Energy Loss Spectroscopy (EELS) were employed to elucidate the distribution of Ti and Mo throughout the sample. EDS mapping was performed across multiple areas (Figure 36, Figure S32 through Figure S35) and showed that the elements Mo and Ti were co-localised throughout the sample, even for randomly chosen, individual particles. EELS analysis (Figure S35) was consistent with the observations from EDS: Ti and Mo are co-localised on each particle in the sample.

Taken together, all the spectroscopic data for the Mo2C1−xNx@TiN sample indicates that it is a nanoparticulate dispersion of TiN coated with crystalline mixed molybdenum carbo-nitride phase. The presence of a Mo–N triple bond from the EXAFS analysis suggests an interaction between the Mo2C and TiN at their interface, with this bond most likely acting as a capping agent against further growth of

Mo2C1−xNx nanoparticles during synthesis by thermal decomposition of the [Mo–melamine] precursor.

The 3 catalysts described and characterised above are all prepared by thermal decomposition of a single precursor complex possessing a fixed Mo:C:N ratio. We hypothesise that the observed variation in the resulting molybdenum (hemi)carbide and carbo-nitride phases is due to the influence of the support on the decomposition pathways that this precursor undergoes.

Melamine is a typical precursor material for the synthesis of carbon nitrides34 and hence it is not surprising that it is also useful as a ligand for the synthesis of N-doped transition metal carbides described here. For molybdenum however, it is not straightforward to predict whether a carbide, nitride or mixed phase will result. Pang et al. were the first to report the use of the [Mo–melamine] complex, which we have used in the current study, as a precursor to unsupported Mo2C.33 In their original work, their elemental analysis of the precursor material yielded an empirical formula of Mo19O66(C3H7N6)18·12H2O in which the melamine ligands are protonated. These authors probe and discuss the specifics of the thermal decomposition mechanism at length.

84 Chapter 3

Discussion:

Although melamine has been shown to afford Mo2C reliably, it is possible to direct the decomposition to the nitride product. In contrast to the above results of Pang et al., Lv et al. successfully heated HNO3-treated melamine and MoO3 at 500–650 °C to prepare 4 nm particulate Mo2N supported on carbon nitride.9

Other nitrogenous ligands such as hexamethylenetetramine (HMT) have also been shown to be useful precursors for the synthesis of Mo2N.

Figure 36. EDS of the Mo2C1−xNx@TiN sample showing: (top left) bright field TEM micrograph, (top right) Mo EDS map, (bottom left) Ti EDS map, (bottom right) composite of all 3 images overlaid.

The complex salt, (HMT)2(NH4)4Mo7O24·2H2O, was prepared by Afanasiev and heated under argon in the range of 550, 650, & 800 °C to yield Mo2N samples with surface areas that decreased from 150 to 70 m2/g with increasing synthesis temperature.35 If excess HMT is present during the synthesis (4–9 molar equivalents), ammonium heptamolybdate will be converted to Mo2C under H2 at 700 °C.36 Decomposition of HMT results in the generation of CH4 in situ and hence the subsequent conversion of molybdenum nitride, formed initially, to the carbide by means of exchange of N for C. Similar nitride to carbide conversion is known under standard carburizing conditions. For example, Kojima & Aika demonstrated the conversion of Mo2N (itself prepared from ammonia and MoO3 at 700 °C) to Mo2C by flowing H2 and

CH4 over the material at this temperature.37

The observation that of the 3 catalysts, only the one containing a nitride support was found to have increased nitridation of the molybdenum phase is interesting. In their XPS study of composite thin films of TiN and MoN, Sanjinés et al.27 observed the direct influence of titanium on the nitridation of molybdenum, noting that the presence of Ti increases the solubility of atomic nitrogen in Mo. A similar increase in nitridation may be in effect during the synthesis of our Mo2C1−xNx@TiN material, notwithstanding the fact that the TiN phase is present at the outset of the catalyst synthesis rather than forming in situ. It remains to be determined how general the effect is: whether the enhancement is unique to TiN, or whether other common nitride supports will also direct molybdenum carbo-nitride composites to self-assemble.

Conclusions

85 Chapter 3

We have compared 2 supported and 1 unsupported molybdenum carbide containing catalysts, all prepared from a melamine–heptamolybdate precursor. The combined spectroscopic and crystallographic evidence suggests that the support plays a significant role in determining the way the precursor decomposes as well as the properties of the resultant heterogeneous catalyst. In the presence of TiN, the precursor decomposes to a highly dispersed molybdenum carbo-nitride phase in which short Mo≡N bonds exist which are more similar to those observed in organometallic complexes than are typically observed in solid state molybdenum nitrides. The variable temperature PXRD experiments, conducted in air, also confirm that within this catalyst, the molybdenum phase is more reactive (as evidenced by oxidation) than the bulkier

Mo2C particles within the other 2 catalysts.

Given the interest in the controlled preparation of nanocomposite molybdenum carbides and nitrides for reductive processes (especially HER), it is interesting that similar strategies can be used to successfully improve catalysts for lignin conversion: requiring depolymerisation, hydrodeoxygenation, and (for Kraft lignin) hydrodesulfurization.

The two sets of processes occur under vastly different conditions (electrochemical, aqueous, ambient temperature vs. supercritical, solvothermal) however moderating the surface interaction with hydrogen through doping of the active-phase and support-interaction appears to be key for successful outcomes in both areas.

This knowledge subsequently drives improvements for treating waste biomass, including lignin, from the pulp and paper industry (amongst others) to produce higher-value platform chemicals and fuel additives: important feedstocks for the chemical industry and thus to mainstream manufacturing sectors.

86 Chapter 3

References 1. Levy, R. B.; Boudart, M., Platinum-like behavior of tungsten carbide in surface catalysis. Science 1973, 181 (4099), 547- 549. 2. Oyama, S. T.; Schlatter, J. C.; Metcalfe, J. E.; Lambert, J. M., Preparation and characterization of early transition metal carbides and nitrides. Industrial & Engineering Chemistry Research 1988, 27 (9), 1639-1648. 3. Chen, Y.-Y.; Zhang, Y.; Jiang, W.-J.; Zhang, X.; Dai, Z.; Wan, L.-J.; Hu, J.-S., Pomegranate-like N,P-Doped Mo2C@C Nanospheres as Highly Active Electrocatalysts for Alkaline Hydrogen Evolution. ACS Nano 2016, 10 (9), 8851-8860. 4. Yan, H.; Xie, Y.; Jiao, Y.; Wu, A.; Tian, C.; Zhang, X.; Wang, L.; Fu, H., Holey Reduced Graphene Oxide Coupled with an Mo2N–Mo2C Heterojunction for Efficient Hydrogen Evolution. Advanced Materials 2018, 30 (2), 1704156. 5. Konda, N. V. S. N. M.; Singh, S.; Simmons, B. A.; Klein-Marcuschamer, D., An Investigation on the Economic Feasibility of Macroalgae as a Potential Feedstock for Biorefineries. BioEnergy Research 2015, 8 (3), 1046-1056. 6. Wei, H.; Xi, Q.; Chen, X. a.; Guo, D.; Ding, F.; Yang, Z.; Wang, S.; Li, J.; Huang, S., Molybdenum Carbide Nanoparticles Coated into the Graphene Wrapping N-Doped Porous Carbon Microspheres for Highly Efficient Electrocatalytic Hydrogen Evolution Both in Acidic and Alkaline Media. Advanced Science 2018, 5 (3), 1700733. 7. Yan, J.; Kong, L.; Ji, Y.; Li, Y.; White, J.; Liu, S.; Han, X.; Lee, S.-T.; Ma, T., Air-stable phosphorus-doped molybdenum nitride for enhanced elctrocatalytic hydrogen evolution. Communications Chemistry 2018, 1 (1), 95. 8. Huang, Y.; Ge, J.; Hu, J.; Zhang, J.; Hao, J.; Wei, Y., Nitrogen-Doped Porous Molybdenum Carbide and Phosphide Hybrids on a Carbon Matrix as Highly Effective Electrocatalysts for the Hydrogen Evolution Reaction. Advanced Energy Materials 2018, 8 (6), 1701601. 9. Lv, Z.; Tahir, M.; Lang, X.; Yuan, G.; Pan, L.; Zhang, X.; Zou, J.-J., Well-dispersed molybdenum nitrides on a nitrogen- doped carbon matrix for highly efficient hydrogen evolution in alkaline media. Journal of Materials Chemistry A 2017, 5 (39), 20932-20937. 10. Rinaldi, R.; Jastrzebski, R.; Clough, M. T.; Ralph, J.; Kennema, M.; Bruijnincx, P. C. A.; Weckhuysen, B. M., Paving the Way for Lignin Valorisation: Recent Advances in Bioengineering, Biorefining and Catalysis. Angew. Chem. Int. Edn. 2016, 55 (29), 8164-8215. 11. Cattelan, L.; Yuen, A. K. L.; Lui, M. Y.; Masters, A. F.; Selva, M.; Perosa, A.; Maschmeyer, T., Renewable Aromatics from Kraft Lignin with Molybdenum-Based Catalysts. ChemCatChem 2017, 9 (14), 2717-2726. 12. Lignin hydrodeoxygenation with molybdenum carbide and carbo-nitride catalysts are part of ongoing studies within our laboratory. Full disclosure of our catalytic results are forthcoming. . 13. Ma, X.; Ma, R.; Hao, W.; Chen, M.; Yan, F.; Cui, K.; Tian, Y.; Li, Y., Common Pathways in Ethanolysis of Kraft Lignin to Platform Chemicals over Molybdenum-Based Catalysts. ACS Catal. 2015, 5 (8), 4803-4813. 14. Chen, M.; Hao, W.; Ma, R.; Ma, X.; Yang, L.; Yan, F.; Cui, K.; Chen, H.; Li, Y., Catalytic ethanolysis of Kraft lignin to small-molecular liquid products over an alumina supported molybdenum nitride catalyst. Catalysis Today 2017, 298, 9-15. 15. Li, C.; Zhao, X.; Wang, A.; Huber, G. W.; Zhang, T., Catalytic Transformation of Lignin for the Production of Chemicals and Fuels. Chem. Rev. 2015, 115 (21), 11559–11624. 16. Newville, M., Fundamentals of XAFS. Rev. Mineral. Geochem. 2014, 78 (1), 33-74. 17. Kopachevska, N.; Melnyk, A.; Bacherikova, I.; Zazhigalov, V.; Wieczorek-Ciurowa, K., Determination of molybdenum oxidation state on the mechanochemically treated MoO3. Хімія, фізика та технологія поверхні 2015, (6,№ 4), 474-480. 18. Farges, F.; Siewert, R.; Brown Jr, G. E.; Guesdon, A.; Morin, G., Structural environments around molybdenum in silicate glasses and melts. I. Influence of composition and oxygen fugacity on the local structure of molybdenum. The Canadian Mineralogist 2006, 44 (3), 731-753. 19. Chen, W. F.; Wang, C. H.; Sasaki, K.; Marinkovic, N.; Xu, W.; Muckerman, J. T.; Zhu, Y.; Adzic, R. R., Highly active and durable nanostructured molybdenum carbide electrocatalysts for hydrogen production. Energy & Environmental Science 2013, 6 (3), 943-951. 20. Liu, P.; Rodriguez, J. A.; Muckerman, J. T., Desulfurization of SO2 and Thiophene on Surfaces and Nanoparticles of Molybdenum Carbide: Unexpected Ligand and Steric Effects. The Journal of Physical Chemistry B 2004, 108 (40), 15662- 15670. 21. Epicier, T.; Dubois, J.; Esnouf, C.; Fantozzi, G.; Convert, P., Neutron powder diffraction studies of transition metal hemicarbides M2C1− x—II. In situ high temperature study on W2C1− x and Mo2C1− x. Acta Metall. 1988, 36 (8), 1903- 1921. 22. Hatnean, J. A.; Johnson, S. A., Diamagnetic molybdenum nitride complexes supported by diligating tripodal triamido- phosphine ligands as precursors to paramagnetic phosphine donors. Dalton Trans. 2015, 44 (33), 14925-14936. 23. C. Cummins, C., Reductive cleavage and related reactions leading to molybdenum–element multiple bonds: new pathways offered by three-coordinate molybdenum(III). Chem. Commun. 1998, (17), 1777-1786. 24. Moulder, J. F.; Chastain, J., Handbook of X-ray Photoelectron Spectroscopy: A Reference Book of Standard Spectra for Identification and Interpretation of XPS Data. Physical Electronics Division, Perkin-Elmer Corporation: 1992. 25. Barr, T. L.; Seal, S., Nature of the use of adventitious carbon as a binding energy standard. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1995, 13 (3), 1239-1246. 26. Ramqvist, L.; Hamrin, K.; Johansson, G.; Fahlman, A.; Nordling, C., Charge transfer in transition metal carbides and related compounds studied by ESCA. J. Phys. Chem. Solids 1969, 30 (7), 1835-1847. 27. Sanjinés, R.; Wiemer, C.; Almeida, J.; Lévy, F., Valence band photoemission study of the Ti-Mo-N system. Thin Solid Films 1996, 290-291, 334-338. 28. Castañeda, S. I.; Montero, I.; Ripalda, J. M.; Dıaz,́ N.; Galán, L.; Rueda, F., X-ray photoelectron spectroscopy study of low-temperature molybdenum oxidation process. J. Appl. Phys. 1999, 85 (12), 8415-8418. 29. Jauberteau, I.; Bessaudou, A.; Mayet, R.; Cornette, J.; Jauberteau, J.; Carles, P.; Merle-Méjean, T., Molybdenum nitride films: crystal structures, synthesis, mechanical, electrical and some other properties. Coatings 2015, 5 (4), 656-687. 30. Inzani, K.; Nematollahi, M.; Vullum-Bruer, F.; Grande, T.; Reenaas, T. W.; Selbach, S. M., Electronic properties of reduced molybdenum oxides. PCCP 2017, 19 (13), 9232-9245. 87 Chapter 3

31. Erdem, B.; Hunsicker, R. A.; Simmons, G. W.; Sudol, E. D.; Dimonie, V. L.; El-Aasser, M. S., XPS and FTIR surface characterization of TiO2 particles used in polymer encapsulation. Langmuir 2001, 17 (9), 2664-2669. 32. Jaeger, D.; Patscheider, J., A complete and self-consistent evaluation of XPS spectra of TiN. J. Electron Spectrosc. Relat. Phenom. 2012, 185 (11), 523-534. 33. Pang, M.; Wang, X.; Xia, W.; Muhler, M.; Liang, C., Mo(VI)–Melamine Hybrid As Single-Source Precursor to Pure-Phase β-Mo2C for the Selective Hydrogenation of Naphthalene to Tetralin. Ind. Eng. Chem. Res. 2013, 52 (12), 4564-4571. 34. Miller, T.; Jorge, A. B.; Suter, T.; Sella, A.; Corà, F.; McMillan, P., Carbon nitrides: synthesis and characterization of a new class of functional materials. PCCP 2017, 19 (24), 15613-15638. 35. Afanasiev, P., New single source route to the molybdenum nitride Mo2N. Inorg. Chem. 2002, 41 (21), 5317-5319. 36. Wang, Z.-Q.; Zhang, Z.-B.; Zhang, M.-H., The efficient synthesis of a molybdenum carbide catalyst via H 2-thermal treatment of a Mo (VI)–hexamethylenetetramine complex. Dalton Trans. 2011, 40 (5), 1098-1104. 37. Kojima, R.; Aika, K.-i., Molybdenum nitride and carbide catalysts for ammonia synthesis. Applied Catalysis A: General 2001, 219 (1), 141-147.

88 Chapter 4

Chapter 4.1 4-Nitrophenol Reduction: Probing the Putative Mechanism of the Model Reaction

This chapter has been published in ACS Catalysis as Strachan, J.; Barnett, C.; Masters, A. F.; Maschmeyer, T., 4-Nitrophenol Reduction: Probing the Putative Mechanism of the Model Reaction. ACS Catalysis 2020, 10 (10), 5516-5521.

Abstract We report a reinterpretation of the reduction of 4- fitting data from our own experiments. nitrophenol catalysed by silver nanoparticles. Mass spectrometry and UV-vis spectroscopy measurements support the existence of 4- nitrosophenol as a stable reaction intermediate. We propose that dissolved oxygen is consumed – both by oxidizing 4-nitrosophenol (an intermediate) and re-oxidising the reduced catalyst surface – resulting in the commonly observed ‘induction period’ in the reaction kinetics. Upon complete consumption of dissolved oxygen, subsequent reduction to 4- aminophenol can occur. A complete kinetic analysis including modelling is presented, conceptually fitting data from recent reports in the literature as well as

Introduction The reduction of 4-nitrophenol is a commonly employed model reaction used to test the activity of relevant catalysts. While nitroarene reduction has been known for over 150 years,1 the use of the reaction to compare catalyst activity was first suggested by Pal and co-workers in 2001.2 Since then, the reaction has been shown to be catalysed by a wide variety of materials: nanoparticulate metals, supported-metals as conventional and photocatalysts and even metal-free powders (such as N-doped graphene).3-6 The ubiquitous use of the reaction7 can be attributed to a number of key elements: (i) the reaction does not take place in the absence of catalyst, (ii) the reaction takes place at ambient temperature/pressure and in aqueous solutions, and (iii) the reaction can be quantitatively monitored using readily available instrumentation (UV-vis spectrometers).

The reduction of 4-nitrophenol (4NP) to 4-aminophenol (4AP, Figure 1) is typically performed in an aqueous solution using sodium borohydride (NaBH4) as a hydride source and a metal catalyst as an electron transfer agent. The formation of 4AP was first supported by UV-vis spectroscopy,2 and the product later isolated and its identity confirmed by NMR, IR and melting point analysis.8 Sodium borohydride undergoes hydrolysis in − solution,9 raising the pH and deprotonating the 4NP to form the 4-nitrophenolate anion (4NP , λ 400 nm)8 – this ion is strongly coloured and can be monitored spectrophotometrically throughout the reaction. Though thermodynamically favourable, the reaction (Eo = 0.76 V)5 does not take place in the absence of a catalyst, − − because the negatively charged 4NP and BH4 are mutually repelling.10 Sodium borohydride, ascorbic acid,

89 Chapter 4 hydrazine, hydrogen gas,8 tetramethyldisiloxane11 and formic acid8 have been screened as reductants, however, − of those candidates, only NaBH4, tetramethyldisiloxane and formic acid successfully reduced 4NP ; consistent 12- with hydride transfer occurring via coordination of BH4¯ (rather than dissolved H2) with the catalyst surface. 13 Other nitroarenes11, 14-15 have been tested (including the 4/3/2-NP series),16 though 4NP is rapidly reduced and, therefore, widely studied.3, 5, 8 Quantitative monitoring of the product 4AP− is rarely reported, likely because

of signal overlap with the catalyst and a significantly smaller molar extinction coefficient than 4NP¯.2, 17-20 Concentration

A B C

Time Figure 37 - Representative kinetic profile of the catalytic reduction of 4-nitrophenol.

Although the reaction is used extensively to compare catalysts, the exact mechanism is contentious.18, 21 The key differences between the competing theories are the interpretation of the induction period (Figure 37, A) in combination with the reaction mechanism. Pal originally ascribed the induction period to a re-reduction of the nanoparticle surface, having noted that the induction period disappears when the reaction was conducted under a nitrogen environment.8 Subsequent studies suggested similar explanations,22-23 however this hypothesis has been refuted within the literature.18, 20 There are other authors who attribute the induction period to the − − diffusion of reactants (namely 4NP and BH4 ) onto the catalyst surface.24-29 However, analysis by Ballauff and co-workers showed that the rate of reactant diffusion is two orders of magnitude higher than the rate of reaction, inconsistent with the diffusion hypothesis.30 These authors proposed an alternative explanation based on surface restructuring of the catalyst and supported this hypothesis with extensive modelling using a Langmuir-Hinshelwood model.19, 30 More recently, Neretina and co-workers have provided compelling evidence that the induction period is related to the depletion of dissolved oxygen in solution.18 This conclusion was reached by observing that a “critical value” of dissolved oxygen coincides with the end of the induction period for several noble metal catalysts.

The hypothesis proposed by Neretina and co-workers involves competing forward and backward reactions between 4NP− and 4AP−, with the backward reaction - oxidation of 4AP− to 4NP− - dominating until the dissolved oxygen reaches a “critical value.”18 The mechanism does not consider intermediates, side reactions

90 Chapter 4 or by-products, which are to be expected based on established nitroarene chemistry.31-33 Specifically, 4AP is widely used as an intermediate in industrial synthesis and has known environmental toxicity.5 Its oxidation - chemically, electrochemically,34-35 and enzymatically34, 36 - to p-benzoquinone and related dyes is known, however, reports of oxidation to the nitro- analogue are rare.37

Given the ubiquity of this reaction and the conflicting theories within the literature, we aim to provide a unifying description of the reduction of 4NP using NaBH4 on nanoparticulate metal catalysts. We report a kinetic model of the reaction, which accounts for the induction period, intermediates, dissolved oxygen and by-products as well as evidence for the key reactive intermediate.

Results and Discussion Plasmonic silver nanoparticles (AgNPs) were chosen so that changes to the catalyst could be monitored throughout the reaction. The plasmon resonance of a nanoparticle is dependent on the shape, size, and importantly, any species bound to the surface.38 When a molecule binds to the surface of a nanoparticle, it can alter the resonant plasmon frequency.39-40 This phenomenon provided a simple method for observing the ‘surface restructuring’ mechanism put forth by Ballauff and co-workers. 18-21, 30 In our system, the nanoparticles were observed before and after catalysis by monitoring the shape and position of the plasmon by UV-vis spectroscopy. Whilst a significant plasmon shift was observed, the AgNP plasmon was found to be strongly influenced by organic species. This method was therefore unable to differentiate between ‘surface restructuring’ and 4NP/4AP adsorption (see Figure S 38 in SI for further discussion). Neretina and co-workers showed that the end of the induction period corresponds with the time at which the dissolved oxygen content drops below a critical threshold – a value dependent on the catalyst.18 Importantly, they noted that to keep the oxygen content below the detection limit of their sensor (0.03 mg·L-1) required all exposed liquid surfaces to be kept under argon. Ambient conditions resulted in dissolved oxygen content well above their observed critical values. This finding suggests that conclusions drawn from experiments in which the solutions were merely sparged beforehand could be erroneously excluding the possibility of dissolved oxygen. When the initial oxygen concentration was below the critical value, there was no induction period.18 They also reported that the consumption of dissolved oxygen was catalytically enhanced and that the rate of hydrogen gas evolution was sufficiently slow so as to preclude self-sparging as an explanation for the loss of oxygen. This suggests that the dissolved oxygen is primarily consumed rather than purged.18

We repeated one of their experiments - the mid-catalysis addition of O2 gas test - to determine whether the oxidation was generalizable to our catalyst. As shown in Figure 38, there is an additional ‘induction period’ post oxygen administration, consistent with Neretina’s report.18 This O2-induced induction period was also observed with nanoparticle catalysts prepared from Ag, Au, S Cu, Ni, Pd, and Pt (Figure S 42 through Figure

S 44) or if an oxygen sparged aliquot of water was used in placed of O2 bubbles (Figure S 45).

91 Chapter 4

1.0 Control 0.8 O2 bubbled 0.6

0.4

0.2 Absorbance (arb.) Absorbance

0.0 0 100 200 300 400 Time (s) Figure 38- Mid-catalysis oxygen sparge on 4NP reduction with AgNP solution. No gas was administered to the control. Arrow indicates time of

O2(g) addition.

To check whether the O2-induced induction period could be explained simply by the presence of gas bubbles in the cuvette, an inert gas (Ar) was added in place of oxygen. A second induction period was not observed, suggesting that it is the addition of oxygen that induces the back-reaction to 4NP− (Figure S 45).

To test Neretina’s assertion that 4AP− is oxidised to 4NP−,18 we subjected an authentic sample of 4AP to the same conditions as during the mid-catalysis oxygen gas addition. However, neither UV-vis nor APCI-mass spectroscopies exhibited evidence of 4NP (Figure S 51). Attempts to convert 4AP− to 4NP− using various conditions and oxygen sources were unsuccessful (see SI for details).

As an authentic sample of 4AP could not be catalytically oxidised to 4NP−, we sought to confirm that the product of the reduction of 4NP− is indeed 4AP−. When the reaction was first reported, the product’s identity was confirmed using UV-vis spectroscopy alone.2 A subsequent paper reported the isolation of 4AP in 73% yield and subjected the material to the standard analytical techniques to confirm the presence of the product.8 Pal and co-workers also noted that the resultant solution turned black within two hours of completion due to oxidation of 4AP− in the alkaline catalysis solution.8 This observation is consistent with that of Lerner.37

To further probe the potential formation of 4AP− by reduction of 4NP−, a catalytic solution was assayed with APCI-MS. There were diagnostic signals at 110.1 and 108.1 m/z assigned to 4AP+/− respectively, confirming that the product of the catalysis (using slightly different nanoparticles to those used by Pal) is indeed 4AP− (Figure S 52). These signals were absent in the mass spectrum of a completed reaction mixture after 24 hours (Figure S 51). When the fresh solution was analysed using GC-MS, no signal attributable to 4AP was present. These findings suggest that 4AP is unstable and decomposes during the GC-MS analysis and in basic solutions.34-37

Aqueous 4AP is known to decompose to quinone derivatives.34-35 Notably, hydro- and benzoquinone have a strong UV-vis spectroscopic absorbance at λ ~300 nm under equivalent conditions to those used during catalysis (Figure 39). Therefore, the assumption that the UV-vis signal at λ 300 nm, which is present during catalysis, is exclusively due to 4AP− is not valid; a consideration not commonly acknowledged.18-21, 41-42 Additionally, there are inconsistencies in the literature concerning the UV-vis signal at λ 300 nm; often this signal intensity increases neither consistently nor proportionally as would be expected by the reaction

92 Chapter 4 stoichiometry.18 In Neretina and co-worker’s 2016 study,18 the decrease in the UV-vis signal intensity at λ 300 − nm concomitant with the administration of a 5 second purge of O2 gas indicates that 4AP is consumed, but does not necessarily point to 4AP− being converted to 4NP−. Rather, 4AP− could be lost by conversion to other oxidised species.34-37

4NO- 1.0 4NP- 4AP-

0.8 BQ- HQ-

0.6

0.4 Absorbance (normalised) Absorbance

0.2

0.0 200 250 300 350 400 450 500 Wavelength (nm) Figure 39 - UV-vis spectra of 4-nitrosophenol (black), 4-nitrophenol (red), 4-aminophenol (blue), benzoquinone (green) and hydroquinone (purple) under equivalent catalysis conditions (Chromophore – 4.8 × 10-5 M, NaOH – 1.0 × 10-2 M).

Several mechanisms have been proposed for the reduction of 4NP− to 4AP− (and the equivalent nitrobenzene to aminobenzene).16, 21, 31-32, 43 Two commonly proposed intermediates are the nitroso- and hydroxylamino- derivatives (Figure 41).21 Since 4AP− was not successfully oxidised to 4NP−, vide supra, oxidations of these intermediates were considered.

21 18, 20 Figure 40 – 4NP reduction mechanism comparison. Direct Route, 4AP Side Reaction, and the mechanisms proposed in this work. “H2” indicates a H–/H+ pair on the catalyst surface.

93 Chapter 4

Analysis of a reaction mixture by APCI-MS (Figure S 51) indicated that the major intermediate was 4- nitrosophenol (unlike previous GC-MS studies,42 the hydroxylaminophenol was not detected in our analysis). Therefore, a commercial sample of 4-nitrosophenol (4NO) was subjected to the equivalent conditions as in the oxygen bubbling test (Figure S 54) except that NaBH4 was replaced with NaOH to exclude the possibility of reduction. No APCI-MS signal attributable to 4NP was present before O2 was introduced into a control solution of 4NO− (Figure S 48). Addition of catalyst resulted in the appearance of mass spectroscopic signals − at +140.0 and −138.0 m/z attributed to 4NPH+/4NP (Figure S 53); after O2 administration, the intensities of these signals significantly increased (Figure S 54). UV-vis spectroscopy could not be used to determine whether 4NO− was oxidised to 4NP as the spectra of 4NO− and 4NP− are nearly indistinguishable (Figure S 39 and Figure 39). Although the two spectra differ in the 220 – 300 nm region, the production of quinones (which also absorb strongly in this region) could not be excluded.

Therefore, we suggest that during the catalysis, there is a build-up of 4NO− which can either be further reduced − − − − to 4AP , or, in the presence of O2, be oxidised to reform 4NP . It is 4NO - not 4AP - that can be oxidised to 4NP (Figure 40). Oxidation of 4AP− does not reform 4NP− and instead forms oligomers and/or quinone derivatives, some of which have absorbances that overlap the spectrum of 4AP− in the UV-vis region. However, since the method commonly used in the literature to follow the reaction monitors only the consumption of 4NP−, this reactivity has not been previously identified.

Our hypothesis can be reconciled with Neretina’s results18 by reinterpreting their proposed oxidation of 4AP− as an oxidation of 4NO− and changes to the UV-vis signal intensity at λ 300 nm as further oxidation of 4AP− to quinone derivatives or changes to their catalyst. Similarly, Ballauff’s kinetic analysis21 can be reinterpreted with 4NO replacing 4-hydroxyaminophenol as the dominant intermediate of their ‘stationary state’ analysis. As the 4-hydroxylaminophenol was not observed by APCI-MS, we propose that it is rapidly and irreversibly reduced to 4AP.

Another mechanism that could give rise to the observed induction period was considered: reduction and oxidation of the catalyst surface by hydride and dissolved oxygen until sufficient depletion of dissolved oxygen (Figure 40). To test this hypothesis, the order of addition of reagents was reversed (Figure 41). By allowing the catalyst and NaBH4 to react with dissolved oxygen before addition of the substrate, the induction time was expected to decrease.

0.06 46 s 15 s 0.06

Water added 0.05 0.04 Control 31 s Substrate added 0 15 30 45 60 75

0.02 Concentration (mM)

0.00 0 30 60 90 120 150 180 210 240 Time (s) Figure 41. Substrate concentration vs time. Each trace shows a different order of addition of reagents. Volume of water and substrate added = 100.0 µL.

94 Chapter 4

When the substrate was added last, at a given time after the induction period of the control had ended, the subsequent induction time substantially decreased (Figure 41, red). To establish the effect that the dissolved oxygen in the substrate aliquot has on the induction period, an equivalent aliquot of water was added to another sample. The addition of this aliquot to the catalysis solution increased the induction time (Figure 41, blue). The increase in induction time was equal to that observed in the experiment when the substrate was added last, indicating that the induction period observed in the ‘substrate added’ test is due to dissolved oxygen and oxygen introduced during aliquot addition. This suggests that catalyst and hydride can react with dissolved oxygen (in the absence of substrate). However, as APCI showed that 4NO can be oxidised to 4NP under our catalysis conditions, we propose that both mechanisms occur together to produce the induction period, with the ‘catalyst surface’ pathway being the dominant contribution.

To test this hypothesis further, we modelled the reaction using DynaFit. As modelling both pathways simultaneously would lead to correlated elementary steps, the mechanisms were modelled individually. In light of Neretina and co-worker’s work, we assumed non-catalytic O2 consumption to be negligible relative to the − rate of O2 consumption by the catalytic reaction with the 4NO so as to keep the number of variables to a minimum.18, 20 The full reaction scheme is given in Figure 40, the elementary steps modelled by Dynafit in Scheme 1 and the modelled data in Figure 6. Alternative mechanisms are presented in Table S 9 and Figure S 41.

Step 1: PreCatalyst + 4NP− → Cat-4NP − − Step 2: 4NO + O2 → 4NP Step 3: 1/2O2 + “H2” → H2O − Step 4: Cat-4NP + “H2” → 4NO + PreCatalyst + H2O

Scheme 1 – Elementary steps that were used to model the ‘4NO oxidation mechanism’ with Dynafit.

Step 1: DeactivatedCatalyst + H- → PreCatalyst Step 2: O2 + PreCatalyst → DeactivatedCatalyst Step 3: PreCatalyst + 4NP− → Cat-4NP − Step 4: Cat-4NP + “H2” → 4NO + PreCatalyst + H2O

Scheme 2– Elementary steps that were used to model the ‘Catalyst Reoxidation Mechanism’ with Dynafit.

Step 1 in Scheme 1 represents the adsorption of the substrate onto the catalyst, analogous to “stage I” in Ballauff’s analysis.21 It also contains the unbound 4NP¯ species, thus is the ‘reporter reaction’. Step 2 is the reduction of the adsorbed 4NP− to form the nitroso- intermediate 4NO−. Step 3 is the ‘backwards’ reaction of − − − − the 4NO with dissolved O2(aq) to reform 4NP . Step 4 is the further reduction of 4NO to form 4AP , the product. The rate constants for the elementary steps are given in Table S 7. Other mechanisms were considered, however, none was appropriate for our system (see SI for further details).

Step 1 in Scheme 2 represents the reduction of the deactivated (oxidised) catalyst surface. Step 2 is the reoxidation of the reduced surface by dissolved oxygen. Step 3 is the adsorption of the substrate onto the catalyst (identical to Step 1, Scheme 1) and Step 4 is the reduction of the substrate.

95 Chapter 4

0.05 Data Scheme 1 Model Scheme 2 Model 0.04

0.03

0.02 Concentration (mM) Concentration 0.01

0.00 0 60 120 180 240 Time (s) Figure 42 – Data (black dots), Scheme 1 Model (red) and Scheme 2 Model (blue) of the catalytic 4NP− reduction fit by Dynafit.

Upon comparison of our data to those in the literature, it was noticed that a pseudo-first order profile was not common to all studies.8, 44-45 Further investigation revealed that the [NaBH4] must be in sufficient excess with respect to [4NP] (greater than a 200 fold excess in our system) for the reaction to proceed in a pseudo-first order regime (Figure S 46). If this condition is not met, the reaction will appear zeroth order in [4NP−]. This finding may be universal, however, attempts to analyse literature data were compromised by obscured or unreported concentrations and conditions – a practice recently highlighted by Baker and coworkers.46

Conclusions We have provided clear evidence that the reduction of 4-nitrophenol to 4-aminophenol involves 4- nitrosophenol as an intermediate that may be oxidised by dissolved oxygen to reform 4-nitrophenol. In addition, catalyst reduction/reoxidation was shown to be a potential mechanism for consuming dissolved oxygen. These processes inhibit the cumulative consumption of 4-nitrophenol before the dissolved oxygen reaches a ‘critical value’18 and result in an induction period in the kinetic profile. Simplified mechanisms were modelled to excellent agreement with experimental data (relative r.m.s <2%). As a full kinetic understanding is required of model reactions,3 this work further establishes the reduction of 4-nitrophenol as a reliable and accepted model reaction in contemporary catalysis and provides an advance in the understanding of the details of the reaction.

96 Chapter 4

References 1. Zinin, N., Beschreibung einiger neuer organischer Basen, dargestellt durch die Einwirkung des Schwefelwasserstoffes auf Verbindungen der Kohlenwasserstoffe mit Untersalpetersäure. Journal für Praktische Chemie 1842, 27 (1), 140-153. 2. Pradhan, N.; Pal, A.; Pal, T., Catalytic reduction of aromatic nitro compounds by coinage metal nanoparticles. Langmuir 2001, 17 (5), 1800-1802. 3. Hervés, P.; Pérez-Lorenzo, M.; Liz-Marzán, L. M.; Dzubiella, J.; Lu, Y.; Ballauff, M., Catalysis by metallic nanoparticles in aqueous solution: model reactions. Chem. Soc. Rev. 2012, 41 (17), 5577-5587. 4. Zhao, P.; Feng, X.; Huang, D.; Yang, G.; Astruc, D., Basic concepts and recent advances in nitrophenol reduction by gold- and other transition metal nanoparticles. Coord. Chem. Rev. 2015, 287, 114-136. 5. Aditya, T.; Pal, A.; Pal, T., Nitroarene reduction: a trusted model reaction to test nanoparticle catalysts. Chem. Commun. 2015, 51 (46), 9410-9431. 6. Kong, X.-k.; Sun, Z.-y.; Chen, M.; Chen, Q.-w., Metal-free catalytic reduction of 4-nitrophenol to 4-aminophenol by N-doped graphene. Energy & Environmental Science 2013, 6 (11), 3260-3266. 7. A search of the concept “catalytic nitrophenol reduction” on Scifinder yielded 4130 hits, over the period, 1933-2020. 8. Pradhan, N.; Pal, A.; Pal, T., Silver nanoparticle catalyzed reduction of aromatic nitro compounds. Colloids and Surfaces A: Physicochemical and Engineering Aspects 2002, 196 (2-3), 247-257. 9. Lo, C.-t. F.; Karan, K.; Davis, B. R., Kinetic studies of reaction between sodium borohydride and methanol, water, and their mixtures. Ind. Eng. Chem. Res. 2007, 46 (17), 5478-5484. 10. Li, M.; Chen, G., Revisiting catalytic model reaction p-nitrophenol/NaBH 4 using metallic nanoparticles coated on polymeric spheres. Nanoscale 2013, 5 (23), 11919-11927. 11. Fountoulaki, S.; Daikopoulou, V.; Gkizis, P. L.; Tamiolakis, I.; Armatas, G. S.; Lykakis, I. N., Mechanistic studies of the reduction of nitroarenes by NaBH4 or hydrosilanes catalyzed by supported gold nanoparticles. ACS Catalysis 2014, 4 (10), 3504-3511. 12. Escano, M. C. S.; Gyenge, E.; Arevalo, R. L.; Kasai, H., Reactivity descriptors for borohydride interaction with metal surfaces. The Journal of Physical Chemistry C 2011, 115 (40), 19883-19889. 13. Guella, G.; Zanchetta, C.; Patton, B.; Miotello, A., New insights on the mechanism of palladium-catalyzed hydrolysis of sodium borohydride from 11B NMR measurements. The Journal of Physical Chemistry B 2006, 110 (34), 17024-17033. 14. Choi, Y.; Bae, H. S.; Seo, E.; Jang, S.; Park, K. H.; Kim, B.-S., Hybrid gold nanoparticle-reduced graphene oxide nanosheets as active catalysts for highly efficient reduction of nitroarenes. J. Mater. Chem. 2011, 21 (39), 15431-15436. 15. Akça, A.; Genç, A. E.; Kutlu, B., BH4 dissociation on various metal (1 1 1) surfaces: A DFT study. Appl. Surf. Sci. 2019, 473, 681-692. 16. Zhang, W.; Xiao, X.; An, T.; Song, Z.; Fu, J.; Sheng, G.; Cui, M., Kinetics, degradation pathway and reaction mechanism of advanced oxidation of 4‐nitrophenol in water by a UV/H2O2 process. Journal of Chemical Technology & Biotechnology: International Research in Process, Environmental & Clean Technology 2003, 78 (7), 788-794. 17. He, X.; Liu, Z.; Fan, F.; Qiang, S.; Cheng, L.; Yang, W., Poly (ionic liquids) hollow nanospheres with PDMAEMA as joint support of highly dispersed gold nanoparticles for thermally adjustable catalysis. J. Nanopart. Res. 2015, 17 (2), 74. 18. Menumerov, E.; Hughes, R. A.; Neretina, S., Catalytic reduction of 4-nitrophenol: a quantitative assessment of the role of dissolved oxygen in determining the induction time. Nano Lett. 2016, 16 (12), 7791-7797. 19. Wunder, S.; Polzer, F.; Lu, Y.; Mei, Y.; Ballauff, M., Kinetic analysis of catalytic reduction of 4-nitrophenol by metallic nanoparticles immobilised in spherical polyelectrolyte brushes. The Journal of Physical Chemistry C 2010, 114 (19), 8814-8820. 20. Neal, R. D.; Inoue, Y.; Hughes, R. A.; Neretina, S., Catalytic Reduction of 4-Nitrophenol by Gold Catalysts: The Influence of Borohydride Concentration on the Induction Time. The Journal of Physical Chemistry C 2019, 123 (20), 12894-12901. 21. Gu, S.; Wunder, S.; Lu, Y.; Ballauff, M.; Fenger, R.; Rademann, K.; Jaquet, B.; Zaccone, A., Kinetic analysis of the catalytic reduction of 4-nitrophenol by metallic nanoparticles. The Journal of Physical Chemistry C 2014, 118 (32), 18618-18625. 22. Zhang, J. T.; Wei, G.; Keller, T. F.; Gallagher, H.; Stötzel, C.; Müller, F. A.; Gottschaldt, M.; Schubert, U. S.; Jandt, K. D., Responsive hybrid polymeric/metallic nanoparticles for catalytic applications. Macromol. Mater. Eng. 2010, 295 (11), 1049- 1057. 23. Sarkar, S.; Sinha, A. K.; Pradhan, M.; Basu, M.; Negishi, Y.; Pal, T., Redox transmetalation of prickly nickel nanowires for morphology controlled hierarchical synthesis of nickel/gold nanostructures for enhanced catalytic activity and SERS responsive functional material. The Journal of Physical Chemistry C 2010, 115 (5), 1659-1673. 24. Khalavka, Y.; Becker, J.; Sönnichsen, C., Synthesis of Rod-Shaped Gold Nanorattles with Improved Plasmon Sensitivity and Catalytic Activity. J. Am. Chem. Soc. 2009, 131 (5), 1871-1875. 25. Mahmoud, M.; El-Sayed, M., Time dependence and signs of the shift of the surface plasmon resonance frequency in nanocages elucidate the nanocatalysis mechanism in hollow nanoparticles. Nano Lett. 2011, 11 (3), 946-953. 26. Wu, K.-L.; Wei, X.-W.; Zhou, X.-M.; Wu, D.-H.; Liu, X.-W.; Ye, Y.; Wang, Q., NiCo2 alloys: controllable synthesis, magnetic properties, and catalytic applications in reduction of 4-nitrophenol. The Journal of Physical Chemistry C 2011, 115 (33), 16268- 16274. 27. Signori, A. M.; Santos, K. d. O.; Eising, R.; Albuquerque, B. L.; Giacomelli, F. C.; Domingos, J. B., Formation of Catalytic Silver Nanoparticles Supported on Branched Polyethyleneimine Derivatives. Langmuir 2010, 26 (22), 17772-17779. 28. Zeng, J.; Zhang, Q.; Chen, J.; Xia, Y., A comparison study of the catalytic properties of Au-based nanocages, nanoboxes, and nanoparticles. Nano Lett. 2009, 10 (1), 30-35. 29. Kuroda, K.; Ishida, T.; Haruta, M., Reduction of 4-nitrophenol to 4-aminophenol over Au nanoparticles deposited on PMMA. J. Mol. Catal. A: Chem. 2009, 298 (1), 7-11.

97 Chapter 4

30. Wunder, S.; Lu, Y.; Albrecht, M.; Ballauff, M., Catalytic activity of faceted gold nanoparticles studied by a model reaction: evidence for substrate-induced surface restructuring. Acs Catalysis 2011, 1 (8), 908-916. 31. Song, J.; Huang, Z.-F.; Pan, L.; Li, K.; Zhang, X.; Wang, L.; Zou, J.-J., Review on selective hydrogenation of nitroarene by catalytic, photocatalytic and electrocatalytic reactions. Applied Catalysis B: Environmental 2018, 227, 386-408. 32. Khan, F. A.; Dash, J.; Sudheer, C.; Gupta, R. K., Chemoselective reduction of aromatic nitro and azo compounds in ionic liquids using zinc and ammonium salts. Tetrahedron Lett. 2003, 44 (42), 7783-7787. 33. Nie, R.; Wang, J.; Wang, L.; Qin, Y.; Chen, P.; Hou, Z., Platinum supported on reduced graphene oxide as a catalyst for hydrogenation of nitroarenes. Carbon 2012, 50 (2), 586-596. 34. Hawley, D.; Adams, R. N., Homogeneous chemical reactions in electrode processes: Measurement of rates of follow-up chemical reactions. Journal of Electroanalytical Chemistry (1959) 1965, 10 (5-6), 376-386. 35. Ahmed, J.; Rakib, R. H.; Rahman, M. M.; Asiri, A. M.; Siddiquey, I. A.; Islam, S. S.; Hasnat, M. A., Electrocatalytic Oxidation of 4‐Aminophenol Molecules at the Surface of an FeS2/Carbon Nanotube Modified Glassy Carbon Electrode in Aqueous Medium. ChemPlusChem 2019, 84 (2), 175-182. 36. Josephy, P.; Eling, T. E.; Mason, R. P., Oxidation of p-aminophenol catalyzed by horseradish peroxidase and prostaglandin synthase. Mol. Pharmacol. 1983, 23 (2), 461-466. 37. Lerner, L., Identity of a purple dye formed by peroxidic oxidation of p-aminophenol at low pH. The Journal of Physical Chemistry A 2011, 115 (35), 9901-9910. 38. Kelly, K. L.; Coronado, E.; Zhao, L. L.; Schatz, G. C., The optical properties of metal nanoparticles: the influence of size, shape, and dielectric environment. The Journal of Physical Chemistry B 2003, 107 (3), 668-677. 39. Rycenga, M.; Cobley, C. M.; Zeng, J.; Li, W.; Moran, C. H.; Zhang, Q.; Qin, D.; Xia, Y., Controlling the synthesis and assembly of silver nanostructures for plasmonic applications. Chem. Rev. 2011, 111 (6), 3669-3712. 40. Willets, K. A.; Van Duyne, R. P., Localized surface plasmon resonance spectroscopy and sensing. Annu. Rev. Phys. Chem. 2007, 58, 267-297. 41. Raghuwanshi, V. S.; Wendt, R.; O’Neill, M.; Ochmann, M.; Som, T.; Fenger, R.; Mohrmann, M.; Hoell, A.; Rademann, K., Bringing Catalysis with Gold Nanoparticles in Green Solvents to Graduate Level Students. J. Chem. Educ. 2017, 94 (4), 510- 514. 42. Drinkel, E. E.; Campedelli, R. R.; Manfredi, A. M.; Fiedler, H. D.; Nome, F., Zwitterionic-surfactant-stabilised palladium nanoparticles as catalysts in the hydrogen transfer reductive amination of benzaldehydes. The Journal of organic chemistry 2014, 79 (6), 2574-2579. 43. Datta, K. J.; Rathi, A. K.; Gawande, M. B.; Ranc, V.; Zoppellaro, G.; Varma, R. S.; Zboril, R., Base‐Free Transfer Hydrogenation of Nitroarenes Catalyzed by Micro‐Mesoporous Iron Oxide. ChemCatChem 2016, 8 (14), 2351-2355. 44. Gao, S.; Zhang, Z.; Liu, K.; Dong, B., Direct evidence of plasmonic enhancement on catalytic reduction of 4-nitrophenol over silver nanoparticles supported on flexible fibrous networks. Applied Catalysis B: Environmental 2016, 188, 245-252. 45. Alzahrani, H. A.; Buckingham, M. A.; Wardley, W. P.; Tilley, R. D.; Ariotti, N.; Aldous, L., Gold nanoparticles immobilised in a superabsorbent hydrogel matrix: facile synthesis and application for the catalytic reduction of toxic compounds. Chem. Commun. 2020, 56 (8), 1263-1266. 46. Larm, N. E.; Bhawawet, N.; Thon, J. A.; Baker, G. A., Best practices for reporting nanocatalytic performance: lessons learned from nitroarene reduction as a model reaction. New J. Chem. 2019, 43 (46), 17932-1

98 Chapter 4

Chapter 4.2 Nanoparticles for Undergraduates: Creation, Characterization, and Catalysis

This chapter has been published in The Journal of Chemical Education as Strachan, J.; Barnett, C.; Maschmeyer, T.; Masters, A. F.; Motion, A.; Yuen, A. K. L., Nanoparticles for Undergraduates: Creation, Characterization, and Catalysis. Journal of Chemical Education 2020, 97 (11), 4166-4172.

Abstract We report a research-enhanced undergraduate nanoscience, catalysis, green chemistry as well as a laboratory practical in which students synthesise range of relevant instrumentation and techniques. and characterise colour-controllable, stable silver nanoparticles for use in a simple catalytic reaction analysed by UV-vis spectroscopy. The practical has been prepared for students using electronic laboratory notebooks, enables collaborative data sharing, and can be altered to suit time, experience and instrumentation constraints. This experiment is conducted as part of our Talented Student Program and introduces students to the themes of

Introduction Nanoparticles In just 150 years, nanoscience has grown from conjecture to commonplace. Its application to technology is now a fundamental part of modern society and is exploited by diverse fields including medicine, electronics, agriculture and defense.1-3 Catalysis is a preeminent nanotechnology; industrial catalysis represents a field where nanomaterials have been applied commercially for more than a century.4 To prepare graduates for the modern workforce, universities have been integrating nanoscience into their undergraduate syllabi.5 Within this context, this journal has published several laboratory experiments that introduce students to core concepts of nanoscience such as scale, size dependent properties, quantum properties, and surface behaviour.6-8 These experiments typically involve the synthesis of colloidal nanoparticles and their subsequent use in a catalytic reduction.9-10

They are designed to draw together multiple concepts that students may view as disparate and in doing so facilitate meaningful learning.11 Unfortunately, these experiments are not suitable for our undergraduate laboratory program given the prohibitive capital outlay of several sputter coaters/vacuum pumps,9 the cost of dendrimer stabilizing agents,10, 12 and the assumed practical skills of the students in our 1st year undergraduate laboratory program. We therefore sought to modify these experiments to design a practical exercise that was suitable for large cohorts, shorter laboratory sessions, and more limited instrumentation whilst still retaining the objective of the experiments designed by Rademann9 and Stevenson.10 The specific constraints we faced were:

- nine hours of laboratory time split into 3 × 3 hour laboratory sessions - a class of approximately 65 students with access to 10 UV-vis spectrometers - a student to demonstrator ratio of 12:1

99 Chapter 4

The practical as presented was designed for our Talented Students Program. We also sought to develop a series of practical experiments as part of a project-based, research-enhanced, laboratory course. The Faculty’s Talented Students Program captures approximately 25% of the State’s top 1.5% of secondary school science students in their first semester of their Bachelor degree coming from a variety of educational backgrounds. The course aims to encourage critical thinking and, through the use of open electronic laboratory notebooks, students are encouraged to share data within their assigned groups and learn skills of collaboration in the laboratory. Additionally, we have integrated authentic assessments as part of this laboratory course; students produce both a research-poster and a video for a high school audience at the end of semester, vide infra. Academics from across our faculty are invited to attend the ‘Film Festival’ and Poster Presentation, and the short films are posted publicly on YouTube.

Our nanoparticle synthesis was modified from a report by Kitaev and co-workers7 which partially addressed some of our specific constraints. The report details the preparation of a stable nanoparticle colloid by the reduction of silver nitrate with sodium borohydride in the presence of citrate and bromide ions as surface capping agents. This bottom-up synthesis is consistent with the principles of green chemistry and is simple, reproducible, quick, and safe.7 The colloids are brightly coloured due to the plasmon resonance of the particles and the colour can be tuned simply by altering the bromide to silver (Br:Ag) ratio during synthesis. Depending on the time and instrumentation constraints of the institution, characterization of the colloidal nanoparticles can range from a simple observation of the Tyndall effect or analysis by UV-vis spectroscopy, through to DLS/XRD for smaller cohorts and then to TEM/AFM for advanced students (example data is reported herein). We also report a number of practical adjustments from Kitaev’s7 preparation to aid the students and instructors. A benefit of this practical is that students can be invited to take their samples to selected research laboratories for analysis, and so get early exposure to the research enterprise of our School of Chemistry.

Catalysis The catalysis component of the practical involves evaluating the activities of different nanoparticle colloids in the reduction of 4-nitrophenol by sodium borohydride, monitored using time-resolved UV-vis spectroscopy as first reported by Pal.13 This reaction is simple; it uses aqueous solutions, ambient conditions, cheap chemicals, and is monitored using readily available instrumentation (UV-vis spectrophotometers and disposable cuvettes).14 The reactant is stable and the reaction will not proceed in the absence of catalyst.15 Conditions have been chosen to facilitate reaction completion within 10 minutes while still providing enough data for meaningful kinetic analyses. As elaborated in our instructor notes, the full reaction can be analysed using a pseudo-first order kinetic approach and a simple understanding of the induction time is provided in light of recent literature.9-10, 16-17

This practical exercise connects synthesis of nanoparticles to their application as catalysts and provides students with hands on experience of the physical chemistry theory (e.g., rate laws and kinetics, XRD, spectroscopy) taught in their lecture courses. Additionally, this practical provides students with the opportunity to engage in inquiry-based learning. This approach is increasingly assimilated into undergraduate curricula to teach students skills in scientific thinking and observation that help to consolidate their understanding of theory and to mould the next generation of scientists.18-21

The aim of the experiment described herein was to give students a broad overview of nanoscience and its links to more traditional chemistry fields, but could be easily recontextualised under themes of green chemistry,22-23 pharmaceuticals,14 or quantum chemistry.6 Additionally, more detail could be devoted to the topics of

100 Chapter 4 spectroscopy (including Beer-Lambert Law), crystallography, kinetics or laboratory technique (e.g. micropipetting) as desired by the instructor (see instructor notes). Alternatively, the practical could be adapted for a general first year cohort by omitting the more advanced aspects, or run in second year as part of the standard laboratory program. The practical, which has been run and refined over 5 years, is presented below.

Experimental Section Chemicals The following chemicals were used in the experiments: trisodium citrate (Univar), silver nitrate (Univar), hydrogen peroxide (30%, Univar), potassium bromide (Univar), sodium borohydride (Aldrich), and 4- nitrophenol (Fluka). All solutions were prepared with Milli-Q water supplied by a Merck Milli-Q Advantage A10. Solutions of sodium borohydride and hydrogen peroxide were kept ice-cold to slow decomposition. Students should prepare or be supplied with the following solutions: 1.25 × 10¯2 M trisodium citrate, 3.75 × 10¯4 M silver nitrate, 5.00 × 10¯1 M hydrogen peroxide, 2.17 × 10¯5 M potassium bromide, 5.00 × 10¯3 M (for synthesis) and 0.250 M (for catalysis) sodium borohydride, 1.05 × 10¯5 M 4-nitrophenol. The synthesis is sensitive to trace contaminants; water must be high purity and reused glassware must be cleaned with acid and washed prior to use. We found it easiest to use fresh sample vials as preparation vessels instead, which additionally facilitated easy sample storage.

Hazards Hydrogen peroxide is an oxidizing agent. Silver nitrate is oxidizing, toxic, and corrosive. Trisodium citrate is an irritant. Sodium borohydride is toxic, corrosive, and releases flammable gas on contact with water. 4- Nitrophenol is toxic and an irritant. 4-Aminophenol is carcinogenic, an irritant, and toxic to aquatic lifeforms. The hazards of silver nanoparticles are not well established and solutions containing them should be handled with care. All solutions handled by students are dilute and risks are low. Students should read the safety data sheets before performing the experiment and standard personal protective equipment (safety glasses, enclosed shoes, and laboratory coats) should be worn.

Synthesis of the Silver Nanoparticles (AgNPs) As has been previously noted,1 the bromide concentration during AgNP formation plays a critical role in the size and shape of the nanoparticles produced. Therefore, the volume of bromide solution added is critical and is the most significant variable. Calibrated micropipettes and adequate training are recommended. Consistent with the literature,24 stirring was found to have minimal influence on the formation of the nanoparticles, provided the solution is sufficiently mixed prior to and after addition of sodium borohydride. Sample vials should be new or cleaned with acid and washed before reuse to remove trace contaminants that can affect the synthesis. Our students are divided into groups of ~10, each group supervised by a demonstrator (lab assistant), and each student prepares four nanoparticle samples.

Into a sample vial - optionally equipped with a stirrer bar - aliquots of trisodium citrate (0.20 mL of 0.0125 M solution), silver nitrate (5.00 mL of 3.75 × 10¯4 M solution), and hydrogen peroxide (5.00 mL of 5.00 × 10¯1 M solution) were pipetted. X µL of aqueous potassium bromide (where 0 < X < 2300 of 2.17 × 10¯5 M solution) and a corresponding volume of water (2300 – X µL) were then added. For stirred samples, stirring was started (500 rpm) and for shaken samples, the solution was mixed by hand. Sodium borohydride (2.50 mL of 5.00 × 10¯3 M solution) was pipetted into the vials and the mixtures were either stirred for fifteen minutes or mixed thoroughly by shaking until the solution colour had stabilised; at ambient temperature. The reaction appears dormant for several minutes before a rapid colour change ( Figure 43, Left). Instructors should encourage

101 Chapter 4 students to be patient, however if no colour change is observed after five minutes the synthesis should be repeated – additional troubleshooting can be found in the Instructor Notes. Approximate volumes of KBr (μL of 2.17 × 10¯5 M solution) to obtain certain colours are as follows:

0: Blue; 800: purple, 1100: orange, 1300: yellow.

Once the colours were constant, the UV-vis spectra of the AgNPs were obtained and the Tyndall effect observed ( Figure 43, Right). Advanced courses can include additional analyses such as transmission electron microscopy (TEM), x-ray diffraction (XRD) and dynamic light scattering (DLS).

180

150

120

90 Time (s) Time 60

30 Absorbance (normalized) Absorbance

0 300 400 500 600 700 800 Wavelength (nm) Figure 43. (Left) Contour plot of UV-Vis spectra with time of nanoparticle spectra during synthesis. (Right) Observation of the Tyndall effect in a silver nanoparticle colloid using a 532 nm laser.

A discussion on the roles of each reagent can be found in a previous practical by Kitaev and co-workers.7

Catalytic reduction of 4-nitrophenol The catalytic reduction of 4-nitrophenol using the silver nanoparticles can be conveniently monitored using UV-vis spectroscopy, by monitoring the absorbance of the intensely coloured 4-nitrophenolate anion at λ = 400 nm as a function of time.

Aqueous solutions of 4-nitrophenol (1.00 mL of 1.05 × 10¯4 M solution) and sodium borohydride (100 µL of 0.25 M solution) were added to a disposable cuvette. The mixture became strongly yellow coloured due to the formation of sodium 4-nitrophenolate. To this solution, an aliquot of the relevant AgNPs (1.00 mL, 8.75 × 10¯6 M) was added at a defined time and the reaction mixture monitored every six seconds between λ = 300– 600 nm (1 nm intervals, 4800 nm/min) or every one second at λ = 400 nm. The reaction was carried out under ambient conditions unless otherwise noted. Typical runs took five to ten minutes for the absorbance at λ = 400 nm to diminish completely – consistent with complete consumption of the starting material.15 Note: sodium borohydride solution will slowly consume any dissolved oxygen in the reaction mixture;16 for best results the time between borohydride and catalyst addition should be consistent – 30 s was found to be practicable.

Results and Discussion Silver Nanoparticles This experiment was inspired by the work of Kitaev7 and Pal.13 Kitaev’s excellent report details a method for controlling the nanoparticle size. Compared to other reported methods,2,3 the use of citrate and bromide as shape modifying agents makes the size control of nanoparticles an achievable goal for the undergraduate chemist.7 Due to the significant effect of bromide on the resultant nanoparticles, a larger volume of lower concentration potassium bromide – relative to the work of Kitaev and co-workers7 – was used to make the experiment less susceptible to pipetting error. Additionally, an extended range of Br:Ag molar ratios was

102 Chapter 4 prepared, resulting in a rainbow spectrum of sample colours (Figure 44). In our laboratory, we allocate different Br:Ag ratios to different groups of students, so that the class observes the full variety of colours (Figure 45; Left). Having all students within a group use the same Br:Ag ratio allows for some redundancy. We also use variations observed in student spectra to discuss errors in measurement, pipetting, or other variables; the spectrum offers a strong visual confirmation of accuracy or problems with experimental integrity that is highly appealing to students.

Figure 44. Silver nanoparticles synthesised with varying amounts of KBr – Br:Ag = 0 (far left), 0.0053, 0.008, 0.0088, 0.0096, 0.0104, 0.0112, 0.0121, 0.0128, 0.0136 and 0.0144:1 (far right).

One of the interesting outcomes of shared data between large cohorts of students over five years of the course delivery is the ability to compile larger data sets. We found that contrary to Kitaev’s report, the λmax was not found to be approximately inversely related to the Br:Ag ratio, but rather that the λmax exhibited a sharp change in the Br:Ag vs. wavelength plot at approximately 0.005:1 Br:Ag (Figure 45, Right). Therefore, only small changes in the Br:Ag ratio near this critical point are required to produce colloidal dispersions displaying a spread of colours.

1.0 KBr:Ag Ratio 0 0.0053 0.008 0.8 0.0088 0.0096 0.0104 0.0112 0.0121 0.6 0.0128 0.0136 0.0144

0.4 Absorbance(Normalized)

0.2

0.0 300 400 500 600 700 800 Wavelength (nm) Figure 45. UV-vis data for silver nanoparticle colloids (shown in Figure 44) prepared using varying amounts of bromide. (Left) normalised UV-vis spectra, (Right) normalised contour plot of Br:Ag ratio vs wavelength.

Increased Br:Ag ratios resulted in a reduction in peak fullwidth at half maximum with a decrease in λmax. This change is consistent with TEM images of the AgNPs (see Figure 46). The change in Br:Ag ratio also results in a change in catalytic activity, vide infra. Particles synthesised using lower Br:Ag ratios are more anisotropic, forming triangular platelets and truncated hexagons (Figure 46, Left, 30 (± 10) nm, blue). At higher Br:Ag mole ratios, the particles appear rounded and isotropic (Figure 46, Right, 19 (± 8) nm, orange). This observation is consistent with the diminishing absorbance associated with the triangular platelet plasmon resonance observed in the UV-vis spectra.25

103 Chapter 4

Figure 46. TEM images of nanoparticles made with a Br:Ag molar ratio of 0:1 (left; blue colloid) and 0.025:1 (right; orange colloid). Insets: Selected area electron diffractograms of nanoparticles indexed to Ag (Left) and AgBr (Right).

The nanoparticles appear triangular prismatic at low Br:Ag ratios (Figure 46, Left), consistent with the (111) face being stabilised by citrate ions, leading to preferential growth of these planes. This explains the strong (111) diffraction reflection at 38° 2θ ( Figure 47).7

Figure 47. PXRD of AgNPs, reflections are consistent with silver metal (ICSD 04-0783) and trace amounts of silver(I) oxide and silver bromide

It was found that the resulting particles made with varying amounts of Br:Ag were highly reproducible – as judged by both the colour and UV-vis spectra of the solutions. Deviations were due to egregious experimental error, poor pipetting technique, or addition of the incorrect solution. The colours of the resulting solutions are so reproducible that they could, if desired, be used as a qualitative indication of the skills of the students.

104 Chapter 4

Figure 48. UV-vis spectra of (grey) fresh and (coloured) aged particles as function of the Br:Ag ratio. Alternating red/blue colour used to aid visual differentiation.

UV-vis studies were also used to confirm that the nanoparticles were stable in solution over at least two months (Figure 48). This stability meant that the synthesis, characterization, and catalysis were able to be performed over three separate laboratory sessions without significantly impacting the results.

Catalysis The silver nanoparticle catalysed reduction of 4-nitrophenol takes approximately five minutes per run, allowing for duplicate or triplicate experiments to be performed – depending on the number of samples, students, and available UV-vis spectrophotometers. A representative profile is shown in Figure 49. The presence of several isosbestic points is clear and could be used in additional teaching (see 26 for further reading). There has been debate about the origin of the induction time, with various explanations given, such as reduction of the oxidised catalyst surface,13 surface restructuring,27-29 and the presence of oxygen in the catalytic mixture.16 Recent work within our laboratory supports the hypothesis that dissolved oxygen causes the induction period. Specifically, simultaneous oxidation of both the intermediate and the catalyst until dissolved oxygen in the system is sufficiently depleted.17

Figure 49. (Left) Representative time course UV-vis spectra of the reduction of 4-nitrophenolate. (Right) Catalytic profiles for the reduction of 4-nitrophenolate monitored at λ = 400 nm, catalyzed by AgNPs of varying Br:Ag ratios. Data are the average of three replicates.

To a good approximation, the data after the induction time can be fitted to a first-order model (Table 1). While the precise catalytic mechanism is contentious,16, 29 applying an assumption of pseudo-first order behaviour makes a simplified kinetic treatment accessible for undergraduate students. Students can compare the activity of the catalysts they synthesised using a slope function on readily available spreadsheet software and relate the activity

105 Chapter 4 to differences between the catalysts themselves. As the Br:Ag ratio increases, the induction time increases and the rate of substrate consumption decreases (Table 1), consistent with fewer active catalyst sites.29 Students are encouraged to compare their results with those of others.

Table 12. Example kinetic profile analysis of the silver nanoparticle catalyzed reduction of 4-nitrophenol Br:Ag Nanoparticle Colloid Colour Induction Time (s) Slope of ln(4-nitrophenol)a r2

0.000 Blue 30 -0.03 0.99

0.014 Orange 70 -0.02 0.98 aLinear fit applied to data collected at times after the induction period. Assessment Our nanoscience module is taught as one component of the first semester of laboratory teaching. Students are assessed on their work in the laboratory, completion of their electronic laboratory notebook, ability to answer questions, and additionally through a video assessment or poster presentation.

The incorporation of an authentic science communication assessment in this unit of study enables students to develop and apply skills in the communication of science to public and academic audiences, and additionally helps to foster a sense of community within the broader Chemistry School.

Conclusions The reported practical is an effective introduction to the fields of nanoscience and catalysis, in particular, for commencing undergraduate chemistry students. The practical is safe, quick, and able to be expanded or recontextualised to fit the requirements of the laboratory instructor and syllabus. It is robust and utilizes readily available instrumentation and chemicals. We have provided an updated interpretation of the catalytic behaviour of the silver nanoparticles and report several significant variations/expansions on the experiments of Kitaev7 and Pal13 to aid laboratory instructors.

106 Chapter 4

References 1. Faraday, M., X. The Bakerian Lecture.—Experimental relations of gold (and other metals) to light. Philos. Trans. R. Soc. London 1857, (147), 145-181. 2. Clunan, A.; Rodine-Hardy, K.; Hsueh, R.; Kosal, M. E.; McManus, I. Nanotechnology in a globalized world: strategic assessments of an emerging technology; Naval Postgraduate School Monterey Ca Center On Contemporary Conflict: 2014. 3. Greenberg, A., Integrating nanoscience into the classroom: Perspectives on nanoscience education projects. ACS Publications: 2009. 4. Chorkendorff, I.; Niemantsverdriet, J. W., Concepts of modern catalysis and kinetics. John Wiley & Sons: 2017. 5. Blonder, R., The story of nanomaterials in modern technology: An advanced course for chemistry teachers. J. Chem. Educ. 2010, 88 (1), 49-52. 6. Yu, J., From coinage metal to luminescent nanodots: the impact of size on silver’s optical properties. J. Chem. Educ. 2014, 91 (5), 701-704. 7. Frank, A. J.; Cathcart, N.; Maly, K. E.; Kitaev, V., Synthesis of silver nanoprisms with variable size and investigation of their optical properties: a first-year undergraduate experiment exploring plasmonic nanoparticles. J. Chem. Educ. 2010, 87 (10), 1098-1101. 8. Mulfinger, L.; Solomon, S. D.; Bahadory, M.; Jeyarajasingam, A. V.; Rutkowsky, S. A.; Boritz, C., Synthesis and study of silver nanoparticles. J. Chem. Educ. 2007, 84 (2), 322. 9. Raghuwanshi, V. S.; Wendt, R.; O’Neill, M.; Ochmann, M.; Som, T.; Fenger, R.; Mohrmann, M.; Hoell, A.; Rademann, K., Bringing Catalysis with Gold Nanoparticles in Green Solvents to Graduate Level Students. J. Chem. Educ. 2017, 94 (4), 510- 514. 10. Feng, Z. V.; Lyon, J. L.; Croley, J. S.; Crooks, R. M.; Vanden Bout, D. A.; Stevenson, K. J., Synthesis and Catalytic Evaluation of Dendrimer-Encapsulated Cu Nanoparticles. An Undergraduate Experiment Exploring Catalytic Nanomaterials. J. Chem. Educ. 2009, 86 (3), 368. 11. Novak, J. D., Application of advances in learning theory and philosophy of science to the improvement of chemistry teaching. J. Chem. Educ. 1984, 61 (7), 607. 12. At the time of submission PAMAM-OH Generation 4 Dendrimer Solution cost almost two times more than gold per gram from Merck. 13. Pradhan, N.; Pal, A.; Pal, T., Catalytic reduction of aromatic nitro compounds by coinage metal nanoparticles. Langmuir 2001, 17 (5), 1800-1802. 14. Zhao, P.; Feng, X.; Huang, D.; Yang, G.; Astruc, D., Basic concepts and recent advances in nitrophenol reduction by gold- and other transition metal nanoparticles. Coord. Chem. Rev. 2015, 287, 114-136. 15. Pradhan, N.; Pal, A.; Pal, T., Silver nanoparticle catalyzed reduction of aromatic nitro compounds. Colloids Surf., A 2002, 196 (2-3), 247-257. 16. Menumerov, E.; Hughes, R. A.; Neretina, S., Catalytic reduction of 4-nitrophenol: a quantitative assessment of the role of dissolved oxygen in determining the induction time. Nano Lett. 2016, 16 (12), 7791-7797. 17. Strachan, J.; Barnett, C.; Masters, A. F.; Maschmeyer, T., 4-Nitrophenol Reduction: Probing the Putative Mechanism of the Model Reaction. ACS Catal. 2020. 18. Allen, J. B.; Barker, L. N.; Ramsden, J. H., Guided inquiry laboratory. J. Chem. Educ. 1986, 63 (6), 533. 19. Wheeler, L. B.; Clark, C. P.; Grisham, C. M., Transforming a traditional laboratory to an inquiry-based course: Importance of training TAs when redesigning a curriculum. J. Chem. Educ. 2017, 94 (8), 1019-1026. 20. Farrell, J. J.; Moog, R. S.; Spencer, J. N., A guided-inquiry general chemistry course. J. Chem. Educ. 1999, 76 (4), 570. 21. Mistry, N.; Fitzpatrick, C.; Gorman, S., Design your own workup: a guided-inquiry experiment for introductory organic laboratory courses. J. Chem. Educ. 2016, 93 (6), 1091-1095. 22. Dorney, K. M.; Baker, J. D.; Edwards, M. L.; Kanel, S. R.; O’Malley, M.; Pavel Sizemore, I. E., Tangential flow filtration of colloidal silver nanoparticles: a “green” laboratory experiment for chemistry and engineering students. J. Chem. Educ. 2014, 91 (7), 1044-1049. 23. Duong, M. H.; Penrod, S. L.; Grant, S. B., Kinetics of p-nitrophenol degradation by Pseudomonas sp.: An experiment illustrating bioremediation. J. Chem. Educ. 1997, 74 (12), 1451. 24. Panzarasa, G., Just what is it that makes silver nanoprisms so different, so appealing? J. Chem. Educ. 2015, 92 (11), 1918- 1923. 25. Kelly, K. L.; Coronado, E.; Zhao, L. L.; Schatz, G. C., The optical properties of metal nanoparticles: the influence of size, shape, and dielectric environment. J. Phys. Chem. B 2003, 107 (3), 668-677. 26. Aditya, T.; Pal, A.; Pal, T., Nitroarene reduction: a trusted model reaction to test nanoparticle catalysts. Chem. Commun. 2015, 51 (46), 9410-9431. 27. Wunder, S.; Lu, Y.; Albrecht, M.; Ballauff, M., Catalytic activity of faceted gold nanoparticles studied by a model reaction: evidence for substrate-induced surface restructuring. ACS Catalysis 2011, 1 (8), 908-916. 28. Herves, P.; Pérez-Lorenzo, M.; Liz-Marzan, L. M.; Dzubiella, J.; Lu, Y.; Ballauff, M., Catalysis by metallic nanoparticles in aqueous solution: model reactions. Chem. Soc. Rev. 2012, 41 (17), 5577-5587. 29. Gu, S.; Wunder, S.; Lu, Y.; Ballauff, M.; Fenger, R.; Rademann, K.; Jaquet, B.; Zaccone, A., Kinetic analysis of the catalytic reduction of 4-nitrophenol by metallic nanoparticles. J. Phys. Chem. C 2014, 118 (32), 18618-18625.

107 Conclusions

Conclusions Summary There still exists a myriad of opportunities for the development of energy conversion materials. Many of the materials currently investigated are scarce and expensive. However, the use of abundant and cheap materials is key to addressing the increasing global energy demand. The work in this thesis demonstrated that several examples of molybdenum-based materials, which already fulfil the requirements of commercial use (heterogeneous, active, inexpensive, and stable), are highly effective at facilitating a range of energy conversion processes.

Within the family of Mo-based materials specifically, the work within this thesis described how Chevrel Phases, polytypic MoS2, and supported molybdenum carbides are promising materials for catalysis and electrochemical energy storage:

Chevrel Phases are not widely used as catalysts because they conventionally exhibit low surface areas as a result of their lengthy syntheses at high temperatures. Chapter 1 described a novel, two-step method that produced active, unsupported, nanoparticulate Chevrel Phases; overcoming the low surface area limitation of these materials. Furthermore, a comprehensive literature review highlighted the unique catalytic features of Chevrel Phases and demonstrated how these features might be exploited by certain catalytic reactions in the context of energy conversion, power production and storage.

The historically murky and scattered literature on the 3R polytype of MoS2 was collected and systematised in Chapter 2 to provide a summary, scope, and guidance for future research. A critical review on the hydrothermal synthesis of 1T-MoS2 was produced to rectify the numerous analytical errors propagated throughout the literature. Finally, the effect of crystal disorder on the electrochemical behaviour of MoS2 was investigated to determine the optimal electrochemical characteristics for use as cathodes in Li-Mg hybrid batteries.

Chapter 3 reported the determination of the structural features of an active and selective Mo2CxNx-1/TiN lignin valorisation catalyst by EXAFS analysis. The final chapter, which centred on silver nanoparticles, reported the re-evaluation of the mechanism of the ubiquitous model catalytic reaction, the reduction of 4-nitrophenol. Additionally, a new mechanism was proposed that united the two previously competing theories, namely catalyst surface restructuring and dissolved oxygen depletion. The above model reaction was then integrated into an undergraduate practical to introduce students to nanoscience, kinetics, and catalysis.

Implications The work in Chapter 1 offers several reasons for researchers to revisit the catalytic potential of Chevrel Phases. It primarily consisted of reviewing the temporally and thematically scattered literature and developing a novel nanoparticulate synthesis. The results are significant in that they challenge several misconceptions about the material and demonstrate that Chevrel Phases can outperform conventional catalysts in industrially relevant reactions such as the hydrogen evolution reaction. Furthermore, the unique features of Chevrel Phases proposed in Chapter 1.1 differentiate Chevrel Phases from conventional catalysts, such as MoS2, and therefore establish a unique place for them in the catalytic toolbox. By exploiting these features, (particularly the adjacent Lewis-acid/base sites) Chevrel Phases are likely to see further use in rationally designed catalytic reactions where selectivity is of key importance.

108 Conclusions

The work in Chapter 2 contributes to the vast and growing literature on polytypic MoS2. By drawing together the entirety of the current literature on catalytic 3R-MoS2, Chapter 2.1 provides structure to the history, syntheses, applications, and future directions of this largely overlooked material. In doing so, this work makes the scattered reports of 3R-MoS2 analyses and applications accessible. As concluded in the chapter, there are very few reports of 3R-MoS2 used as catalysts. However, by summarising the salient features of 3R-MoS2, this work can be expected to stimulate further studies investigating the catalytic behaviour of this widely overlooked polytype.

The subsequent discussion on 1T-MoS2 in Chapter 2.2, has a significant and timely position within the hydrothermal MoS2 literature. The 1T-MoS2 field is in its infancy, with a large portion of the literature having been published only in the last 5 years. It is critical that these publications are founded on robust science and, to this end, identification and rectification of the numerous errors in analyses was necessary. The instructions contained in this section clarify analysis procedures and interpretations, and should therefore help to accelerate research within this field. Given that hydrothermally produced 1T-MoS2 has demonstrated one of the highest activities for the hydrogen evolution reaction, this fundamental work may yet assist the realisation of commercial electrocatalysts based on the 1T polytype.

Structural analysis of Mo2CxNx-1/TiN in Chapter 3 using XAS in air and under ethanol at high temperature led to the observation of Mo≡N motifs, that are normally observed in organometallic complexes rather than in a solid-state material. Furthermore, the catalyst is highly selective for combined lignin hydrodeoxygenation and depolymerisation. This structural feature is unprecedented for catalysts known for lignin depolymerisation. The finding provides key information for computational mechanistic studies and is a starting point for the design of other active transition metal carbo-nitride catalysts. The conclusions offered within this chapter aid the development of waste biomass treatment processes, which in turn mitigates our reliance on conventional fossil fuels.

Chapter 4 begins with an investigation into the mechanism of silver-catalysed 4-nitrophenol reduction and translates these findings into a laboratory experiment that utilises this reaction as a practical example of nanoscience, kinetics, and catalysis. The work further establishes the reliability of this model reaction, thus enabling researchers to consistently and confidently benchmark catalysts. In effect, the work presented helps accelerate the testing throughput of novel catalysts; the development of which is a core goal of green chemistry and integral to facilitating sustainable chemical transformations. It does this by providing an understanding of the reaction mechanism, such that a consistent reaction protocol is developed. The integration of this model reaction into an undergraduate laboratory experiment was an important contribution to the educational literature on catalysis and kinetics. By making significant improvements over preceding methodologies, it is expected that this teaching experiment will be widely utilised. The resulting publication helps to teach green chemistry in an approachable and understandable manner – which is key for influencing the researchers of tomorrow.

Future Work The Chevrel Phase catalysts in Chapter 1.2 were not optimised. Further improvements can be investigated by employing alternative synthesis strategies (such as hydrothermal methods), supporting the catalysts (e.g. by wetness impregnation), and incorporating more stable elements in the framework (such as the ‘large-ion’ Pb2+ used previously in HDS catalysts). These procedures have improved the surface area, conductivity, and activity of catalysts in other systems and similar enhancements are expected for Chevrel Phases. It is anticipated that the catalysts reported in this chapter would exhibit favourable kinetics in electrochemical energy storage

109 Conclusions applications due to the negligible path length of ion migration during intercalation. Improvements to the conductivity of these materials, by metal-intercalation or otherwise, would enhance the kinetics of electron- transfer in both electrochemical energy storage applications and, perhaps more importantly, electrocatalysis. Given an increasing trend in the field of catalysis towards the design and evaluation of electrocatalytic processes, the suggested research avenues will likely lead to an improved utilisation of the very promising Chevrel Phases.

Despite the large number of publications devoted to MoS2 (Chapter 2), much remains to be explored. The catalytic behaviour of 3R-MoS2 is largely unknown. Given that screw-dislocation grown crystals are electrically conductive, electrochemical applications are a promising avenue for development. However, greater control over particle growth during screw dislocation synthesis is a likely prerequisite for the optimisation of their catalytic characteristics. Substitution of Mo by, e.g. Re, has been established as one synthetic mechanism for producing 3R-(Rex)Mo1-xS2. However, the effect of this substitution on the band structure has yet to be investigated. Exploring the scope of dopants that can induce the 3R phase, identifying the resultant effect on the band structure, and applying the findings to rationally design a conductive electrocatalyst is an attractive path to realising the catalytic potential of 3R-MoS2.

With regard to 1T-MoS2, whether the Mo centre is d2 or d3, and whether a small, charge-compensating ion is present between the layers in crystalline samples are key questions that need to be answered to reach a more complete understanding of this material. In single crystal samples synthesised by a solid-state method, this small ion may be H+, which is small enough to have avoided detection by diffraction measurements thus far. By using the deuterated reagent, D2O, it is expected that neutron diffraction experiments could detect the presence of this ion. Still, it remains unclear why [MoS2]− may be isolated from the hydrothermal synthesis yet the analogous sample synthesised by solid-state methods appears unstable in water. Electron paramagnetic resonance spectroscopy and magnetic susceptibility measurements should be performed to better understand the electron configuration of 1T-MoS2 in comparison with a 2H-MoS2 sample. It is still unclear what relationship exists between the oxidation state of the Mo centre and it’s geometric configuration. Regarding the synthesis of 1T-

MoS2, we proposed that a reverse Claus process may be responsible for the consumption of elemental sulfur in the reaction vessel (given there exists precedent in the literature). However, it is currently unknown whether this process may be catalysed by the nascent MoS3 and MoS2 in the reaction vessel, thus future studies may investigate this system further. These studies would provide an improved understanding of the reductant, currently assumed to be S2− species, and therefore lead to greater control over the reaction parameters and products. The heat-induced decomposition pathways of MoS2 samples intercalated with non-volatile species (such as alkali metals) is currently unknown. It can be expected that an endothermic heat flow would be observed without a corresponding mass loss (as seen for intercalants with volatile decomposition products such as NH4+). DFT calculations on the 1T-MoS2 polytype have used a Mo(IV) centre, although more accurate models might be obtained if experimental data on the oxidation state informs future modelling studies (i.e. if a

Mo(III) species is used). The 1T polytype has been shown to outperform 2H-MoS2 in HER reactions; whether this is because of the active site geometry differences between the samples, the presence of Lewis-basic sites in

1T-[MoS2]−, the improved conductivity facilitates electron transfer, or other yet unknown reasons is unclear.

1T-MoS2 may be formed by doping with transition metals. Whether these metals substitute for Mo, the degree, spatial variance and elemental scope to which they do, the possibility of oxidation state change and the ranges of potential oxidation states are unknown. Attempts to synthesise the range of possible structures, in combination with high resolution elemental mapping (by HR-STEM-EDS/EELS) are simple first steps to gain insights into these questions.

110 Conclusions

The crystalline disorder of the MoS2 electrodes produced in Chapter 2.3 had a clear influence on the electrochemical properties of the material. What remains to be elucidated is exactly which sites are active, what the energies of these sites are, and whether specific sites might be engineered in order to optimise the desirable electrocatalytic characteristics of these materials.

Computational mechanistic studies of lignin model compounds over the Mo≡N active site may validate the findings reported in Chapter 3. The study reported in this thesis was limited by the challenges of recording genuinely in-situ EXAFS data, i.e. under realistic operating conditions. Future work could involve the design of an in-situ apparatus to analyse the structure of the catalyst under more representative reaction conditions. Additionally, there exists large scope for variation of the support, precursor, stoichiometry, and synthesis conditions of the Mo2CxNx-1/TiN catalyst – it is not yet known whether the catalyst reported in this thesis is optimal, or whether there are other thermo- or electrocatalytic processes for which this material is well suited. Electrocatalytic decomposition of lignin under mild conditions could allow for similar product streams to that reported in this chapter, though with drastically lower energy input.

An in-situ analysis of the oxidation state of the nanoparticulate silver catalyst would help discriminate between the two induction period activation steps that were proposed in Chapter 4. Continuous flow EXAFS or EPR analysis could provide the necessary information at a sufficient time resolution to determine the oxidation state of the catalyst as a function of time (and therefore as a function of reaction progression). These data could then be incorporated into the kinetic models used in Chapter 4.1 to achieve this goal. Once this is established, a comprehensive model that combines Langmuir-Hinshelwood theory with the two activation mechanisms would be possible. Such a model is the ultimate objective of research into this reaction.

Concluding Remarks The main work presented in this thesis is united by the aim to advance our understanding of sustainable energy conversion materials. To this end, several candidate molybdenum sulfides and carbides were analysed as catalysts and electrodes. Through a combination of literature reviews and fundamental studies it was shown that there exist several underutilised molybdenum-based materials, which exhibit highly promising and unique behaviours, leading to a well-founded expectation that they will perform well in a number of critical energy conversion reactions. It is anticipated that the findings will stimulate further research into these materials and advance the processes, which they aid, namely: the hydrogen evolution reaction, electrochemical energy storage, lignin valorisation, and the discovery of novel catalysts. Such advancements are integral to shifting the current energy landscape towards decarbonised and sustainable systems. These significant transformations are vital components of our collective response to the climate crisis that lies before us.

111

Blank Page Appendices

Appendix to Chapter 1.2 Experimental Chemicals and Equipment

Chemicals used as received were Na2MoO4.H2O (Ajax), CuCl (Merck), 1-butyl-3-methylimidazolium trifluoromethanesulfonate (Sigma), Mo (Ajax), MoS2 (Sigma, “90 nm diameter”), CuS (Sigma-Aldrich), 5 wt% Nafion solution (Sigma-Aldrich), ZnSO4.7H2O (Ajax), MnSO4.4H2O (Hopkin and Williams),

FeSO4.H2O (Ajax), CoNO3.6H2O (Merck), HCl (Sigma-Aldrich) and NaCl (Ajax).

Electrochemical measurements were performed on an eDaq ER466 Integrated Potentiostat System. X-ray photoelectron spectroscopy was performed on a Thermo Fisher Scientific K-Alpha+ equipped with an aluminium X-ray source. Data were collected using a step size of 0.100 eV and pass energy of 50 eV then analysed using Thermo Avantage v5.9902. X-ray diffraction measurements were performed on a PANalytical X'Pert Powder diffractometer equipped with a copper X-ray source. Raman analysis was performed on a Renishaw InVia Qontor Dual Raman System using a 532nm laser. Bright field Transmission Electron Microscopy was recorded using a JEM-2100 at 200 kV. Impedance Spectroscopy was performed using a BioLogic SP300 potentiostat at -200 mV (vs NHE) between 5 MHz and 5 mHz with a 20 mV sinusoidal perturbation. Faradaic efficiency was calculated using:

FE (%) = VH2, m/VH2, c

Where VH2, m is the measured volume of H2 and VH2, c is the volume of H2 calculated using:

VH2, c = (I (A) × t (s) × Vm)/(F (C/mol) × z)

Where Vm is the molar volume (24 L/mol) and z is the number of electrons consumed per mol H2 produced

(for HER = 2). The VH2, m was measured using a Shimadzu GC-2014 equipped with a pulsed discharge ionization detector (Vici PDID D-4-I-SH14-R). The electrodes were contained in a sealed quartz reactor comprising a Nafion separator, Pt counter electrode, Ag/AgCl reference electrode, and a 1 M HCl electrolyte. Syntheses

Ammonium tetrathiomolybdate, (NH4)2MoS4, was synthesised by a published procedure and

Cu(NH4)MoS4 was synthesised using variation of this synthesis as follows.1 Na2MoO4.H2O (2 g, 8.3 mmol) and CuCl (272 mg, 2.7 mmol) were dissolved in water (4 mL). This solution was added to (NH4)2S (30 mL) and held at 35 °C over night. The solution was left to cool to room temperature, then an ice-cold mixture of 1:1 ethanol/ether was added and the solution left to cool on ice to precipitate the remaining

Cu(NH4)MoS4. The product was vacuum filtered and washed with 10 mL: cold water, 1:1 ethanol/ether, and ether before collection and storage in an argon filled glove box. The identity of the product was analysed by UV–Vis spectroscopy (λmax = 317, 468 nm, ((NH4)2MoS4)ε317 nm = 17400 Lmol-1cm-1)2 and XRD (I-4, a=b=8, c=5.409)3.

Bulk-CP

The bulk Mo6S8 was synthesised by conventional methods.4-5 In brief, a stoichiometric ratio of Mo, MoS2 and CuS was ground with a mortar and pestle into a macroscopically homogeneous powder. The powder mixture was heated in a tube furnace at 5 °C/min to 1000 °C and held for 36 h under 5% H2/Ar. Once cooled, the Chevrel Phase was treated with 8 M HCl and bubbling O2(g) for 12 h. The powder was washed with water and acetone, then dried in an oven at 120 °C.

Solvothermal (Ionic Liquid) Synthesis (n-CuCP) Copper ammonium tetrathiomolybdate, Cu(NH4)MoS4, (0.4 g, 1.3 mmol) was dissolved in 1-butyl-3- methylimidazolium trifluoromethanesulfonate (4 g). The mixture was sonicated for 30 min before being loaded into a two neck round bottom flask. The flask was equipped with a stirrer bar, a 5% H2/Ar inlet, and a gas outlet that bubbled into a CuSO4 trap (for evolved H2S). The flask was purged for 1 h, heated

113 Appendices to 300 °C, and held at temperature for 3 h. Once cooled, the mixture was washed thoroughly with dichloromethane, dried in air, then heated in a tube furnace at a 2 °C/min ramp to 700 °C for 60 h under a flow of 5% H2 in Ar.

Electrochemistry

5 mg of the Chevrel Phase powder was dispersed in 0.965 mL of 3:1 water/isopropanol, to which 45 µL of 5 wt% Nafion solution was added.6 The mixture was sonicated for 1 h then 1.6 µL was drop cast onto a 3 mm glassy carbon electrode and dried at 120 °C.

Cyclic voltammetry was performed using an ER466 Integrated Potentiostat System (eDAQ) in a conventional three electrode configuration following best practice.7-8 The working, counter and reference electrodes were, respectively, drop-cast Chevrel phase on a 3 mm polished glassy carbon surface, graphite rod and Ag/AgCl (3 M NaCl, eDaq). The working electrode was polished in an alumina slurry (0.3 µm), rinsed with acetone, sonicated in acetone, rinsed thoroughly with acetone and ethanol then dried under a stream of nitrogen gas.9 The scan rate was 50 mV/s and the electrolyte was deoxygenated 1 M HCl, unless otherwise specified. All electrolyte solutions were purged thoroughly with N2 gas prior to use, and during use the electrolyte was kept under a blanket of positive pressure N2.

Electrochemical potentials were converted to RHE using:

ERHE = E(Ag/AgCl) + 0.210 + 0.059 × pH

The electrochemical surface area was determined by the double layer capacitance method.10 Given the difficulties of estimating the specific capacitance,9 especially when comparing different materials10-11 (and more problematically, Nafion bound Chevrel phases)12 the double layer capacitance of the drop cast electrodes was measured relative to that of glassy carbon (which was estimated to have a surface area of 0.071 cm2 based on the electrode dimensions).

Electrocatalyst preparation / CP (de)intercalation Prior to electrocatalysis, the n-CuCP was electrochemically deintercalated. The potential was scanned from 0.25–0.5 V (vs Ag/AgCl) until the cathodic current at 0.33 V reached zero. The peak area of the cathodic deintercalation current indicates a stoichiometry of Cu0.2Mo6S8, which suggests that most of the copper is very quickly chemically leached by the HCl when placing the electrode into the electrolyte, as per the commonly reported chemo-oxidation pathway:13

2CuxMo6S8 + 8xHCl + xO2 → 2Mo6S8 + 2xH2O + 2x[CuCl4]2− +4xH+

Scheme S 1 – Chemical oxidative deintercalation of CuCP

114 Appendices

0.5 4 Cu Mo S → Mo S + 2xe- + xCu2+ Zn Mo S → Mo S + xZn2+ + 2xe- x 6 8 6 8 3 x 6 8 6 8 0.4

) 2 2 0.3 1

0.2 0

-1 0.1

Current (mA/cm Current -2 0.0 -3 2+ - Mo6S8 + xZn + 2xe → ZnxMo6S8 -0.1 -4 0.0 0.1 0.2 0.3 0.4 0.5 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 Potential (vs Ag/AgCl) Potential (vs Ag/AgCl) Figure S 1 – (Left) Cyclic voltammogram of copper deintercalation from the n-CuCP electrode under hydrodynamic conditions in oxygenated 1 M HCl. (Right) Cyclic voltammogram of the n-CP in 0.05 M ZnSO4, 1 M NaCl showing de/intercalation currents. All scans at 50 mV/s, initial potential 0 V (cathodic, L) and −0.4 V (anodic, R).

Cyclic voltammetry was then performed on the fully deintercalated electrodes to determine the potential at which a transition metal, zinc, could be inserted into the framework. As shown in Figure S 1, Zn could be successfully intercalated (zinc plating occurs at ~−1 V vs Ag/AgCl thus was not responsible for the observed current). The product is formed via the reaction (M = Zn):

Mo6S8 + M2+(aq) + 2e− → MMo6S8

Scheme S 2 – Chemical equation of metal ion intercalation into Mo6S8

The area ratio of the anodic and cathodic peak currents was close to 1 indicating that the de/intercalation was fully reversible.

The Mo6S8 was intercalated by bulk electrolysis. A voltage of −0.7 V was applied for 20 min under hydrodynamic conditions in a 0.05 M ZnSO4 + 1 M NaCl electrolyte.

Sample Preparation for analyses Samples were prepared for various analyses as follows:

TEM – samples were sonicated in ethanol, then drop cast onto holey carbon grids.

SEM, XPS, XRD (of electrode materials) – the samples that were drop-cast on electrodes were removed by adhesion with, then analysed on, carbon tape.

Supplementary Data

115 Appendices

Electron Microscopy

Figure S 2 – Transmission electron micrographs of IL-MoS2 at different magnifications displaying short and curly MoS2 layers (right).

Figure S 3 – Transmission electron micrographs of Annealed IL-MoS2 at different magnifications (showing crystallisation of the disordered

MoS2 layers within the spheres relative to the IL-MoS2 in Figure S 2).

Figure S 4 – SEM of the n-CP electrode (active material + Nafion on carbon).

116 Appendices

Figure S 5 – SEM of the bulk-CP electrode (active material + Nafion on carbon).

Linear Sweep Voltammetry

Polarisation curves for the comparison catalysts are shown in Figure S 6. After annealing the IL-MoS2 at 700 °C for 60 h under a reducing atmosphere (equivalent to heat treatment in the n-CP synthesis), the activity decreased—as determined by the increased overpotential of the Ann-IL-MoS2 sample relative to the un-annealed IL-MoS2. The decrease in current density of the IL-MoS2 sample (Figure S 3) upon annealing is explained by a restructuring of the S–Mo–S layers from their delaminated form into stacks of sheets containing fewer defects. This morphological restructure decreases the ratio of edge to basal sites of the MoS2 which in turn decreases its activity.14 The analogous annealing process used to convert the delaminated IL-CuxMoSy into n-CP had the opposite effect on the activity. While the total surface area may be partially lowered by the loss of disordered nanosheets, the Chevrel Phases possess active sites at every facet of the crystal, rather than just edge sites, thus HER activity is retained.

117 Appendices

0 n-CP

) IL-MoS2 2 -5 Bulk-CP Glassy Carbon

Ann-IL-MoS2 -10 IL-CuxMoSy

Commercial-MoS2

-15 Current (mA/cm Current

-20 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 Potential (vs Ag/AgCl) Figure S 6 – Polarisation curves for various electrocatalysts performing HER in 1 M HCl normalised to geometric surface area.

X-ray Diffraction

Figure S 7 – Rietveld refinement of the n-CuCP XRD pattern. Pattern refined to ICSD: 602374 (Cu2Mo6S8, R -3 H) and 52267 (Mo, I m -3 m). Shown are the observed data (crosses), refined model (red), and residual (blue).

118 Appendices

Figure S 8– Rietveld refinement of the bulk-CuCP XRD pattern. Pattern refined to ICSD: 602374 (Cu2Mo6S8, R -3 H) and 52267 (Mo, I m -3 m). Shown are the observed data (crosses), refined model (red), and residual (blue).

We propose that the IL-CuxMoSy to n-CuCP topotactic transformation proceeds in-line with a previously identified mechanism15 (with a Mo + CuS intermediate) due to the presence of Mo reflections in the XRD (Figure 7).16 Trace Mo impurities remained in the sample after heat treatment at 700 °C, though were dissolved by subsequent treatment of the electrodes in acidic media (Figure S 10).

6.0

5.0

4.0 IL-CuxMoySz 3.0 IL-MoS2 MoS2 (ISCD: 24000) 2.0

1.0 Intensity (counts ×1000) (counts Intensity

0.0 20 40 60 80 °2Theta (Cu)

Figure S 9 – XRD diffractogram of the n-CP precursor (IL-CuxMoSy) and the IL-MoS2.

Ex-situ XRD analysis of the electrodes was hampered by both the limited amount of sample (8 ng), and the broad background features. Even so, the strong intercluster (0, 1, −1) reflection is clearly visible in each diffractogram (Figure S 10).

119 Appendices

Cu2Mo6S8

ZnMo6S8

Mo6S8

25 30 35 40 45 Intensity (Normalised) Intensity

10 20 30 40 50 60 70 80 90 °2Theta (Cu) Figure S 10 – XRD diffractograms of n-CP electrodes on carbon tape. Broad features at 20–25° 2-theta are due to the carbon substrate.

X-ray Photoelectron Spectroscopy XPS spectra of the nano CP samples were consistent with successful deintercalation. The XPS spectra were charge referenced to the adventitious C1s peak at 284.8 eV. The C 1s, S 2s and Mo 3d regions exhibit features characteristic of Chevrel Phases (

IL-CuxMo6S8

IL-ZnxMo6S8

90 IL-Mo S 80 6 8 Adventitious C 70 CF2 60 COOH COH 50 Backgnd. 40 Envelope

30 Counts / s / Counts 1000) (× 20

294 292 290 288 286 284 282 Binding Energy (E) (eV)

Figure S 11, Figure S 12, Figure S 13, and Figure S 15).6, 17-20 The Mo 3d5/2 signal is positioned at 227.9 eV, consistent with a reduced Mo oxidation state relative to common Mo4+ catalysts such as MoS2. The presence of Mo6+ may be due to surface oxidation of the catalyst due to handling in air, and is typical of XPS analysis.20 The Cu 2p region of the n-CuCP exhibits intense signals that are significantly diminished in intensity in the deintercalated Mo6S8 and ZnMo6S8 samples. The copper is present in a mix of oxidation states, however, these cannot be unambiguously assigned as the copper 2p peaks do not shift significantly as a function of oxidation state.21 The amount of copper in the deintercalated samples in all cases was ~

120 Appendices

0.7 atomic %, indicating that the electrochemical deintercalation process is reproducible, though not completely effective (even after the oxidative deintercalation current reached zero). The ZnxMo6S8 sample exhibits intense Zn peaks (2p3 binding energy = 1022.3, Δ=23.0) but again, analogous to the copper case, the oxidation states cannot be unambiguously assigned.22 It is highly likely that the intercalated transition metals exist in a 2+ oxidation state as prior studies on MgxMo6S8 report that Mg2+ and Mo2+ are observed upon intercalation.17

The XPS analysis is consistent with successful intercalation; a small, reduced Mo contribution is observed in the Mo 3d XPS spectrum for the intercalated samples (Figure S 12, ~226.5 eV). Unfortunately, limited deconvolution from the S 2s, other Mo–S, and Mo–O signals prevent the confident assignment of the species contributing to the spectrum.

IL-CuxMo6S8

IL-ZnxMo6S8

90 IL-Mo S 80 6 8 Adventitious C 70 CF2 60 COOH COH 50 Backgnd. 40 Envelope

30 Counts / s / Counts 1000) (× 20

294 292 290 288 286 284 282

Binding Energy (E) (eV)

Figure S 11 – XPS spectra of the C 1s region for the nano CP samples. Note: Fluorine peak arises from the Nafion binder.

121 Appendices

IL-ZnxMo6S8

IL-CuxMo6S8

70 IL-Mo6S8 MoS 3d5 60 MoO 3d5 50 S 2s Backgnd. 40 Envelope 30

20 Counts / s (×1000) 10

236 234 232 230 228 226 224 222 Binding Energy (E) (eV) Figure S 12 - XPS spectra of the Mo 3d region for the nano CP samples.

IL-ZnxMo6S8

IL-Mo6S8

90

85 IL-CuxMo6S8 80 Cu0 CuO 75 Backgnd. 70 Envelope

65 Counts / s (× 1000) Counts(× s / 60

965 960 955 950 945 940 935 930 925 Binding Energy (E) (eV) Figure S 13 – XPS spectra for the Cu 2p region of the nano CP samples.

122 Appendices

20 IL-ZnxMo6S8 Zn 2p Backgnd. 18

16

14 Counts / s (×1000) s / Counts 12

1050 1045 1040 1035 1030 1025 1020 1015 Binding Energy (E) (eV)

Figure S 14 – XPS spectra for the Zn 2p region of the n-ZnCP sample.

IL-ZnxMo6S8

IL-Mo6S8

26 24 IL-Cu Mo S 22 x 6 8 Mo-SH 20 Mo-SO3 18 Mo-S 16 Backgnd. 14 Envelope 12 10 Counts / s Counts (×1000) 8 6 174 172 170 168 166 164 162 160 158 Binding Energy (E) (eV) Figure S 15 - XPS spectra of the S 2s region for the nano CP samples.

Table S 1 – Representative XPS fit data for the nano CPs (ZnxMo6S8) with references used for peak assignment. Element Assignment Peak Binding Energy (eV) Peak separation (eV) FWHM (eV) Reference

123 Appendices

Adventitious 23 C C 284.8 1.5 Alcohols 286.3 1.6 Acids 288.7 1.5 24 C-F (Binder) 291.7 1.7

Cu Cu 2p 932.7 19.3 2.0 21, 25

Mo Mo-S 228.1 3.2 1.0 6, 17-20 26 Mo-O 229.2 3.3 2.0

S S-Mo 161.8 1.1 0.6 27 S-H 161.7 1.3 1.7 S-O 168.7 1.3 2.0 2s 226.1 2.0

Zn Zn 2p 1022.3 23.1 2.0 25

Figure S 16 – TEM micrographs of n-CP electrode material after the stability test (i.e. 1000 cycles). Right micrograph is a magnified image of the boxed area in the left micrograph. Note presence of crystallites of Mo6S8, as identified by the channeled structure. Sample prepared by sonicating coated electrode from glassy carbon surface in acetone, then drop casting onto a holey-carbon copper grid.

124 Appendices

Pre HER

Post HER Counts (normalised) Counts

100 200 300 400 500 600 -1 Wavenumber (cm ) Figure S 17 – Raman spectra of n-CP electrodes before and after the stability test (i.e. 1000 cycles).

750 V = -195 mV nCP bulk-CP

500 -Im(Z)/Ohm 250

0 0 250 500 750 1000 1250 1500 1750 Re(Z)/Ohm

Figure S 18 – Nyquist plot of impedance data for the nano and bulk Chevrel Phase samples.

110

100

90 Faradaic Efficiency (%) Efficiency Faradaic

0 50 100 150 200 250 300 Time (min)

Figure S 19 – Faradaic efficiency of the n-CP electrodes as measured by H2 gas chromatography.

125 Appendices

References 1. Walton, R. I.; Dent, A. J.; Hibble, S. J., In Situ Investigation of the Thermal Decomposition of Ammonium Tetrathiomolybdate Using Combined Time-Resolved X-ray Absorption Spectroscopy and X-ray Diffraction. Chem. Mater. 1998, 10 (11), 3737-3745. 2. Reid, R. S.; Clark, R. J.; Quagraine, E. K., Accurate UV–visible spectral analysis of thiomolybdates. Can. J. Chem. 2007, 85 (12), 1083-1089. 3. P. Binnie, W.; J. Redman, M.; J. Mallio, W., Preparation, properties, and structure of cuprous ammonium thiomolybdate. Inorg. Chem. 1970, 9. 4. Chevrel, R.; Sergent, M.; Prigent, J., Un nouveau sulfure de molybdene: Mo3S4 preparation, proprietes et structure cristalline. Mater. Res. Bull. 1974, 9 (11), 1487-1498. 5. Chevrel, R.; Sergent, M.; Prigent, J., Sur de nouvelles phases sulfurées ternaires du molybdène. J. Solid State Chem. 1971, 3 (4), 515-519. 6. Jiang, J.; Gao, M.; Sheng, W.; Yan, Y., Hollow Chevrel‐Phase NiMo3S4 for Hydrogen Evolution in Alkaline Electrolytes. Angew. Chem. 2016, 128 (49), 15466-15471. 7. Wei, C.; Rao, R. R.; Peng, J.; Huang, B.; Stephens, I. E.; Risch, M.; Xu, Z. J.; Shao‐Horn, Y., Recommended practices and benchmark activity for hydrogen and oxygen electrocatalysis in water splitting and fuel cells. Adv. Mater. 2019, 31 (31), 1806296. 8. Elgrishi, N.; Rountree, K. J.; McCarthy, B. D.; Rountree, E. S.; Eisenhart, T. T.; Dempsey, J. L., A practical beginner’s guide to cyclic voltammetry. J. Chem. Educ. 2018, 95 (2), 197-206. 9. McCrory, C. C.; Jung, S.; Ferrer, I. M.; Chatman, S. M.; Peters, J. C.; Jaramillo, T. F., Benchmarking hydrogen evolving reaction and oxygen evolving reaction electrocatalysts for solar water splitting devices. J. Am. Chem. Soc. 2015, 137 (13), 4347-4357. 10. Trasatti, S.; Petrii, O., Real surface area measurements in electrochemistry. Pure Appl. Chem. 1991, 63 (5), 711-734. 11. McCrory, C. C.; Jung, S.; Peters, J. C.; Jaramillo, T. F., Benchmarking heterogeneous electrocatalysts for the oxygen evolution reaction. J. Am. Chem. Soc. 2013, 135 (45), 16977-16987. 12. Alonso-Vante, N.; Schubert, B.; Tributsch, H., Transition metal cluster materials for multi-electron transfer catalysis. Mater. Chem. Phys. 1989, 22 (3), 281-307. 13. Lancry, E.; Levi, E.; Mitelman, A.; Malovany, S.; Aurbach, D., Molten salt synthesis (MSS) of Cu2Mo6S8—new way for large-scale production of Chevrel phases. J. Solid State Chem. 2006, 179 (6), 1879-1882. 14. Lau, V. W. h.; Masters, A. F.; Bond, A. M.; Maschmeyer, T., Promoting the Formation of Active Sites with Ionic Liquids: A Case Study of MoS2 as Hydrogen‐Evolution‐Reaction Electrocatalyst. ChemCatChem 2011, 3 (11), 1739-1742. 15. Gershinsky, G.; Haik, O.; Salitra, G.; Grinblat, J.; Levi, E.; Daniel Nessim, G.; Zinigrad, E.; Aurbach, D., Ultra fast elemental synthesis of high yield copper Chevrel phase with high electrochemical performance. J. Solid State Chem. 2012, 188, 50-58. 16. Cheng, Y.; Parent, L. R.; Shao, Y.; Wang, C.; Sprenkle, V. L.; Li, G.; Liu, J., Facile synthesis of Chevrel phase nanocubes and their applications for multivalent energy storage. Chem. Mater. 2014, 26 (17), 4904-4907. 17. Richard, J.; Benayad, A.; Colin, J.-F.; Martinet, S., Charge Transfer Mechanism into the Chevrel Phase Mo6S8 during Mg Intercalation. The Journal of Physical Chemistry C 2017, 121 (32), 17096-17103. 18. McCarty, K. F.; Schrader, G. L., Hydrodesulfurization by reduced molybdenum sulfides: activity and selectivity of Chevrel phase catalysts. Industrial & Engineering Chemistry Product Research and Development 1984, 23 (4), 519-524. 19. Ooi, S.; Zhang, H.; Hinode, H., The hydrodesulfurization activity and characterization of cobalt Chevrel phase sulfides. React. Kinet. Catal. Lett. 2004, 82 (1), 89-95. 20. Naik, K. M.; Sampath, S., Cubic Mo6S8-Efficient Electrocatalyst Towards Hydrogen Evolution Over Wide pH Range. Electrochim. Acta 2017, 252, 408-415. 21. Biesinger, M. C., Advanced analysis of copper X‐ray photoelectron spectra. Surf. Interface Anal. 2017, 49 (13), 1325- 1334. 22. Hammer, G.; Shemenski, R., The oxidation of zinc in air studied by XPS and AES. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1983, 1 (2), 1026-1028. 23. ISO/TC 201/SC 7, Surface chemical analysis‐X‐ray photoelectron spectroscopy‐Reporting of methods used for charge control and charge correction. In ISO 19318:2004, 2004. 24. Chen, C.; Levitin, G.; Hess, D. W.; Fuller, T. F., XPS investigation of Nafion® membrane degradation. J. Power Sources 2007, 169 (2), 288-295. 25. Biesinger, M. C.; Lau, L. W.; Gerson, A. R.; Smart, R. S. C., Resolving surface chemical states in XPS analysis of first row transition metals, oxides and hydroxides: Sc, Ti, V, Cu and Zn. Appl. Surf. Sci. 2010, 257 (3), 887-898. 26. Choi, J. G.; Thompson, L. T., XPS study of as-prepared and reduced molybdenum oxides. Appl. Surf. Sci. 1996, 93 (2), 143-149. 27. McCarty, K. F.; Anderegg, J.; Schrader, G. L., Hydrodesulfurization catalysis by Chevrel phase compounds. J. Catal. 1985, 93 (2), 375-387.

126 Appendices

Appendix to Chapter 2.3 Supplementary Information

Experimental

2 µm MoS2, 90 nm MoS2, Lithium chloride (LiCl) and Lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) were purchased from Sigma-Aldrich. THF was purchased from Merck and distilled before use to bring the moisture content below 15 ppm. Coin cell components including spacers, springs and cages were purchased from TOB New Energy, China. Whatman glassy fibre separators (GF/D) were purchased from Sigma Aldrich. All chemicals were used without further purification.

The electrode slurry was prepared by mixing 80% (mass ratio) MoS2, 10% polyvinylidene fluoride (as a binder) and 10% Super P (conductive agent) in N-Methyl-2-pyrrolidone (NMP, solvent). The slurry was subsequently coated on stainless-steel foil at 80 °C. It was further dried at 120 °C overnight before use.

The electrolyte was prepared by adding 2.0 M EtMgCl/THF dropwise to a flask containing t-amyl alcohol (t- AmOH). An ice bath was used to cool the strongly exothermic reaction. Gas evolution occured immediately upon adding EtMgCl/THF, as ethane gas is generated. EtMgCl/THF was added in slight excess (1.05:1,

EtMgCl: t-AmOH) to ensure the t-AmOH was fully consumed. The mixture was stirred for 12 h under an N2 atmosphere. AlCl3 powder was then slowly added into the resultant solution. The milky solution cleared after stirring overnight. The solution was stirred for a further 8 hours and transferred to glovebox for storage.

Morphological characterization of the MoS2 was carried out by a Zeiss Scanning Electron Microscope (SEM). Powder X-ray Diffraction (XRD) was collected using a PANalytical X'Pert Powder instrument using Cu X-ray source. Raman spectra were collected using a Bruker MultiRam FT-Raman Spectrometer with a 512 nm laser.

N2-sorption was performed using a Micromeritics ASAP 2020 Accelerated Surface Area and Porosity.

Ball-milling was performed using a Retsch PM 100 planetary ball mill. MoS2, 1 g, was loaded into an agate jar with 3 agate balls (diameter = 20 mm) and subject to milling at 450 rpm for the stated time.

Electrochemical operations were carried out inside a glovebox (oxygen and moisture concentrations were maintained below 1 ppm). Voltammetric experiments were carried out on an ADAQ ER466 Integrated Potentiostat System (Edaq, Australia). Cyclic voltammograms were obtained using a 1 mm diameter platinum working electrode (Edaq) with a magnesium wire as both counter and reference electrodes.

Coin cell tests were performed with a Neware Coin Battery Cycler (BTS - 5V 5 mA, Shenzhen, China) using constant current charging/discharging parameters.

Electrolyte development The dual metal ion electrolyte shows increased reduction and oxidation currents compared to the 0.8 M Mg electrolyte (at all concentrations tested, Figure S 20a). The current continues to increase as the concentration of LiCl increases from 0.25 M to 1.0 M. This is due to the increased conductivity of the electrolyte, as reflected in the slope of the stripping curve.1 The overpotential of reduction of the active Mg species ([Mg2Cl3]+) reduction and the anodic stability of this dual metal ions electrolyte remains unchanged, suggesting that no additional complexes have been generated by reaction of the Mg electrolyte and LiCl. The role of LiCl is therefore solely a Li+ source and supporting electrolyte.

127 Appendices

The anodic stability of the electrolyte (2.75V vs Mg) is sufficient for Li+ intercalation (which occurs below 2.0 V vs Mg (Figure S 20c). Cycling tests were performed using coin cells containing a single (Mg2+) or dual ion electrolyte (Mg2+/Li+) to compare the electrolytes (Figure S 20b). The single ion electrolyte cells show a discharge capacity of just 32 mAh/g even under a very small current density of 28.5 mA/g; the dual ion electrolyte cells exhibit significantly improved capacity and power, reaching 138 mAh/g capacity under a higher current density of 250 mA/g. The improved performance was also confirmed by the cyclic voltammetry (Figure S 20c), where the two intercalation peaks (at 1.4 and 1.8 V)have a greater current than when LiCl is omitted.

Because of its suitable performance, the 0.8 M Mg + 1.0 M LiCl was chosen for use in the current study.

Figure S 20 – (a) Cyclic voltammograms obtained with Pt working electrode in electrolyte containing 0.8 M Mg (black) and LiCl with concentration of 0.25 M (green) and 1.0 M (red), respectively. (b) The cycling performance of the battery containing 0.8 M Mg electrolyte only with current density of 28.5 mA/g (black) and battery containing 0.8 M Mg electrolyte and 1.0 M LiCl with current density of 250 mA/g (red). (c) Cyclic voltammogram collected with as-prepared coin cells containing single Mg electrolyte (black) and dual (Mg and Li) metal ion (red) with scan rates of 1.0 mV/s. (d) Illustration of the configuration of the Mg cell containing dual metal ions in the electrolyte.

Supplementary Data

128 Appendices

Figure S 21 – Scanning electron micrographs of MoS2 samples before and after ball-milling for various durations. a) 2µm-MoS2; b) 90nm-MoS2;

c) 4h-MoS2; d) 24h-MoS2; e) 60h-MoS2

Figure S 22 – Transmission electron micrographs of 90nm-MoS2 showing large, crystalline sheets of MoS2 that exhibit conventional hexagonal features.

1. Chen, L.; Zhao, S.; Liu, Y.; Horne, M.; Bond, A. M.; Zhang, J., Room temperature electrodeposition of metallic magnesium from ethylmagnesium bromide in tetrahydrofuran and ionic liquid mixtures. J. Electrochem. Soc. 2016, 163 (4), H3043-H3051.

129 Appendices

Appendix to Chapter 3 Catalyst Synthesis

Mo2C and Mo2C@Al2O3 were synthesised according to procedures described previously.1 Mo2C1−xNx@TiN was synthesised in an analogous fashion using the melamine–[Mo] coordination polymer: prepared according to procedures originally reported by Pang et al.2 The resulting precursor was co-ground with 20 nm cubic TiN (US Research Nanomaterials Inc.; Product# US2060) and subjected to heat treatment under flowing H2 in Ar (70 mL min-1; 2:5 ratio): ramp rate 5 °C/min to 650 °C, held for 1 h, allowed to cool to room temperature, then gas flow ceased and back diffusion of air allowed for several h before removal of sample from the furnace. Catalyst samples were stored in air at room temperature. All supported catalysts were prepared similarly.

In brief, several batches of unsupported Mo2C were synthesized by thermal decomposition of the melamine–[Mo] coordination polymer as described above. The averaged ratio of the mass of starting material to product was then used as a guide for the mass- ratio of the catalyst support to the melamine–[Mo] coordination polymer in the synthesis mixture, such that the resultant molybdenum loading of the catalyst was 30 wt%.

Catalysis As we have previously reported,1 batch reactions were carried out in a Parr reactor system (Hastelloy C), equipped with a mechanical stirrer, cooling loop and thermowell. The reactor was charged with powdered lignin (1.0 g), catalyst (0.5 g for supported catalysts, or 0.16 g for unsupported Mo2C), and absolute ethanol (100 mL). The reactor was sealed, purged with nitrogen thrice (pressurization to 50 bar, followed by release) and finally ca 20 psi pressure of nitrogen was introduced. The reaction was heated to 280 °C or 330 °C over a period of approximately 1.5 h with 700 rpm stirring, held at the desired temperature for up to 6 h, then cooled to room temperature. The cooling step was assisted by use of a water bath. The reactor was depressurized, unsealed and the mixture filtered at the pump. The residue was washed with ethanol several times and dried at 80 °C overnight. o-Cresol (20 µL) was added and the volume of the filtrate was adjusted to 200 mL volumetrically with ethanol. In between catalytic runs, the reactor was cleaned using supercritical methanol. In brief, the reactor was charged with 100 mL of methanol, purged with nitrogen, and heated to 250 °C (autogenous pressure >82 bar) and stirred for 30 min, then cooled, pressure let down, the reactor emptied and the process repeated with fresh methanol.

Reaction analysis Compound identification by GC-MS was performed on a Shimadzu GCMS-QP2010 equipped with Rtx-5MS 0.25 µm × 30 m × 0.25 mm column. The temperature program had an isothermal period of 3 min at 50 °C, followed by a heating ramp of 10 °C min–1 to 330 °C, followed by an isothermal period of 10 min. Compounds were identified by comparing the EI-MS spectra with those in the system’s database (NIST05). o-Cresol was used as an external standard. The calculated yield is reported as mg of analyte per g lignin.

X-ray Absorption Spectroscopy Molybdenum k-edge XAS data were collected in transmission mode at the XAS beamline at the Australian Synchrotron. The photon delivery system comprised a 1.9 T wiggler and step scanned Si(311) monochromator. A sample of molybdenum foil was simultaneously measured as a reference to account for beamline energy drift. Samples were packed into open-ended quartz capillaries with glass wool which allowed the flow of gases through the sample during analysis. To observe the reducing effect of ethanol at high temperature on the catalyst, N2 was bubbled through ethanol and hence the atmosphere passing over the samples contained ethanol vapour. The portion of the sample in the beam path was heated to 300 °C with a hot air blower (measured using a thermocouple). The XAS data was analysed using the IFEFFIT packages of Athena and Artemis.

Variable Temperature X-ray Diffraction Samples were loaded into 0.5 mm open ended glass capillaries and analysed using a Stoe Stadi P diffractometer equipped with a Mo X-ray source, operating in transmission mode (Debye–Scherrer geometry) and equipped with an Oxford Cryosystems Cryostream (80–500 K) and FMB Oxford hot air blower (up to 1000 °C). The data were analysed by the Rietveld structure refinement program GSAS II. Variable temperature XRD were plotted using Origin 9.6.5.

130 Appendices

X-ray Photoelectron Spectroscopy XPS was performed on a ThermoFisher K-Alpha X-ray Photoelectron Spectrometer equipped with a Al Ka source and a 180° double focusing hemispherical analyser-128-channel detector. Samples were prepared by first pelletising the powders, then loading onto a sample holder with carbon tape. Preparation took place in air. The binding energy was adjusted to 284.8 eV based on the adventitious aliphatic carbon 1s peak.

Microscopy and Microanalysis Bright field Transmission Electron Microscopy was taken using a JEM-2100 at 200 kV.

Energy Dispersive X-ray Spectroscopy was taken on a JEM-2200FS (200 kV) equipped with a Bruker EDS SDD, with an energy resolution of 129 eV, we applied a hard X-ray aperture to improve the quality of the EDS.

Electron energy loss spectroscopy was taken on a JEM-2200FS (200kV) equipped with an in-column omega filter, at an energy resolution of 1 eV, using condenser lens #3, with a convergence angle of approximately 38 mrad and a collection angle of approximately 170 mrad.

Supplementary Data

131 Appendices

Figure S23. Mo K-edge oxidation state vs edge energy comparison of the literature with this work. Kopachevska et al.3, Farges et al.4

Figure S24. Displaying the fitting of k3-weighted Mo K-edge Fourier transform EXAFS spectra (left with imaginary part) and extracted EXAFS ° (right) for Mo2C at 25 C. (Blue – experimental data; Red – fitting analysis given in Table S1).

° Table S2. Full details of the Mo K-edge EXAFS fitting analysis of Mo2C at 25 C. Variables indicated by standard deviation in brackets. Path Distance (Å) σ2 3 Mo–C 2.03(1), 2.09(1), 2.13(1) 0.000(1) 12 Mo–Mo 2.91(1)–3.00(1) 0.004(1) 6 Mo–Mo 4.15(1)–4.19(1) 0.001(3) 8 Mo–C–Mo 4.20(1)–4.22(1) 0.001(2) −1 S02 = 0.85. E0 = −6(1) eV. K-weight = 3. k-range = 2–15 Å . R-range 1.2–4.5 Å.

132 Appendices

Figure S25. Displaying the fitting of k3-weighted Mo K-edge Fourier transform EXAFS spectra (left with imaginary part) and extracted EXAFS

(right) for Mo2C at 300 °C. (Blue – experimental data; Red – fitting analysis given in Table S2).

° Table S3. Full details of the Mo K-edge EXAFS fitting analysis of Mo2C at 300 C. Variables indicated by standard deviation in brackets. Path Distance (Å) σ2 3 Mo–C 2.03(1), 2.09(1), 2.12(1) 0.001(1) 12 Mo–Mo 2.91(6)–3.01(6) 0.009(1) 6 Mo–Mo 4.1(7)–4.1(6) 0.020(18) 8 Mo–C–Mo 4.20(6)–4.22(6) 0.010(3) −1 S02 = 0.85. E0 = −6(1) eV. K-weight = 3. K-range = 2–10 Å . R-range 1.2–4.5 Å.

Figure S26. Displaying the fitting of k3-weighted Mo K-edge Fourier transform EXAFS spectra (left with imaginary part) and extracted EXAFS ° (right) for Mo2C1−xNx@TiN at 25 C. (Blue – experimental data; Red – fitting analysis given in Table S3).

° Table S4. Full details of the Mo K-edge EXAFS fitting analysis of Mo2C1−xNx@TiN at 25 C. Variables indicated by standard deviation in brackets. Path Distance (Å) σ2 1 Mo–N 1.66(1) 0.001(1) 3 Mo–C 2.12(1) 0.002(3) 6 Mo–Mo 2.92(1)–3.02(1) 0.005(1) 3 Mo–Mo 4.25(1), 4.27(1), 4.29(1) 0.003(1) 8 Mo–C–Mo (MS) 4.28(2)–4.30(2) 0.009(2) −1 S02 = 0.85. E0 = −5(1) eV. K-weight = 3. K-range = 2–15 Å . R-range 1.2–4.5 Å.

133 Appendices

Figure S27. Displaying the fitting of k3-weighted Mo K-edge Fourier transform EXAFS spectra (left with imaginary part) and extracted EXAFS °C (right) for Mo2C1−xNx@TiN at 300 . (Blue – experimental data; Red fitting analysis given in Table S4).

° Table S5. Full details of the Mo K-edge EXAFS fitting analysis of Mo2C1−xNx@TiN at 300 C. Variables indicated by standard deviation in brackets. Path Distance (Å) Debye-Waller Factor 1 Mo–N 1.63(4) 0.002(3) 3 Mo–C 2.08(2) 0.001(1) 4 Mo–Mo 2.91(2)–3.00(2) 0.003(1) 8 Mo–C–Mo (MS) 4.30(2) 0.016(3) −1 S02 = 0.85. E0 = −5(4) eV. K-weight = 3. K-range = 2–10 Å . R-range 1.2–4.5 Å.

Figure S28. Rietveld fit of the Mo2C1−xNx@TiN sample at 25 °C

134 Appendices

Figure S29. Rietveld fit of Mo2C@Al2O3 sample

Table S6. XPS fit details for the Mo2C1−xNx@TiN sample Binding Energy (eV) Component Mo 3d5/2 Mo 3d3/2 C 1s N 1s Mo 3p Ti 2p3/2 Ti 2p1/2 Reference 228.7 231.9 282.4 5-7 Mo2C (1.2) (1.2) (1.7) 394.9 229.5 232.7 397.5 (2.5) MoN (1.5) (1.5) (0.79) 7-8 231.2 234.5 MoO2 (1.6) (1.6) 398.8 7, 9 232.8 235.8 (3.5) MoO3 (1.8) (1.8) 7 397.11 455.3 461.0 TiN (1.7) (1.1) (1.1) 10-11 459.0 464.8 TiO2 (1.3) (1.3) 11-12 471.7 TiO2 Satellite (3) 12 457.9 464.1 TiN Shake up (3.0) (3) 10 284.6 Aliphatic Carbon (1.8) 13 286.3 Alcohol/Ether (1.8) 14 289.2 Ester/Acid (1.8) 14 287.7 Carbonyl (1.8) 14

135 Appendices

Figure S30. XPS plot of the Mo2C1−xNx@TiN sample in the C 1s region showing the effect of etching on the relative intensity of species associated with surface oxidation (290–286 eV) to aliphatic and carbide species (286–281 eV)

Figure S31. Bright field TEM image of the Mo2C1−xNx@TiN sample. Scale bar (100 nm) shown in bottom right.

136 Appendices

Figure S32. EDS analysis of the Mo2C1−xNx@TiN sample showing a single particle adjacent to a few overlaying particles.

Figure S33. EDS analysis of the Mo2C1−xNx@TiN sample at lower magnification than Figure S32.

137 Appendices

Figure S34. EDS analysis of the Mo2C1−xNx@TiN sample.

Figure S35. Electron Energy Loss Spectroscopy (EELS) of Mo2C1−xNx@TiN sample.

138 Appendices

Figure S36. Variable Temperature XRD of unsupported Mo2C heated in air. Note the absence of any significant change over the 25–500 °C temperature range

Figure S37. Variable Temperature XRD of Mo2C@Al2O3

139 Appendices

References 1. Cattelan, L.; Yuen, A. K. L.; Lui, M. Y.; Masters, A. F.; Selva, M.; Perosa, A.; Maschmeyer, T., Renewable Aromatics from Kraft Lignin with Molybdenum-Based Catalysts. ChemCatChem 2017, 9 (14), 2717-2726. 2. Pang, M.; Wang, X.; Xia, W.; Muhler, M.; Liang, C., Mo(VI)−Melamine Hybrid As Single-Source Precursor to Pure-Phase β‐ Mo2C for the Selective Hydrogenation of Naphthalene to Tetralin. Ind. Eng. Chem. Res. 2013, 52, 4564–4571. 3. Kopachevska, N.; Melnyk, A.; Bacherikova, I.; Zazhigalov, V.; Wieczorek-Ciurowa, K., Determination of molybdenum oxidation state on the mechanochemically treated MoO3. Хімія, фізика та технологія поверхні 2015, (6,№ 4), 474-480. 4. Farges, F.; Siewert, R.; Brown Jr, G. E.; Guesdon, A.; Morin, G., Structural environments around molybdenum in silicate glasses and melts. I. Influence of composition and oxygen fugacity on the local structure of molybdenum. The Canadian Mineralogist 2006, 44 (3), 731-753. 5. Brainard, W. A.; Wheeler, D. R., An XPS study of the adherence of refractory carbide silicide and boride rf‐sputtered wear‐ resistant coatings. Journal of Vacuum Science and Technology 1978, 15 (6), 1800-1805. 6. Castañeda, S. I.; Montero, I.; Ripalda, J. M.; Dıaz,́ N.; Galán, L.; Rueda, F., X-ray photoelectron spectroscopy study of low- temperature molybdenum oxidation process. J. Appl. Phys. 1999, 85 (12), 8415-8418. 7. Choi, J.-G., Influence of surface composition on HDN activities of molybdenum nitrides. Journal of Industrial and Engineering Chemistry 2002, 8 (1), 1-11. 8. Sanjinés, R.; Wiemer, C.; Almeida, J.; Levy, F., Valence band photoemission study of the Ti Mo N system. Thin Solid Films 1996, 290, 334-338. 9. Tenretnoel, C.; Verbist, J.; Gobillon, Y., Oxidation-States of Molybdenum Oxides and Mixed-Oxide C2Mo1 by Photoelectron-Spectroscopy. Journal de Microscopie et de Spectroscopie Electroniques 1976, 1 (2), 255-262. 10. Jaeger, D.; Patscheider, J., A complete and self-consistent evaluation of XPS spectra of TiN. J. Electron Spectrosc. Relat. Phenom. 2012, 185 (11), 523-534. 11. Chan, M.-H.; Lu, F.-H., X-ray photoelectron spectroscopy analyses of titanium oxynitride films prepared by magnetron sputtering using air/Ar mixtures. Thin Solid Films 2009, 517 (17), 5006-5009. 12. Erdem, B.; Hunsicker, R. A.; Simmons, G. W.; Sudol, E. D.; Dimonie, V. L.; El-Aasser, M. S., XPS and FTIR surface characterization of TiO2 particles used in polymer encapsulation. Langmuir 2001, 17 (9), 2664-2669. 13. Moulder, J. F.; Chastain, J., Handbook of X-ray Photoelectron Spectroscopy: A Reference Book of Standard Spectra for Identification and Interpretation of XPS Data. Physical Electronics Division, Perkin-Elmer Corporation: 1992. 14. Barr, T. L.; Seal, S., Nature of the use of adventitious carbon as a binding energy standard. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1995, 13 (3), 1239-1246.

140 Appendices

Appendix to Chapter 4.1

Experimental Chemicals – The following chemicals were used in the experiments: trisodium citrate (Univar), silver nitrate (Univar), hydrogen peroxide (30%, Univar), potassium bromide (Univar), sodium borohydride (Aldrich), 4- nitrophenol (Fluka), Oxygen (BOC), Argon (BOC) and 4-Nitrosophenol (TCI), and were used as received. 1,4- Benzoquinone (Aldrich) was sublimed before use. 4-Aminophenol (Merck) was recrystallised from ethanol under a nitrogen atmosphere. All solutions were prepared with Milli-Q water supplied by a Merck Milli-Q Advantage A10. Solutions of sodium borohydride and hydrogen peroxide were kept ice-cold to slow decomposition.

Instrumentation – UV-vis data were collected using a UV-vis – Cary 60. ACPI-MS was conducted using a Bruker amazon SL with no manipulation of samples prior to analysis.

Synthesis – Nanoparticle preparation was adapted from a report by Kitaev et al..1 Into a sample vial, aliquots of trisodium citrate (0.20 mL, 0.0125 M), silver nitrate (5.00 mL, 3.75 × 10−4 M), water (2.30 mL) and hydrogen peroxide (5.00 mL, 5.00 × 10−2 M) were pipetted. Stirring was started (500 RPM) then sodium borohydride (2.50 mL, 5.00 × 10−3 M) was added. The sample was stirred for 30 minutes or until the solution colour had stabilised; at ambient temperature. Experiments performed in this manuscript were performed with one of two batches, as noted in Table S 4.

Catalysis – An aqueous solution of 4-nitrophenol (1.00 mL, 1.05 × 10−4 M) and aqueous sodium borohydride (100 µL, 0.25 M) were added to a disposable cuvette. To this solution, an aliquot of the AgNPs (1.00 mL, 8.75 × 10−6 M) was added and the reaction mixture monitored over a range of wavelengths (1 or 2 nm intervals, 4800 nm/min) or at λ 400 nm. The reactions were carried out at ambient conditions. The time between borohydride addition and catalyst addition was kept constant at 30 seconds. Gases (Ar/O2) was bubbled into the cuvettes using a balloon equipped with a needle. In other tests, H2O2 (0.05 mL, 5.5 mM) was added to the reaction mixture. For systems using Pt, Pd, Cu or Ni, the catalysts were generated in-situ by reduction of a metal salt: Pt – PtCl2 (10 μM), Pd – PdCl2 (0.2 μM), Cu – Cu(NO3)·3H2O (0.2 μM), Ni – Ni(NO3)2 (10 μM). Concentrations of all other species were kept constant. See Table S4 for details.

Work-up – Reaction mixture was worked up by one of three methods: 1. Extraction into ethyl acetate, 2. Neutralization of the reaction mixture followed by extraction into ethyl acetate, 3. Removal of water by rotary evaporation followed by redissolution in ethyl acetate.

Additional characterization TEM, UV-Vis and SAED (Selected Area Electron Diffraction) data for the AgNPs shown in Figures S19-21.

Modelling Our proposed mechanism was modelled using Dynafit.2 This program performs nonlinear least squares regression on differential equations derived from a user supplied mechanism. The mechanism and rate constants are given in Table S 1.

Table S 7 – Rate constants for the modelled reaction (Figure 6, Scheme 1)

141 Appendices

Elementary step Rate constant (k) Fit? Standard Error Notes Scheme 1 PreCatalyst + 4NP− → Cat-4NP 3.9 × 106 Y 4200 Reporter reaction (see 3) − − 6 4NO + O2 → 4NP 1.6 × 10 Y 4200 -40 1/2O2 + “H2” → H2O 1.1 × 10 Y 4200 4 Cat-4NP + “H2” → 4NO− + PreCatalyst 1.7 × 10 Y 55 + H2O -5 -5 4 -9 -2 Initial concentrations: 4NP – 5.3e M, O2 – 3.3e M (from ), PreCatalyst – 9.3e M, 4NO – 0 M, H – 1.0e M.

Scheme 2 DeactivatedCatalyst + H− → PreCatalyst 6.3 × 104 Y 9600 10 O2 + PreCatalyst → DeactivatedCatalyst 1.6 × 10 Y 9600 PreCatalyst + 4NP− → Cat-4NP 6.6 × 107 Y 9600 Reporter reaction (see 3) − 10 Cat-4NP + “H2” → 4NO + PreCatalyst 1.4 × 10 Y 9600

+ H2O Initial concentrations: 4NP – 5.3e-5 M, O2 – 9.8e-6 M (from 4), PreCatalyst – 3.7e-10 M, 4NO – 0 M, H – 1.0e-2 M.

Steps regarded as insignificant or irrelevant were omitted from the modelling so as to keep the number of variables to a minimum. These are listed in Table S 2.

Table S 8 - Steps omitted from the Dynafit Model Step Notes NaBH4 + 2H2O → NaBO2 + Rate of decomposition is negligible compared to catalytic consumption of 2H2 NaBH4 4NO + Cat + 3H- → 4AP + Cat This step was omitted because it occurs after the reporter reaction. + H2O

Table S 9 – Variant mechanisms tested. Data shown in Figure S 3. Mechanism A Mechanism B Mechanism C Cat + 4NP → 4NP-Cat Cat + 4NP → 4NP-Cat Cat + 4NP → 4NP-Cat - ½O2 + 4NO → 4NP 4NP-Cat + 3H → Cat + 4NO + 4NP-Cat + H- → Cat + 4NO 4NP-Cat + 3H- → Cat + 4NO + H2O 4NO + Cat + 3H- → Cat + 4AP H2O + H2O 4NO + Cat + 3H- → Cat + 4AP 4AP + O2 → 4NP + H2O 4AP + O2 → 4NP Notes: reoxidation of 4NO Notes: 4NO oxidation removed Notes: additional step added. cf. 4 removed. Other models were considered during modelling, notably those proposed by Finke et al.,3, 5 however our mechanism does not include any autocatalytic steps (a key feature of the work of Finke et al.). Therefore, although our data may appear visually similar to Finke’s cyclohexene hydrogenation reaction profile, we interpret the two reactions to be fundamentally different.

Table S 10 – Reaction Conditions Reaction [4NP] [Ag NPs] [NaBH4] Atmosphere Temp Other (M) (M) (M) (°C)

142 Appendices

Fig 2 5.0 ×10-5 1 .0×10-8 1.0 ×10-2 Air 25 ± 2 Catalyst batch 1 Fig 5 5.0 ×10-5 1 .0×10-8 1.0 ×10-2 Air 25 ± 2 Catalyst batch 2 Fig S1 5.0 ×10-5 1 .0×10-8 1.0 ×10-2 Air 25 ± 2 Catalyst batch 1 Fig 5.0 ×10-5 1 .0×10-8 2.5 ×10-3 Air 25 ± 2 Catalyst batch 1 S3(L) Fig 5.0 ×10-5 1 .0×10-8 1.0 ×10-2 Air 25 ± 2 Catalyst batch 1 S3(R) Fig S5 5.0 ×10-5 - 1.0 ×10-2 Air 25 ± 2 See Exptl for [metal salts] Fig S6 5.0 ×10-5 - 1.0 ×10-2 Air 25 ± 2 See Exptl for [metal salts] Fig S7 5.0 ×10-5 - 1.0 ×10-2 Air 25 ± 2 See Exptl for [metal salts] Fig S8 5.0 ×10-5 1 .0×10-8 1.0 ×10-2 Air 25 ± 2 Catalyst batch 1 Fig S9 5.0 ×10-5 varied 1.0 ×10-2 Air 25 ± 2 Catalyst batch 1 Fig S10 5.0 ×10-5 - - Air 25 ± 2 [NaOH] = 1.0 ×10-2 Fig S11 - - - Air 25 ± 2 [4NO] = 5.0 ×10-5, [NaOH] = 1.0 ×10-2 Fig S12 - - - Air 25 ± 2 [4AP] = 5.0 ×10-5, [NaOH] = 1.0 ×10-2 Fig S13 - 1 .0×10-8 1.0 ×10-2 Air 25 ± 2 [4AP] = 5.0 ×10-5 M, Catalyst batch 2 Fig S14 5.0 ×10-5 1 .0×10-8 1.0 ×10-2 Air 25 ± 2 Catalyst batch 2 Fig S15 5.0 ×10-5 1 .0×10-8 1.0 ×10-2 Air 25 ± 2 Catalyst batch 2 Fig S16 - 1 .0×10-8 - Air 25 ± 2 [4NO] = 5.0 ×10-5, [NaOH] = 1.0 ×10-2, Catalyst batch 2 Fig S17 - 1 .0×10-8 - Air 25 ± 2 [4NO] = 5.0 ×10-5, [NaOH] = 1.0 ×10-2, Catalyst batch 2

Supplementary Figures

143 Appendices

180 A B 150

120

90 4NP Time (s) Time 60 NaBH4

30 AgNPs

0 300 400 500 600 700 800 500 550 600 650 700 750 180

C D Intensity(arb.) 150

120

90 4AP Time (s) Time 60 NaBH4

30 AgNPs

0 300 400 500 600 700 800 500 550 600 650 700 750 Wavelength (nm) Wavelength (nm)

Figure S 38 – Time course UV-vis contour plots of preparation and stages of 4NP reduction (A, B) and 4AP addition (C, D) at standard conditions to monitor the AgNP plasmon. Addition of AgNPs, NaBH4 and either 4NP (A) or 4AP (C) at 30, 60 and 90 s respectively; normalised subregions (500 – 750 nm, B and D). Sharp peaks along the dotted lines are due to physical beam interference by the pipette during aliquot addition.

A change in the plasmon resonance was observed upon sequential addition of NaBH4 and 4NP, then again after reaction completion (~ 125 s). During the catalytic reduction of 4NP the lambda max of the plasmon does not shift significantly. After complete consumption of 4NP, there is a rapid and significant blue shift of the plasmon lambda max. This suggests that changes to the catalyst are insignificant during catalysis. When 4AP was added (instead of 4NP; C and D), a similar shift in the plasmon was observed, indicating that the final plasmon lambda max in both tests is due to 4AP-bound AgNPs.

144 Appendices

Figure S 39–UV-vis spectra of 4-nitrosophenol (black), 4-nitrophenol (red), 4-aminophenol (blue), benzoquinone (green) and hydroquinone (purple) in water.

1.2 2 1:50 4NP:NaBH4 1:200 4NP:NaBH4 1.0 Linear fit 1 ln(x) Linear fit 0 0.8 -1 0.6 -2 0.4 -3

Absorbance (arb.) Absorbance 0.2 -4 0.0 -5 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 Time (min) Time (min)

Figure S 40- Representative kinetic profile plots of different catalysis regimes. Left, zero order behaviour; right, first order behaviour.

145 Appendices

Figure S 41 – Data (black circles) and model (red) of the catalytic 4NP reduction fit. Left, mechanism A; Centre, mechanism B; Right, mechanism C in Table S 3. Every 4th data point shown for clarity.

1.0 1.00 Ag Ag 0.8 0.75

0.6 0.50

0.4 Absorbance (arb.) Absorbance Absorbance (arb.) Absorbance 0.25 0.2

0.0 0.00 300 350 400 450 0 30 60 90 120 150 180 210 Wavelength (nm) Time (s) 1.00 0.8 Au Au

0.75 0.6

0.50

0.4 Absorbance (arb.) Absorbance Absorbance (arb.) Absorbance 0.2 0.25

0.0 0.00 300 350 400 450 0 30 60 90 Wavelength (nm) Time (s) 1.0 1.00 Cu Cu 0.8 0.75

0.6 0.50

0.4 Absorbance (arb.) Absorbance Absorbance (arb.) Absorbance 0.25 0.2

0.0 0.00 300 350 400 450 0 30 60 90 120 150 180 210 Wavelength (nm) Time (s)

Figure S 42 – Time course UV-vis spectra and kinetic profile data for 4-nitrophenol reduction using silver, gold and copper nanoparticles. Arrow indicates introduction of O2(g) into reaction solution. Left and right column are separate catalytic runs for each metal to optimise spectrum range (left) and time resolution (right).

146 Appendices

0.8 1.0 Ni Ni 0.8 0.6 0.6

0.4 0.4

0.2 0.2 0.0

0.0 -0.2

Absorbance (arb.) adjusted (arb.) Absorbance -0.4 -0.2 -0.6 300 350 400 450 nm 400 at ) (arb. absorbance Adjusted 0 30 60 90 120 150 Wavelength (nm) Time (s) 1.0

0.8 Pd Pd 0.8

0.6 0.6

0.4

0.4 Absorbance (arb.) Absorbance

0.2 0.2 Absorbance (arb.) at 400 nm 400 at (arb.) Absorbance

0.0 0.0 300 350 400 450 0 30 60 90 120 150 180 210 240 270 300 Wavelength (nm) 0.8 0.8 Time (s) Pt Pt 0.6 0.6 0.4

0.2 0.4

0.0 0.2

Absorbance (arb.) adjusted (arb.) Absorbance -0.2

-0.4 0.0 300 350 400 450 nm 400 at ) (arb. absorbance Adjusted 0 30 60 90 120 150 180 210 240 270 300 330 Wavelength (nm) Time (s) Figure S 43- Time course UV-vis spectra and kinetic profile data for 4-nitrophenol reduction using nickel, palladium and platinum nanoparticles. Arrow indicates introduction of O2(g) into reaction solution. Adjusted data are aligned to the isosbestic point at 340 nm to account for the changing catalyst absorbance. Original data for Ni and Pt are shown inFigure S 44. Left and right column are separate catalytic runs for each metal to optimise spectrum range (left) and time resolution (right).

1.0

1.0 Ni (original) Ni (original)

0.8 0.8

0.6

0.6

0.4 Absorbance (arb.) Absorbance

0.2 Absorbance (arb.) at 400 nm at 400 (arb.) Absorbance 0.4

0.0

300 350 400 450 0 30 60 90 120 150 Wavelength (nm) Time (s) 1.6 1.0 Pt (original) Pt (original) 1.4

1.2 0.8

1.0

0.8 0.6

0.6 Absorbance (arb.) Absorbance 0.4

0.4 Absorbance (arb.) at 400 nm 0.2

0.0 0.2 300 350 400 450 0 30 60 90 120 150 180 210 240 270 300 330 360 Wavelength (nm) Time (s) Figure S 44- Time course UV-vis spectra and kinetic profile data for 4-nitrophenol reduction using nickel and platinum nanoparticles. Arrow

indicates introduction of O2(g) into reaction solution. Left and right column are separate catalytic runs for each metal to optimise spectrum range (left) and time resolution (right).

147 Appendices

1.0

N2 Aliquot

O2 Aliquot 0.8 Ar Bubbled

0.6

0.4 Absorbance(arb.)

0.2

0 60 120 180 240 300 Time (s) Figure S 45 –Kinetic profile data for 4-nitrophenol reduction using silver nanoparticles. Arrow indicates introduction of gas sparged aliquot of water or gas bubbles into reaction solution.

0.06 -0.5 Mol% Catalyst y = -0.0173x - 0.17 32 -1.1 16 8 -1.6

0.04 4 )

2 5 -2.2

(×10 -2.7

app app k 0.02 -3.2

4NP Concentration 4NP (mM) -3.8

-4.3 0 0 2 4 6 8 0 5 10 15 20 25 30 35 Time (min) Mol% Catalyst

Figure S 46– Apparent rate constant as a function of catalyst loading.

148 Appendices

Figure S 47- APCI-MS spectra of a control sample of 4NP– (positive – blue, negative – purple).

149 Appendices

Figure S 48- APCI-MS spectra of a control sample of 4NO– (positive – red, negative – green).

150 Appendices

Figure S 49- APCI-MS spectra of a control sample of 4AP– (positive – purple, negative – orange).

151 Appendices

– – Figure S 50- APCI-MS spectra of a sample of 4AP + OH + AgNPs post-O2 administration.

152 Appendices

Figure S 51- APCI-MS spectra of a one day old sample of a completed reaction mixture (positive – green, negative – blue). Notably absent, signals at 110, 124 and 140 m/z due to 4AP, 4NO and 4NP.

153 Appendices

Figure S 52- APCI-MS spectra of sample taken partway (10 seconds) through catalysis (positive – red, negative – green).

154 Appendices

– – Figure S 53- APCI-MS spectra of a control sample of 4NO + OH + AgNPs pre-O2 administration (positive – yellow, negative – teal). Presence of 4NP– (–138 m/z), likely due to the reaction of 4NO– and dissolved oxygen on catalyst surface.

155 Appendices

Figure S 54 – APCI-MS spectra of 4NO + OH- + AgNPs after administration of oxygen (positive – purple, negative – green).

156 Appendices

Figure S 55– UV-Vis of as-synthesised AgNPs.

Figure S 56– TEM of as-synthesised AgNPs (left) face-on perspective, (right) edge on perspective of stacked platelets.

157 Appendices

Figure S 57– Selected Area Electron Diffraction of as-synthesised AgNPs indexed to Fm-3m (225) silver.

Figure S 58– TEM of post-catalysis AgNPs (left) 200 nm scale, (right) magnified region, 100 nm scale. Note: particles appeared slightly distorted as they were covered in quinoid polymer (decomposition products of 4AP) due to having not being washed between catalysis and TEM sample preparation.

The citrate and peroxide promote shape selective nanoparticle synthesis. Triangular platelet formation is favoured due to faster growth along planar twinning defects compared with the citrate/stabilised (111) facet.7

Ag2O crystallites were not detected by SAED analysis, consistent with literature reports of stable Ag0 particles in aqueous solutions.8

158 Appendices

References 1. Frank, A. J.; Cathcart, N.; Maly, K. E.; Kitaev, V., Synthesis of silver nanoprisms with variable size and investigation of their optical properties: a first-year undergraduate experiment exploring plasmonic nanoparticles. J. Chem. Educ. 2010, 87 (10), 1098- 1101. 2. Kuzmič, P., Program DYNAFIT for the analysis of enzyme kinetic data: application to HIV proteinase. Anal. Biochem. 1996, 237 (2), 260-273. 3. Besson, C.; Finney, E. E.; Finke, R. G., Nanocluster nucleation, growth, and then agglomeration kinetic and mechanistic studies: a more general, four-step mechanism involving double autocatalysis. Chem. Mater. 2005, 17 (20), 4925-4938. 4. Menumerov, E.; Hughes, R. A.; Neretina, S., Catalytic reduction of 4-nitrophenol: a quantitative assessment of the role of dissolved oxygen in determining the induction time. Nano Lett. 2016, 16 (12), 7791-7797. 5. Watzky, M. A.; Finke, R. G., Transition metal nanocluster formation kinetic and mechanistic studies. A new mechanism when hydrogen is the reductant: Slow, continuous nucleation and fast autocatalytic surface growth. J. Am. Chem. Soc. 1997, 119 (43), 10382-10400. 6. Zheng, X.; Zhou, X.; Ji, X.; Lin, R.; Lin, W., Simultaneous determination of ascorbic acid, dopamine and uric acid using poly (4-aminobutyric acid) modified glassy carbon electrode. Sensors and Actuators B: Chemical 2013, 178, 359-365. 7. Frank, A. J.; Cathcart, N.; Maly, K. E.; Kitaev, V., Synthesis of silver nanoprisms with variable size and investigation of their optical properties: a first-year undergraduate experiment exploring plasmonic nanoparticles. J. Chem. Educ. 2010, 87 (10), 1098- 1101. 8. Gallardo, O. A. D.; Moiraghi, R.; Macchione, M. A.; Godoy, J. A.; Pérez, M. A.; Coronado, E. A.; Macagno, V. A., Silver oxide particles/silver nanoparticles interconversion: susceptibility of forward/backward reactions to the chemical environment at room temperature. RSC Advances 2012, 2 (7), 2923-2929.

159 Appendices

Appendix to Chapter 4.2 Resources for Instructors Recommended Reading Nanoparticles – References of main text, especially A. J. Frank, N. Cathcart, K. E. Maly and V. Kitaev, J. Chem. Educ., 2010, 87, 1098–1101 and references therein.

Optical properties of Nanoparticles – K. L. Kelly, E. Coronado, L. L. Zhao and G. C. Schatz, J. Phys. Chem. B, 2003, 107, 668–677

Beer-Lambert law – Most first year textbooks e.g. A. Blackman, S. Bottle, S. Schmid, M. Mocerino and U. Wille, John Wiley & Sons 2006.

Catalysis and Kinetics – Most first year textbooks e.g. A. Blackman, S. Bottle, S. Schmid, M. Mocerino and U. Wille, John Wiley & Sons 2006.

Pseudo Rate Laws – C. Vallance, Morgan & Claypool Publishers 2017 for a primer or J. W. Moore and R. G. Pearson, Kinetics and mechanism, John Wiley & Sons 1981 for more detail.

Safety Information Risk Assessment and Control Although some reagents used in this synthesis are hazardous, their dilution in water significantly reduces risk. Follow best practices for laboratory safety (wearing safety glasses and a lab coat).

This experiment is low risk.

Very low concentrations of silver nitrate, hydrogen peroxide and sodium borohydride are used in this experiment but contact of any reagent with skin/eyes should be avoided and the skin washed thoroughly with water in case of skin contact.

Vial lids should not be tightly fastened as H2 gas may evolve from decomposing sodium borohydride.

All electronic devices should be used well away from any experimental chemistry.

Chemical Hazard Identification silver nitrate – hazardous, contact with combustible material may cause fire, skin contact can result in burns, low to moderate toxicity, very toxic to aquatic organisms. Non-hazardous at the concentrations used in this practical. trisodium citrate tribasic dihydrate – non-hazardous. hydrogen peroxide – oxidizing. Non-hazardous at the concentrations used in this practical (<3%). potassium bromide – non-hazardous. sodium borohydride – solid material produces flammable gas in contact with water, causes burns and is toxic if swallowed. Non-hazardous at the concentrations used in this practical. silver nanoparticles – hazards not well established and should be handled with care. 4-nitrophenol – harmful by inhalation, in contact with skin and if swallowed. Danger of cumulative effects. Non-hazardous at the concentrations used in this practical.

160 Appendices

Part A: Synthesis of Silver Nanoparticles with Addition of Bromide Ions Part A of this experiment is to be performed individually but results are shared and discussed within the group. Each group member should synthesise one sample of silver nanoparticles containing no potassium bromide and three samples that contain between 0 and 2300 μL of KBr solution. Your demonstrator will help you to determine a suitable range of volumes.

IMPORTANT: This experiment is highly sensitive to contamination. Do not reuse pipettes or measuring cylinders, and ensure that everything you use is clean and dry – this week you are working with precise concentrations, so all pipetted volumes must be measured accurately, most importantly KBr(aq).

Apparatus: Chemicals:

Glassware must be pristine as any Solutions prepared with Milli-Q water: impurity impacts the formation of 0.2 mL (1.25 × 10-1 M) aqueous sodium citrate AgNPs 5.0 mL (3.75 × 10-4 M) aqueous silver nitrate Safety glasses and lab coat 5.0 mL (5.00 × 10-2 M) hydrogen peroxide 4 x 22 mL vials X μL (2.175 × 10-5 M) potassium bromide Micropipettes (5, 1 mL) 2300 – X μL Milli-Q water 2.5 mL (5.00 × 10-3 M) sodium borohydride

Experimental Procedure:

To 4 × 22 mL vials, add in order:

Sodium citrate (1.25 × 10-1 M, 0.20 mL), silver nitrate (3.75 × 10-4 M, 5.0 mL) and hydrogen peroxide (5.00 × 10-2 M, 5.0 mL). Take three of your four vials. To each one add a different volume of potassium bromide (2.175 × 10-5 M, X µL) using a micropipette. Add sufficient water (2300-X) µL that the total volume of KBr solution and water is 2300 µL. Discuss the selected volume with your group or demonstrator to ensure that your group synthesises a range of samples. Next, add freshly prepared sodium borohydride (5.00 × 10-3 M, 2.5 mL). Once all the reagents are combined, place a cap on the vials and carefully swirl to fully mix the reactants. Record any observations in your ELN, consider taking a video of your synthesis. Note any changes that occur until the colour is stable (approximately ten minutes).

NB: Do not close the lids of the vials tightly due to possible gas release from residual sodium borohydride.

Questions

1. What species is causing the nanoparticle solutions to appear coloured? 2. For the nanoparticle solution that does not contain any KBr, which wavelengths of light are being absorbed and which are being transmitted? 3. An Australian five-cent coin has a diameter of approximately 2 cm. Assuming your nanoparticles have a diameter of 50 nm, how many orders of magnitude smaller is the diameter of the nanoparticles you have synthesised?

161 Appendices

4. Why does a coin appear silver and your nanoparticles appear coloured? 5. What effect did increasing the bromide concentration have on the colour of the solutions? 6. How might the bromide be influencing the nanoparticles? 7. Why is the order of reagent addition important?

162 Appendices

Part B: UV-Vis and Beer-Lambert’s Law Investigation Part B of this experiment is to be performed in groups of 2-3 by some members of the class, but results should be shared and discussed within the class. A different sample should be measured/analysed by different groups and shared with the whole class.

Apparatus: Chemicals:

5 × 10 mL volumetric pipette Nanoparticle sample(s) 4 × 22 mL vials 5 × disposable UV-Vis cuvette 1 × UV-Vis Spectrophotometer

Procedure: Preparing the Dilutions 1. Sequentially dilute AgNPs four times by making a series of 1 in 2 dilutions as follows. a. Using a volumetric pipette, take 10.0 mL of your sample of nanoparticles (Solution A) and place in another clean and dry vial. Add 10.0 mL of Millipore water with a clean volumetric pipette and swirl well. Label this sample Solution B. b. Place 10.0 mL of Solution A into a clean and dry vial and then dilute with 10.0 mL of Millipore water, swirl to mix. Label this sample Solution C. c. Repeat two more times to prepare Solutions D and E.

Measuring the Absorbances 2. Zero the spectrophotometer by inserting a cuvette filled with Milli-Q water (your ‘blank’) and pressing the appropriate button. 3. Replace the ‘blank’ cuvette in the instrument with a cuvette containing Solution A. 4. Measure the UV-Vis spectra of your samples. 5. Repeat absorbance measurements for solutions B – E.

6. In your notebook, record the wavelength of maximum absorbance (λmax) and the absorbance at that wavelength for each sample.

Plotting your data

7. Using spreadsheet software, plot the absorbance at λmax of each sample (Solutions A – E) against its concentration. 8. Note whether the relationship is linear. The plot should pass through (0,0). 9. Find the slope of this line. 10. Upload a table of your measurements and your plot into your ELN.

Questions 1. Why is it necessary to record a baseline (the ‘blank’) on a spectrophotometer before taking a measurement? 2. The Beer-Lambert law is A = ϵcl, where A = Absorbance, ϵ = the molar extinction coefficient, c = the concentration of the solution, and l = the path length. a. What is the path length of your sample? Include units. b. What does the molar extinction coefficient represent?

163 Appendices

c. Do your solutions obey the Beer-Lambert law? d. When might a solution not obey the Beer-Lambert law?

e. What is the extinction coefficient of your nanoparticles at λmax?

3. Absorbance is given by A = log10(T), where T = I0/I, I0 = incident light intensity, I = transmitted light intensity. What %T corresponds to an absorbance of 1? 4. Why are data where the absorbance is greater than 3 considered unreliable? a. If your sample gives an absorbance > 3, how might you lower the absorbance to give a more reliable value? (hint: consider the Beer-Lambert law)

164 Appendices

Part C: Catalytic Reduction of 4-Nitrophenol In this experiment you will use your silver nanoparticle solutions to investigate their catalytic properties.

Apparatus: Chemicals:

NB: Glassware and cuvettes must be pristine as any Milli-Q Water impurity prevents formation of NPs 1.0 mL (1.2 × 10-7 M) aqueous 4-nitrophenol Safety glasses, lab coat and gloves if desired 100 µL (0.25 M) aqueous sodium borohydride Cuvettes 1.0 mL (1 × 10-5 M) diluted silver nanoparticles Micropipettes Stopwatch pH paper Experimental Procedure:

Preparing your sample for analysis 1. Ensure you have all necessary samples and equipment accessible, and have understood all of the following steps before starting. 2. Dilute your nanoparticles by 1 in 12.5 by taking 1.6 mL of nanoparticle solution and mixing it with 18.4 mL water in a 21 mL vial. 3. Add aqueous 4-nitrophenol (1.0 mL, 1.2 × 10-7 M) to a cuvette.

Preparing your instrument 4. Set instrument into ‘kinetics mode’ (where one wavelength is measured at a set time interval). 5. Set the measurement wavelength of a spectrophotometer to 400 nm. 6. Zero the spectrophotometer by inserting a cuvette filled with Milli-Q water (your ‘blank’) and pressing the appropriate button.

Monitoring the reaction NB: Keep the time between aliquot additions (Δt) consistent between runs. Recommended: 30 s. 7. Replace the blank with your cuvette of 4-nitrophenol 8. Add 100 µL (0.25 M) aqueous sodium borohydride and start a timer upon addition. 9. After your chosen time (Δt), add 1.0 mL (1.0 × 10-5 M) diluted silver nanoparticles and simultaneously start data collection.

Measuring the pH

10. Measure the pH of your 4-nitrophenol solution before and after addition of NaBH4. a. Use a pipette to sample your solutions and drop them onto pH paper.

Record any observable changes seen in your samples in your ELN.

Questions

1. What was the pH of your starting solutions before and after NaBH4 addition? What, if anything, does this suggest about the nature of the 4-nitrophenol present at the beginning of the reaction? (Hint: the pKa of 4-nitrophenol is 7.15)

165 Appendices

2. Using Excel, plot the graphs below for your data (post induction period), do any of the plots possess a reasonable linear trend? a. [A] vs t b. ln[A] vs t c. 1/[A] vs t 3. What does this result tell you about the order of your reaction with respect to 4-nitrophenol (4NP)? 4. Write the rate law for this reaction in general terms. 5. Why, in this context, is the rate constant considered a PSEUDO-rate constant? (hint: think about the other species involved in the reaction and whether their concentrations changed significantly during the course of the reaction). 6. A linear trendline can be described by y = mx+c, what does ‘m’ represent in your analysis? 7. Compare the rate constant between the different KBr concentrations, sharing the data with other students as necessary to compare the complete range of concentrations. 8. Which samples exhibited the fastest rate of reaction? a. Hypothesise why this might be the case 9. Does every atom in the nanoparticle participate in catalysis? a. If not, which atoms participate, and which do not? 10. The bulk density of silver is 10.49 g/cm3, assume your particles are cylinders with diameter = 50 nm and height = 5 nm: a. What is the surface area of each nanoparticle? b. What is the mass of each nanoparticle? c. What is the total surface area of all the nanoparticles in your catalysis solution? d. Silver crystallizes in a face centred cubic structure with 1 Ag atom per 0.1115 nm2. How many surface silver atoms are present in your reaction solution? e. Assuming complete reaction of your substrate, 4NP, how many catalytic cycles did each surface silver atom undergo? 11. If each oxygen is lost as water, how many H–/H+ pairs are required per reduction of 4NP to 4- aminophenol? 12. Write a balanced equation for the catalytic reduction of 4-nitrophenol to 4-aminophenol.

166 Appendices

Answers to Student Questions Part A: Synthesis of Different Sized Silver Nanoparticles by Addition of Bromide Ions What species is causing the nanoparticle solutions to appear coloured?

The plasmonic nanoparticles are absorbing light to make the solution appear coloured.

For the nanoparticle solution that does not contain any KBr, which wavelengths of light are being absorbed and which are being transmitted?

The solution appears blue which indicates that orange light is absorbed by the sample. (note for instructors: provide a colour wheel for the students).

An Australian five-cent coin has a diameter of approximately 2 cm, assuming your nanoparticles have a diameter of 50 nm, how many orders of magnitude smaller is the diameter of the nanoparticles you have synthesised?

2 cm = 2 x 10-2 m, 50 nm = 5x10-8 m; therefore, the diameter of the nanoparticles is 6 orders of magnitude (1 million times) smaller than a 5 cent coin.

Why does a coin appear silver and your nanoparticles appear coloured?

Bulk silver (as could be found in old coins) reflects, rather than absorbs most of the light that hits it.1 Some metal nanoparticles (like those in this experiment) will absorb light if the frequency of the incident light matches the surface electron resonance frequency of the nanoparticle.2

What effect did increasing the bromide concentration have on the colour of the resulting solutions?

Increasing the bromide concentration shifted the colour of the nanoparticle solutions from blue to orange. This corresponds to an absorption blue shift from ~700 nm to ~350 nm.

How might the bromide be influencing the nanoparticles?

The plasmon resonance frequency is sensitive to particle shape and size,2 it is therefore likely that the bromide has changed these features. Silver cations and bromide anions interact strongly and hence silver bromide has a high lattice enthalpy and is insoluble in water. By binding to the surface of the silver NPs, bromide effectively 'caps' the growing crystal facets and hence controls the shape and size of the NPs – both of which affect the plasmon resonance frequency.2

Why is the order of reagent addition important?

It is critical that the reducing agent (NaBH4) be added last. If it is added before the stabilising agents (sodium citrate and potassium bromide), then particle growth will be uncontrolled and the silver may precipitate out of solution.

Part B: Measuring UV-Vis Data and Investigating Beer’s Law Why is it necessary to record a baseline (the ‘blank’) on a spectrophotometer before taking a measurement?

So that the absorbance due to the air, cuvette, and solvent is accounted for and subtracted from the absorbance of the sample.

The Beer-Lambert law is A = ϵcl, where A = Absorbance, ϵ = the molar extinction coefficient, c = the concentration of the solution, and l = the path length.

What is the path length of your sample? Include units.

1.0 cm - the internal length of the cuvette

167 Appendices

What does the molar extinction coefficient represent?

The molar extinction coefficient is wavelength dependent and represents how strongly a molecule will absorb light of a given wavelength.

Do your solutions obey the Beer-Lambert law?

Yes.

In what situation might a solution not obey the Beer-Lambert law?

If the behaviour of the molecules changes as a function of concentration, e.g. surfactants reaching the critical micelle concentration or methylene blue dimerising in solution as the concentration is increased.

What is the extinction coefficient of your nanoparticles at λmax?

Typical value: ~10000 L mol-1 cm-1

Absorbance is given by A = log10(I0/I), what %transmittance corresponds to an absorbance of 1?

10%, i.e. 1/10th of intensity of the incident light being detected.

Why are data where the absorbance is greater than 3, considered unreliable?

Because less than 0.1% of the incident light is reaching the detector and hence the absorbance value has a large degree of error.

If your sample gives an absorbance > 3, how might you lower the absorbance to give a more reliable value? (hint: consider the Beer-Lambert law)

Accurately diluting the sample until the absorbance reaches a value of between 0.1 and 3 (an absorbance close to 1 is often considered ideal).

Part C: Silver Nanoparticles Experiment Three - Catalytic reduction of 4-nitrophenol by silver nanoparticles in the presence of NaBH4

What was the pH of your starting solutions before and after NaBH4 addition? What, if anything, does this suggest about the nature of the 4-nitrophenol present at the beginning of the reaction? (Hint: the pKa of 4- nitrophenol is 7.15)

Typical starting pH: ~5.5

After NaBH4 addition, pH: ~9

The 4-nitrophenol is deprotonated when the pH is increased upon NaBH4 addition.

Using Excel, plot the graphs below for your data (post induction period), do any of the plots possess a reasonable linear trend?

[A] vs t - This is usually closely linear but systematically deviates

ln[A] vs t - This is usually the most reasonable linear fit

1/[A] vs t - This is the least reasonable fit

168 Appendices

What does this result tell you about the order of your reaction with respect to 4-nitrophenol (4NP)?

The reaction appears to be first order with respect to 4-nitrophenol

Write the rate law for this reaction in general terms.

푑[4푁푃] = −푘 [4푁푃] 푑푡 푎푝푝

Why, in this context, is the rate constant considered a PSEUDO-rate constant? (hint: think about the other species involved in the reaction and whether their concentrations changed significantly over the course of the reaction)

The rate constant, kapp, is considered a PSEUDO-rate constant because the rate of reaction is also proportional to the concentration of NaBH4. However, under these conditions (a large excess of NaBH4), the concentration of NaBH4 changes so little throughout the course of the reaction that it can be considered constant and therefore the [NaBH4] term is included within the rate constant kapp. Because of this, we use the term pseudo-rate constant.

A linear trendline can be described by y = mx + c, what does ‘m’ mean in your analysis?

‘m’ represents the pseudo rate constant

Compare the rate constant between the different KBr concentrations, sharing the data as necessary to compare the complete range of concentrations.

Which samples exhibited the fastest rate of reaction?

The samples synthesised without KBr exhibited the fastest rate of reaction.

Hypothesise why this might be the case.

These samples may contain nanoparticles with a greater number of catalytic sites, or these sites may be more active.

Does every atom in the nanoparticle participate in catalysis?

No.

If not, which atoms participate, and which do not?

Since this reaction occurs on the surface of the AgNPs, only surface sites are involved in catalysis. The reagents cannot gain access to the silver atoms below the surface of the nanoparticles.

The bulk density of silver is 10.49 g/cm3, assume your particles are cylinders with diameter = 50 nm and height = 5 nm:

i) What is the surface area of each nanoparticle?

A = 2πrh+2πr2 = 2π*25x10-9*5x10-9 +2π(25x10-9)2 = 4.7x10-15 m2/particle = ~4700 nm2/particle

ii) What is the mass of each nanoparticle?

Vol = πr2h; density = 10.49x106 g/m3 -9 2 -9 6 -16 Mass = Vol*density = π*(25x10 ) *5x10 *10.49x10 = ~1x10 g/particle

169 Appendices

iii) What is the total surface area of all the nanoparticles in your catalysis solution?

1x10-8 mol of Ag in reaction solution 1x10-8 mol * 107.8682 g/mol = 1x10-6 g -6 -16 10 1x10 g / 1x10 g/particle = 1x10 particles 1x1010 particles * 4700 nm2/particle = 5x1013 nm2 total surface area

iv) Silver crystallizes in a face centred cubic structure with 1 Ag atom per 0.1115 nm2. How many surface silver atoms are present in your reaction solution?

5x1013 nm2/0.1115 nm2 = 4.5x1014 surface atoms (i.e. ~7.5% of all silver atoms, assuming 1x10-8 mol of Ag atoms in total)

v) Assuming complete reaction of your substrate, 4NP, how many catalytic cycles did each surface silver atom undergo?

-7 16 1.2x10 mol 4NP * NA = 7.2x10 molecules of 4NP 7.2x1016 molecules / 4.5x1014 surface atoms = 160 catalytic cycles (assuming a catalytic site requires only 1 silver atom)

If each oxygen is lost as water, how many H–/H+ pairs are required per reduction of 4NP to 4AP?

3 H–/H+ pairs are required

Write a balanced equation for the catalytic reduction of 4-nitrophenol to 4-aminophenol.

– + 3H + 3H + 4NP ⟶ 4AP + 2H2O

For further discussion of the mechanism refer to 3.

170 Appendices

Common Errors / Troubleshooting Silver nitrate should be kept in glass and away from light (wrapping with foil will suffice) to avoid decomposition.

Sodium borohydride will decompose significantly within 5 minutes of dissolution at room temperature. Prepare with ice cold water and store the solution cold to minimise the decomposition. Our results show that decomposition over 2 h is insignificant at 0 °C even for the 0.5 M solution.

If students are having trouble producing a ‘rainbow’ of nanoparticles, check that the glassware is clean (best to acid soak then rinse beforehand).

If the profile of the catalytic reaction shows humps/bulges, these are likely caused by unavoidable bubble build up in the cuvette. Removing and flicking the cuvette will help, but the students may need to rerun the catalysis if the effect is large.

When comparing the reaction profiles of multiple nanoparticle samples students may find that the absorbance is not always 1. This is due to the superposition of nanoparticle plasmons and the absorbance of 4-nitrophenol at 400 nm. The students can subtract the effect in a post-analysis, or run a baseline with the correct concentration of their nanoparticles in the cuvette.

If running TEM, concentrate the particles beforehand. Centrifuging the colloid until a pellet forms, followed by decanting of the supernatant and resuspending the pellet by sonication is sufficient.

Suggested Laboratory Schedule Session / Week Time Activity 1 3 h Part A: Synthesis of Silver Nanoparticles with Addition of Bromide Ions Part B: UV-Vis and Beer-Lambert’s Law Investigation 2 3 h Part C: Catalytic Reduction of 4-Nitrophenol 3 3 h Additional Characterisation* (XRD, DLS, TEM, AFM) *Details regarding additional characterization are not provided in this work and are left to the instructor.

1. Guerrero, A. H.; Fasoli, H. J.; Costa, J. L., Why gold and copper are colored but silver is not. J. Chem. Educ. 1999, 76 (2), 200. 2. Rycenga, M.; Cobley, C. M.; Zeng, J.; Li, W.; Moran, C. H.; Zhang, Q.; Qin, D.; Xia, Y., Controlling the synthesis and assembly of silver nanostructures for plasmonic applications. Chem. Rev. 2011, 111 (6), 3669-3712. 3. Strachan, J.; Barnett, C.; Masters, A. F.; Maschmeyer, T., 4-Nitrophenol Reduction: Probing the Putative Mechanism of the Model Reaction. ACS Catal. 2020.

171

End of Document