STRUCTURE-PROPERTY RELATIONSHIPS IN MIXED-METAL OXIDES AND (OXY)HYDROXIDES FOR ENERGY APPLICATIONS by LISA J. ENMAN A DISSERTATION Presented to the Department of Chemistry and Biochemistry and the Graduate School of the University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy September 2018 ii DISSERTATION APPROVAL PAGE Student: Lisa J. Enman Title: Structure-Property Relationships in Mixed-Metal Oxides and (Oxy)Hydroxides for Energy Applications This dissertation has been accepted and approved in partial fulfillment of the requirements for the Doctor of Philosophy degree in the Department of Chemistry and Biochemistry by: Catherine Page Chairperson Shannon Boettcher Advisor Georgy Nazin Core Member Benjamin McMorran Institutional Representative and Janet Woodruff-Borden Vice Provost and Dean of the Graduate School Original approval signatures are on file with the University of Oregon Graduate School. Degree awarded September 2018 iii © 2018 Lisa J. Enman iv DISSERTATION ABSTRACT Lisa J. Enman Doctor of Philosophy Department of Chemistry and Biochemistry September 2018 Title: Structure-Property Relationships in Mixed-Metal Oxides and (Oxy)Hydroxides for Energy Applications Metal oxides and (oxy)hydroxides, particularly those containing two or more metals have many uses as electronic materials and catalyst, especially in energy applications. In this dissertation, the structure-property relationships of these mixed- metal materials are explored in order to understand how these materials work and to guide design of materials with even higher efficiency for a given application. Chapter I introduces the materials and studies undertaken. Chapter II presents a fundamental analysis of the electronic and local atomic properties of mixed-transition-metal aluminum oxide thin films. The final three chapters focus on water electrolysis for hydrogen production, which is limited in part by the slow kinetics of the oxygen evolution reaction (OER). Nickel- iron and cobalt-iron (oxy)hydroxides have been shown to be the most active in alkaline conditions. Although it is evident that Fe is essential for high activity, its role is still unclear. Chapter III investigates the role of Fe in NiOOH by comparing the effects of Ti, Mn, La, and Ce incorporation on the OER activity of NiOOH in base. Chapter IV evaluates the OER activity and Tafel behavior of Fe3+ impurities on different noble metal substrates. Chapter V describes the results of in situ and in operando X-ray spectroscopy v experiments, which shows that the local structure around Fe atoms in Co(Fe)OOH changes during OER while that of Co stays the same. This work adds to the growing body of literature that suggests Fe is essential to the catalytic active site for the OER on transition-metal (oxy)hydroxides. This dissertation contains previously published and un-published coauthored material. vi CURRICULUM VITAE NAME OF AUTHOR: Lisa J. Enman GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon, Eugene Worcester State University, Worcester, Massachusetts Worcester Polytechnic Institute, Worcester, Massachusetts DEGREES AWARDED: Doctor of Philosophy, Chemistry, 2018, University of Oregon Master of Science, Chemistry, 2015, University of Oregon Bachelor of Science, Chemistry, 2013, Worcester State University AREAS OF SPECIAL INTEREST: Materials Chemistry; Electrochemistry; Catalysis PROFESSIONAL EXPERIENCE: Research Assistant, Boettcher Laboratory, University of Oregon, 2014-2018 Teaching Assistant, University of Oregon, 2013-2014 Organic Chemistry Lab, General Chemistry Lab, General Chemistry Lecture LEADERSHIP EXPERIENCE: Executive Board, Women in Graduate Science, University of Oregon, 2016-2018 Treasurer, Public Relations Chair Peer-elected Representative, Chemistry and Biochemistry Graduate Representative Advisory Team, University of Oregon, 2016-2017 Laboratory Safety Officer, Boettcher Lab, University of Oregon, 2015-2017 vii PUBLICATIONS: 10) Enman, L. J.; Kast, M. G.; Cochran, E. A.; Pledger, E.; Stevens, M. B.; Boettcher, S. W. Transition-Metal-Incorporated Aluminum Oxide Thin Films: Towards Electronic Structure Design in Amorphous Mixed-Metal Oxides. J. Phys. Chem. C. 2017, DOI: 10.1021/acs.jpcc.8b00239 9) Cochran E. A.; Park D. H.; Kast M. G.; Enman L. J.; Perkins, C. K.; Mansergh, R. H.; Keszler, D. A.; Johnson, D. W.; Boettcher S. W. Role of Combustion Chemistry in Low- Temperature Deposition of Metal Oxide Thin Films from Solution. Chem. Mater. 2017, 29, 9480-9488. 8) Woods, K.N.; Plassmeyer, P.N.; Park, D.H.; Enman, L.J.; Grealish, A. K.; Kirk, B. L.; Boettcher, S. W.; Keszler, D. A.; Page, C.J. Low-Temperature Steam Annealing of Metal Oxide Thin Films from Aqueous Precursors: Enhanced Counterion Removal, Resistance to Water Absorption, and Dielectric Constant. Chem Mater. 2017, 29, 8531-8538. 7) Stevens, M.B.; Trang, C.D.; Enman, L. J.; Deng, J; Boettcher, S. W. Reactive Fe-sites in Ni/Fe (oxy) hydroxide are responsible for exceptional oxygen electrocatalysis activity. J. Am. Chem. Soc. 2017, 139, 11361–11364. 6) Kast, M. G.; Cochran, E. A.; Enman, L.J.; Mitchson, G.; Ditto, J; Siefe, C; Plassmeyer, P. N.; Greenaway, A. L.; Johnson, D. C.; Page, C. J.; Boettcher, S. W. Amorphous Mixed-Metal Oxide Thin Films from Aqueous Solution Precursors with Near Atomic Smoothness. J. Am. Chem. Soc. 2016, 138, 16800–16808. 5) Stevens, M. B.; Enman, L. J.; Batchellor, A. S.; Cosby, M. R.; Vise, A. E.; Trang, C. D. M.; Boettcher, S. W. Measurement Techniques for the Study of Heterogeneous Water Oxidation Electrocatalysts. Chem Mater. 2016, 29, 120-140. 4) Enman, L. J.; Burke, M. S.; Batchellor, A. S.; Boettcher, S. W. Effects of Intentionally Incorporated Metal Cations on the Oxygen Evolution Electrocatalytic Activity of Nickel (Oxy)hydroxide in Alkaline Media. ACS Catal. 2016, 6, 2416-2423. 3) Burke, M. S.; Enman, L. J.; Batchellor, A. S.; Zou, S.; Boettcher, S. W. Oxygen- evolution-reaction electrocatalysis on transition metal oxides and (oxy)hydroxides: Activity trends and design principles. Chem. Mater. 2015, 27, 7549-7558. 2) Burke, M. S.; Zou, S.; Enman, L. J.; Kellon, J. K.; Gabor, C. A.; Pledger, E.; Boettcher, S. W. Revised oxygen-evolution-reaction activity trends for first-row transition metal (oxy)hydroxides in alkaline media. J. Phys. Chem Lett. 2015, 6, 3737- 3742. 1) Kast, M. G.; Enman, L. J.; Gurnon, N. J.; Nadarajah, A.; Boettcher, S. W. Solution- Deposited F:SnO2/TiO2 as a Base-Stable Protective Layer and Antireflective Coating for Microtextured Buried-Junction H2-Evolving Si Photocathodes. ACS Appl. Mater. Interfaces 2014, 6, 22830–22837. viii ACKNOWLEDGMENTS This work was funded in part by the National Science Foundation through GOALI Grants CHE-1301461 and CHE-1566348 and the Center for Sustainable Materials Chemistry Grants CHE-1102637 and CHE-1606982. This work made use of the CAMCOR shared instrument facilities, supported by grants from the W. M. Keck Foundation, the M. J. Murdock Charitable Trust, ONAMI, the Air Force Research Laboratory, NSF, and the University of Oregon. I am grateful to the CAMCOR staff, particularly Dr. Stephen Golledge for maintaining the instruments and providing guidance. The Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DEAC02-06CH11357 and of the Advanced Light Source, which is supported by the Director, Office of Science, Office of Basic Energy Sciences, of the U.S. DOE under Contract No. DE-AC02-05CH11231. I thank Dr. Sungsik Lee and Dr. Sirine Fakra for help performing experiments. To my advisor, Prof. Shannon Boettcher, thank you for your enthusiasm for science and your never-ending supply of ideas. Your high standards have made me a better scientist. Also thank you to my committee, especially Prof. Cathy Page for contributing to my scientific development. To my undergraduate professors, Prof. Anne Falke, Prof. Eihab Jaber, and Prof. Joseph Quattrucci, thank you for the chemistry education you gave me, for your advice, and for telling me about this thing called grad school and convincing me it was something I could do. ix Importantly I would like to thank all Boettcher lab members past and present for their encouragement, but also constructive criticism. Thank you for making the lab feel more like a family. I would particularly like to thank past Team OER members: Dr. Adam Batchellor and Dr. Adam Smith for their thought-provoking conversations, and especially Dr. Michaela Stevens for challenging me scientifically and for your constant positive encouragement. To Dr. Ann Greenaway, thank you for always being available to talk when I needed it. To Lizzie Cochran and Mike Nellist, thank you for maintaining the spirit of the Boettcher lab. To Ashlee Vise, thank you for your hard work, dedication, enthusiasm, and for being a great mentee. And to Dr. Matthew Kast, the best mentor I have had in my science career, and a great friend, words cannot express my gratitude. I would like to thank my friends for their emotional support. To Ashley Bernier, thank you for moving across the country with me and for all your support, especially at home those first two years. To Susan Cooper, thank you for being my best friend throughout my PhD experience, your support has saved me from falling apart on many occasions. To all the friends I have made here, especially Andrea, Huiying, Loni, Sam, Hillary, Peter, Brandi, and James, thank you for distracting me from the hard times of graduate school. And to those back home, Jamie, Brianna, Melissa, and Callen, thank you for effortlessly continuing our friendship despite my moving 3,000 miles away. To my sister, Julia, thank you for always being there for me, especially when things in life got difficult. To my aunt Chrissy, thank you for your never-ending support and understanding. To my mother, thanks for pushing me to be better. And to my dad, thank you for everything: for coaching me, for always telling me how smart I am, for understanding who I am and supporting me even when my path got murky. x TABLE OF CONTENTS Chapter Page I. INTRODUCTION .................................................................................................... 1 Electronic Structure Design in Amorphous Oxides ............................................... 2 Effects of Transition-Metal Cations on Nickel (Oxy)hydroxide Activity ............. 3 Metal Electrode Substrate Effects on FeOxHy Activity ......................................... 4 The Role of Fe in the Oxygen Evolution Reaction on Co(Fe)OxHy...................... 5 II. TRANSITION-METAL-INCORPORATED ALUMINUM OXIDE THIN FILMS: TOWARDS ELECTRONIC SRUCTURE DESIGN IN AMORPHOUS MIXED-METAL OXIDES ........................................................... 6 Introduction ............................................................................................................ 6 Experimental .......................................................................................................... 8 Film Deposition ............................................................................................... 8 X-ray Absorption Spectroscopy ....................................................................... 9 X-ray Photoelectron Spectroscopy .................................................................. 10 X-ray Diffraction ............................................................................................. 10 Thickness Characterization .............................................................................. 10 Optical Absorption Characterization ............................................................... 11 Results and Discussion .......................................................................................... 12 d0 Transition-Metal Cations ............................................................................. 13 d5 Transition-Metal Cations ............................................................................. 17 d6,7,8,9 Transition-Metal Cations ....................................................................... 20 d10 Transition-Metal Cations ............................................................................ 25 Tuning the Electronic Structure and Optical Properties ........................................ 28 xi Chapter Page Number of d Electrons ..................................................................................... 28 Cation Concentration and Oxidation state ....................................................... 29 Conclusion and Bridge ........................................................................................... 30 III. EFFECTS OF INTENTIONALLY INCORPORATED METAL CATIONS ON THE OXYGEN EVOLUTION ELECTROCATALYTIC ACTIVITY OF NICKEL (OXY)HYDROXIDE IN ALKALINE MEDIA .................................... 32 Introduction ............................................................................................................ 32 Experimental .......................................................................................................... 35 Spin Cast Film Deposition ............................................................................... 35 Electrochemical Film Deposition .................................................................... 36 Electrochemical Characterization .................................................................... 36 Physical Characterization ................................................................................. 37 Results and Discussion .......................................................................................... 37 Dependence of Film Morphology on Added Cation Identity .......................... 37 Effects of Added Cations on OER Activity ..................................................... 39 Transient Activity Enhancement from Ce Addition ........................................ 41 Mechanisms of Ni(Ce)OxHy Deactivation ....................................................... 42 Effects of Ti, Mn, and La on NiOxHy OER Activity ....................................... 45 Analysis of Fe Impurities ................................................................................. 46 Tafel Behavior and Stability ............................................................................ 47 Cation Influence on Electronic Structure ......................................................... 48 Comparison of Film Preparation Methods ....................................................... 49 Conclusion and Bridge ...................................................................................... 50 xii Chapter Page IV. ELECTRODE SUBSTRATE DEPENDENCE OF IRON (OXY)HYDROXIDE ELECTROCATALYTIC ACTIVITY FOR THE OXYGEN EVOLUTION REACTION ................................................................. 52 Introduction ............................................................................................................ 52 Experimental .......................................................................................................... 54 Electrode Fabrication ....................................................................................... 54 Electrochemical Characterization .................................................................... 54 X-Ray Photoelectron Spectroscopy (XPS) ...................................................... 55 Results and Discussion .......................................................................................... 55 Electrolyte Impurities ....................................................................................... 55 Substrate Dependence of FeOxHy OER Activity ............................................ 58 Au Enhancement of FeOxHy Activity ............................................................. 61 Conclusion and Bridge ........................................................................................... 63 V. OPERANDO X-RAY ABSORPTION SPECTROSCOPY SHOWS FE OXIDATION IS CONCURRENT WITH OXYGEN EVOLUTION IN COBALT-IRON (OXY)HYDROXIDE ELECTROCATALYSTS ....................... 65 Introduction ............................................................................................................ 65 Experimental .......................................................................................................... 66 Electrochemical Cell Fabrication ..................................................................... 66 Catalyst Film Deposition ................................................................................. 66 Electrochemical Characterization .................................................................... 67 XAS Measurements ...................................................................................... 67 XAS Data Analysis .......................................................................................... 68 Results and Discussion .......................................................................................... 68 xiii Chapter Page Conclusion ............................................................................................................. 75 APPENDICES ............................................................................................................. 76 A. CHAPTER II SUPPORTING INFORMATION .............................................. 76 B. CHAPTER III SUPPORTING INFORMATION ............................................. 89 C. CHAPTER IV SUPPORTING INFORMATION ............................................. 104 D. CHAPTER V SUPPORTING INFORMATION .............................................. 106 REFERENCES CITED ................................................................................................ 114 xiv LIST OF FIGURES Figure Page 1.1. Schematic depiction of UV-vis and X-ray spectroscopies ................................... 3 1.2. Comparison of Ni2+/3+ oxidation peak potential .................................................... 4 2.1. Analysis of VyAl1-yOx thin films .......................................................................... 15 2.2. Analysis of CryAl1-yOx thin films ......................................................................... 17 2.3. Analysis of FeyAl1-yOx thin films ......................................................................... 18 2.4. Analysis of MnyAl1-yOx thin films ....................................................................... 20 2.5. Analysis of CoyAl1-yOx thin films ........................................................................ 21 2.6. Analysis of NiyAl1-yOx thin films ........................................................................ 23 2.7. Analysis of CuyAl1-yOx thin films ........................................................................ 25 2.8. Analysis of ZnyAl1-yOx thin films ........................................................................ 26 2.9. Analysis of GayAl1-yOx thin films ........................................................................ 28 3.1. SEM images of spin-cast films ............................................................................. 39 3.2. Voltammetry of spin-cast Ni1-zMzOxHy films ...................................................... 40 3.3. OER turnover frequency of spin-cast films .......................................................... 41 3.4. a) Tafel plots for spin cast Ni1-zMzOxHy films ..................................................... 43 3.5. O 1s XP spectra for Ni1-zMzOxHy films ............................................................... 45 3.6. TOFtm vs. E0' of Ni2+/3+ for Ni1-zMzOxHy ............................................................. 49 4.1. Cyclic voltammetry at 20 mV s-1 of electrode substrates ..................................... 57 4.2. Cyclic voltammetry at 20 mV s-1 for FeOxHy ...................................................... 59 4.3. Tafel analysis of different electrode substrates ..................................................... 61 4.4. Fe XP spectra ........................................................................................................ 62 xv Figure Page 4.5. Metal electrode XP spectra ................................................................................... 63 5.1. A) Cyclic voltammetry data on Co(Fe)OxHy films .............................................. 69 5.2. Operando Co K-edge XANES data ...................................................................... 70 5.3. Operando Fe K-edge XANES data ....................................................................... 71 5.4. FT-EXAFS data for Co(Fe)OxHy films ................................................................ 72 5.5. Schematic OER mechanism for Fe-doped CoOOH .............................................. 73 A.1. GIXRD patterns for MyAl1-yOx films where M is not d0, d5, or d10 .................... 76 A.2. GIXRD patterns from MyAl1-yOx films where M is d0, d5, or d10 ....................... 77 A.3. C 1s and O 1s XP spectra of a-Al2O3 film .......................................................... 78 A.4. XP spectra of VyAl1-yOx films ............................................................................. 78 A.5. XP spectra of CryAl1-yOx films ............................................................................ 79 A.6. XP spectra of MnyAl1-yOx films .......................................................................... 79 A.7. XP spectra of FeyAl1-yOx films ............................................................................ 80 A.8. XP spectra of CoyAl1-yOx films ........................................................................... 81 A.9. XP spectra of NiyAl1-yOx films ............................................................................ 81 A.10. XP spectra of CuyAl1-yOx films ......................................................................... 82 A.11. XP spectra of ZnyAl1-yOx films .......................................................................... 83 A.12. XP spectra of GayAl1-yOx films ......................................................................... 83 A.13. VB-XPS of ZnyAl1-yOx films ............................................................................. 84 A.14. k2-weighted χ(k) data MyAl1-yOx films .............................................................. 85 B.1. Number of electrons removed per Ni atom .......................................................... 90 B.2. Morphology of spin-cast Ni1-zMzOxHy films ...................................................... 91 xvi Figure Page B.3. Morphology of electrodeposited Ni1-zMzOxHy films .......................................... 91 B.4. Initial voltammograms (Cycle 1) ......................................................................... 93 B.5. Representative cyclic voltammograms ................................................................. 94 B.6. Tafel plots ............................................................................................................. 96 B.7. O 1s XP spectra .................................................................................................... 97 B.8. Ce 3d XP spectra .................................................................................................. 98 B.9. Cyclic voltammetry of a representative Ni0.7Ti0.3OxHy film ............................... 98 B.10. XP spectra of Fe 2p region ................................................................................. 99 B.11. Voltammetry of bare Au substrates .................................................................... 100 B.12. Voltammetry comparison of spin-cast Ni0.85Ce0.15OxHy ................................... 100 B.13. Voltammetry (Cycle 5) of bare Au substrate and CeOx film ............................. 101 B.14. Change in film loading ....................................................................................... 101 B.15. Voltammetry of electrodeposited Ni1-zMzOxHy films ....................................... 102 B.16. Comparison of OER activity .............................................................................. 103 B.17. TOFtm vs. E0’ of electrodeposited Ni1-zMzOxHy films ....................................... 103 C.1. Dependence of current response on scan rate ...................................................... 104 C.2. Plot of current density at 450 mV overpotential .................................................. 104 C.3. Cyclic voltammetry at 20 mV s-1 of Pt ................................................................. 105 D.1. Co K-edge XANES on Co(Fe)OxHy films .......................................................... 107 D.2. Fe K-edge XANES ............................................................................................... 107 D.3. k3-weighted chi(k) data for Co(Fe)OxHy films .................................................... 108 D.4. Atomic position labelling for Fe-doped CoOOH ................................................. 110 xvii LIST OF TABLES Table Page 4.1. OER current density at 0.45 V .............................................................................. 58 A.1. Metal K-edge energy ............................................................................................ 84 A.2. EXAFS fitting results for VyAl1-yOx films .......................................................... 86 A.3. EXAFS fit results for CryAl1-yOx films ............................................................... 86 A.4. EXAFS fitting results for MnyAl1-yOx films ....................................................... 87 A.5. EXAFS fitting results for FeyAl1-yOx films ......................................................... 87 A.6. EXAFS fitting results for CoyAl1-yOx films ........................................................ 87 A.7. EXAFS fitting results for NiyAl1-yOx films ......................................................... 87 A.8. EXAFS fitting results for CuyAl1-yOx films ........................................................ 87 A.9. EXAFS fitting results for ZnyAl1-yOx films ........................................................ 88 A.10. EXAFS fitting results for GayAl1-yOx films ...................................................... 88 B.1. Electrodeposition conditions for the electrodeposited Ni1-zMzOxHy films ......... 89 B.2. Overpotential at 0.5 mA cm-2 for spin-cast Ni1-zMzOxHy films .......................... 92 B.3. Tafel slope and TOFtm for spin-cast Ni1-zMzOxHy films ..................................... 95 D.1. Results of FT-EXAFS fitting for Co .................................................................... 108 D.2. Results of FT-EXAFS fitting for Fe .................................................................... 109 D.3. Fe-O bond distances around the Fe active site ..................................................... 111 D.4. Co-O bond distances far away from the active site ............................................. 111 D.5. Co-O bond distances close to the active site ........................................................ 111 D.6. Co-O bond distances around the Co active site ................................................... 112 xviii Table Page D.7. Atomic magnetization and estimated oxidation state .......................................... 112 D.8. Free energies for water oxidation reaction intermediates .................................... 113 D.9. Zero-point energy (ZPE) and entropy corrections ............................................... 113 1 CHAPTER I INTRODUCTION Section “Electronic Structure Design in Amorphous Oxides” contains co-authored material previously published as J. Phys. Chem. C. 2018, DOI: 10.1021/acs.jpcc.8b00239 reproduced with permission. Copyright 2018 American Chemical Society. Prof. Boettcher, M. Kast, and I conceived of the project. I performed and directed experiments, collected data, and analyzed data with help from M. Kast, E. Cochran, E. Pledger, and M. Stevens. I wrote the paper with help from Prof. Boettcher and M. Kast along with editorial assistance from all authors. Section “Effects of Transition-Metal Cations on Nickel (Oxy)hydroxide Activity” contains co-authored material previously published as ACS Catal. 2016, 6(4), 2416-2423 reproduced with permission. Copyright 2016 American Chemical Society. M. Burke and I conceived of the project. I performed the experiments and analyzed data with help from M. Burke and A. Batchellor. I wrote the paper with help from Prof. Boettcher and editorial assistance from all authors. Section “Metal Electrode Substrate Effects on FeOxHy Activity” contains co- authored unpublished material. Prof. Boettcher and I conceived of the project. I performed experiments and analyzed data as well as directed experiments and analysis performed by undergraduate A. Vise. I wrote the paper with help from Prof. Boettcher and A. Vise along with editorial assistance from all authors. It is prepared in the format of the Journal of the American Chemical Society. Section “The Role of Fe in the Oxygen Evolution Reaction on Co(Fe)OxHy” contains co-authored unpublished material. M. Stevens, Prof. Boettcher, and I conceived of the project. I performed experiments along with M. Stevens and M. Nellist with technical assistance from S. Fakra. M C. Toroker and M Dahan performed calculations. I analyzed data and wrote most of the paper with computational sections written by M. 2 Toroker and with editorial assistance from M. Stevens and S. Boettcher. It is prepared in the format for submission to the journal Angewandte Chemie International Edition. Electronic Structure Design in Amorphous Oxides Amorphous metal oxides have numerous applications spanning from electronics to catalysis. The ability to rationally design the electronic structure of such materials would thus have significant impact. Previous work has shown that the amorphous structure allows for uniformly incorporating a large number of different metal cations into ultra-smooth thin films. Here we investigate how the atomic structure, electronic structure, and optical absorption properties of amorphous aluminum oxide (a-Al2O3) thin films are modulated when metal cations of V, Cr, Mn, Fe, Co, Ni, Cu, Zn, or Ga are added to form MyAl1-yOx where y varies from 0.1 to 0.7. We use X-ray absorption spectroscopy (XAS) to analyze oxidation state and local bonding around the incorporated cations (e.g. coordination number), valence-band X-ray photoelectron spectroscopy (VB- XPS) to assess the impact of the cations on the valence band electronic structure, and optical absorption spectroscopy to assess the presence and energies of new unoccupied electronic states (schematically depicted in Figure 1.1.). Cations with partially filled d- orbitals (e.g. Fe3+, Cr3+, Co2+, Ni2+, Cu2+, etc.) add filled 3d electronic states at the valence band edge of a-Al2O3 as well as unfilled 3d electronic states within the a-Al2O3 band gap. Cations with the d10 electronic configuration, such as Zn2+ and Ga3+, introduce filled 3d electronic states at energies well below the valence band edge in addition to new unfilled electronic states in the a-Al2O3 band gap. Cations with unfilled d-orbitals (e.g. those with a d0 configuration such as V5+ and Cr6+) do not modify the valence band of a- Al2O3, but affect the optical properties by inserting unfilled 3d electronic states within the band gap. These results provide guidance for the rational design of the electronic structure of amorphous mixed-metal oxides and are discussed further in Chapter II. 3 Figure 1.1. Schematic depiction of UV-vis and X-ray spectroscopies used to probe the density of states of transition-metal incorporated aluminum oxide thin films. Chapter II contains co-authored material previously published as J. Phys. Chem. C. 2018, DOI: 10.1021/acs.jpcc.8b00239 reproduced with permission. Prof. Boettcher, M. Kast, and I conceived of the project. I performed and directed experiments, collected data, and analyzed data with help from M. Kast, E. Cochran, E. Pledger, and M. Stevens. I wrote the paper with help from Prof. Boettcher and M. Kast along with editorial assistance from all authors. Effects of Transition-Metal Cations on Nickel (Oxy)hydroxide Activity Fe-doped Ni (oxy)hydroxide, Ni(Fe)OxHy, is one of the most-active oxygen- evolution-reaction (OER) catalysts in alkaline conditions, while Fe-free NiOxHy is a poor OER catalyst. One approach to better understand the role of Fe, and enable the design of catalysts with higher activities, is to find other cations that behave similarly and compare the common chemical features between them. In Chapter III, we evaluate the effects of La, Mn, Ce, and Ti incorporation on the OER activity and redox behavior of NiOxHy in rigorously Fe-free alkaline solution using cyclic voltammetry and electrochemical quartz- crystal microgravimetry. We use X-ray photoelectron spectroscopy and time-of-flight secondary-ion-mass spectrometry to confirm that measurements are free from relevant levels of trace Fe contamination. We find that only Ce leads to increased activity in NiOxHy (about a factor of ten enhancement), but this effect is transient, likely due to phase separation. We further find no clear correlation between activity and the nominal Ni2+/3+ redox potential, shown in Figure 1.2, suggesting that the “oxidizing” power of the 4 Ni is not directly correlated with the OER activity. These findings are discussed further in Chapter III highlight the uniqueness of Fe and are consistent with it being the active site in Ni(Fe)OxHy. Figure 1.2. Comparison of Ni2+/3+ oxidation peak potential and OER turnover frequency (TOF) on Ni1-zMzOxHy thin film electrocatalysts. Chapter III contains co-authored material previously published as ACS Catal. 2016, 6(4), 2416-2423 reproduced with permission. M. Burke and I conceived of the project. I performed the experiments and analyzed data with help from M. Burke and A. Batchellor. I wrote the paper with help from Prof. Boettcher and editorial assistance from all authors. Metal Electrode Substrate Effects on FeOxHy Activity FeOxHy and Fe-containing materials are the some of the most active catalysts for the OER in alkaline solution. However, its activity is highly dependent on the electrode substrate material and/or the elemental composition of the matrix in which it is embedded. A fundamental understanding of these interactions that modulate the OER activity of FeOxHy is lacking. In Chapter IV, we use cyclic voltammetry and chronopotentiometry to assess the substrate-dependent activity of FeOxHy on a number of commonly used electrode substrates, Au, Pt, Pd, Cu, and C. We also evaluate the OER activity and Tafel behavior of these metallic substrates in 1 M KOH aqueous solution with Fe3+ and other electrolyte impurities. We find that the OER activity of FeOxHy varies by substrate in the order Au > Pd ≈ Pt ≈ Cu > C. The trend is likely caused by differences in the adsorption strength of metal-oxo-ion and metallic substrate, where a 5 stronger adhesion results in decreased charge-transfer resistance at the FeOxHy-substrate interface. These results also suggest that the local atomic and electronic structure of [FeO6] units play an important role in catalysis of the OER as it can be tuned by substrate interactions, and is further discussed in Chapter IV. Chapter IV contains co-authored unpublished material. Prof. Boettcher and I conceived of the project. I performed experiments and analyzed data as well as directed experiments and analysis performed by undergraduate A. Vise. I wrote the paper with help from Prof. Boettcher and A. Vise along with editorial assistance from all authors. It is prepared in the format of the Journal of the American Chemical Society. The Role of Fe in the Oxygen Evolution Reaction on Co(Fe)OxHy In alkaline conditions, Fe-incorporated Ni- and Co-based (oxy)hydroxides (Ni(Fe)OxHy, Co(Fe)OxHy) are among the best electrocatalysts for the oxygen evolution reaction (OER). Fe cations are essential for the high activities measured, but the role of Fe in the catalytic mechanism remains controversial. In Chapter V we use operando X- ray absorption spectroscopy (XAS) to demonstrate partial Fe oxidation and a shortening of the Fe-O bond length during oxygen evolution in Co(Fe)OxHy. Co oxidation is only observed in the absence of Fe. These results support the hypothesis that Fe is the active site in high-activity Co(Fe)OxHy OER catalysts, and the partial oxidation is consistent with the idea that only some of the Fe sites are highly reactive. These results are further discussed in Chapter V. Chapter V contains co-authored unpublished material. M. Stevens, Prof. Boettcher, and I conceived of the project. I performed experiments along with M. Stevens and M. Nellist with technical assistance from S. Fakra. I analyzed data and wrote most of the paper with computational sections written by M. Toroker and with editorial assistance from M. Stevens and S. Boettcher. It is prepared in the format for submission to the journal Angewandte Chemie International Edition. 6 CHAPTER II TRANSITION-METAL-INCORPORATED ALUMINUM OXIDE THIN FILMS: TOWARDS ELECTRONIC SRUCTURE DESIGN IN AMORPHOUS MIXED- METAL OXIDES Chapter II contains co-authored material previously published as J. Phys. Chem. C. 2018, DOI: 10.1021/acs.jpcc.8b00239 reproduced with permission. Copyright 2018 American Chemical Society. Prof. Boettcher, M. Kast, and I conceived of the project. I performed and directed experiments, collected data, and analyzed data with help from M. Kast, E. Cochran, E. Pledger, and M. Stevens. I wrote the paper with help from Prof. Boettcher and M. Kast along with editorial assistance from all authors. Introduction Amorphous and polycrystalline mixed-metal oxides are employed in applications that range from thin-film transistors1–4 and solar cells,5,6 to catalysts for reactions including hydrocarbon oxidation,7–9 hydrogen production via water-splitting,10 and oxygenate reforming.11–13 The versatility of mixed-metal oxide materials is due to their wide range of electronic properties, which depend on chemical composition and physical structure (e.g. crystalline vs. amorphous). For example, the band gap of bismuth- containing mixed-transition-metal oxides, used for oxidation catalysis, varies based on the identity of the transition metal(s) due to the dependence of the lowest unoccupied energy levels on empty d-orbitals.9 Electron-selective contacts to organic photovoltaics have been made by tuning the conduction band position of amorphous In-Ga-O and Ga- Zn-Sn-O.14 Efficient water oxidation catalysts have nominally amorphous mixed-metal- oxide phases on the surface that are responsible for the catalytic activity.8,15 Designing mixed-metal oxides for these, and other, applications requires the ability to systematically vary structural and electronic properties. The fundamental understanding of correlations between composition, local atomic bonding, and electronic structure essential for this goal, however, is lacking – particularly for amorphous and mixed-amorphous-polycrystalline thin films. Much of this knowledge gap is due to the lack of long range order in these systems, which makes typical structural probes (e.g. X- 7 ray diffraction, XRD) inadequate, necessitating the use of synchrotron techniques (e.g. X- ray absorption spectroscopy, XAS).16 The electronic structure is also complex for amorphous and multicomponent systems.17,18 Electron correlation in transition metal oxides can make computation of electronic structure difficult19 for amorphous materials where large structure units are required for calculations.20 For such amorphous oxides, the electronic structure might be predicted based on the experimental atomic and electronic structures of the parent crystalline binary oxides. However, the coordination environment in amorphous films can differ substantially from purely crystalline oxides21– 24 and the electronic structure is dependent on the metal cation coordination.25,26 Here we aim to understand how the electronic structure of mixed-metal-oxide thin films can be controlled to design new materials. We systematically compare the electronic and local atomic structure of MyAl1-yOx thin films, where M = V, Cr, Mn, Fe, Co, Ni, Cu, Zn, or Ga and y varies from 0.1 to 0.7. The materials can be considered mixtures of a wide-band-gap insulating oxide, e.g. amorphous Al2O3 (alumina), and metal cations that have varying number of 3d electrons. Alumina is chosen as the host material because it can readily incorporate many different cations as an amorphous film27 and has a wide band-gap26 that allows for directly measuring the energetics of the electronic states added into the band gap. Additionally, transition-metal-incorporated alumina is of interest for many applications, e.g. catalysis12,28–32 and photovoltaics,33 where the electronic structure determines its utility. We previously showed such multicomponent MyAl1-yOx thin films are easily synthesized from solution precursors and are ultra-smooth over a wide composition range.27 Here we focus on understanding the local atomic and electronic structure of the films, and how they depend on composition. Except for Co, Ni, and Cu at y ≥ 0.2, the MyAl1-yOx films remain amorphous up to y = 0.4 or higher. This composition space allows for an assessment of periodic trends based on number of d-electrons and the metal cation oxidation state. The transition metal concentration range also allows for the study of both amorphous alloy and mixed-amorphous-polycrystalline compositions, giving access to multiple coordination environments for most of the transition metals. 8 We use X-ray absorption spectroscopy to determine the coordination around the transition metal, the metal-oxygen bond length, and the transition metal oxidation state. UV-vis absorption spectroscopy is used to determine the optical charge-transfer band gap and further confirm coordination environment around the transition metals. The electronic structure in the valence band is determined using X-ray photoelectron spectroscopy (XPS). By empirically probing the electronic and local atomic structure, we determine that the d-electron count of the transition metal cation, which is linked with the choice of transition metal, oxidation state, and concentration, influences the final electronic structure and optical properties of the film. Experimental Film Deposition. Aqueous precursor solutions were prepared from transition-metal nitrate (or chloride in the case of V) salts and aq. Al(NO3)x(OH)3-x solutions, as described previously.27 The aq. Al(NO3)x(OH)3-x solution was prepared via the process described by Wang et al.34 Briefly, approximately 40 mL aliquots of 1 M aq. Al(NO3)3 were subject to 1.8 kC of reductive charge (-100 mA for 5 h) while stirring in a glass container using an 80 cm2 Pt-mesh working electrode, a Pt-coil counter electrode (separated by a glass frit), and a DC power supply. This results in an increase of the pH of the solution, causing oligomerization and formation of soluble “flat” [Al13(μ3-OH)6(μ- OH)18(H2O)24](NO3)15 clusters, which we will refer to as the f-Al13 precursor solution. To account for any evaporation, the f-Al13 solution was then standardized to a 1 M [Al] via gravimetric analysis. The f-Al13 precursor was then mixed with the 1 M aq. transition-metal nitrate solutions in appropriate ratios to achieve the desired film compositions (the solution metal ratios are maintained in the film, as shown previously).27 Substrates for film deposition varied by experiment. 500-μm-thick (100) Si was used for XAS and ellipsometry. For XPS measurements, 500 μm-thick (100) Si substrates were coated with 25 nm of Ti (adhesion layer) followed by 50 nm of Ir via electron beam evaporation prior to film deposition (to minimize charging phenomena). Fused-quartz substrates were used for UV-vis and X-ray reflectivity (XRR) measurements. All substrates were cut into 2 × 2 cm squares and cleaned via sonication 9 in aq. detergent solution (Contrad 70, 6.25%) for 10 min. Substrates were then rinsed with 18.2 MΩ·cm water for 30 s and dried for 30 s on a spin coater at 3000 rpm. To render them hydrophilic, the substrates were treated with O2/N2 plasma for 10 min, and finally water-rinsed for 10 s and spin-dried for 20 s at 3000 rpm before deposition. Approximately 250 μL of precursor solution (enough to cover the substrate) was dropped from a plastic syringe through a 0.2 μm filter onto the substrate, which was then spun at 3000 rpm for 30 s. The spun film was immediately placed on an Al block on a hot plate at 150 °C for 5 min to set the film and then moved onto an Al block at 450 °C to anneal for 10 min for multi-layer films or 30 min for single-layer films. To fabricate multi-layer films the substrates were subject to successive water rinses, solution depositions and 150 °C/450 °C annealing steps until the desired number of layers was reached, at which point the film was annealed at 450 °C for 30 min. X-ray Absorption Spectroscopy. XAS measurements were performed on Beamline 12- BM,B (bending magnet) of the Advanced Photon Source at Argonne National Lab. Data were collected at V, Cr, Mn, Fe, Co, Ni, Cu, Zn, and Ga K-edges. Si(111) was used as a monochromator and the incoming beam was detuned 60-80% for harmonic rejection with a 1.5 mm × 0.5 mm spot size. Samples were approximately 300-nm-thick films deposited on Si(100) substrates and were measured in air at room temperature in total fluorescence mode using a 13-element Ge detector (Canberra). The energy was calibrated by measuring the absorption at the K-edge of metallic foils of the same element (except for the L-edge of Ta used for Ga calibration) in transmission mode directly before and after each set of samples (2-5 samples) to verify there was no shift in the energy axis. The edge energy was selected as the highest point in the first derivative of μ(E) and shifted to the metal foil value as listed in the Hephaestus software.35 All sample spectra for a given metal edge were shifted by this same amount before determining E0 for the sample. The energy of the first peak in the first derivative of μ(E) (after the pre-edge feature) was taken as E0 for each sample (Table A.1). All XAS data was analyzed using the Athena/Artemis software package.35 Absorption spectra were merged (between two and six scans) and normalized to an edge step of 1.0 using the Athena software. FT-EXAFS spectra were fit using Artemis 10 software, in which theoretical scattering paths of crystalline oxides (Shcherbinaite- V2O5,36 α-Cr2O3,37 CrO3,38 α-Mn2O3,39 α-Fe2O3,40 Co3O4,41 NiO,42 CuO,43 hexagonal ZnO,44 and β-Ga2O345) were computed using FEFF646,47 with crystal structures obtained from the Inorganic Crystal Structure Database (ICSD). Fits were performed over an R- space of 0.9 – 2 Å in the Fourier transform using k-ranges of 2.5 to 11 Å-1 (V, Fe, Co, Ni, Cu, Zn) or 2.5 to 12 Å-1 (Cr, Mn, Ga) with a Hanning window and k2-weighting. EXAFS spectra plotted in k-space can be found in Figure A.14. The amplitude reduction factor, S02, values for each metal edge were calculated from fits to spectra of polycrystalline films (Shcherbinaite-V2O5, α-Cr2O3, α-Mn2O3, α-Fe2O3, Co3O4, NiO, CuO, hexagonal- ZnO, and β-Ga2O3) and were constrained between 0.7 and 1.0. The obtained S02 value was then used in the fits to sample spectra in order to calculate oxygen coordination around each transition metal. The values of the difference in energy alignment between sample and standard, E0, the difference in path length between sample and standard, ΔR, and the mean square disorder parameter, σ2, were also allowed to vary in the fits (see Eq. A.1).35 The number of independent points ranged from 5.3 to 6, determined by the Nyquist criterion. X-ray Photoelectron Spectroscopy. Valence band XP spectra were collected on a ThermoScientific EscaLab 250 with a monochromatic Al Kα (1486 eV) X-ray source using a 500 μm spot size, 40 eV pass energy and 0.2 eV step size. Core-level spectra (Figures A.3 – A.12) were collected with a pass energy of 20 eV and step size of 0.1 eV. An in-lens electron flood gun was used for charge neutralization. All spectra were shifted so that the adventitious C-H peak was centered at 248.8 eV binding energy. X-ray Diffraction. XRD patterns (shown in Figures A.1 and A.2) were collected on the same samples used for XAS measurements. Data were collected in a grazing-incidence (GIXRD) geometry on a Rigaku SmartLab with a Cu Kα source. The incident beam angle was 0.5° (relative to sample) and reflected signal was collected with a scintillation point detector over a range of 10 – 70° 2θ in step sizes of 0.1° with a 30 s integration time. Ni foil was used to filter out Cu Kβ radiation. Thickness Characterization. The thickness of films on fused-quartz substrates used for optical measurements was measured directly using XRR except for films that were too 11 rough (frequently found for compositions know to phase segregate) to yield Kiessig fringes that could be fit. For these compositions, an identical film was deposited onto a (100) Si substrate and the thickness measured with ellipsometry. XRR was carried out on a Bruker D8-Discover with a Cu Kɑ source. Spectra were collected from 0 – 6° 2θ after sample alignment. The thickness of each sample was determined with Fourier transform fits using GlobalFit software (Rigaku). Ellipsometry measurements were carried out on a Woollam M44 Spectroscopic Ellipsometer. Data were fit with a single Cauchy-model layer to determine the thickness. Optical Absorption Characterization. Transmittance and reflectance measurements were collected on a Perkin Elmer Lambda-1050 UV/Vis/NIR spectrophotometer utilizing a 150 mm integrating sphere. Deuterium and tungsten halogen light sources were used with a Si photomultiplier tube. The transmittance and reflectance measurements were taken at the transmittance and reflectance ports on the integrating sphere, respectively. This allowed for collection of all transmitted (or reflected) light, including both specular and diffuse components. A Speclatron® reflectance standard was used. The optical absorption coefficient was calculated using 𝛼𝛼 = −1 𝑙𝑙 �ln � 𝑇𝑇s 1−𝑅𝑅s∙𝐶𝐶r � − ln � 𝑇𝑇q 1−𝑅𝑅q∙𝐶𝐶r ��, (Eq. 2.1.) where α (cm-1) is the absorption coefficient, l is the film thickness (cm), Ts is the transmittance of a sample, Rs is the reflectance of a sample, Cr is the reflectance correction factor, Tq is the transmittance and Rq is the reflectance of the fused-quartz substrate, respectively.48 The Cr parameter accounts for the difference in the collection efficiency of the integrating sphere as a function of the amount of specular and diffuse reflected light. As reported previously a Cr of 0.8 is needed when a Speclatron® reflectance standard is used with samples that are very smooth and whose reflectance is mostly specular.48 We verified this as a reasonable factor by calculating the absorbance of our fused-quartz substrate with and without using the correction factor. If the correction factor is not used negative non-physical absorbance values are found. This is due to the higher collection 12 efficiency of an integrating sphere with regards to specular reflected light over diffusely reflected light. Thus the “1-R” term is too small resulting in absorbance that is too low (negative if in a region of the spectrum where no absorption is observed). With the correction factor applied no non-physical negative absorbance is observed. Effective optical band gaps were determined by the energy where the absorption coefficient α was ≥ 10-5 cm-1 to avoid tail state transitions that typically occur below this value.49 Results and Discussion As discussed in the following sections, we find that the transition-metal d-electron count, which dictates the crystal-field stabilization energy (CFSE) and therefore preference for octahedral (Oh) or tetrahedral (Td) coordination (based on the octahedral site preference energy, OSPE),50 is the best predictor of the resulting film’s structural and electronic properties. The electronic structure of the different transition-metal- incorporated amorphous aluminum oxide (a-Al2O3) thin films can be generalized into four different groups: metal cations with no d-electrons (d0), those with exactly five d- electrons (d5, high spin), those with partially filled orbitals (dn, but not d5) and cations with completely filled d-orbitals (d10). From our previous work,51 the local structure of a-Al2O3 deposited via spin- casting from aqueous f-Al13 solutions was shown by 27Al solid-state nuclear magnetic resonance spectroscopy to consist of a mixture of 4-, 5-, and 6-coordinate Al species. The average Al coordination number is about 4.6 over a wide annealing temperature range (400 – 800 °C). Only about 10% of the Al in these films exists as 6-coordinate species. This average coordination gives a low oxygen coordination number, (CO), consistent with an open amorphous network. CO is dependent on the metal coordination, CM, and oxidation state, QM, by 𝐶𝐶O = 𝐶𝐶𝑀𝑀|𝑄𝑄𝑂𝑂||𝑄𝑄𝑀𝑀| , (Eq. 2.2) where QO is the charge on the oxygen anions.52 Previously we showed that dn (where n ≠ 0, 5, or 10) cations phase segregate from a-Al2O3 into transition metal oxide particles at relatively low concentration (>10% total metal).27 Here we show that when 13 incorporated into the a-Al2O3 matrix at lower concentrations, these dn cations may adopt a different local coordination or oxidation state from their parent transition metal oxide (i.e. that prepared in this work under the same conditions from the pure transition metal salt, y = 1). This effect is presumably driven by the amorphous network in which a low oxygen coordination (CO) is more energetically favorable for an open network and an amorphous structure.52,53 This phenomenon makes predictions of the electronic structure for mixed-oxide amorphous films difficult, as most non-computational predictions are based on the assumption that bonding motifs are similar between the parent and the mixed-metal oxides.54 Transition-metal cations that are d0, d5 high-spin, or d10 incorporate into a-Al2O3 up to much higher concentrations, likely due to their isotropic d-electron configurations (i.e. no CSFE) allowing them to maintain a low coordination number.27 The average coordination numbers of the transition metal in amorphous films in this work are similar to that of a-Al2O3, ~4.5, suggesting that the transition metals are replacing Al atoms in the amorphous structure. Additionally, there is no diffraction from crystalline Al2O3 or MAl2O4 phases, indicating the films are amorphous alloys. In the next sections we discuss the electronic structure, local atomic structure, and optical properties of each class of metal cations in more detail. d0 Transition-Metal Cations. Both VyAl1-yOx and CryAl1-yOx films contain d0 cations, V5+ and Cr6+, respectively, as determined by the position of the metal K-edge as well as V and Cr 2p XPS (Figures A.4 and A.5). The lack of d-electrons on these cations results in a zero crystal field stabilization energy (CFSE) and therefore no preference for Td or Oh coordination.55 The V5+ and Cr6+ cations readily adopt the lower cation coordination of the a-Al2O3 matrix (~4.5), allowing the mixed films to remain amorphous up to high concentrations (y = 0.5). Fits to the FT-EXAFS data in Figure 2.1b indicate that V5+ cations have an average coordination of 4.5 (Figure 2.1c) in VyAl1-yOx films, presumably a mix of 4- and 5-coordinate V5+. The presence of 4- and 5-coordinate V is further supported by the XANES data in Figure 2.1a, where the intensity of the pre-edge feature (Figure 2.1a inset) is high for VyAl1-yOx films, indicative of a non-centrosymmetric environment. The narrower distribution of V – O bond lengths in VyAl1-yOx films (< 0.2 Å), determined from fits to the FT-EXAFS shown in Figure 2.1c, suggests the V 14 geometry is different than in V2O5 films, despite the similar coordination number. The ability of V5+ cations to incorporate into the a-Al2O3 film is maintained up to high concentration, y = 0.5, as shown previously.27 At y = 0.7 a crystalline VOx phase different from V2O5 is present27 (Figure A.2) however, by XAS the film still appears very disordered; the XANES region (Figure 2.1a) of V0.7Al0.3Ox is more similar to amorphous VyAl1-yOx films than the crystalline V2O5, with no sharp features past the absorption edge. The unique coordination of V5+ cations in VyAl1-yOx films compared to V2O5 is also apparent from their UV-vis absorption spectra in Figure 2.1e. The UV-vis spectrum of V2O5 shows absorption bands centered around 480, 420, 375, and 260 nm, representing transitions from occupied O 2p into empty V 3d orbitals. These appear to correspond to the four d-orbital energy levels created by the crystal field of the square pyramidal geometry of the V5+ cation in the Shcherbinaite-V2O5 phase.56 Additionally, the band gap of the V2O5 film found here (~ 2.1 eV) agrees well with previous reports.56,57 The spectra for VyAl1-yOx films, where y ≤ 0.5, appear to have only two distinct absorption bands, 280 and 225 nm. These likely correspond to O 2p – V 3d transitions into the empty t2 and e orbitals, respectively, of a tetrahedral V5+ cation. The contraction of the d-splitting (as expected in going from square pyramidal to tetrahedral geometry) results in a greater energy difference between the O 2p orbitals and the lowest V 3d orbitals (t2 in distorted-octahedra and e in tetrahedra). This pair of strong absorption peaks has been seen for tetrahedral V5+ in other systems, notably on Al2O3/Al(OH)3 surfaces.58 The broadness of the absorption peaks is likely reflective of the broad range of d-band energy levels due to variation in V-O bond distances and O-V-O bond angles in the disordered amorphous film. It is also notable that the spectrum for V0.7Al0.3Ox more closely resembles those of amorphous VyAl1-yOx compositions, again indicating the similar V – O local coordination that is different from that in V2O5. Due to the lack of d-electrons, the valence band position of the VyAl1-yOx films remains relatively constant with increasing V content, as shown in Figure 2.1d. The optical band gap (Figure 2.1e), however, decreases substantially upon addition of V at y = 15 0.1 to a-Al2O3, suggesting modification of the conduction band. This decrease in effective band gap is thus the result of insertion of V 3d orbitals above the O 2p valence band, but below the Al 3s conduction band of a-Al2O3. The V0.5Al0.5Ox and V0.7Al0.3Ox films have the same effective band gap which is smaller than that of the amorphous VyAl1-yOx films, despite the similar local V coordination, likely due to full delocalization of the V 3d orbitals. In whole, the data thus show it is possible to systematically tune the conduction band edge position of VyAl1-yOx via the concentration of V cations while maintaining an essentially constant valence band position. The conduction band edge position, dictated by the energy of the unfilled 3d orbitals of V, is modulated by the change in orbital overlap of the V5+ 3d orbitals with V5+ concentration. Figure 2.1. Analysis of VyAl1-yOx thin films. a) V K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) V-O bond lengths (closed circles) and coordination numbers (open squares) from fits to the FT-EXAFS, d) valence band XP spectra, e) thin film UV-vis spectra, and f) schematic density of states (DOS) diagram of amorphous VyAl1-yOx films. The grey shaded region in (d) indicates the amorphous composition range and lines between points are meant to guide the eye. The data show that the conduction band edge – and hence optical band gap – can be tuned by V concentration while the position of the valence band edge is kept constant. 16 In our analysis, we find that CryAl1-yOx films contain d0 Cr6+ cations as well as d3 Cr3+ cations. The ratio of Cr6+ to Cr3+ decreases with increasing Cr concentration. The presence of Cr6+ despite starting with a Cr3+ precursor is likely due to oxidation of Cr3+ to HCrO4- by molecular oxidation during spin-casting/annealing. The Cr3+/HCrO4- reduction potential becomes more negative with increasing pH and decreasing concentration.59 We find higher concentrations of Cr6+ in films that have lower total Cr concentration and a higher precursor solution pH (f-Al13 precursor has a higher pH than aq. 1 M Cr(NO3)3). The presence of both Cr6+ and Cr3+ was confirmed by multiple techniques, shown in Figure 2.2. The Cr K-edge in Figure 2.2a gradually shifts to lower energy as y goes from 0.1 to 0.5, indicating a decrease in the average oxidation state of Cr in the films. Additionally, the intensity of the pre-edge feature in the XANES region (Figure 2.2a inset) of CryAl1-yOx films continually decreases with increasing Cr, indicating a decrease in tetrahedral Cr6+ and an increase in octahedral Cr3+.60 The oxidation states of Cr are further confirmed by Cr 2p XPS (Figure A.5). The FT-EXAFS in Figure 2.2b has two distinct first shell peaks for y = 0.2 – 0.5, suggesting there are two different M – O bond lengths; Cr6+ has tetrahedral coordination with a short bond length around 1.6 Å, while Cr3+ is octahedral with a bond length of about 1.9 Å.37,38 Absorption bands from both Cr species can also be seen in the UV-vis spectra in Figure 2.2e for films with y ≥ 0.5 (the symmetry-forbidden Cr3+ d-d transitions in films with lower Cr concentration are too weak to be observed).61 The O 2p – Cr 3d transitions of Cr6+ are centered around 3.4 and 4.5 eV, while the d – d transitions of Cr3+ are centered at 2.0 and 2.6 eV.62 The energy of Cr3+ d-d transitions in CryAl1-yOx films is similar to crystalline α-Al1-xCrxO3.63,64 As the amount of Cr increases, so does the ratio of Cr3+ to Cr6+ resulting in overlapping O 2p – Cr 3d transitions. The presence of Cr3+ also results in a shift of the valence band maximum to higher energies due to the occupied Cr 3d states as shown in the valence band XP spectra in Figure 2.2d. Overall, the data on CryAl1-yOx films show that the electronic structure can be systematically tuned by modulating the valence band edge position through the Cr3+/Cr6+ 17 ratio and the related local coordination geometry. Synthetically, the Cr3+/Cr6+ ratio can likely be modified via the precursor transition metal concentration and solution pH. Figure 2.2. Analysis of CryAl1-yOx thin films. a) Cr K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) Cr-O bond lengths (closed circles) and coordination numbers (open squares) from FT-EXAFS fits, d) valence band XP spectra, e) UV-vis spectra, and f) schematic DOS diagram of amorphous CryAl1-yOx thin films showing both Cr6+ (Td) and Cr3+ (Oh) coordination. The grey shaded region in (d) indicates the amorphous composition range and lines are meant to guide the eye. The data show how the valence band edge can be tuned to higher energy as the octahedral Cr3+ concentration is increased. d5 Transition-Metal Cations. The FeyAl1-yOx films contain d5 Fe3+ ions, determined from the energy of the main edge on the Fe K-edge absorption spectra in Figure 2.3a and further confirmed by XPS (Figure A.7). The d5 electron count and the weak field of the oxygen ligands leads to isotropy in the crystal field, resulting in no difference in CFSE for Td or Oh coordination. Fitting of the FT-EXAFS of the FeyAl1-yOx films (Figure 2.3b and 3c) indicates that the average oxygen coordination around Fe atoms is ~ 4.5, suggesting more tetrahedral as opposed to octahedral Fe3+ ions. Td coordination is also consistent with the increased intensity of the pre-edge XANES feature (Figure 2.3a inset) compared to α-Fe2O3, in which Fe3+ has Oh coordination. It is likely that the ability of 18 the Fe3+ ion to readily adopt a lower coordination, e.g. Td, allows the FeyAl1-yOx films to remain amorphous up to high amounts of Fe.27 The presence of Fe in FeyAl1-yOx films shifts the UV-vis absorption onset to lower energies compared to a-Al2O3, as seen in Figure 2.3e. This decrease in effective band gap is the result of a change in the charge transfer gap from an O 2p – Al 3s transition in a-Al2O3 to O 2p – Fe 3d in FeyAl1-yOx films, as the effective band gap estimated from the UV-vis data, 3.0 eV (Figure 2.3e), agrees well with the charge transfer gap reported for α-Fe2O3.65,66 The forbidden Fe3+ d – d transitions are likely too weak to be observed in the optical absorption spectra for films with low Fe content,61 but are visible in Fe0.7Al0.3Ox and α-Fe2O3 (although the amorphous nature of the former causes significant broadening of the features). Although d-d transitions are not observed in the optical spectra of FeyAl1-yOx films with low Fe content, VB-XPS in Figure 2.3d shows (via the shift in the valence band edge) that there are filled Fe 3d states at the top of the valence band, above the O 2p states. The addition of filled 3d states above the valence band of a-Al2O3 is expected from the five additional valence electrons of Fe3+ compared to Al3+. Figure 2.3. Analysis of FeyAl1-yOx thin films. a) Fe K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) Fe-O bond lengths 19 (closed circles) and coordination numbers (open squares) from fits to FT-EXAFS, d) valence band XP spectra, e) UV-vis spectra, and f) schematic DOS diagram of FeyAl1- yOx films. The grey shaded region in (d) indicates the amorphous composition range and lines between points are meant to guide the eye. The data show how Fe is largely tetrahedral in coordination and introduces occupied states at the top of the valence band that tune the VB edge energy along with unfilled states that give rise to visible optical transitions. MnyAl1-yOx films also contain some cations with a d5 electron configuration, Mn2+, as determined by the slightly lower energy of the Mn K-edge in Figure 2.4a compared to that of the α-Mn2O3 standard. However, a large percentage of Mn in the film exists in the d4 Mn3+ configuration, based on the similarity of the XANES region to the α-Mn2O3 spectrum and the shake-up features of the Mn 2p peaks in the XP spectra (Figure A.6).67 The single peak in the pre-edge, Figure 2.4a inset, for MnyAl1-yOx films where y ≤ 0.4 suggests that Mn is not present in a centro-symmetric Oh geometry, but is more disordered.68 This distortion is possibly due to the presence of multiple oxidation states and the amorphous a-Al2O3 network lowering the coordination number for Mn. Fits to the FT-EXAFS, shown in Figures 2.4b and 2.4c, indicate a low coordination around Mn atoms, ~ 4.0, however the extracted values may be artificially low due to the high amount of disorder in the local geometry (i.e. bond lengths and coordination number) of Mn atoms.69,70 The disorder of the Mn local geometry is also apparent from the UV-vis spectra in Figure 2.4e, where the absorption edge of each spectrum is difficult to determine due to significant tailing. The onset of absorption is at a similar energy (~ 2 eV) to that of MnAl2O4,71 however the spectral features in amorphous MnyAl1-yOx films are significantly broader due to multiple Mn and Al geometries and the overall amorphous structure. Crystalline Mn2O3 is present for films where y > 0.4, but the absorption spectra of these films more closely resemble those of the amorphous MnyAl1-yOx films, indicating much of the Mn is still present in the a-Al2O3 phase. The VB-XPS data in Figure 2.4d show an increase in the valence band edge to higher energy, likely due to the addition of filled 3d states from Mn2+/3+. Given the d-electron counts of Mn2+ and Mn3+, unfilled 3d Mn states are presumably added as well, giving rise to overlapping Mn 3d – 20 3d and O 2p – Mn 3d transitions that account for the tailing absorption in MnyAl1-yOx films. This data shows that, like adding Fe3+, the addition of Mn2+/3+ adds new filled electronic states at the valence band edge and unfilled states in the band gap of a-Al2O3. The data also indicate that precise tuning of the electronic structure of MnyAl1-yOx films is difficult due to the multiple oxidation states and local geometries of Mn cations resulting in many overlapping electronic states. Figure 2.4. Analysis of MnyAl1-yOx thin films. a) Mn K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) Mn-O bond lengths (closed circles) and coordination numbers (open squares) from FT-EXAFS fits, d) valence band XP spectra, (e) UV-vis spectra, and f) schematic DOS diagram of MnyAl1- yOx films. The grey shaded region in (d) indicates the amorphous composition range and lines between points are meant to guide the eye. The data show that the addition of Mn cations in MnyAl1-yOx adds both occupied and unoccupied states of Mn2+ and Mn3+ giving rise to broad optical absorption features. d6,7,8,9 Transition-Metal Cations. Transition-metal cations that are not d0, d5 high spin, or d10 contain partially filled degenerate d-orbitals which leads to a greater crystal-field stabilization energy and a strong preference for octahedral vs. tetrahedral coordination. As we have shown previously, these cations tend to readily phase separate to form the binary transition metal oxides rather than an amorphous alloy with a-Al2O3.27 21 In the XANES edge analysis of CoyAl1-yOx films here, we find Co in both the 2+ and 3+ oxidation states, with increased 2+ content at lower Co concentrations. These oxidation states are also confirmed by Co 2p XPS (Figure A.8). In the binary oxide, Co3O4, Co3+ ions have a d6 electron count and an Oh coordination while Co2+ ions have a d7 electron count in a Td geometry. CoyAl1-yOx films with y = 0.1, which are X-ray amorphous (Figure A.1), have Co in a lower average coordination with longer Co – O bond lengths than in Co3O4 (shown in Figure 2.5c) consistent with more Co2+ in a Td geometry.41,72 The higher Co2+ content in Co0.1Al0.9Ox films is also evident from the lower energy of the main edge in the XANES region of Figure 2.5a. With higher Co content, Co3O4 nanoparticles readily phase segregate from the amorphous alumina network,27 confirmed by XRD (Figure A.1) and the appearance of the second coordination sphere in the FT-EXAFS in Figure 2.5b. The optical spectra of CoyAl1-yOx films in Figure 2.5e are consistent with the atomic structure of the films. In the Co0.1Al0.9Ox film, the optical absorption onset is shifted to lower energies. The shift is likely due to either O 2p – Co 3d or Co 3d – Al 3s transitions; VB-XPS in Figure 2.5d shows the presence occupied of 3d states above those of O 2p in the Co0.1Al0.9Ox film. The thinness and low Co content of this film prevent the observation of Co2+ d – d transitions that are seen in CoAl2O4.73 CoyAl1-yOx films with higher Co content have an absorption onset at 1.5 eV due to Co3+ 3d ‒ Co2+ 3d transition while the feature at 2.0 eV is from O 2p – Co 3d transitions.74 Each film with Co content at y > 0.1 has an absorption spectrum that closely resembles that of the binary oxide, Co3O4, indicating that phase-segregated Co3O4 nanoparticles dominate the optical response.74,75 22 Figure 2.5. Analysis of CoyAl1-yOx thin films. a) Co K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) Co-O bond lengths (closed circles) and coordination numbers (open squares) from EXAFS fits, d) valence band XPS spectra, e) UV-vis spectra, and f) schematic DOS diagram of amorphous CoyAl1-yOx films. The grey shaded region in (d) indicates the amorphous composition range. Lines between points are meant to guide the eye. The data show that at y ≤ 0.1 filled and unfilled 3d states of Co2+ are added to the band gap of a-Al2O3. At higher concentrations Co3O4 particles phases separate and dominate the optical response. Based on the energy of the Ni K-edge XANES spectra in Figure 2.6a, Ni cations in NiyAl1-yOx films are present in the 2+ oxidation state. The Ni oxidation state is further confirmed by Ni 2p XPS (Figure A.9). At low Ni content, y ≤ 0.1, the d8 Ni cations adopt a Td coordination (i.e. instead of the Oh coordination that would be predicted based solely on the CFSE), apparent from the fits to the FT-EXAFS (Figures 2.6b and 2.6c) and the increased intensity of the pre-edge feature in the XANES region, Figure 2.6a inset. This unique Ni geometry is likely due to the low concentration and the influence of the low coordination of the surrounding a-Al2O3 network. Additionally, the Ni – O bond length in the Ni0.1Al0.9Ox film is 2.05 Å, about 0.1 Å shorter than in NiO, consistent with a similar oxidation state, but a lower coordination number. These Ni – O bond lengths are still longer than those of Td Ni2+ in crystalline NiAl2O4, likely due to the amorphous structure of the films. At this low Ni content, the films are also known to be 23 amorphous,27 confirmed by XRD (Figure A.1) and the lack of the second coordination sphere in the FT-EXAFS shown in Figure 2.6b. At higher Ni content (y > 0.1), NiO phase separates from the a-Al2O3 (Figure A.1). The phase separation of NiO is also seen in the FT-EXAFS in Figure 2.6b, where the second coordination sphere intensity becomes apparent above 2 Å. Evidence of NiO in the films with Ni content at y > 0.1 is also clear from the UV-vis spectra in Figure 2.6e, where the absorption edge shifts to ~3.5 eV, representing O 2p – Ni 3d transitions.76,77 VB-XPS data in Figure 2.6d show that there are occupied Ni 3d states above the O 2p states for all NiyAl1-yOx films. There are no clear Ni d-d transitions present in the UV-vis spectra, likely because the films are too thin to see these low intensity transitions. Interestingly, the UV-vis spectra of NiyAl1-yOx films containing NiO particles show a different spectral shape than that of pure NiO, with increased absorption at lower wavelengths, beginning at ~4.5 eV. This may indicate that some of the Ni atoms are still present in the a-Al2O3 phase, resulting in O 2p – Ni 3d transitions that are higher in energy than those of NiO due to the localization of the electronic states on spatially- separated Ni2+ cations. Other evidence of the persistence of Ni in the a-Al2O3 phase is that the average Ni – O bond length in Figure 2.6c shows a gradual increase from Ni0.1Al0.9Ox to NiO. An abrupt change to the Ni – O bond length in NiO would be expected with y > 0.1 if all the Ni atoms were present as NiO, as is seen for CoyAl1-yOx films in Figure 2.5d. 24 Figure 2.6. Analysis of NiyAl1-yOx thin films. a) Ni K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) Ni-O bond lengths (closed circles) and coordination numbers (open squares) from EXAFS fits, d) valence band XP spectra, e) UV-vis spectra, and f) schematic DOS diagram of amorphous NiyAl1-yOx films. The grey shaded region in (d) indicates the amorphous composition range and lines between points are meant to guide the eye. These data show that Ni incorporates as both dispersed Ni2+ cations as well as phase-segregated NiO nanoparticles within the NiyAl1-yOx, both giving rise to new optical features and substantial new filled electronic states at the valence band edge. Based on the energy of the Cu K-edge in Figure 2.7a and the presence of shake-up features in the Cu 2p XPS (Figure A.10), Cu is present as Cu2+ in CuyAl1-yOx films. Similar to films containing Ni or Co, we find CuyAl1-yOx films are only X-ray amorphous up to low concentrations of Cu, i.e. y = 0.2. At y > 0.2, tenorite CuO phase separates from the a-Al2O3 (Figure A.1), a slightly higher concentration than our previous work.27 Unlike Ni or Co, Cu appears to have a slightly larger composition range for forming an amorphous alloy with a-Al2O3, likely due to its lower metal coordination number of ~ 4.2 (Figure 2.7c), and therefore a lower oxygen coordination, CO. Despite being 4- coordinate, Cu still has a smaller amorphous alloy range than d0, d5 or d10 cations due to its square planar geometry. This geometry is indicated by the similarity of the XANES region in Figure 2.7a to the tenorite CuO phase, particularly the lack of increase in the 25 intensity of the pre-edge feature compared to CuO, which would suggest distortion of the square planar geometry.43 For y = 0.1 the optical absorption edge of CuyAl1-yOx films decreases to 3.5 eV (Figure 2.7e). Simultaneously, there is a shift of the valence band maximum to higher energy, Figure 2.7d, indicating the addition of occupied states to the top of the valence band. The shift of the valence band maximum is approximately equal to the shift of the absorption onset (~1.5 eV), possibly indicating that the new charge-transfer process is from occupied Cu 3d states to unoccupied Al 3s or Cu 4s states. The new absorption process also agrees well with reported values of ligand to metal charge transfer (O 2p to Cu 3d) of CuO.78 The absorption spectra of amorphous CuyAl1-yOx films are different than crystalline CuAl2O4,79 likely due to the different Cu2+ and Al3+ geometries in the two types of films. At higher Cu contents, y ≥ 0.5, the optical absorption edge shifts to about 1.5 eV, in agreement with other reports for CuO.80–82 The transition associated with the absorption edge for CuO has been debated;82 our data agree well with reported Cu 3d – 3d transitions of square planar Cu in CuO.78,83 Figure 2.7. Analysis of CuyAl1-yOx thin films. a) Cu K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) Cu-O bond lengths (closed circles) and coordination numbers (open squares) from EXAFS fits, d) valence band XP spectra, e) UV-vis spectra, and f) schematic DOS diagram of CuyAl1- yOx films. The grey shaded region in (d) indicates the amorphous composition range and 26 lines between points are meant to guide the eye. The data show that Cu is maintained in a 4-coordinate square-planar geometry and adds substantial densities of electronic states to the valence band edge that lead to a tunable effective bandgap. d10 Metal Cations. As with d5 cations, the isotropic d-electron configuration of d10 transition metal cations leads to no preference for the octahedral coordination. Therefore, both Zn2+ and Ga3+ can easily adopt a lower coordination that is conducive to forming an amorphous structure. The presence of Zn2+ in ZnyAl1-yOx films is confirmed by the Zn K-edge energy in Figure 2.8a as well as Zn 2p XPS (Figure A.11). Zn can incorporate into the a-Al2O3 matrix and remain amorphous up to y = 0.4 (Figure A.2).27 This is also demonstrated by the lack of coordination spheres beyond the first shell in the FT-EXAFS in Figure 2.8b. Additionally, the UV-vis spectra for y ≤ 0.4 show the absorption onset shifts to energies lower than for a-Al2O3, but higher than for ZnO, seen in Figure 2.8e. This absorption process for amorphous ZnyAl1-yOx films is likely charge transfer from O 2p to Zn 4s orbitals that are localized rather than delocalized as they would be in crystalline ZnO; i.e. there are weak electronic interactions between the spatially well- separated Zn2+ ions. The energy of the absorption onset is also different from that of crystalline ZnAl2O4,84 presumably due to variations in Zn2+ and especially Al3+ geometries in the two types of films. At y > 0.4, ZnO particles segregate from the a-Al2O3 (Figure 2.8b) and second coordination sphere (Zn-Zn distances) appear in the FT-EXAFS data. The crystalline nature of the ZnO is also apparent in the UV-vis spectra of films with y > 0.4 (Figure 2.8e) where the absorption onset shifts to ~ 3.2 eV, consistent with O 2p – Zn 4s charge transfer in ZnO. Similar to NiyAl1-yOx films, there is still increased absorption at lower wavelengths for films containing ZnO particles, likely indicating that some amount of Zn remains in the a-Al2O3 phase. At both high and low Zn concentration, the Zn 3d states make no contribution to the occupied states near the top of the valence band, seen in Figure 2.8d. Instead, occupied Zn 3d states are present in the lower valence band of ZnyAl1-yOx films. Due to the increased nuclear charge and decreased shielding per electron in the d10 configuration, these states are at lower energy than other first row d- orbitals.85 27 Figure 2.8. Analysis of ZnyAl1-yOx thin films. a) Zn K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) Zn-O bond lengths (closed circles) and coordination numbers (open squares) from FT-EXAFS fits, d) valence band XP spectra, e) UV-vis spectra, and f) schematic DOS diagram of ZnyAl1- yOx films. The grey shaded region in (d) indicates the amorphous composition range and lines between points are meant to guide the eye. GayAl1-yOx films also contain a metal cation with a d10 configuration, Ga3+, determined from the Ga K-edge energy in Figure 2.9a and Ga 2p XPS (Figure A.12). Unlike ZnyAl1-yOx films, GayAl1-yOx films remain amorphous at all concentrations of Ga when films are annealed at 450°C. Thermal annealing at 800°C is required to crystallize an all Ga-containing film to β-Ga2O3. The amorphous structure of GayAl1-yOx films is apparent from both XRD (Figure A.2) and FT-EXAFS, shown in Figure 2.9b, where a second coordination sphere cannot be detected. The higher charge on the Ga3+ cation compared to the Zn2+ cation allows for a lower oxygen coordination, CO (i.e. the average number of metal cations bound to each oxygen), which likely contributes to the persistence of an amorphous phase. The UV-vis spectra of the GayAl1-yOx films in Figure 2.9e show that adding Ga up to y = 0.5 causes almost no change in the optical absorption onset. This likely indicates that Ga 4s states are localized and similar in energy to the more delocalized Al 3s states. At higher concentration of Ga, the 28 absorption onset shifts closer to that of a-Ga2O3, indicating greater delocalization of Ga 4s orbitals. These data show that the conduction band position of a-Al2O3 can be tuned via the concentration of d10 cations such as Zn2+ and Ga3+, which add unoccupied TM 4s states to the band gap. Figure 2.9. Analysis of GayAl1-yOx thin films. a) Ga K-edge XANES, b) FT-EXAFS with fits (dashed lines) to the first coordination sphere (solid black line), c) Ga-O bond lengths (closed circles) and coordination numbers (open squares) from FT-EXAFS fits, d) valence band XP spectra, e) UV-vis spectra, and f) schematic DOS diagram of GayAl1- yOx films. The grey shaded region in (d) indicates the amorphous composition range and lines between points are meant to guide the eye. Tuning the Electronic Structure and Optical Properties Amorphous and poorly crystalline mixed-metal oxides have garnered much attention recently, in part due to their ability to stabilize metal cations in unique coordination environments that can be useful for applications including catalysis15,86,87 and controlling charge transport in photovoltaics.14,88 As interest and utility of these materials grow, it is important to have a guided approach to designing the properties of these materials. From the data presented here, we propose that the chemical properties of the transition metal 29 cations in mixed-metal oxide thin films can be used to help predict the resulting atomic and, hence, electronic structures. Number of d Electrons. The d-electron filling of the transition metal can be used to predict the resulting electronic structure and optical properties. Cations with partially filled d-orbitals (e.g. Fe3+, Cr3+, Co2+, Ni2+, Cu2+, etc.) add filled electronic states at the valence band edge of a-Al2O3 as well as unfilled electronic states within the a-Al2O3 band gap. These added cations in a-Al2O3 give rise to strong optical absorption features associated with excitation between the band states of the parent oxide (O 2p and Al 3s) and the new cation states (TM 3d), as well as weaker d-d optical transitions of the incorporated cations. Cations with unfilled d-orbitals (e.g. d0 configuration such as V5+ and Cr6+) do not modify the valence band edge of a-Al2O3, but do modify the optical properties by introducing unfilled electronic states below the conduction band of a- Al2O3. Cations with the d10 electronic configuration, such as Zn2+ and Ga3+, introduce new filled electronic states in the valence band, but these are well below the valence band edge in energy. They also introduce new electronic states in the band gap. The energy of these states depends on the concentration of the Zn2+ and Ga3+. For example, when Zn2+ is at a low concentration the Zn2+ cations are well isolated from each other leading to localized 4s-derived electronic states that are much higher in energy than when Zn2+ segregates to form ZnO domains in the film. The d-electron count of a transition-metal cation also serves as a predictor for the crystal field stabilization energy and the octahedral site preference energy, and therefore the ability of the metal to maintain an amorphous structure.50 When predicting the electronic structure of mixed-metal oxide films made from solution processing, cations with d0, d5 or d10 electron counts can be incorporated into amorphous main-group metal oxides such as a-Al2O3 up to high concentrations, modifying the electronic structure while maintaining the overall amorphous atomic structure. Cations with other d-electron counts incorporated into mixed-metal oxides at high concentrations are likely to form phase-separated transition metal oxides rather than modifying the existing atomic and electronic structure. This result is useful for designing films in which maintaining the amorphous structure and hence ultra-smooth films and interfaces is important for the 30 application, for example in thin-film semiconducting oxide devices,3,25 interface layers,14 or for certain catalytic reactions.15 Cation Concentration and Oxidation state. The concentration of the different cations in the films has a large influence on the resulting atomic structure and therefore the electronic structure. At low concentrations of transition metal ions, films are more likely to form solid solutions or alloys that maintain the local structure of the amorphous host material, here a-Al2O3. As in the case of low concentration CoyAl1-yOx and NiyAl1-yOx films, this may mean the transition metal cation will adopt an oxidation state and/or coordination environment that is different than that of the parent transition metal oxide. For example, in films with y = 0.1, Co remains in the 2+ oxidation state with a tetrahedral coordination and in CryAl1-yOx films more Cr is present as Cr6+ (rather than Cr3+) when the total Cr concentration is low. As we have shown, these different local cation coordination environments result in different electronic structure and optical properties. Exploiting this concentration-dependence of cation oxidation state and local structure could be advantageous for designing thin films with specific electronic structures. For example, Cr6+ in CrOx/Al2O3 catalysts has been shown to be more selective for the desired product, isobutene, in oxidative dehydrogenation of isobutane.89 Catalysts containing Cr6+ have also exhibited lower catalytic activity than Cr3+ for the same reaction as well as for the combustion of CO and C3H6.90 The ability to control the ratio of Cr6+ to Cr3+ in these types of catalysts could lead to more-optimal catalytic properties. Conclusion and Bridge We have assessed the electronic, optical, and atomic structure of a large range of amorphous and amorphous/polycrystalline mixed-metal oxide thin films. The work forms the basis for applications of such materials, for example in catalysis, as interfacial layers to control charge transport, and more broadly in functional coatings. The ability of the film to maintain an amorphous structure can be predicted based largely on d-electron count of the metal cation. By judicious choice of the metal cations included in the oxide, both the electronic structure and optical properties can be systematically tuned – for 31 example by adding filled electronic states at the valence band edge, or localized unfilled states in the bandgap of a-Al2O3. The results here should also be generalizable to other host oxide systems. For example, both In2O3 and Ga2O3 can form amorphous structures at low processing temperatures91,92 and should be amenable to analogous electronic structure modulation by doping with transition metal cations. These oxides have smaller band gaps than Al2O3 studied here and therefore offer a different range of applications. The spectroscopic techniques and methods used in this work are further utilized for characterizing mixed-transition-metal (oxyhydroxides), a similar class of materials. The themes and guiding principles developed in this chapter are likewise used in the following chapters to guide the analysis of these mixed-transition-metal (oxyhydroxides) as oxygen evolution reaction electrocatalysts in aqueous alkaline solution. 32 CHAPTER III EFFECTS OF INTENTIONALLY INCORPORATED METAL CATIONS ON THE OXYGEN EVOLUTION ELECTROCATALYTIC ACTIVITY OF NICKEL (OXY)HYDROXIDE IN ALKALINE MEDIA Chapter III “Effects of Transition-Metal Cations on Nickel (Oxy)hydroxide Activity” contains co-authored material previously published as ACS Catal. 2016, 6(4), 2416-2423 reproduced with permission. Copyright 2016 American Chemical Society. M. Burke and I conceived of the project. I performed the experiments and analyzed data with help from M. Burke and A. Batchellor. I wrote the paper with help from Prof. Boettcher and editorial assistance from all authors. Introduction Solar or electricity driven water electrolysis for the production of hydrogen fuel is a promising renewable energy storage technology.1–3 The efficiency of water-splitting devices is limited in part by the high kinetic overpotential needed to drive the oxygen- evolution reaction (OER), which is coupled to the production of hydrogen gas.4–7 The oxidation of water to O2 proceeds through four electron transfer steps, each of which involves adsorbed intermediates, making the reaction dependent on multiple parameters and the catalytic mechanism difficult to study.8 Generally, the reaction proceeds through oxygen-based intermediates adsorbed to metal-cation active sites.9,10 The binding strength of these intermediates to the metal active site is thought to dictate catalytic activity, and thus the M – O bond strength is used as an activity descriptor.6,11 If the M – O bond is too strong, the OER is rate-limited by product desorption, and if too weak, by precursor or intermediate binding.10,12 Considerable effort has been made over several decades to find earth-abundant electrocatalysts with optimal M-O bond strengths that can decrease both the OER overpotential and the cost of water electrolysis systems.13–20 We have previously identified Ni(Fe)OxHy as the most-active known OER catalyst in alkaline media,13 and found that NiOxHy without Fe is a poor OER catalyst (note we use a general formula here so as not to specify a specific oxidation or 33 protonation state).14,21,22 Further studies in the rigorous absence of Fe impurities of the single-transition-metal (oxy)hydroxides showed new activity trends, FeOxHy> CoOxHy > NiOxHy > MnOxHy,22 with Fe present in all high activity OER catalysts, e.g. Ni(Fe)OxHy and Co(Fe)OxHy).21 Ni(Fe)OxHy can be further optimized and deposited with high loading using a pulsed electrodeposition that ensures uniform Fe incorporation and compact morphology with good electrical interconnection23 and with similar performance to more-complicated composite architectures.24–27 One way to better understand the OER activity enhancement upon Fe addition to NiOxHy is to study other incorporated cations. If other cations have similar properties to Fe, or analogous electrochemical behavior in NiOxHy, this would provide insight into the mechanism of the activity enhancement by Fe. Additional cations may also provide avenues to further enhance activity, electrical conductivity, or chemical stability. Such binary Ni1-zMzOxHy OER catalysts (M is a metal cation) have been studied to identify synergistic effects, particularly for metals such as Co and Fe.13,28–38 Previous studies also report a decrease in the OER overpotential upon La or Ce addition to NiOxHy.28,34 Corrigan reported that the addition of La or Ce dopants induce a similar or lower overpotential to that of Fe in Ni1-zMzOxHy at 16 mA cm-2.28 A challenge in studying other cations in NiOxHy is the potential for Fe contamination. Nickel oxides and (oxy)hydroxides have long been studied for OER catalysis,13,39–42 but, as discussed above, unintentional incorporation of Fe from even the purest commercially available electrolytes is responsible for the high catalytic activity reported.14,16,43 For example, the increased catalytic activity reported39,44–46 for β-Ni(OH)2/β-NiOOH relative to α-Ni(OH)2/γ-NiOOH has been shown to be due to Fe impurity incorporation during the aging step to form the β-Ni(OH)2/β-NiOOH phase.14,47 Due to the ease of Fe incorporation into NiOxHy, previous studies on Ni-based catalysts with other intentionally incorporated cations in alkaline conditions may have Fe contamination, convoluting the true trends in activity. The effects of incorporation of other cations into NiOxHy must be studied in rigorously Fe-free conditions. In some previous studies of Ni1-zMzOxHy, it has been suggested that an anodic shift of the Ni2+/3+ oxidation peak potential is connected to increased OER activity.14,30,48 One hypothesis is that if it is more difficult to oxidize Ni2+, possibly due to charge 34 transfer to Fe cations, the oxidized Ni cation will be able to better oxidize water.14,30,49 This “electron-withdrawing effect” hypothesis assumes the host matrix cation (Ni or Co) is the active site. In support of this view, Corrigan and Bendert found that the addition of La and Ce increased the Ni2+/3+ oxidation peak potential and decreased the OER overpotential,28 but the degree to which these samples contain Fe impurities is unclear. CoOxHy also incorporates Fe to become more active (on a total-metal-cation basis) with a concurrent anodic shift in the Co2+/3+ oxidation peak potential.15 Recent work, however, has provided compelling evidence that Fe cations are the active sites within Ni/CoOxHy matrices and adding Fe largely simply increases the number of active sites.15,22,50–52 In this case, other cations that increase the Ni2+/3+ oxidation potential would not lead to higher activity unless they too are more-active than the Ni sites. Here we report effects of Ti, Mn, La, and Ce cation incorporation on the OER activity and Ni2+/3+ oxidation potential of Ni1-zMzOxHy in Fe-free alkaline solution using cyclic voltammetry in combination with electrochemical quartz-crystal microgravimetry. Metal cations were chosen based on either previous reports of high activity,26,28,34 i.e. La and Ce, similar d-orbital filling to Fe,17,18,38,53 i.e. Mn, or anticipated ability to shift the Ni2+/3+ oxidation peak to higher potentials without increasing activity,54,55 i.e. Ti. We prepared mixed Ni1-zMzOxHy films by both solution spin-casting and electrodeposition. While we focus our analysis on the spin-cast films, the same behavior is observed on the electrodeposited films indicating the activity trends are intrinsic to the compositions and not dependent on synthesis method. The thin films (~ 20 nm thick, or a loading of 4 µg cm-2) used allow for estimation of intrinsic activities of the mixed metal hydroxide films with minimal complication from other factors (e.g. electronic conductivity, mass transport, and inefficient catalyst utilization).13,22,51,56 We show that only Ce enhances the OER activity of NiOxHy in 1 M Fe-free KOH, although the enhancement is much less than with Fe and is also transient. We also find no strong correlation between activity and the Ni2+/3+ redox potential. These results highlight the uniqueness of Fe and are consistent with Fe being the active site for the OER in Ni(Fe)OxHy. Understanding the effects of dopants on the activity of OER catalysts will help guide the design of improved catalysts. 35 Experimental Spin Cast Film Deposition. NiOxHy films were fabricated by both spin coating from precursor solutions and from cathodic electrodeposition. Films were spin cast from solutions of 0.1 M Ni(NO3)2·6H2O (99.9985% Strem) in 1:1 (v/v) mixture of ethanol and water onto Au/Ti or Pt/Ti 5 MHz AT-cut quartz crystal microbalance (QCM) substrates (1.38 cm2) from Stanford Research Systems. QCM substrates were sonicated in a solution of Contrad 70 (5% v/v) at 40 °C for 20 min followed by rinsing with water (18.2 MΩ·cm). The electrode surface was then rendered hydrophilic with a 10 min O2 plasma etch. About 0.25 mL of metal nitrate salt solution was deposited onto the substrates through a 0.1 µm filter and films were spun at 3000 rpm for 30 s. Substrates were then immediately transferred to a hot plate at 150 °C and annealed for 10 min. Spin-cast Ni1- zMzOxHy films with Ti, Mn, La, Fe and Ce were prepared in the same manner by mixing the appropriate amount of TiOSO4·xH2O, Mn(NO3)2·xH2O, Fe(NO3)3·6H2O La(NO3)3·6H2O, or Ce(NO3)3·6H2O with the Ni(NO3)2 to obtain added metal concentrations of nominally 10 and 30 at. % (on a total metals basis). The as-deposited mass of each film was obtained by measuring the frequency of the crystal before and after deposition and converting the change in frequency to mass using the Sauerbrey equation57 Δf = -Cf × Δm, where Δf is the observed frequency change (Hz), Cf is the sensitivity factor of the 5 MHz AT-cut quartz crystal (56.6 Hz μg−1 cm2) and Δm is the change in mass per unit area (μg cm−2). To ensure viscoelastic effects of film damping were not affecting the validity of the Sauerbrey relationship we measured the crystals’ resonance behavior using impedance spectroscopy before and after film deposition in air and in 1 M KOH – the resonator quality factor was unaffected by the addition of the film. We have also previously shown that the mass loadings determined by QCM measurements are consistent with those determined by integration of the Ni redox wave for wide range of film thicknesses.23 Total metal content was determined by assuming complete Ni(OH)2 formation, accounting for added cations substituting for Ni. Typical films loadings were 4 – 5 µg cm-2 which is ~10-20 nm in thickness, based on calculations using the mass loading and density of Ni(OH)2. 36 Electrochemical Film Deposition. Electrodeposited NiOxHy films were deposited directly onto the QCM substrates from 0.1 M aq. Ni(NO3)2. Deposition was performed in a two-electrode glass cell with a carbon-cloth counter electrode (Fuel Cell Earth) by applying 1 mA cm2 of cathodic current for 30 to 90 s (the time was varied to achieve the same mass loading as the spin-cast films). Ni1-zMzOxHy films were electrodeposited in the same fashion by mixing the appropriate amount of transition-metal-nitrate salts (or the chloride salt in the case of Fe) in the deposition solution to obtain 10% in the films (see Table B.1) and varying the cathodic current density (-0.1 to -1 mA cm-2) if necessary to obtain full coverage. Ti-incorporated NiOxHy films could not be electrodeposited due to formation of a gel-like film with poor adhesion to the electrode surface. The as- deposited mass for each of the electrodeposited films was determined based on the frequency change of the QCM crystal in a beaker of water (18.2 MΩ·cm) measured directly before and after deposition. Electrochemical Characterization. Following deposition, NiOxHy and Ni1-zMzOxHy films were characterized electrochemically by cycling in 1.0 M KOH (semiconductor grade 99.99% trace metal basis) that had been cleaned of Fe impurities using high-purity precipitated Ni(OH)2 as an Fe-absorbent following our reported method.14 Cyclic voltammograms (CVs) at 20 mV s-1 and steady-state chronopotentiometry measurements (Tafel analysis) were recorded using a BioLogic SP200 or SP300 potentiostat in a polymethylpentene three-electrode cell with a Pt-wire counter electrode, Hg/HgO (CH Instruments) reference electrode, and QCM working electrode (Stanford Research Systems). The electrolyte was continuously stirred to remove bubbles from the working electrode surface and high-purity O2 was continuously bubbled through the solution before and during the measurements. All results shown here have been corrected for uncompensated series resistance of the electrolyte based on electrochemical impedance measurements taken after cycling each film. Uncompensated resistances were 4-7 Ω. Current density was calculated based on the 1.38 cm2 geometric surface area of Au on the QCM electrodes. Turnover frequencies (TOFs) are calculated based on the average of the forward and reverse current from the CV at a specific overpotential and the total number of metal centers (TOFtm) obtained from microbalance measurements accounting 37 for the percentage of added transition metal. TOFtm reported implicitly assume all metal cations are active sites, and thus represent lower limits. To determine whether all the films had similar internal electrolyte-accessible surface area we compared the integrated charge in the nominally Ni2+/3+ redox wave (Figure B.1). Most of the films showed variations of less than a factor of 2 and 75 +/- 25% of the Ni cations electrochemically accessible. The Ni(Ce)OxHy films showed the lowest fraction of accessible Ni Cations ~30%, indicating that the TOFtm are slightly under estimated for the Ni(Ce)OxHy. In situ mass measurements were based on the frequency change of the crystal starting with the point at which the crystal was first immersed in the electrolyte. Physical Characterization. The morphology of each film was examined by scanning electron microscopy (SEM) on a Zeiss Ultra 55 at 5 keV with an in-lens detector. X-ray photoelectron spectroscopy (XPS) was used to determine the overall composition of the films. XPS was performed on a ThermoScientific ESCALAB 250 using a monochromatic Al Kα (150 W, 20 eV pass energy, 500 µm spot size) or non- monochromated Mg Kα (200 W, 40 eV pass energy) source. An in-lens electron flood gun was used for charge neutralization of the samples. Spectra were analyzed using ThermoScientific Avantage 4.75 software and the binding energy was calibrated using the adventitious C 1s peak at 284.8 eV. Stoichiometries of the films are those obtained from XPS measurements on one representative sample of each composition. Fe impurity concentrations were also determined for NiOxHy and Ni0.95Ce0.15OxHy films using an ION-TOF Model IV time-of-flight secondary ion mass spectrometer (ToF-SIMS) with a Bi+ primary ion, O2 sputter source operating at 1 keV, and a flood gun for charge neutralization, all operating in interlaced mode. Concentrations were calculated based on comparisons to the signal intensity in ToF-SIMS of a spin-cast Ni0.98Fe0.02OxHy control film (2% Fe intentionally added in the deposition solution which was confirmed by XPS of the film). Results and Discussion Dependence of Film Morphology on Added Cation Identity. SEM imaging of the solution spin-cast films shows similar morphology (Figure 3.1). The films are 10 to 20 nm in 38 thickness (estimated from the as-deposited mass). The NiOxHy, Ni0.9Mn0.1OxHy, Ni0.93La0.07OxHy, and Ni0.9Ti0.1OxHy films all show a rough, sponge-like morphology, with small cracks revealing the gold substrate underneath the film. The Ni0.85Ce0.15OxHy and Ni0.9Fe0.1OxHy films have a more platelet-like morphology and better coverage of the underlying substrate. It is unlikely that these small differences account for order(s) of magnitude differences in activity (discussed below). The morphologies of spin-cast Ni1- zMzOxHy films at z ≈ 0.1 are similar to those at z ≈ 0.3 (Figure B.2) and they are also similar to the electrodeposited films (Figure B.3). This similarity to the electrodeposited films (which form by local pH gradients driving metal hydroxide precipitation) is explained by the fact that the spin-cast films are annealed only at a low temperature of 150 ºC. This temperature is not sufficient to completely decompose the salt counter ions and form crystalline oxides. Upon immersion of the precursor film in KOH, the hydroxides likely exchange with the remaining precursor counter ions leading to local precipitation in the film of the (oxy)hydroxide phases. Electrochemical cycling likely speeds this transformation process. 39 Figure 3.1. SEM images of spin-cast films after 50 cycles between 0.2 and 0.85 V vs. Hg/HgO in Fe-free 1 M KOH at 20 mV s-1. Effects of Added Cations on OER Activity. The voltammetry in Figure 3.2 shows that Fe has the most pronounced effect on Ni1-zMzOxHy electrochemical behavior. Ni1-zFezOxHy has the lowest OER onset potential and the lowest overpotential at 3 mA cm-2 at z ~ 0.1 and 0.3. The Ni0.7Fe0.3OxHy films have the best performance, consistent with previous reports.14,47,50,58 Besides Fe, Ce is the only other cation tested that significantly increases the OER activity, but this activity decays with time, as discussed below. Ti, Mn, and La have little effect on the activity of Ni1-zMzOxHy, which is different than previously reported.26,28,33,34,38 We have previously shown NiOOH to be sufficiency electrically conductive under OER conditions (0.1 – 0.2 S cm-2), such that the electrical iR drop 40 across the thin films studied here would be negligible. Given that the films here are largely composed of NiOOH, we thus expect the observed differences in electrochemical behavior to not be due to differences in film electrical conductivity. Figure 3.2. Voltammetry of spin-cast Ni1-zMzOxHy films, where M is a metal cation, at approximately a) z ≈ 0.1 and b) z ≈ 0.3 in Fe-free 1 M KOH at 20 mV s-1. For the initial cycle 1 scans, see Figure B.4. In Figure 3.3, the OER activities of the different films are quantified at an overpotential of 400 mV. The activity is calculated as a total metal turnover frequency (TOFtm), which we define as moles of O2 per mole of metal per second. Here, both Ni and added metal cation are included in the moles of metal for the TOFtm calculation, and the number of moles is determined from the in situ mass from the QCM working electrode. The moles of O2 are determined from the amount of current, through 41 Faraday’s constant and assuming a four electron transfer, taken from the average of the forward and reverse scan of the cyclic voltammetry at 20 mV s-1 in 1 M KOH. We note that this is not a steady-state TOFtm, but it is reasonable for comparisons of the films to each other. Figure 3.3 shows that the activities of Ni0.85Ce0.15OxHy and Ni0.65Ce0.35OxHy films are initially (i.e. after cycle 5) about a factor of 8 and 4 higher than that of NiOxHy, respectively. However, after 50 cycles, the performance decreases, and the activity of Ni0.85Ce0.15OxHy is only a factor of 2 higher than NiOxHy and Ni0.65Ce0.35OxHy is statistically equivalent to NiOxHy. This decrease in activity over the course of 50 cycles for Ni(Ce)OxHy films is double that of the other films, which, on average, decay in activity by only a factor of 2. By contrast, Ni1-zFezOxHy films retain their activity throughout 50 cycles. The activity enhancement from Ce incorporation into NiOxHy is a transient effect. The same is observed for the electrodeposited Ni(Ce)OxHy films. This is consistent with the work of McCrory et al.,33,34 which shows a decrease in the OER current density of electrodeposited Ni(Ce)OxHy over time, but different in that we find high initial activity for Ni(Ce)OxHy. Figure 3.3. OER turnover frequency of spin-cast films at 400 mV overpotential at cycle 5 (solid) and cycle 50 (pattern) in Fe-free 1 M KOH from voltammetry data collected at 20 mV s-1. In the case of Ni0.9Fe0.1OxHy there is no activity decrease over 50 cycles and the error bars small and not visible on this scale . TOFtm are calculated assuming all metal cations are active (and thus are lower limits) and the OER current is taken as the average of the forward and reverse scans. Values reported are the average and error bars are the standard deviation of 3 samples. Note the different scale on the y-axis for Ni1-zFezOxHy films. Transient Activity Enhancement from Ce Addition. At 15% concentration, Ce initially decreases the overpotential for the OER on Ni0.85Ce0.15OxHy, with an OER onset 42 potential only 40 mV higher than that of Ni0.9Fe0.1OxHy (Table B.2). However, the current does not rise as sharply for Ni0.85Ce0.15OxHy films (see Figure 3.2) resulting in decreased performance at higher overpotentials compared to Ni0.9Fe0.1OxHy. Unlike 30% Fe addition to Ni1-zMzOxHy, higher Ce concentration (~35%) does not result in further performance improvement and in fact decreases it compared to 15%. These results are inconsistent with those of Corrigan and Bendert, who report a lower overpotential at 16 mA cm-2 for Ni(Ce)OxHy than Ni(Fe)OxHy.28 The results are consistent with McCrory et al.33,34 who showed that Ni(Ce)OxHy has a higher overpotential at 10 mA cm-2 compared to NiOxHy, but exhibit higher current densities at lower overpotentials (~350 mV) on glassy carbon electrodes. We show that initially both Ni0.85Ce0.15OxHy and Ni0.65Ce0.35OxHy are more active than pure NiOxHy. After 50 cycles only Ni0.85Ce0.15OxHy remains more active, by about a factor of 2 (Figure B.5). Continued cycling would likely decrease the OER activity of Ni0.85Ce0.15OxHy until it is equal to or lower than that of NiOxHy, as observed during the steady-state Tafel analysis (Figure 3.4a). Mechanisms of Ni(Ce)OxHy Deactivation. In order to understand the cause of the transient enhancement from Ce addition, Ni(Ce)OxHy films were characterized by XPS before and after 50 CVs in 1 M KOH. One hypothesis is that gradual dissolution of Ce from the Ni(Ce)OxHy films may be the cause of the decrease in activity with cycling. However, the as-deposited and post-electrochemistry films have comparable Ce concentrations (by XPS), 18% and 15% relative to Ni, respectively. The lack of Ce dissolution is also corroborated by the minimal change in film mass after 1 h under OER conditions (Figure 3.4b). Ce is not expected to be highly soluble under alkaline OER conditions based on thermodynamic Pourbaix diagrams, which show CeO2 as the predominant phase.59,60 43 Figure 3.4. a) Tafel plots for spin cast Ni1-zMzOxHy films along with b) in situ mass change over the course of hour-long Tafel measurements. Each chronopotetniometry step, starting at low current density, is held for 3 min with the exception of the first two steps, which are held for 10 min each. Dashed lines are meant to guide the eye and be representative of the Tafel slopes calculated. Data points are the average and error bars are the standard deviation of 3 individual samples. Some error bars are smaller than the symbols. Actual Tafel slopes are listed in Table B.3 and individual plots shown in Figure B.6. Another hypothesis for the decaying activity of Ni(Ce)OxHy films is CeO2 phase segregation in the film during electrochemical measurements such that active sites relying on Ni-O(H)-Ce motifs are eliminated. There is evidence to support this hypothesis from the O 1s XP spectra of the Ni(Ce)OxHy films. The as-deposited films (Figure B.7) show only hydroxide character where-as after cycling the films with Ce exhibit oxide character, evident from the O 1s peak shape and binding energy in the XP spectra (Figure 3.5). This suggests that the Ce3+ ions (which are present in the deposition solution and as-deposited films, Figure B.8) are better distributed in the Ni(OH)2 lattice at the start of electrochemical characterization. After cycling, phase-segregated CeO2 nanoparticulates in the Ni(OH)2 film may form, which have been reported before for 44 related Ni-based films.61,62 Haber et al. found that CeO2 segregated from NiCoFeOx in mixed NiCeCoFeOx films, although we note that those films were annealed at higher temperatures (300 °C) than here (150 °C) and thus are composed primarily of oxides.63 Of all the films studied, Ni(Ce)OxHy has the most oxide character (Figure 3.5). Since CeO2 is not apparent in the as-deposited films, phase segregation (likely driven by Ce3+ oxidation) is probably responsible for the decrease in performance of the Ni(Ce)OxHy with cycling. More analytical work is needed to confirm this hypothesis, for example by synchrotron x-ray absorption measurements, as traditional analyses by electron microscopy or diffraction are challenged by the largely amorphous and disordered thin- film structures. We also note that our hypothesis on Ce-enhancement is different than that of Haber et al., where the nominally high OER activity of NiCeCoFeOx films was attributed to the presence of CeO2 nanoparticles.63 However, those films show CeO2 phase segregation as-deposited and also have high concentrations of Fe. The electrodeposited NiCeCoFeOx films in that work (which are less likely to have CeO2 phase segregation as-deposited) show a similar decrease in activity during cycling, consistent with our hypothesis that phase segregation of CeO2 accounts for the transient OER activity of Ni(Ce)OxHy films. 45 Figure 3.5. O 1s XP spectra for Ni1-zMzOxHy films after 50 cycles in Fe-free 1 M KOH at 20 mV s-1 showing the increased oxide character of Ni(Ce)OxHy films compared to the other mixed metal films. The grey shading indicates the range of metal hydroxide O 1s binding energy and the yellow shading indicates that of metal oxides.64 Effects of Ti, Mn, and La on NiOxHy OER Activity. In contrast to the three and one order(s) of magnitude activity enhancement from Fe and Ce, respectively, the Ti, Mn, and La cations have little effect on the OER activity of NiOxHy. Addition of 7% La results in only a 20 mV decrease of the OER onset potential compared to NiOxHy (Table B.2). The performance of the Ni0.93La0.07OxHy catalyst is considerably worse than that reported by Corrigan,28 but slightly better than that reported by McCrory et al.33,34 The OER activity of Ni0.93La0.07OxHy films are about a factor of two higher than NiOxHy at both cycle 5 and 50. Addition of 38% La results in a decrease in OER performance with the films exhibiting higher overpotentials than NiOxHy films at all current densities. 46 Addition of Mn to NiOxHy did not result in a significant increase in OER activity, consistent with some previous reports for electrodeposited films,22,28 but in contrast to that of Diaz-Morales et al.38 With ~10% Mn there is only a slight decrease in overpotential. As with La and Ce, a higher concentration of Mn, ~ 30%, results in decreased performance compared to NiOxHy. This is in contrast to higher concentrations (~30%) of Fe, which lead to improved OER activity. Unlike the other added cations, incorporation of 10% Ti resulted in higher overpotentials and lower activity than NiOxHy films, shown in Figures 3.2 and 3.3, respectively. Interestingly, higher amounts of Ti slightly increased performance, with Ni0.7Ti0.3OxHy films exhibiting higher activity than Ni0.9Ti0.1OxHy films, but not higher than NiOxHy films. Unlike the other films tested, Ni0.7Ti0.3OxHy films increase in activity with cycling. We attribute this to the denser nature of the as-deposited film (likely due to crosslinking from the reactive Ti precursor) blocking access to some of the Ni sites. With cycling, the film becomes hydrated and more-permeable to electrolyte, allowing for catalytic participation from more of the metal sites. This hypothesis is supported by the fact that the charge in the Ni2+/3+ wave increases with cycling for Ni0.7Ti0.3OxHy films (Figure B.9). Analysis of Fe Impurities. A number of experiments were performed to show that the results presented here are due to the intentionally incorporated cations in NiOxHy rather than Fe impurities that could originate from precursors salts or electrolytes. All compositions were found to be free of Fe contamination by XPS (Mg Kα excitation) after electrochemical characterization (Figure B.10). Spin-cast NiOxHy and Ni0.85Ce0.15OxHy films were further checked for Fe contamination using ToF-SIMS and were found to contain 0.025 and 0.05% Fe, respectively, relative to total metal content. The same films showed no Fe by XPS, confirming the lower sensitivity limit of ToF-SIMS. We do not believe that this small difference in Fe content between the two catalysts accounts for the ten-fold OER activity difference.14,43 First, the transient nature of the activity enhancement from Ce suggests that it is not due to the presence of Fe, because the Fe enhancement is stable. Second, Corrigan conducted an analysis of Fe impurities in Ni oxide electrodes and did not find a tenfold activity enhancement until >1% Fe.43 Third, 47 the Tafel slopes of Ni(Ce)OxHy films are not lower than that of NiOxHy. We also note that trace Fe has been shown to dramatically activate Au substrates for OER, but we do not see any activation of the Au substrate (Figure B.11), indicating it is not affecting our results.14,15,51,65 We found the same increased OER activity of Ni(Ce)OxHy compared to NiOxHy on Pt substrates (which are not enhanced by Fe impurities)51 suggesting that substrate activation of the catalyst material is not playing a significant role in the observed trends (Figure B.12). We also find that CeOx alone mildly activates Au for OER, but is substantially less active than Ni(Ce)OxHy or NiOxHy (Figure B.13). Tafel Behavior and Stability. Steady-state Tafel analysis was performed on the films with 10% added metal cations as well as the pure NiOxHy (Figure 3.4a). Only the addition of Fe lowers the Tafel slope of NiOxHy. We find Ni0.9Fe0.1OxHy has a slope of ~40 mV dec-1, which is consistent with other reports23,58 for ~10% Fe in NiOxHy. NiOxHy alone has a higher Tafel slope of ~93 mV dec-1. This suggests a difference in the rate- limiting step of oxygen evolution on Ni0.9Fe0.1OxHy and NiOxHy catalyst films.66,67 The other cations (La, Ti, Mn, and Ce) have little effect on the Tafel slope. For La, Ti, and Mn, the Tafel slope is not statistically different from that of NiOxHy. Determining the Tafel slope of Ni0.85Ce0.15OxHy is complicated due to the rapidly decreasing activity (Figure B.6). Ni0.85Ce0.15OxHy is marginally better than NiOxHy at low overpotentials, but worse at high overpotentials (Figure 3.4a). Because the Tafel data was collected from low to high overpotential these results are explained by the instability of the Ni0.85Ce0.15OxHy discussed above. The chemical stability of the films was determined by monitoring the film mass during chronopotentiometry steps of the Tafel analysis (Figure 3.4b) as stability during cycling is difficult to ascertain due to movement of ions/ solvent into and out of the film during oxidation/reduction (Figure B.14). All the films were stable and insoluble during the ~1 h of chronopotentiometry measurements. NiOxHy along with La, Mn, and Ce- incorporated films have a mass change of less than ± 3%. The Ni0.9Fe0.1OxHy films were more variable, but exhibited an overall mass change of ± 10%. We note that the differences in mass change could be the result of variable amounts of NO3- (and SO42- for 48 Ti films) left in the films after spin coating, which may not be fully removed before the Tafel analysis as well as Ni2+ ions left in the 1 M KOH after cleaning with Ni(OH)2. Cation Influence on Electronic Structure. It has been hypothesized that an electron withdrawing effect on NiOxHy, either by added cations or a noble metal substrate, increases the ability of NiOxHy to catalyze water oxidation.14,30,39 For example, one hypothesis for Fe enhancement is that charge transfer from Ni to Fe makes it more difficult to oxidize NiOxHy (evident from the anodic shift of the Ni2+/3+ oxidation peak) therefore giving it more oxidizing power and faster OER kinetics.14,30 The formal potential E0' for the Ni redox wave may thus be a activity descriptor. Recently, Haber et al. correlated the Tafel slope of catalysts in the Ni-Fe-Co-Ce-Ox composition space to the redox potential of the catalysts, likely a combination of the Co2+/3+ and Ni2+/3+ redox potentials.48 However, the lower Tafel slopes found for catalysts with higher redox potentials may only be due to Fe incorporation which is known to both lower the Tafel slope and anodically shift the redox peak for both Ni and Co based catalysts.14,15,28,43 Figure 3.6 shows TOFtm for each Ni1-zMzOxHy catalyst versus the formal potential E0' of the Ni redox wave, where E0' is taken as the average of the anodic and cathodic peak potentials, Epa and Epc. There is no strong correlation between E0' and TOF, with the exception of Fe showing both increased E0' and TOFtm. The data show that 10% Fe induces an 18 mV anodic shift of the formal potential along with two orders of magnitude increase in activity. At 30% total metal, cations such as Ce and Ti induce similar anodic peak shifts, 13 and 36 mV, respectively, but do not result in comparable increases in activity. This suggests that the electronics of the Ni cation, as reported by the formal potential of the Ni couple, are not affecting the OER activity. This result is consistent with Ni not being the active site in Ni(Fe)OxHy. We do note that the measured formal potentials can be influenced by kinetic and ion-transfer resistances in the film, which add some uncertainty to the above analysis. The fact that the Ni2+/3+ redox potential is not a good descriptor of catalytic activity suggests that Fe or Ce may be the OER active site in the doped films. Ce, like Fe, shows 3+/4+ redox activity. We find that increasing Fe content in Ni1-zFezOxHy films from 10 to 30% results in increased activity, consistent with Fe acting as the active site 49 within the NiOxHy matrix. Increasing Ce content from 10 to 30% decreases activity. This is likely explained by a low solubility of Ce4+ in NiOxHy, due to the difference in cation size. Figure 3.6. TOFtm vs. E0' of Ni2+/3+ for Ni1-zMzOxHy after 5 (top) and 50 (bottom) cycles in 1 M Fe-free KOH at 20 mV s-1 at low concentrations (closed squares) and higher concentrations (open circles) of added cations. The formal potential, E0', is taken as the average of the cathodic and anodic peak potentials for the Ni2+/3+ redox wave. Comparison of Film Preparation Methods. In heterogeneous electrocatalysis, activity is often dependent on synthesis method due to changes in surface area or electrical connectivity.13,23,56 We investigated if activity trends determined for spin-cast Ni1- zMzOxHy films also hold for their electrodeposited counterparts of similar composition. The OER activity, electrochemical cycling behavior and morphology are similar for the electrodeposited and spin-cast films (Figures B.3 and B.15-17). The solution spin-casting method likely contributes to the similarities between spin-cast and electrodeposited films, 50 as discussed in Section 3.1. The low annealing temperature of the spin-cast films, 150°C, prevents formation of the oxide phases. The as-deposited films have more hydroxide character than spin-cast films annealed at higher temperatures above 300 ºC,13 which is evident from the Ni2+/3+ oxidation wave that remains largely unchanged throughout 50 cycles for most of the spin-cast films (Figure B.5). These low-temperature spin-cast hydroxide films are thus easier to convert to the (oxy)hydroxide phases that are the active phase present during the OER.50,68–70 They also have the advantage over the electrodeposited films of precise composition and thickness control. We note that for both preparation methods the degree of cation mixing is unknown due to the difficulty in obtaining local structural data in thin, disordered films from conventional x-ray diffraction. However, because these films are electrolyte- permeated at the molecular level, added cations should be electrochemically accessible. The short (10 min) low-temperature (150 ºC) processing makes complete phase segregation during synthesis unlikely. Additionally, the observed electronic changes to the NiOxHy films with added cation (i.e. changes in the Ni2+/3+ redox wave) suggest interaction between cations on the molecular level. Further studies such as X-ray absorption measurements are needed to understand local structural differences which may affect catalytic activity. Conclusion and Bridge We studied Ni1-zMzOxHy films, where M is Ti, Mn, Fe, La, or Ce, and found that only Fe increases the OER electrocatalytic activity. Although adding Ce induces a ten- fold OER activity increase for NiOxHy, the enhancement decays quickly. Ti, Mn, and La cations have almost no effect on NiOxHy catalytic activity. The results also inform the discussion of which cation(s) are the active site for the OER. We showed that there is no strong correlation between the formal potential of the Ni2+/3+ redox couple and the electrocatalytic OER activity, suggesting that Ni site electronics are not of primary importance to the OER mechanism. This is consistent with Fe being the active site in Ni(Fe)OxHy. We also demonstrate that these conclusions are intrinsic to the Ni1-zMzOxHy by comparing films fabricated with two different deposition methods, solution spin- casting and electrodeposition. In sum, our results illustrate the apparently unique effects 51 of Fe cations, and point to other Ni1-zMzOxHy compositions as analogous systems for understanding the mechanisms behind those synergistic effects. Given that Fe appears to be uniquely important in transition metal (oxy)hydroxide electrocatalysts for the OER, further studies of its properties are warranted. In the following Chapter further investigation of the intrinsic activity of FeOxHy will be performed to understand the exceptional activity of Fe and Fe-incorporated catalysts. 52 CHAPTER IV ELECTRODE SUBSTRATE DEPENDENCE OF IRON (OXY)HYDROXIDE ELECTROCATALYTIC ACTIVITY FOR THE OXYGEN EVOLUTION REACTION Chapter IV contains co-authored unpublished material. Prof. Boettcher and I conceived of the project. I performed experiments and analyzed data as well as directed experiments and analysis performed by undergraduate A. Vise. I wrote the paper with help from Prof. Boettcher and A. Vise along with editorial assistance from Boettcher and M. B. Stevens. It is prepared in the format of the Journal of the American Chemical Society. Introduction The oxygen evolution reaction (OER, in aqueous alkaline solution: 4OH- → O2 + 2H2O + 4e) is a source of efficiency loss for several renewable energy technologies including water electrolysis and CO2 reduction.1–4 Significant effort has been made to understand and design electrocatalysts for this reaction. Iron (oxy)hydroxide (denoted as FeOxHy due to potential-dependent oxidation and protonation state) is one of the most intrinsically-active earth-abundant single-metal materials for catalyzing the OER in base, however, its activity is highly dependent on conductivity, electrode substrate, and structure.5–11 Additionally, when Fe is embedded into NiOOH or CoOOH, the OER activity is even higher.6,12,13 These effects are poorly understood and a better understanding would allow for more rational design within the field of OER electrocatalysis. Understanding electrode substrate interactions with the OER catalyst is important for determining the intrinsic activity of a catalyst.14 We have shown that the OER activity of FeOxHy at low overpotentials (350 mV) is nearly two orders of magnitude higher on Au substrates compared to Pt substrates.5 Computational efforts by Bell et al. have shown that an Fe atom atop an Au2O3 matrix is more active than pure Au2O3,15 53 which is not surprising given the intrinsic activity of FeOxHy.6 There has been no investigation in the literature so far of the unique relationship between FeOxHy and Au/Au2O3 compared to other noble metals and their oxides. The enhancement of the OER activity of other transition metal (oxy)hydroxides (Mn, Co, Ni, NiFe) by Au and other noble metals (Pt, Ag, Cu, Pd) has been investigated previously.16–25 For example, Bell et al. showed enhanced OER activity of Au-supported Ni and Co oxides, which they attributed to charge transfer from transition metal to the highly electronegativity Au substrate.26,27 Additionally, they found that the activity of CoOx correlated with the electronegativity of the metal substrate in the order Au > Pt > Pd > Cu > Co. Jaramillo and coworkers have shown that MnOx embedded with noble metal nanoparticles (Au, Pt, Pd, Ag) is more active than MnOx on glassy carbon due to an increase in the amount of MnIV, observed by X-ray absorption spectroscopy (XAS).28,29 In this work, the activity of MnOx with noble metal nanoparticles varied in the order Au ≈ Pd > Pt > Ag, in contrast with the electronegativity trend found by Bell. Investigation of the effects of metal substrate on the OER activity of FeOxHy is necessary for several reasons. Firstly, it is possible that studies of other noble-metal enhancement of transition metal oxides are affected by Fe impurities, which have been shown to readily incorporate into Ni, Co, and Mn oxides/oxyhydroxides.6,12,13,30 Secondly, the two orders of magnitude difference in OER activity between Au and Pt- supported FeOxHy is much higher than other transition metals, indicating a different enhancement mechanism may be in effect. Finally, since noble metals are often used as substrates for heterogeneous electrocatalysis, it is important to understand the effects of Fe impurities on these substrates in order to make accurate measurements of intrinsic activities of catalytic materials. Here we use cyclic voltammetry (CV) and chronopotentiometry to assess the OER activity of FeOxHy on different substrates (Au, Pt, Pd, C, and Cu) in 1 M or 0.1 M KOH. CV was also used to examine the effects of electrolyte impurities on the activity of the substrates themselves. We show that the enhanced OER activity of FeOxHy on Au electrode substrates is unique, likely due to a greater adsorption strength that also leads to better charge transport at the Au-FeOxHy interface. Additionally, we show that the 54 presence of transition metal impurities in KOH electrolyte greatly increase the OER current of Au electrode substrates, making it a questionable substrate choice for fundamental studies of thin film OER catalysts. Experimental Electrode Fabrication. The electrodes had surface areas of ~ 0.1-1.0 cm2 and were constructed by depositing 25 nm of Ti adhesion layer and 50 nm of Au, Cu, Pd, or Pt onto glass microscope slides. Au and Pd were deposited via thermal evaporation and Ti, Pt, and Cu by electron beam evaporation (Amod, Angstrom Engineering). The films were connected to Sn-Cu wire via Ag paint (Ted Pella). Epoxy (Loctite Hysol E-60HP or 9640) covered any exposed wire and Ag paint. The entire perimeter and back of slide was covered with hot glue to prevent leaching of Fe from glass or epoxy components.31 The C electrode was purchased from CH Instruments (glassy carbon, 3 mm diameter). FeOxHy films were deposited onto the metal electrodes from aqueous solutions containing 0.1 M FeCl2 (Alfa Aesar, 98%) and 0.05 M NaNO3 (Mallinckrodt) for 20 seconds at -0.2 mA cm-2 according to previous procedures.5 Electrochemical Characterization. All electrodes and films were analyzed by CV at 20 mV s-1 and steady-state chronopotentiometry measurements (Tafel analysis) using a BioLogic potentiostat (SP300 or SP200) in a polymethylpentene 3-electrode cell with a Pt-wire counter electrode and Hg/HgO reference electrode (CH Instruments). Tafel analysis was completed from low to high current densities (first two steps held for 10 minutes and all consecutive steps were held for 3 minutes). Two CVs and an impendence measurement were taken before the first step and after the last step. Throughout electrochemical characterization, O2 was bubbled and a magnetic stir bar was used to prevent bubbles from accumulating on the electrode surface. Electrochemical characterization was carried out in four types of 1 M KOH except for Cu electrodes which were analyzed in 0.1 M KOH due to lower Cu solubility at this electrolyte concentration. ACS Grade KOH was purchased from Alfa Aesar. Semiconductor-grade (SC) KOH electrolyte was purchased from Sigma Aldrich, (semiconductor grade 99.999% trace metal basis). Fe-free electrolyte was SCG KOH cleaned of Fe impurities using Ni(OH)2 following our previously reported method.13 Fe-spiked electrolyte was 55 achieved by adding 1 ppm of Fe(NO3)3 (Ricca, ICP-MS standard, 1000 ppm Fe) to SC grade KOH during the first CV cycle for each film. All data were corrected after collection for uncompensated series resistance based on impedance measurements taken before and after cycling. Uncompensated series resistance ranged from 3 – 15 Ω in 1 M KOH and 25 – 45 Ω in 0.1 M KOH. All potentials are given in overpotential, referenced to the thermodynamic potential for the OER, 𝐸𝐸𝑂𝑂𝑂𝑂−/𝑂𝑂2 , taken as 1.23 V anodic of the hydrogen evolution reaction potential at a Pt electrode in H2-saturated 1 M or 0.1 M KOH. X-Ray Photoelectron Spectroscopy (XPS). XP spectra were collected on a Thermo EscaLab 250 using a monochramtic Al Kα X-ray with a 20 eV pass energy and 0.1 eV step size. Data on Cu samples were collected with a non-monochromatized Mg Kα source with a 40 eV pass energy and 0.2 eV, due to Cu Auger interference at the Fe 2p binding energy region. An electron flood gun was used during all measurements for charge neutralization. The binding energy scale for each sample was calibrated by setting the C-H peak in the C 1s spectra to 284.8 eV. Results and Discussion Electrolyte Impurities. It is known that electrolyte impurities can affect the catalytic ability of OER catalysts,12,13,30 but relatively little attention has been given to impurity effects on the underlying electrode substrate.15,32 We tested commonly used electrode substrates (Au, Pt, C, Pd, and Cu) in KOH electrolyte of different purity. We find that some substrates are more affected by electrolyte impurities than others, shown in Figure 4.1, where electrolyte impurities increase in the order Fe-free < SC < ACS. Cu appears to be the least affected by solution impurities, as it has the same OER current at 450 mV overpotential in all three electrolytes, Table 4.1, although this is likely due to the lower electrolyte concentration, 0.1 M rather than 1 M. Pd substrates are also little affected by solution impurities; the OER current at 450 mV overpotential for a Pd electrode is the same in both Fe-free and ACS grade and double in SC grade. The OER current of a Pt electrode roughly doubles each time as electrolyte purity decreases. The insets in each of the CVs in Figure 4.1 show that the capacitive regions on each of the substrates vary little 56 between the different electrolytes. Additionally, differences in electrochemically active surface areas are unlikely to contribute to the activity trends as they are similar (Figure C.1). The C electrode is least affected by solution impurities and has the lowest OER current at 450 mV overpotential in all three types of electrolyte tested here. However, the lack of activity increase in metal ion-contaminated electrolyte likely implies weak adhesion/adsorption of metal oxide (MOx) species to the glassy carbon surface. This poor adhesion has been demonstrated previously to affect charge transfer in catalysts intentionally deposited onto glassy carbon as an electrode substrate.16,33,34 Au is the most affected by solution impurities, with its OER current increasing by roughly an order of magnitude as the electrolyte purity decreases. The onset of OER is as low as 325 mV in ACS grade and 400 mV in SC grade 1 M KOH. The order of magnitude increase in OER current density from Fe-free to SC grade suggests a strong interaction between Au and Fe, since only Fe impurities are removed in the cleaning process. The interaction between Au and Ni impurities, however, is surprisingly weak; OER current in Fe-free KOH is lower than all substrates, despite the presence of residual Ni impurities from the cleaning procedure.13,31,32 The large increases in OER current density for an Au electrode as metal ion impurities increase indicates a greater MOx adsorption strength and better charge transfer between Au and adsorbed MOx species. 57 Figure 4.1. Cyclic voltammetry at 20 mV s-1 of electrode substrates in A) ACS grade, B) semiconductor grade, and C) Fe-free KOH. Cycle 5 of 50 is shown for each material which is representative of at least three replicates. All CVs were run in 1 M KOH except Cu in 0.1 M KOH. 58 Table 4.1. OER current density at 0.45 V vs 𝐸𝐸𝑂𝑂𝑂𝑂−/𝑂𝑂2 for electrodes in KOH electrolyte of different purity. Current Density at η = 450 mV Type of 1 M KOH Electrode Substrate Fe-free Semiconductor grade (99.99%) ACS Grade 1 ppm Fe SCG FeOxHy film Au 0.040 (0.006) 0.20 (0.07) 1.8 (0.6) 33 (4) 8 (2) C 0.032 (0.009) 0.042 (0.005) 0.11 (0.02) 0.5 (0.4) 0.24 (0.08) Cu* 0.086 (0.002) 0.08 (0.02) 0.08 (0.03) 4 (1) 0.6 (0.2) Pd 0.34 (0.02) 0.72 (0.07) 0.3 (0.1) 2.9 (0.3) 2.0 (0.8) Pt 0.09 (0.01) 0.20 (0.09) 0.41 (0.02) 0.8 (0.1) 4.5 (2) Substrate Dependence of FeOxHy OER Activity. Due to the chemical instability of FeOxHy at high pH and anodic potentials, determining the intrinsic OER activity on different substrates is inherently difficult.5 Here, we test the substrate dependence of the OER activity of FeOxHy in two different ways: 1) by adding 1 ppm of Fe(NO3)3 to the electrolyte solution to effectively deposit a monolayer of FeOxHy on the electrode surface and 2) by intentionally depositing a thin film (~ 1 μg cm-2) of FeOxHy on the electrode surface. In general, we find the same trends for both Fe-deposition methods indicating we are measuring the intrinsic activity of FeOxHy on each substrate. In both Fe-deposition methods we find that the OER activity of FeOxHy on a given substrate increases in the order C < Cu ≈ Pt ≈ Pd < Au, shown in Figure 4.2. The OER current density at 450 mV overpotential of each deposition method is greater than the bare electrodes in ACS grade KOH, indicating that we are indeed measuring FeOxHy on the substrates and not the bare substrates themselves. FeOxHy on Au has a low onset potential for the OER, ~ 300 mV by ~ 100 mV. Figure 4.2 shows the OER activity of FeOxHy on glassy carbon substrates in lower than all the other substrates tested here. The overpotential at 1 mA cm-2 of FeOxHy on C is 100 mV higher than on Pt, Pd, or Cu 59 and more than 200 mV higher than on Au. This discrepancy has likely made fundamental studies of FeOxHy challenging and is further compounded in high-loading studies where low conductivity of FeOOH limits charge transport.5 Figure 4.2. Cyclic voltammetry at 20 mV s-1 for FeOxHy on different substrates deposited from (top) 1 ppm Fe(NO3)3 in KOH and (bottom) 0.1 M FeCl2 in 0.5 M NaNO3. Cycle 5 of 50 is shown for each and is representative of at least three replicates. All CVs were run in 1 M SC grade KOH except Cu in 0.1 M SC grade KOH. The presence of Fe on the surface for the measurement is also further confirmed by contrasting the Tafel slopes of FeOxHy from both deposition methods to that of the bare electrodes in Fe-free KOH, Figure 4.3. The Tafel slopes, shown in Figure 4.3, are similar for the two deposition methods, although there is some variability. The Tafel slopes found in 1 M KOH with 1 ppm Fe are likely to be closer to the true value, as there was a constant amount of Fe present in solution and the error bars at each point are smaller in Figure 4.3. The higher variability in Tafel measurements for the cathodically- 60 deposited FeOxHy may indicate that these films do not adhere as well to the substrate as FeOxHy adsorbed from solution. The Tafel slopes of FeOxHy on Cu and Au substrates are similar, ~ 45 mV dec-1 (regardless of Fe-deposition method) indicating the OER mechanism and rate- determining step may be the same. FeOxHy on Pt and Pd also have similar Tafel slopes, ~ 60 mV dec-1 indicating similar OER mechanisms on both these substrates. Doyle and Lyons have also found Tafel slopes of 45 and 60 mV dec-1 for FeOxHy grown on metallic Fe electrodes in NaOH under different solution and cycling conditions.11 This suggests that Tafel slopes found here for FeOxHy on different substrates is due to differences in the FeOxHy atomic structure on those substrates. The Tafel analysis on C substrates has high variability in both FeOxHy deposition methods, but low variability (smaller error bars) in the Fe-free Tafel measurements, likely due to poor adhesion of MOx species to the glassy carbon surface. The accurate determination of the Tafel slope for FeOxHy on glassy carbon is therefore difficult and the results here should be taken with caution. 61 Figure 4.3. Tafel analysis of different electrode substrates in (top) Fe-free KOH, (middle) KOH with 1 ppm Fe, and (bottom) in semiconductor grade KOH with FeOxHy films deposited on the electrode surface. Au Enhancement of FeOxHy Activity. It is clear from the data presented here that the enhancement of the OER activity of FeOxHy from Au is both large and unique. This is consistent with other reports of Au enhancing the OER activity of NiOx,13,21,24,26 MnOx,28,29,35 and CoOx17,20,24,25,27,36 catalysts, indicating a strong interaction between Au and first-row transition metal oxides/(oxy)hydroxides. The cause of the enhancement from Au is suggested to be the increased oxidation state of the transition metal (Mn, Fe, 62 Co, Ni) due to the electron-withdrawing ability of Au. Indeed, Au has a high electronegativity, but comparison with other highly-electronegative elements shows no correlation between substrate electronegativity and OER current of FeOxHy catalyst (Figure C.2). This suggests there is an additional factor contributing to the OER activity enhancement from Au. The electron-withdrawing ability of the substrate is dependent on the electronegativity of the substrate, but also the ability to transfer charge at the interface between substrate and catalyst. Studies comparing NiOxHy and Ni(Fe)OxHy on GC and Au substrates have shown that Ni2+/3+ oxidation is more complete on Au than GC substrates,16,33 indicating better charge transfer at the interface. The improved charge transfer at the Au-transition metal oxide interface is likely due to an increased strength of adsorption or adhesion of the oxide to the Au surface. This increased adhesion strength is confirmed by XPS analysis after performing electrochemistry in KOH with 1 ppm Fe, shown in Figure 4.4. Despite that fact that electrochemical analysis indicates OER activity from FeOx species, after removing the electrodes from solution and rinsing with H2O, Fe is only present on Au and Cu substrates, not Pt or Pd (Figure 4.4). This is evidence of strong adhesion of FeOxHy to the Au surface, and the Cu surface as well. The lower current density from Cu substrates despite strong FeOxHy adhesion can be attributed to the lower efficiency of the OER in 0.1 M KOH compared to 1 M KOH. Figure 4.4. Fe XP spectra on A) Au, B) Pd, C) Cu, and D) Pt electrodes after cycling in KOH spike with 1 ppm Fe. Au, Pt, and Pd electrode spectra were collected with an Al 63 Kα source and Cu with Mg Kα due to Cu Auger peak at Fe 2p binding energy. The Fe 2s region is shown for Pt electrode due to overlap with Fe 2p and Pt 3s. Interestingly, we also find that of the metallic substrates, the Au surface is the only one to undergo completely reversible oxidation in solution. Redox features can be seen for all substrates except for C (Figures 4.1, 4.2, and C.3), and XPS analysis after electrochemistry shows evidence of substrate oxidation for Cu, Pd, and Pt, Figure 4.5. The Au surface, while it clearly undergoes oxidation and reduction, evident from the redox features in Figures 4.1 and 4.2, returns to a completely metallic state with no evidence of residual oxide. Figure 4.5. Metal electrode XP spectra for A) Au, B) Pd, C) Cu, and D) Pt substrates. Au, Pd, and Pt were collected with Al Kα and Cu with Mg Kα sources. Conclusion and Bridge The OER activity of FeOxHy has been shown here to vary based on electrode substrate material in the order Au > Pd ≈ Cu ≈ Pt > C. The differences in activity are likely due to differences in the strength of adhesion of FeOxHy to the electrode substrate. The variation in adsorption strength lead to differences in charge transport at the FeOxHy- electrode interface, contributing to the observed activity trend. The electrode substrate material also likely affects the atomic structure of FeOxHy and is responsible for differences in activity. Further investigation into differences in the atomic and electronic structure of FeOxHy on different electrode substrates in operando would necessary to confirm this hypothesis. 64 The sensitivity of the OER activity of FeOxHy on different substrates and in different matrices (eg. NiOOH explored in Chapter III) suggests it is important for the active site in OER catalysis. To further investigate the role of Fe in the active site for OER catalysis, in operando experiments on CoFeOxHy are performed and analyzed in the next chapter, Chapter V. 65 CHAPTER V OPERANDO X-RAY ABSORPTION SPECTROSCOPY SHOWS FE OXIDATION IS CONCURRENT WITH OXYGEN EVOLUTION IN COBALT- IRON (OXY)HYDROXIDE ELECTROCATALYSTS This chapter contains co-authored unpublished material. M. Stevens, Prof. Boettcher, and I conceived of the project. I performed experiments along with M. Stevens and M. Nellist with technical assistance from S. Fakra. M C. Toroker and M Dahan performed calculations. I analyzed data and wrote most of the paper with computational sections written by M. Toroker. Editorial assistance was provided by M. Stevens and S. Boettcher. It is prepared for submission to the peer-reviewed journal Angewandte Chemie International Edition. Introduction The efficiency of the oxygen evolution reaction (OER), particularly in alkaline solution, limits several prospective renewable energy-conversion technologies.[1–3] In order to improve the efficiency, a better understanding of catalysts for this reaction is necessary. In alkaline conditions, Fe-incorporated Ni- and Co-based (oxy)hydroxides (Ni(Fe)OxHy, Co(Fe)OxHy) are some of the best catalysts for the OER, but still operate at higher overpotentials than their HER catalyst counterparts.[4,5] Understanding the mechanism for activity enhancement upon Fe incorporation would lead to the rational design of better catalysts. Operando XAS has been used previously to probe atomic and electronic structure of catalysts during the OER.[6–12] Friebel et al.[6] determined that in Ni(Fe)OxHy the Fe – O bond length during the OER was similar to that of NiOOH, which is shorter than in FeOOH. Using this structural information, they performed DFT calculations to show that catalysis on Fe – O sites has a lower theoretical overpotential than on Ni – O sites, indicating that Fe was likely the active site. Despite the Fe – O bond shortening, they claim no oxidation of Fe to Fe4+. Conversely, Chen et al.[13] find evidence of Fe4+ in 66 Ni(Fe)OxHy during the OER using Mössbauer spectroscopy. Hunter et al.[14] have shown evidence of Fe6+ during OER catalysis in non-aqueous electrolytes. Co(Fe)OxHy is structurally and electrochemically similar to Ni(Fe)OxHy.[15,16] Both are comprised of layers of edge-sharing MO(H)6 octahedra and show increased OER activity and a more-positive (nominal) M2+/3+ reduction potential upon Fe- incorporation. Smith et al.[9] examined amorphous photochemically-deposited Fe100- yCoyOx films using operando electrochemical XAS and observed oxidation of Co, but not Fe during the OER. Gong et al.[11] found that CoOx with Fe3+ ions adsorbed from solution was more active than co-deposited Fe4.4Co95.6Ox. Operando XAS suggests oxidation of Co during the OER to a similar extent for both samples. Fe oxidation was not found in either sample, but adsorbed Fe3+ exhibited a lower oxygen coordination and was concluded to be responsible for the sample’s higher activity. Under-coordinated Fe-sites or edge-sites have also been hypothesized to be the active site in Ni(Fe)OxHy[13,17–19] and CoOOH.[20,21] Here we use operando XAS to study Co(Fe)OxHy. We find clear evidence for Fe oxidation and Fe – O bond shortening at OER relevant potentials, consistent with the hypothesis that Fe is the active site[22] Experimental Electrochemical Cell Fabrication. Si3N4 windows (2 x 2 mm, 1 μm thick) embedded in Si (5 x 5 mm, 200 μm thick) were purchased from Norcada. The windows were then treated with an O2/N2 plasma for 10 min to promote adhesion. Ti (30 nm) and Au (25 nm) were then deposited sequentially on the windows via evaporation with an electron- beam and thermal source, respectively (Amod, Angstrom Engineering). Electrical connection to the Au surface was made using Sn-Cu wire (McMaster-Carr) and Ag paint (Ted Pella). The windowed-electrode was then embedded in a 60-mL polypropylene using hot glue. All metal components except the Au surface were covered in hot glue to prevent contact with solution. A new cell was fabricated for each sample tested. Catalyst Film Deposition. Films were deposited from 0.1 M transition metal aqueous solutions (18.2 M∙cm H2O). Co(NO3)2 (Alfa Aesar, 98%) and FeCl2 (Alfa Aesar, 98%) 67 were used as the Co and Fe sources respectively. All Co and Fe compositions in the text are referenced to the deposition solution ratios as characterization of the deposited films was not possible. Previous deposition under similar conditions shows that the Fe content in lower in the films than in the deposition solution by ~ 5%.[15] N2 was bubbled through the solution for 20 min before addition of FeCl2. Using a carbon cloth counter electrode, -1 mA cm-2 was applied to the working electrode for 30 s to precipitate a Co or Co-Fe hydroxide film on the electrode surface. The use of electro-chemical precipitation leads to highly porous films where nearly every Co/Fe cation is exposed to the electrolyte[15]; these films are easily completely transformed into the active oxyhydroxide phase under OER conditions making them superior samples for study compared to oxide phases where only the surface is electrochemically active.[4,23–25] The deposition solution was immediately emptied from the cell to prevent dissolution of the deposited film and rinsed with water. Films were cycled in 1 M KOH (10 mV s-1) to confirm film deposition from -0.3 to 0.4 V vs 𝐸𝐸𝑂𝑂𝑂𝑂−/𝑂𝑂2. Electrochemical cells with deposited films were then shipped to the Advanced Light Source (ALS) for combined electrochemical-XAS experiments. Electrochemical Characterization. Electrochemical analysis was performed using a Biologic SP-200 potentiostat in 1 M KOH (Sigma Aldrich, 99.99% trace metals) with an Hg/HgO reference electrode and Pt counter electrode. Co-cleaned KOH was used for CoOxHy films.[15] Chronoamperometry measurements were performed at potentials negative of the Co2+/3+ reduction wave, positive of the oxidation wave, and during OER, for the duration of the XAS measurements. Cyclic voltammetry at 10 mV s-1 was performed before and after XAS measurements on each sample. All potentials in Appendix D and in Chapter V are reference in overpotential, vs 𝐸𝐸𝑂𝑂𝑂𝑂−/𝑂𝑂2 . XAS Measurements. XAS was performed at beamline 10.3.2 (bending magnet) at the ALS with a ring current of 500 mA. A Si(111) monochromator was used to monochromatize the incident beam, I0. The energy axis was aligned using a glitch in I0 near the main edge which was calibrated on Co (7708.78 eV edge) and Fe (7110.75 eV edge) foils measured in transmission mode using an ionization chamber. The CoFeOxHy samples were oriented at a 45° angle from the incident beam and fluorescence signal from the films was detected using a 7-element UltraLEGe detector (Canberra) at a 90° 68 angle from the incident beam. XAS data was collected under ambient conditions in the electrochemical cell while a constant potential was applied to the working electrode. XAS Data Analysis. All absorption spectra, μ(E), were deglitched and corrected for detector dead time using the LABVIEW software at beamline 10.3.2. Further data reduction and fourier transformed extended X-ray absorption fine structure (FT-EXAFS) fitting was performed using IFEFFIT-based Athena and Artemis software programs.[26] For each measurement, between 5 and 10 scans were merged to obtain the final spectrum. All spectra were edge step normalized using pre and post-edge backgrounds. The edge energy, E0, was set to 1.0 and determined by the energy of the highest peak in the first derivative of μ(E). FT-EXAFS spectra were fit using Artemis software. Theoretical scattering paths were calculated using FEFF6[27,28] for CoOOH[29] (ICSD #56288). Fits were performed over an R space range of 0.9 – 2.0 Å in the Fourier transform using k ranges of 2.0 to 11.0 Å-1 (Fe) or 2.0 – 11.5 Å-1 (Co) with a Kaiser-Bessel window. Results and Discussion Cyclic voltammetry of the CoFeOxHy films in Figure 5.1 shows that the addition of Fe improves the OER activity of CoOxHy and shifts the nominal Co2+/3+ reduction potential anodic, consistent with previous work.[15] In this way, the Co(Fe)OxHy system is similar to that of Ni(Fe)OxHy. An important difference, however, is that the nominal 2+/3+ reduction potential is well-separated from the onset of OER current in the Co(Fe) system while in the Ni(Fe) system the OER onset current occurs at a potential near the 2+/3+ oxidation wave for high Fe content. Thus for the Co(Fe) system we can compare XAS spectra for Co and Fe in a potential range prior to the OER, but where the sample is already in the oxidized “active” oxyhydroxide state, with spectra where the catalyst remains in the active oxyhydroxide state but is now at potentials sufficient to drive the OER (see dotted lines in Figure 5.1). This comparison is complicated in the Ni(Fe) system where the oxidation of Ni is often coincident with the onset of OER current.[30] 69 Figure 5.1. A) Cyclic voltammetry data on Co(Fe)OxHy films collected at 10 mV s-1 immediately prior to XAS data collection. Vertical dashed lines represent potentials at which chronoamperometry was performed while collecting operando XAS measurements. B) Schematic of electrochemical cell used for operando electrochemical- XAS data collection. Incident X-rays and outgoing fluorescence from the catalyst went travelled through the 1 μm Si3N4 window and the Ti (30 nm)/Au (25 nm) layers. For Fe-free CoOxHy, the Co K-edge energy at an overpotential of 0.1 V vs. 𝐸𝐸𝑂𝑂𝑂𝑂−/𝑂𝑂2 (anodic of the Co oxidation wave but cathodic of the OER current onset) indicates that Co is mostly present as Co3+ (Figure D.1A). This is consistent with conversion from the hydroxide to oxyhydroxide phase during the Co2+/3+ oxidation wave. Comparison of the X-ray absorption near edge spectroscopy (XANES) data collected at a potential negative of the reduction wave (-0.3 V) to Co(OH)2 and CoOOH show that only a small fraction of the Co can be reduced back to Co2+ and that the oxyhydroxide structure dominates rather than the hydroxide (Figure D.1C), consistent with previous reports.[15] Based on the higher K-edge energy shown in Figure 5.2A, a small amount of oxidation occurs (0.5 eV increase in edge energy) when the catalyst potential is stepped from an overpotential of 0.1 V to 0.4 V where OER occurs. There is no change in the pre-edge feature. This oxidation likely represents a fraction of Co being oxidized from Co3+ to Co4+ (Figure D.1A), consistent with other reports that show presence of Co4+ during the OER.[23,31,32] It is possible that these oxidized Co sites represent those performing the OER, perhaps at reactive edge sites.[17,18,21] Upon Fe addition, Co oxidation during the OER is suppressed, evidenced by the lack of change in the Co K-edge energy when stepping the potential to OER conditions (Figure 5.2B and 5.2C). There is thus little to no Co4+ present in mixed Co(Fe)OxHy films during the OER. This is further confirmed by the slightly higher (0.3 eV) Co K-edge energy of 70 CoOxHy during the OER compared to Co(Fe)OxHy (Figure D.1B). However, the pre- edge feature shown in the inset of Figures 5.2B and 5.2C does increase intensity during OER, suggesting a distortion of the CoO6 octahedra that increases the transition probability. This distortion is likely caused by changes to the Fe – O environment during the OER, which will be discussed below. Figure 5.2. Operando Co K-edge XANES data collected under applied electrochemical potential for A) CoOxHy, B) Co0.8Fe0.2OxHy, and C) Co0.6Fe0.4OxHy films. Pre-OER spectra were collected at potentials higher than the nominal Co2+/3+ oxidation wave and prior to OER onset. OER spectra were collected at potentials necessary to maintain 3 – 4 mA cm-2 of OER current density. Fe K-edge spectra on Co(Fe)OxHy films collected at an overpotential of 0.15 V (anodic of the nominal Co2+/3+ oxidation wave, but prior to the onset of OER) show that Fe is nominally Fe3+ in the films. This edge position is higher than at -0.3 V vs. 𝐸𝐸𝑂𝑂𝑂𝑂−/𝑂𝑂2 71 (cathodic of the Co2+/3+ reduction peak) the indicating that Fe is also oxidized in the Co2/3+ redox wave (Figure D.2A-B). In the OER region, Fe3+ is further oxidized (Figures 5.3A and 5.3B); the Fe K-edge energy shifts positive by 0.7 eV relative to at 0.15 V vs. 𝜀𝜀O2/OH− (prior to the OER onset) and is also similarly higher in energy than the α-Fe2O3 and γ-FeOOH standards (Figure D.2C). This increase in Fe edge energy upon driving OER represents oxidation of only a fraction of the Fe3+ sites. XAS spectra from La1- xSrxFeO3-δ show that a change in Fe oxidation state from Fe3+ to Fe4+ results in an increase in K-edge energy of ~ 1.3 eV. The increase in Fe K-edge energy could also be the result of oxidation of Fe sites to Fe6+, which has recently been suggested to be the active site in Ni(Fe)OxHy,[14] however Fe6+ sites would be highly reactive and are unlikely to be present in substantially high concentrations. Balasubramanian et al.[33] have reported Fe oxidation in anodically polarized NiOOH similar to that found here. The smaller than expected Fe K-edge shift was attributed to the higher degree of covalency in the Fe(IV) – O bond resulting in charge transfer back to Fe.[34] The fact that incomplete oxidation of Fe3+ is observed may be because only a fraction of Fe (perhaps at reactive edge/defect sites) are participating in catalysis.[11,17,18] Figure 5.3. Operando Fe K-edge XANES data collected under applied electrochemical potential for A) CoOxHy, B) Co0.8Fe0.2OxHy, and C) Co0.6Fe0.4OxHy films. Post- oxidation spectra were collected at potentials higher than the nominal Co2+/3+ oxidation wave and prior to OER onset. OER spectra were collected at potentials necessary to maintain 3 – 4 mA cm-2 of OER current density. 72 Figure 5.4. FT-EXAFS data for Co(Fe)OxHy films at pre-OER and OER potentials in 1 M KOH. D) Metal – oxygen bond lengths for Co(Fe)OxHy films before (pre-OER) and during the OER determined from FT-EXAFS fitting analysis. We fit the Fourier transformed extended X-ray absorption fine structure (FT- EXAFS) (Figure 5.4) data to evaluate the metal-oxygen bond lengths (Figure 5.4D). We find no change in Co – O bond length when stepping the potential from 0.1 – 0.15 V vs 𝐸𝐸O2/OH− to the OER regime at 0.3 – 0.4 V vs 𝐸𝐸O2/OH−. At both potentials the Co – O bond length is 1.90 Å, with or without Fe present. This is as expected for Co(Fe)OxHy films since there is no change in the Co oxidation state. For Fe-free CoOxHy, there is no apparent change in the Co – O bond length change despite the oxidation of Co observed by the positive shift in the Co K-edge. This is likely because the fraction of Co changing oxidation state is small. Conversely, the Fe – O bond length changes during the OER, shortening from 1.93 Å at 0.15 V vs 𝜀𝜀O2/OH− (prior to OER but more anodic than the Co oxidation wave) to 1.90 Å during the OER. This average Fe – O bond length is shorter than in Fe phases such as α-Fe2O3 (2.03 Å)[35] and γ-FeOOH (2.04 Å)[36], indicating that Fe is dispersed in the CoOOH matrix. While all Fe – O bonds are shortened during the OER, only a fraction of the Fe sites undergo oxidation, likely the reactive sites.[18] 73 The observed bond distance and oxidation state changes at the proposed active Fe sites were also studied with first principles calculations. We used Density Functional Theory +U (DFT+U) implemented in the VASP simulation package[37–39] to calculate the free energies required for the OER mechanism[40] shown in Figure 5.5 (see Appendix D for details). For Fe-doped CoOOH, the Fe-O average bond when the surface is adsorbed with water (*OH2 intermediate) is ~1.95 Å and reduces to ~1.90 – 1.94 Å when OER proceeds through *OH and *O intermediates. This correlates to an increase in the calculated Fe oxidation state from +3 to nominally +4/+5. In contrast, the Co-O bond lengths in Fe-doped CoOOH both near and far from the active site do not change for the different OER intermediates maintaining an average value of ~1.91 Å. In the case of pure CoOOH, the Co-O bond lengths vary between 1.9 and 1.92 Å for the different intermediates as the Co oxidation state changes from +3 for the *OH2 intermediate to nominally +4 for the *OH and *O intermediates. While there are certainly substantial differences between the simple structures and reaction mechanisms used for the DFT+U calculations and experiment, the central result that the Fe bond length and oxidation change for the different OER intermediates appears consistent. Figure 5.5. Schematic OER mechanism for Fe-doped CoOOH. The values for the free energies of reaction intermediates are indicated. Red, blue, brown, and white spheres represent O, Co, Fe, and H atoms. Created with VESTA visualization software.[41] 74 The experimental findings of this work also agree well with theory. The higher activity of Co(Fe)OxHy compared to CoOxHy films has been demonstrated by considering the mechanisms proposed by Selloni and Norskov. In both cases, lower overpotential was demonstrated on Fe-doped CoOOH compared to pure CoOOH. Additionally, calculations also predicted a shortening of the Fe – O bond length during the OER, consistent with our experimental observations. Our results are also consistent with the experimental and theoretical study by Friebel et al. on NiFeOxHy, which is likely relevant due to the structural and electronic similarities to CoFeOxHy. Their DFT analysis showed lower overpotential on Fe-sites embedded in a NiOOH matrix, where the Fe – O bond lengths were the same as those for Ni – O.[6] Most other studies have not reported an increase in Fe oxidation state along with potential-dependent shortening of the Fe – O bond in Co(Fe)OxHy, or Ni(Fe)OxHy, OER catalysts in alkaline solution. This could in part be due to differences in catalyst preparation. The Fe sites in catalyst films here are well-integrated into the CoOOH structure, evidenced by the similarity of the M – O and M – M shell distances in the FT- EXAFS (Figure 5.4). Since CoOOH is electronically conductive at all potentials, unlike FeOOH, the Fe sites in Co(Fe)OxHy could be expected to undergo potential-dependent structural changes. The samples used for previous studies on Co(Fe)OxHy may include oxide phases or separated Fe-phases that reduce the amount of electrochemically accessible Co(Fe)OxHy, making the changes seen here difficult to detect.[9,11,24,25] Interestingly, our potential-dependent Fe – O bond lengths are similar to those of Fe3+ ions adsorbed on CoOxHy, reported by Gong et al., which showed a decrease of the Fe – O bond length from 2.01 to 1.94 Å when changing the potential from open-circuit to OER conditions.[11] However in that case no concurrent oxidation of Fe was observed and bond length changes due to Co or Fe oxidation during the Co2+/3+ oxidation wave rather than OER were not separated. Calvillo et al. has reported surface Fe oxidation on CoxFe3-xO4 spinels at OER potentials after extended cycling.[25] Balasubramanian et al. reported Fe oxidation and Fe – O bond shortening ex situ in anodically polarized Ni(Fe)OxHy films.[33] 75 Conclusion These data taken together support the hypothesis that Fe – O species are the active site in Co(Fe)OxHy films. The films contain Fe3+ at all potentials prior to the onset of OER. Upon driving Co(Fe)OxHy into the OER regime, Fe3+ is further oxidized while Co3+ observed at potentials just cathodic of the OER is not. In Fe-free CoOxHy Co3+ is oxidized during OER; indicating a fundamentally different mechanism compared to the Co(Fe)OxHy. Structurally, there are no differences in Co – O environment upon Fe addition, but large changes in activity, making it further unlikely that Co is the active site in the more-active CoFeOxHy. Finally, the small amount of oxidation observed for both Co in CoOxHy and Fe in Co(Fe)OxHy indicates that only a fraction of these sites undergo oxidation, suggesting that some sites are more reactive than others. 76 APPENDIX A CHAPTER II SUPPORTING INFORMATION Figure A.1. GIXRD patterns for MyAl1-yOx films where M is not d0, d5, or d10. Diffraction patterns were collected on the highest M-containing amorphous composition and on the Al-free MOx films to ensure crystallinity. 77 Figure A.2. GIXRD patterns from MyAl1-yOx films where M is d0, d5, or d10. The highest M-containing amorphous composition and lowest M-containing crystalline composition studied are shown. 78 Figure A.3. C 1s and O 1s XP spectra of a-Al2O3 film with fits. The binding energy was calibrated by setting the C-H peak to 284.8 eV. Figure A.4. XP spectra of VyAl1-yOx films C 1s (top), Al 2p, O 1s, and V 2p (bottom). The binding energy was calibrated by setting the C-H peak of each film to 284.8 eV. The V 3s peak (located at 66 eV for metallic V) can be seen on the low binding energy side of the Al 2p peak for films containing V.1 The binding energy of the V 2p3/2 peak is indicative of V5+ species.2 The broadness of the V 2p peaks for films containing Al likely indicate the lack of uniform V coordination environment. 79 Figure A.5. XP spectra of CryAl1-yOx films C 1s (top), Al 2p, O 1s, and Cr 2p (bottom). The binding energy was calibrated by setting the C-H peak of each film to 284.8 eV. The Cr 3s peak (located at 75 eV for Cr0) can be seen on the high binding energy side of the Al 2p peak for films containing Cr. The Cr 2p spectra for films containing Al indicate the presence of both Cr3+ and Cr6+ species.3 The Cr 2p spectrum for α-Cr2O3 agrees well with previous reports.4 Figure A.6. XP spectra of MnyAl1-yOx films C 1s (top), Al 2p, O 1s, and Mn 2p (bottom). The binding energy was calibrated by setting the C-H peak of each film to 80 284.8 eV. The presence of shake-up features in the Mn 2p spectra of films containing Al indicate the presence of Mn2+ in addition to Mn3+.5 Figure A.7. XP spectra of FeyAl1-yOx films C 1s (top), Al 2p, O 1s, and Fe 2p (bottom). The binding energy was calibrated by setting the C-H peak of each film to 284.8 eV. The difference in energy between the main 2p peaks and their shake-up features is indicative of Fe3+ rather than Fe2+.6 81 Figure A.8. XP spectra of CoyAl1-yOx films C 1s (top), Al 2p, O 1s, and Co 2p (bottom). The binding energy was calibrated by setting the C-H peak of each film to 284.8 eV. The position of the shake-up features in the Co 2p spectrum of the Co0.1Al0.9Ox film indicates Co is present as Co2+. Co 2p spectra for films with Co content greater than y = 0.1 are indicative of Co3O4.5 Figure A.9. XP spectra of NiyAl1-yOx films C 1s (top), Al 2p, O 1s, and Ni 2p (bottom). The binding energy was calibrated by setting the C-H peak of each film to 284.8 eV. Ni 3p peaks (located at 67 eV for Ni0) can be seen on the low binding energy side of the Al 82 2p peak for films containing both Ni and Al.1 The binding energy of the Ni 2p3/2 peak further confirms that Ni is present as Ni2+.7 Figure A.10. XP spectra of CuyAl1-yOx films C 1s (top), Al 2p, O 1s, and Cu 2p (bottom). The binding energy was calibrated by setting the C-H peak of each film to 284.8 eV. Cu 3p peaks (located at 77 and 75 eV for Cu0) can be seen on the high binding energy side of the Al 2p peak for films containing both Cu and Al. The presence of shake-up features in the Cu 2p XP spectra indicate the presence of Cu2+.2 83 Figure A.11. XP spectra of ZnyAl1-yOx films C 1s (top), Al 2p, O 1s, and Zn 2p (bottom) XP spectra of ZnyAl1-yOx films. The binding energy was calibrated by setting the C-H peak of each film to 284.8 eV. Figure A.12. XP spectra of GayAl1-yOx films C 1s (top), Al 2p, O 1s, and Ga 2p (bottom). The binding energy was calibrated by setting the C-H peak of each film to 284.8 eV. The binding energy of the Ga 2p3/2 peaks indicate that Ga is present as Ga3+.1 84 Figure A.13. VB-XPS of ZnyAl1-yOx films. Due to the crystalline nature of the ZnO film, the Ir substrate was not completely covered. Contributions from Ir to the valence band can be seen in the spectrum for ZnO on the low binding energy side. XANES Analysis Table A.1. Metal K-edge energy evaluated at the first peak in the first derivative of normalized μ(E) spectrum (after pre-edge feature). Sample E0 (eV) Sample E0 (eV) Sample E0 (eV) V foil 5465 Fe foil 7712 Cu foil 8979 V2O5 5481.1 Fe2O3 7123.4 CuO 8983.5 V0.7Al0.3Ox 5483.3 Fe0.7Al0.3Ox 7123 Cu0.7Al0.3Ox 8983.7 V0.5Al0.5Ox 5483.1 Fe0.3Al0.7Ox 7122.8 Cu0.5Al0.5Ox 8983.8 V0.3Al0.7Ox 5483.6 Fe0.2Al0.8Ox 7122.6 Cu0.3Al0.7Ox 8984.1 V0.2Al0.8Ox 5483.3 Fe0.1Al0.9Ox 7122.7 Cu0.2Al0.8Ox 8984.4 V0.1Al0.9Ox 5483.5 Co foil 7709 Cu0.1Al0.9Ox 8984.6 Cr foil 5989 Co3O4 7717.3 Zn foil 9659 Cr2O3 5999.6 Co0.3Al0.7Ox 7717.1 ZnO 9661.5 Cr0.5Al0.5Ox 6005.8 Co0.2Al0.8Ox 7717 Zn0.7Al0.3Ox 9662.2 Cr0.3Al0.7Ox 6006.3 Co0.1Al0.9Ox 7716.4 Zn0.4Al0.6Ox 9662.3 Cr0.2Al0.8Ox 6006.4 Ni foil 8333 Zn0.3Al0.7Ox 9662.3 Cr0.1Al0.9Ox 6006.4 NiO 8341 Zn0.2Al0.8Ox 9662.4 Mn foil 6539 Ni0.3Al0.7Ox 8341.1 Zn0.1Al0.9Ox 9662.4 Mn2O3 6548.1 Ni0.2Al0.8Ox 8341 Ta foil 9881 Mn0.7Al0.3Ox 6548.1 Ni0.1Al0.9Ox 8341.5 Ga2O3 10375.5 Mn0.4Al0.6Ox 6548.2 Ga0.5Al0.5Ox 10375.5 Mn0.3Al0.7Ox 6547.9 Ga0.3Al0.7Ox 10375.6 Mn0.2Al0.8Ox 6548 Ga0.2Al0.8Ox 10375.7 Mn0.1Al0.9Ox 6547.6 85 Figure A.14. k2-weighted χ(k) data MyAl1-yOx films. Colors correspond to amounts of M as follows: y = 0.1 (magenta), 0.2 (orange), 0.3 (dark yellow), 0.4 (green), 0.5 (blue), 0.7 (violet), 1.0 crystalline (dark grey), and 1.0 amorphous (light grey). 86 EXAFS Fitting All EXAFS data were fit using the Artemis software.8 The EXAFS equation can generally be described by the sum of all scattering paths, i, according to 𝜒𝜒(𝑘𝑘) = ∑ 𝑆𝑆02𝑁𝑁𝑖𝑖𝑓𝑓𝑖𝑖(𝑘𝑘) 𝑘𝑘𝑅𝑅𝑖𝑖 2𝑖𝑖 𝑒𝑒 − 2𝑅𝑅𝑖𝑖 𝜆𝜆(𝑘𝑘)𝑒𝑒−2𝑘𝑘2𝜎𝜎𝑖𝑖2 sin(2𝑘𝑘𝑘𝑘𝑖𝑖 + 𝛿𝛿𝑖𝑖(𝑘𝑘)) (Eq. A.1) where S02 – amplitude reduction factor (intrinsic to absorber) N – coordination number 𝑘𝑘 = 𝑘𝑘0 + ∆𝑘𝑘 , where R0 is the distance between scatterer and absorber of the theoretical standard and ΔR is the change in distance (i.e. difference between standard and sample) σ2 – mean-square disorder parameter f(k) – effective scattering factor δ(k) – effective scattering phase shift λ(k) – photoelectron mean free path 𝑘𝑘2 = 2𝑚𝑚𝑒𝑒(𝐸𝐸 − 𝐸𝐸0)/ħ, where E0 is difference in energy between standard and sample data. The values of f(k), δ(k), λ(k), and R0 were all calculated using theoretical standards (i.e. CIF files from ICSD) in the Artemis software/IFEFFIT.9 The values of N, ∆𝑘𝑘, E0, and σ2 were fit by the software. Uncertainties and r-factor values are directly from Artemis. Table A.2. EXAFS fitting results for VyAl1- yOx films. y in VyAl1-yOx V – O (Å) N (atoms) σ2 (Å2) r-factor 1 1.60 (0.01) 2.0 (0.1) 0.003 (0.001) 0.007 1 1.85 (0.02) 2.9 (0.3) 0.003 (0.001) 0.007 0.7 1.71 (0.01) 4.4 (0.5) 0.004 (0.002) 0.008 0.5 1.70 (0.01) 4.5 (0.5) 0.006 (0.001) 0.008 0.3 1.70 (0.01) 4.5 (0.4) 0.004 (0.001) 0.006 0.2 1.71 (0.01) 4.3 (0.4) 0.003 (0.001) 0.006 0.1 1.71 (0.01) 4.5 (0.5) 0.003 (0.001) 0.009 S02 value used for all fits was 0.73. Table A.3. EXAFS fit results for CryAl1- yOx films. y in CryAl1-yOx Cr – O (Å) N (atoms) σ2 (Å2) r-factor fraction Cr 1 1.99 (0.01) 6 (0.4) 0.002 (0.001) 0.006 1 0.5 1.60 (0.02) 4 0.008 (0.002) 0.026 0.7 (0.05) 0.5 1.88 (0.02) 6 0.003 (0.002) 0.026 0.2 (0.05) 0.3 1.61 (0.02) 4 0.008 (0.002) 0.02 0.75 (0.05) 0.3 1.88 (0.02) 6 0.001 (0.002) 0.02 0.25 (0.05) 0.2 1.62 (0.01) 4 0.006 (0.001) 0.0045 0.8 (0.05) 0.2 1.89 (0.02) 6 0.006 (0.003) 0.0045 0.2 (0.05) 0.1 1.62 (0.01) 4.0 (0.4) 0.008 (0.002) 0.01 1 S02 value used for all fits was 0.76. 87 Table A.4. EXAFS fitting results for MnyAl1- yOx films. y in MnyAl1-yOx Mn – O (Å) N (atoms) σ2 (Å2) r-factor 1 1.90 (0.01) 6.0 (0.8) 0.004 (0.002) 0.009 0.7 1.90 (0.01) 4.4 (0.4) 0.005 (0.002) 0.01 0.4 1.88 (0.01) 3.2 (0.5) 0.005 (0.002) 0.01 0.3 1.88 (0.01) 4.0 (0.3) 0.004 (0.001) 0.004 0.2 1.88 (0.01) 4.0 (0.3) 0.004 (0.001) 0.004 0.1 1.88 (0.01) 3.9 (0.4) 0.004 (0.001) 0.008 S02 value used for all fits was 0.7. Table A.5. EXAFS fitting results for FeyAl1- yOx films. y in FeyAl1-yOx Fe – O (Å) N (atoms) σ2 (Å2) r-factor 1 1.93 (0.01) 3 0.003 (0.001) 0.008 1 2.10 (0.01) 3 0.003 (0.001) 0.008 0.7 1.93 (0.01) 4.4 (0.4) 0.007 (0.001) 0.001 0.3 1.92 (0.01) 4.6 (0.4) 0.007 (0.002) 0.002 0.2 1.93 (0.02) 4.1 (0.5) 0.005 (0.002) 0.003 0.1 1.93 (0.01) 4.4 (0.4) 0.006 (0.002) 0.002 S02 value used for all fits was 0.75. Table A.6. EXAFS fitting results for CoyAl1- yOx films. y in CoyAl1-yOx Co – O (Å) N (atoms) σ2 (Å2) r-factor 1 1.91 (0.02) 5.4 (0.8) 0.003(0.002) 0.018 0.3 1.91 (0.02) 5.3 (1.0) 0.004 (0.002) 0.03 0.2 1.92 (0.01) 5.8 (0.6) 0.005 (0.002) 0.01 0.1 1.97 (0.01) 4.9 (0.2) 0.006 (0.001) 0.002 S02 value used for all fits was 0.75. Table A.7. EXAFS fitting results for NiyAl1- yOx films. y in NiyAl1-yOx Ni – O (Å) N (atoms) σ2 (Å2) r-factor 1 2.07 (0.01) 5.6 (0.3) 0.004 (0.002) 0.01 0.3 2.04 (0.02) 6 (0.6) 0.009 (0.002) 0.002 0.2 2.03 (0.01) 5.7 (0.5) 0.009 (0.002) 0.002 0.1 2.01 (0.01) 3.7 (0.4) 0.006 (0.002) 0.003 S02 value used for all fits was 0.85. Table A.8. EXAFS fitting results for CuyAl1- yOx films. y in CuyAl1-yOx Cu – O (Å) N (atoms) σ2 (Å2) r-factor 1 1.94 (0.01) 3.9 (0.4) 0.002 (0.001) 0.008 0.7 1.94 (0.01) 4.2 (0.2) 0.002 (0.001) 0.01 0.5 1.93 (0.01) 3.7 (0.4) 0.004 (0.002) 0.009 0.3 1.92 (0.01) 4.0 (0.4) 0.004 (0.002) 0.01 0.2 1.92 (0.01) 4.0 (0.4) 0.005 (0.002) 0.01 0.1 1.91 (0.01) 4.2 (0.2) 0.006 (0.002) 0.013 S02 value used for all fits was 0.8. 88 Table A.9. EXAFS fitting results for ZnyAl1- yOx films. y in ZnyAl1-yOx Zn – O (Å) N (atoms) σ2 (Å2) r-factor 1 1.97 (0.01) 4.0 (0.2) 0.004 (0.001) 0.0015 0.7 1.96 (0.01) 3.9 (0.2) 0.003 (0.001) 0.003 0.4 1.95 (0.01) 3.8 (0.4) 0.005 (0.002) 0.0035 0.3 1.95 (0.01) 3.9 (0.4) 0.005 (0.002) 0.005 0.2 1.97 (0.01) 3.9 (0.4) 0.005 (0.002) 0.003 0.1 1.97 (0.02) 4.2 (0.5) 0.006 (0.002) 0.007 S02 value used for all fits was 0.8. Table A.10. EXAFS fitting results for GayAl1- yOx films. y in GayAl1-yOx Ga – O (Å) N (atoms) σ2 (Å2) r-factor 1 1.89 (0.01) 4 0.003 (0.001) 0.011 1 2.08 (0.03) 6 0.007 (0.007) 0.011 0.5 1.88 (0.01) 4.3 (0.4) 0.005 (0.002) 0.008 0.3 1.88 (0.01) 4.2 (0.3) 0.004 (0.001) 0.004 0.2 1.88 (0.01) 4.2 (0.2) 0.004 (0.001) 0.002 S02 value used for all fits was 0.8. 89 APPENDIX B CHAPTER III SUPPORTING INFORMATION Table B.1. Electrodeposition conditions for the electrodeposited Ni1- zMzOxHy films. Film ratio (XPS) Ni:M Solution ratio Ni:M Current density for deposition (mA cm-2) Time of deposition (s) Ni 87 : Ce 13 Ni 99 : Ce 1 -0.5 45 Ni 86 : La 14 Ni 90 : La 10 -1.0 55 Ni 92 : Mn 8 Ni 70 : Mn 30 -0.5 120 Ni 87 : Fe 13 Ni 97 : Fe 3 -0.1 38 Ni 60 : Ce 40 Ni 90 : Ce 10 -1.0 150 Ni 68 : La 32 Ni 70 : La 30 -2.0 90 90 Figure B.1. Number of electrons removed per Ni atom during Ni2+/3+ oxidation for spin- cast films, calculated from integration of the anodic peak. Total number of Ni atoms is obtained from the in situ mass determined by QCM. We find that for all the films synthesized, most only vary by a factor of 2, except films with higher content of Ce and Ti, which vary by a factor of 4. Since these differences are less than an order of magnitude, the films are nominally the same and differences in electrochemically active surface area do not affect the activity trends established in this work. We also note this calculation may be and underestimate of the number of electrons per Ni atom as the films likely suffer from lack of complete reversibility of oxidation, as has been shown for Ni- based (oxy)hydroxides.1,2 91 Figure B.2. Morphology of spin-cast Ni1-zMzOxHy films where z ~ 0.3 after 50 cycles in 1 M Fe-free KOH at 20 mV s-1. Figure B.3. Morphology of electrodeposited Ni1-zMzOxHy films after 50 cycles in 1 M Fe-free KOH at 20 mV s-1. 92 Table B.2. Overpotential at 0.5 mA cm-2 for spin-cast Ni1- zMzOxHy films; this is referred to as the “OER onset potential” in Chapter III. Composition η @ j = 0.5 mA cm-2 (mV) cycle 5 η @ j = 0.5 mA cm-2 (mV) cycle 50 η @ j = 0.5 mA cm-2 (mV) Tafel NiOxHy 389 ± 5 426 ± 2 419 ± 7 Ni0.85Ce0.15OxHy 315 ± 7 392 ± 6 430 ± 2 Ni0.93La0.7OxHy 369 ± 6 401 ± 5 428 ± 3 Ni0.9Mn0.1OxHy 385 ± 8 417 ± 7 432 ± 4 Ni0.9Ti0.1OxHy 402 ± 5 420 ± 7 434 ± 11 Ni0.9Fe0.1OxHy 277 ± 6 299 ± 3 297 ± 1 Ni0.65Ce0.35OxHy 322 ± 11 382 ± 11 - Ni0.62La0.38OxHy 417 ± 7 436 ± 13 - Ni0.7Mn0.3OxHy 402 ± 10 432 ± 10 - Ni0.7Ti0.3OxHy 413 ± 12 403 ± 6 - Ni0.7Fe0.3OxHy 248 279 - 93 Figure B.4. Initial voltammograms (Cycle 1) of all spin-cast and electrodeposited film compositions taken at 20 mV s-1 in 1 M Fe-free KOH. Labels that do not specify preparation method are for spin-cast films. For most compositions, electrodeposited films show better redox reversibility, indicated by the similar area and sharpness of the anodic and cathodic redox waves. 94 Figure B.5. Representative cyclic voltammograms of spin-cast Ni1-zMzOxHy films at 20 mV s-1 in Fe-free 1 M KOH. Comparing cycle 5 (Figure 3.1 in the main text) to cycle 50 (Figure B.5) it is clear that all of the films exhibit a decrease in current density. As noted in the main text, the reduction in current density is most pronounced for Ni(Ce)OxHy films. 95 Table B.3. Tafel slope and TOFtm for spin-cast Ni1- zMzOxHy films. Values marked with an asterisk (*) are extrapolated. Composition Tafel slope (mV dec-1) TOFtm (s-1) @ η = 400 mV (Tafel) TOFtm (s-1) @ η = 400 mV (CV cycle 5) TOFtm (s-1) @ η = 400 mV (CV cycle 50) NiOxHy 93 ± 2 0.02 ± 0.007 0.036 ± 0.004 0.016 ± 0.002 Ni0.85Ce0.15OxHy 126 ± 3 0.015 ± 0.006 0.26 ± 0.03 0.04 ± 0.01 Ni0.93La0.7OxHy 101 ± 2 0.022 ± 0.003 0.07 ± 0.01 0.04 ± 0.01 Ni0.9Mn0.1OxHy 92 ± 1 0.018 ± 0.002 0.05 ± 0.02 0.03 ± 0.01 Ni0.9Ti0.1OxHy 105 ± 2 0.021 ± 0.007 0.024 ± 0.003 0.019 ± 0.005 Ni0.9Fe0.1OxHy 40 ± 3 20 ± 6 7.02 ± 0.06* 7.01 ± 0.07* Ni0.65Ce0.35OxHy - - 0.15 ± 0.05 0.03 ± 0.02 Ni0.62La0.38OxHy - - 0.035 ± 0.008 0.018 ± 0.006 Ni0.7Mn0.3OxHy - - 0.03 ± 0.01 0.02 ± 0.01 Ni0.7Ti0.3OxHy - - 0.019 ± 0.006 0.03 ± 0.01 Ni0.7Fe0.3OxHy - - 174 ± 86* 19 ± 5* 96 Figure B.6. Tafel plots for individual spin-cast Ni1-zMzOxHy films along with in situ mass change. Symbols are the average and error bars represent the standard deviation of 3 individual samples. 97 Figure B.7. O 1s XP spectra taken with Al Kα excitation of Ni0.85Ce0.15OxHy films before and after 50 cycles at 20 mV s-1 in Fe-free 1 M KOH. All spectra were shifted so that the C 1s peak was centered at 284.8 eV with the exception of the spin-cast as- deposited spectrum which had significant charging issues likely due to residual H2O and NO3- left in the film. It has been arbitrarily shifted in the figure (bottom) in order to compare peak shape. 98 Figure B.8. Ce 3d XP spectra taken with Al Kα excitation for spin-cast (left) and electrodeposited (right) Ni0.85Ce0.15OxHy films before and after 50 cycles at 20 mV s-1 in Fe-free 1 M KOH. The as-deposited films show evidence of both Ce3+ and Ce4+, while the films after electrochemical analysis have only Ce4+. Specifically, the distinct presence of the peak at ~905 eV corresponding to Ce3+ final states in as deposited Ni0.85Ce0.15OxHy films suggests the film has a significant amount of Ce3+. Conversely, the lack of this peak and the presence of a peak at ~889 eV, which corresponds to Ce4+ final states, indicate mainly Ce4+ in the Ni0.85Ce0.15OxHy films after electrochemical cycling. Some of the distinct spectral features unique to Ce3+ (blue) or Ce4+ (purple) have been indicated by dashed lines.3 Figure B.9. Cyclic voltammetry of a representative Ni0.7Ti0.3OxHy film in Fe-free 1 M KOH at 20 mV s-1. Inset shows a magnification of the Ni2+/3+ redox peaks. The increase in OER current from cycle 5 to cycle 50 is due to the increase in electrochemically active surface area as evident from the increase in the area of the Ni2+/3+ oxidation peak. 99 Figure B.10. XP spectra of Fe 2p region using Mg Kα excitation showing that there is no detectable Fe in the samples that do not intentionally contain Fe. The use of Mg Kα excitation is essential for Fe quantitation given interference from Ni auger peaks when the Al source is used. 100 Figure B.11. Voltammetry of bare Au substrates at 20 mV s-1 in Fe-free 1 M KOH. Figure B.12. Voltammetry comparison of spin-cast Ni0.85Ce0.15OxHy on Au and Pt substrates. Ni0.85Ce0.15OxHy shows similar current density at cycle 5 and 50 on either Pt or Au substrates. The activities at cycle 5 and 50 are also comparable: TOFtm cycle 5 = 0.26 s-1 (Au), 0.19 s-1 (Pt) TOFtm cycle 50 = 0.044 s-1 (Au), 0.038 s-1 (Pt). We note that the Ni2+/3+ oxidation wave is significantly broader on Pt and attribute this to poorer 101 electrical connection of the film to Pt as well as higher resistance of the Pt QCM substrates. Figure B.13. Voltammetry (Cycle 5) of bare Au substrate and CeOx film on Au substrate in 1 M Fe-free KOH at a scan rate of 20 mV s-1. Figure B.14. Change in film loading during Ni2+/3+ oxidation and reduction for a representative spin-cast NiOxHy film during cycle 5 at 20 mV s-1 in 1 M Fe-free KOH. Films in this work exhibit nominally β-Ni(OH)2 behavior in that there is mass loss upon oxidation and mass gain upon reduction.4 The exact ion/solvent movement responsible for these small mass changes is beyond the scope of this work. 102 Figure B.15. Voltammetry of electrodeposited Ni1-zMzOxHy films in 1 M Fe-free KOH at a scan rate of 20 mV s-1. 103 Figure B.16. Comparison of OER activity of electrodeposited (open circles) and spin- cast (solid squares) Ni1-zMzOxHy films. Figure B.17. TOFtm vs. E0’ of electrodeposited Ni1-zMzOxHy films after 5 (solid squares) and 50 (open circles) cycles in Fe-free 1 M KOH at 20 mV s-1. 104 APPENDIX C CHAPTER IV SUPPORTING INFORMATION Electrochemical Results Figure C.1. Dependence of current response on scan rate for electrode substrates. Electrodes were cycled between 0.1 and 0.2 V vs Hg/HgO until current response was stable. The current plotted is the average of the absolute value of the anodic and cathodic current at 0.15 V vs Hg/HgO. Calculated surface areas are shown and are based on an average general capacitance of 0.035 mF cm-2, adapted from Jaramillo et al.1 Figure C.2. Plot of current density at 450 mV overpotential vs Pauling's electronegativity of the electrode substrate material. 105 Figure C.3. Cyclic voltammetry at 20 mV s-1 of Pt (1 M KOH SC grade) and Cu (0.1 M KOH SC grade) electrodes showing irreversible oxidation of the surface. Cycle 1 is plotted for both. 106 APPENDIX D CHAPTER V SUPPORTING INFORMATION Extended X-ray Absorption Fine Structure (EXAFS) Details The EXAFS equation is described by the sum of all scattering paths, i, according to 𝜒𝜒(𝑘𝑘) = ∑ 𝑆𝑆02𝑁𝑁𝑖𝑖𝑓𝑓𝑖𝑖(𝑘𝑘) 𝑘𝑘𝑅𝑅𝑖𝑖 2𝑖𝑖 𝑒𝑒 − 2𝑅𝑅𝑖𝑖 𝜆𝜆(𝑘𝑘)𝑒𝑒−2𝑘𝑘2𝜎𝜎𝑖𝑖2 sin(2𝑘𝑘𝑘𝑘𝑖𝑖 + 𝛿𝛿𝑖𝑖(𝑘𝑘)) (Eq. D.1) where 𝑆𝑆0 2 – amplitude reduction factor (intrinsic to absorber) N – coordination number 𝑘𝑘 = 𝑘𝑘0 + ∆𝑘𝑘 , where R0 is the distance between scatterer and absorber of the theoretical standard and ΔR is the change in distance (i.e. difference between standard and sample) σ2 – mean-square disorder parameter f(k) – effective scattering factor δ(k) – effective scattering phase shift λ(k) – photoelectron mean free path 𝑘𝑘2 = 2𝑚𝑚𝑒𝑒(𝐸𝐸 − 𝐸𝐸0)/ħ, where E0 is difference in energy between standard and sample data. The values of f(k), δ(k), λ(k), and R0 were all calculated using theoretical standards (i.e. CIF file from ICSD #556288) in the Artemis software/IFEFFIT. The amplitude reduction factor, 𝑆𝑆02, was set to 0.78 for Co and 0.75 for Fe, based on previous reports. The values of N, ∆𝑘𝑘, E0, and σ2 were fit by the software. Uncertainties and r-factor values are directly from Artemis. The number of independent points ranged from 6.1 to 6.5, determined by the Nyquist criterion. 107 Figure D.1. Co K-edge XANES on Co(Fe)OxHy films. A) Comparison of CoOxHy films at various potentials with CoOOH standard. B) Comparison of films during the OER with CoOOH standard. C) Comparison of films held at potentials negative of the Co2+/3+ oxidation wave (0 V or 0.05 V vs Hg/HgO) along with Co(OH)2 and CoOOH standards. Figure D.2. Fe K-edge XANES of A) Co0.8Fe0.2OxHy at various Eapp, B) Co0.6Fe0.4OxHy at various Eapp, and C) Co(Fe)OxHy films during the OER along with Fe standards. 108 Figure D.3. k3-weighted chi(k) data for Co(Fe)OxHy films at A) the Co K-edge pre-OER, B) the Co-K-edge during OER, C) the Fe K-edge pre-OER, and D) the Fe K-edge during the OER in 1 M KOH. Table D.1. Results of FT-EXAFS fitting for Co in Co(Fe)OxHy films under applied potential in 1 M KOH. pre-OER OER Film Composition Co – O (Å) N σ2 (Å-2) R- facto r Co – O (Å) N σ2 (Å-2) R- factor CoOxHy 1.90 (0.01) 5.5 (0.7) 0.003 (0.001) 0.014 1.90 (0.01) 5.7 (0.9) 0.002 (0.001) 0.021 Co0.8Fe0.2OxH y 1.90 (0.01) 5.3 (0.8) 0.002 (0.001) 0.019 1.90 (0.01) 5.4 (0.8) 0.002 (0.001) 0.018 Co0.6Fe0.4OxH y 1.90 (0.01) 5.6 (0.8) 0.003 (0.001) 0.018 1.90 (0.01) 5.6 (0.8) 0.002 (0.001) 0.014 109 Table D.2. Results of FT-EXAFS fitting for Fe in Co(Fe)OxHy films under applied potential in 1 M KOH. pre-OER OER Film Composition Fe – O (Å) N σ2 (Å-2) R- factor Fe – O (Å) N σ2 (Å-2) R- factor Co0.8Fe0.2OxH y 1.93 (0.01) 5.0 (0.7) 0.005 (0.001) 0.012 1.90 (0.01) 4.7 (1.0) 0.002 (0.001) 0.018 Co0.6Fe0.4OxH y 1.93 (0.01) 5.4 (0.6) 0.007 (0.002) 0.009 1.90 (0.01) 4.8 (0.6) 0.005 (0.002) 0.011 Computational Details The Vienna Ab-initio Simulation Package (VASP) program was used to perform spin-polarized DFT calculations[6,7] with the Perdew-Burke-Ernzerhof (PBE)6 functional and the DFT+U formalism of Dudarev et al.[8] at a U-J term of 3.0 eV and 3.3 eV for Co and Fe, respectively, as usually for modeling catalysis of oxyhydroxides.[9–13] Projected- augmented wave (PAW) potentials replaced the core electrons of Co 1s2s2p3s3p, Fe 1s2s2p3s3p , and O 1s.[14,15] The unit cell of CoOOH[10,16–19] was cleaved at the (01�5) facet and the dopant was located at the active site as done previously for similar oxyhydroxides.[10,20] The energy cutoff and k-point Gamma-centered grid of 600 eV and 2×2×1, respectively were converged to within < 1 meV per atom. The ion positions convergence threshold was set until the force components on ions were less than 0.03 eV Å-1 while lattice vectors were held fixed for the surface in order to maintain the bulk geometry below. The more commonly used reaction mechanism that has been proposed for several metal oxides includes these reaction steps:[21] *OH2  *OH + H+ + e- (1) *OH  *O + H+ + e- (2) *O + H2O  * + 2H+ + O2 + 2e- (3) * + H2O  *OH2 (4) where intermediate "*OH2" is the surface with a monolayer of adsorbed water molecules and intermediate "*OH" is the surface with an adsorbed hydroxyl group. Here intermediate "*O" has an oxygen termination, and "*" has a vacant termination. The 110 overpotential is defined as the applied voltage needed so that all reaction free energies are negative. The free energy for each reaction was calculated by subtracting reactants and products total energies and adding our calculated Zero Point Energy (ZPE) corrections and entropic contributions. Our results support previous theoretical investigations that considered a different model and mechanism for pure and Fe-doped CoOOH.[19,22] Those models contained different number of transition metal ion layers and different facets. In addition, we did not include the *OOH surface reaction intermediate since we found in previous work that this reaction intermediate is unstable in some cases.[23] Nevertheless, we accounted for the *OH2 surface since this intermediate was proposed for other oxides.[24] The free energy for the direct reaction of ∗ ∗ 𝑂𝑂𝑂𝑂, which was obtained by summing the free energies of reactions 1 and 4 equals -0.57 eV and 0.26 eV for pure and Fe-doped CoOOH, respectively; hence, this route also a possibility that is included in the model. In agreement with previous studies,[22] we also find lower overpotential upon doping with Fe for CoOOH as detailed below. Computational Results Below is additional information on the calculation data, including bond lengths, atomic magnetization, free energies and overpotential, and optimized structures. Figure D.4. Atomic position labelling for Fe-doped CoOOH demonstrated for intermediate *OH2. 111 Table D.3. Fe-O bond distances around the Fe active site for Fe-doped CoOOH. * *O *OH *OH2 Termination 1.875 1.920 1.933 1.923 d_Fe1-O1 [Å] 1.933 2.058 1.913 1.962 d_Fe1-O9 [Å] 1.876 1.885 1.900 1.921 d_Fe1-O13 [Å] 1.892 1.904 1.911 1.964 d_Fe1-O18 [Å] 1.905 1.928 1.897 1.959 d_Fe1-O6 [Å] 1.896 1.939 1.911 1.946 Average [Å] 0.022 0.061 0.013 0.019 Standard deviation [Å] Table D.4. Co-O bond distances far away from the active site for Fe-doped CoOOH. * *O *OH *OH2 Termination 1.926 1.974 1.972 1.951 d_Co4-O4 [Å] 1.898 1.915 1.918 1.917 d_Co4-O23 [Å] 1.882 1.899 1.909 1.895 d_Co4-O17 [Å] 1.912 1.898 1.904 1.903 d_Co4-O12 [Å] 1.918 1.894 1.884 1.893 d_Co4-O5 [Å] 1.926 1.914 1.907 1.916 d_Co4-O11 [Å] 1.910 1.915 1.916 1.913 Average [Å] 0.015 0.030 0.030 0.021 Standard deviation [Å] Table D.5. Co-O bond distances close to the active site for Fe-doped CoOOH. * *O *OH *OH2 Termination 1.905 1.909 1.920 1.918 d_Co7-O9 [Å] 1.912 1.912 1.912 1.913 d_Co7-O3 [Å] 1.918 1.915 1.897 1.915 d_Co7-O15 [Å] 1.927 1.865 1.892 1.916 d_Co7-O13 [Å] 1.910 1.919 1.916 1.914 d_Co7-O21 [Å] 1.915 1.963 1.957 1.926 d_Co7-O22 [Å] 1.915 1.914 1.916 1.917 Average [Å] 0.008 0.020 0.011 0.002 Standard deviation [Å] 112 Table D.6. Co-O bond distances around the Co active site for pure CoOOH. * *O *OH *OH2 Termination 1.913 1.917 1.885 1.903 d_Co12-O1 [Å] 1.897 1.911 1.895 1.909 d_Co12-O6 [Å] 1.878 1.886 1.854 1.913 d_Co12-O18 [Å] 1.867 1.893 1.910 1.891 d_Co12-O13 [Å] 2.021 2.007 1.929 1.881 d_Co12-O9 [Å] 1.915 1.923 1.895 1.900 Average [Å] 0.017 0.102 0.022 0.073 Standard deviation [Å] The atomic oxidation states at the active site were estimated according to the calculated atomic magnetization. The magnetic moments on Fe2+, Fe3+, Fe4+, and Fe5+ are anticipated to be M = 4, 5, 4, and 3, respectively, at high-spin valence configurations of 3d6, 3d5, 3d4, and 3d3. The magnetic moments on Co2+, Co3+, and Co4+ are anticipated to be M = 3, 4, and 5, respectively, at high-spin valence configurations of 3d7, 3d6, and 3d5. The corresponding low-spin configurations for Co2+, Co3+, and Co4+ can be associated with M = 1, (0 or 2), and (1 or 3), respectively. At DFT+U calculations the magnetic moments are typically smaller due to the slightly delocalized nature of the atomic orbitals. The expected oxidation state of Fe or Co in CoOOH is 3+ according to the material's stoichiometry, but the metals may have a distribution of oxidation states. Furthermore, during the reaction the oxidation state of the active site may increase upon deprotonation. Taking all of these into consideration, we provide estimate the oxidation state of the active site as given in Table D.7. Table D.7. Atomic magnetization and estimated oxidation state (in parenthesis) at the Fe and Co active sites for pure and Fe-doped CoOOH, respectively. * *O *OH *OH2 3.6 (3+) 2.6 (5+) 3.3 (4+) 3.9 (3+) Fe-doped CoOOH 1.9 (3+) 1.2 (4+) 1.0 (4+) 0.0 (3+) pure CoOOH 113 Table D.8. Free energies for water oxidation reaction intermediates for pure and Fe- doped CoOOH. The free energies are normalized per electron participating in the reaction. Hence, the free energy of reaction *O * is divided by two. The largest free energy is indicated in bold. Reaction * ->*OH2 -1.97 ∆G4 [eV] pure CoOOH *OH2 ->*OH 1.40 ∆G1 [eV] *OH ->*O 2.29 ∆G2 [eV] *O ->* 1.37 ∆G3/2 [eV] 1.19 Overpotential [eV] *->*OH2 -1.22 ∆G4 [eV] Fe-doped CoOOH *OH2->*OH 1.48 ∆G1 [eV] *OH ->*O 1.60 ∆G2 [eV] *O ->* 1.30 ∆G3/2 [eV] 0.59 Overpotential [eV] Table D.9. Zero-point energy (ZPE) and entropy corrections at each reaction step for pure and Fe-doped CoOOH. Pure Doped �∆(𝑍𝑍𝑍𝑍𝐸𝐸) − 𝑇𝑇∆(𝑆𝑆)� 1 -0.385 -0.364 �∆(𝑍𝑍𝑍𝑍𝐸𝐸) − 𝑇𝑇∆(𝑆𝑆)� 2 -0.362 -0.360 �∆(𝑍𝑍𝑍𝑍𝐸𝐸) − 𝑇𝑇∆(𝑆𝑆)� 3 -0.636 -0.665 �∆(𝑍𝑍𝑍𝑍𝐸𝐸) − 𝑇𝑇∆(𝑆𝑆)� 4 0.792 0.790 114 REFERENCES CITED CHAPTER II (1) Hennek, J. W.; Kim, M. G.; Kanatzidis, M. G.; Facchetti, A.; Marks, T. J. Exploratory Combustion Synthesis: Amorphous Indium Yttrium Oxide for Thin- Film Transistors. J. Am. Chem. Soc. 2012, 134, 9593–9596. (2) Nadarajah, A.; Carnes, M. E.; Kast, M. G.; Johnson, D. W.; Boettcher, S. W. Aqueous Solution Processing of F-Doped SnO2 Transparent Conducting Oxide Films Using a Reactive Tin(II) Hydroxide Nitrate Nanoscale Cluster. Chem. Mater. 2013, 25, 4080–4087. (3) Nadarajah, A.; Wu, M. Z. B.; Archila, K.; Kast, M. G.; Smith, A. M.; Chiang, T. H.; Keszler, D. A.; Wager, J. F.; Boettcher, S. W. Amorphous In–Ga–Zn Oxide Semiconducting Thin Films with High Mobility from Electrochemically Generated Aqueous Nanocluster Inks. Chem. Mater. 2015, 27, 5587–5596. (4) Kim, S.; Gil, Y.; Choi, Y.; Kim, K.-K.; Yun, H. J.; Son, B.; Choi, C.; Kim, H. Carrier Transport at Metal/Amorphous Hafnium–Indium–Zinc Oxide Interfaces. ACS Appl. Mater. Interfaces 2015, 7, 22385–22393. (5) Leong, W. L.; Ren, Y.; Seng, H. L.; Huang, Z.; Chiam, S. Y.; Dodabalapur, A. Efficient Polymer Solar Cells Enabled by Low Temperature Processed Ternary Metal Oxide as Electron Transport Interlayer with Large Stoichiometry Window. ACS Appl. Mater. Interfaces 2015, 7, 11099–11106. (6) Qin, P.-L.; Lei, H.-W.; Zheng, X.-L.; Liu, Q.; Tao, H.; Yang, G.; Ke, W.-J.; Xiong, L.-B.; Qin, M.-C.; Zhao, X.-Z.; et al. Copper-Doped Chromium Oxide Hole-Transporting Layer for Perovskite Solar Cells: Interface Engineering and Performance Improvement. Adv. Mater. Interfaces 2016, 3, No. 1500799. (7) Moriceau, P.; Lebouteiller, A.; Bordes, E.; Courtine, P. A New Concept Related to Selectivity in Mild Oxidation Catalysis of Hydrocarbons: The Optical Basicity of Catalyst Oxygen. Phys. Chem. Chem. Phys. 1999, 1, 5735–5744. (8) Routray, K.; Zhou, W.; Kiely, C. J.; Wachs, I. E. Catalysis Science of Methanol Oxidation over Iron Vanadate Catalysts: Nature of the Catalytic Active Sites. ACS Catal. 2011, 1, 54–66. (9) Getsoian, A.; Zhai, Z.; Bell, A. T. Band-Gap Energy as a Descriptor of Catalytic Activity for Propene Oxidation over Mixed Metal Oxide Catalysts. J. Am. Chem. Soc. 2014, 136, 13684–13697. (10) Smith, R. D. L.; Prévot, M. S.; Fagan, R. D.; Zhang, Z.; Sedach, P. A.; Siu, M. K. J.; Trudel, S.; Berlinguette, C. P. Photochemical Route for Accessing Amorphous Metal Oxide Materials for Water Oxidation Catalysis. Science 2013, 340, 60–63. (11) Velu, S.; Suzuki, K.; Osaki, T. Oxidative Steam Reforming of Methanol over 115 CuZnAl(Zr)-Oxide Catalysts; a New and Efficient Method for the Production of CO-Free Hydrogen for Fuel Cells. Chem. Commun. 1999, 0, 2341–2342. (12) Subramanian, V.; Gnanakumar, E. S.; Jeong, D.-W.; Han, W.-B.; Gopinath, C. S.; Roh, H.-S. A Rationally Designed CuFe2O4–mesoporous Al2O3 Composite towards Stable Performance of High Temperature Water–gas Shift Reaction. Chem. Commun. 2013, 49, 11257–11259. (13) Li, D.; Li, X.; Gong, J. Catalytic Reforming of Oxygenates: State of the Art and Future Prospects. Chem. Rev. 2016, 116, 11529–11653. (14) Zhou, N.; Kim, M.-G.; Loser, S.; Smith, J.; Yoshida, H.; Guo, X.; Song, C.; Jin, H.; Chen, Z.; Yoon, S. M.; et al. Amorphous Oxide Alloys as Interfacial Layers with Broadly Tunable Electronic Structures for Organic Photovoltaic Cells. Proc. Natl. Acad. Sci. 2015, 112, 7897–7902. (15) May, K. J.; Carlton, C. E.; Stoerzinger, K. A.; Risch, M.; Suntivich, J.; Lee, Y. L.; Grimaud, A.; Shao-Horn, Y. Influence of Oxygen Evolution during Water Oxidation on the Surface of Perovskite Oxide Catalysts. J. Phys. Chem. Lett. 2012, 3, 3264–3270. (16) Trudel, S.; Daryl Crozier, E.; Gordon, R. A.; Budnik, P. S.; Hill, R. H. X-Ray Absorption Fine Structure Study of Amorphous Metal Oxide Thin Films Prepared by Photochemical Metalorganic Deposition. J. Solid State Chem. 2011, 184, 1025– 1035. (17) Arhammar, C.; Pietzsch, A.; Bock, N.; Holmstrom, E.; Araujo, C. M.; Grasjo, J.; Zhao, S.; Green, S.; Peery, T.; Hennies, F.; et al. Unveiling the Complex Electronic Structure of Amorphous Metal Oxides. Proc. Natl. Acad. Sci. 2011, 108, 6355–6360. (18) Medvedeva, J. E.; Buchholz, D. B.; Chang, R. P. H. Recent Advances in Understanding the Structure and Properties of Amorphous Oxide Semiconductors. Adv. Electron. Mater. 2017, 3, No. 1700082. (19) Zimmermann, R.; Steiner, P.; Claessen, R.; Reinert, F.; Hüfner, S.; Blaha, P.; Dufek, P. Electronic Structure of 3d-Transition-Metal Oxides: On-Site Coulomb Repulsion versus Covalency. J. Phys. Condens. Matter 1999, 11, 1657–1682. (20) Biswas, P.; Tafen, D. N.; Inam, F.; Cai, B.; Drabold, D. A. Materials Modeling by Design: Applications to Amorphous Solids. J. Phys. Condens. Matter 2009, 21, No. 084207. (21) Lee, S. K.; Lee, S. B.; Park, S. Y.; Yi, Y. S.; Ahn, C. W. Structure of Amorphous Aluminum Oxide. Phys. Rev. Lett. 2009, 103, 4–7. (22) Lee, S. K.; Park, S. Y.; Yi, Y. S.; Moon, J. Structure and Disorder in Amorphous Alumina Thin Films: Insights from High-Resolution Solid-State NMR. J. Phys. Chem. C 2010, 114, 13890–13894. (23) Sarou-Kanian, V.; Gleizes, A. N.; Florian, P.; Samélor, D.; Massiot, D.; Vahlas, C. 116 Temperature-Dependent 4-, 5- and 6-Fold Coordination of Aluminum in MOCVD- Grown Amorphous Alumina Films: A Very High Field 27Al-NMR Study. J. Phys. Chem. C 2013, 117, 21965–21971. (24) Machala, L.; Zboril, R.; Gedanken, A. Amorphous Iron(III) Oxide - A Review. J. Phys. Chem. B 2007, 111, 4003–4018. (25) Nomura, K.; Kamiya, T.; Ohta, H.; Uruga, T.; Hirano, M.; Hosono, H. Local Coordination Structure and Electronic Structure of the Large Electron Mobility Amorphous Oxide Semiconductor In-Ga-Zn-O: Experiment and Ab Initio Calculations. Phys. Rev. B - Condens. Matter Mater. Phys. 2007, 75, 2–6. (26) Filatova, E. O.; Konashuk, A. S. Interpretation of the Changing the Band Gap of Al2O3 Depending on Its Crystalline Form: Connection with Different Local Symmetries. J. Phys. Chem. C 2015, 119, 20755–20761. (27) Kast, M. G.; Cochran, E. A.; Enman, L. J.; Mitchson, G.; Ditto, J.; Siefe, C.; Plassmeyer, P. N.; Greenaway, A. L.; Johnson, D. C.; Page, C. J.; et al. Amorphous Mixed-Metal Oxide Thin Films from Aqueous Solution Precursors with Near-Atomic Smoothness. J. Am. Chem. Soc. 2016, 138, 16800–16808. (28) Velu, S.; Suzaki, K.; Gopinath, C. S. Photoemission and in Situ XRD Investigations on CuCoZnAl-Mixed Metal Oxide Catalysts for the Oxidative Steam Reforming of Methanol. J. Phys. Chem. B 2002, 106, 12737–12746. (29) Velu, S.; Suzuki, K.; Vijayaraj, M.; Barman, S.; Gopinath, C. S. In Situ XPS Investigations of Cu1-xNixZnAl-Mixed Metal Oxide Catalysts Used in the Oxidative Steam Reforming of Bio-Ethanol. Appl. Catal. B Environ. 2005, 55, 287–299. (30) Fu, L. J.; Li, X. Y.; Liu, M. Z.; Yang, H. M. Insights into the Nature of Cu Doping in Amorphous Mesoporous Alumina. J. Mater. Chem. A 2013, 1, 14592–14605. (31) Fu, L.; Yang, H.; Hu, Y.; Wu, D.; Navrotsky, A. Tailoring Mesoporous γ-Al2O3 Properties by Transition Metal Doping: A Combined Experimental and Computational Study. Chem. Mater. 2017, 29, 1338–1349. (32) Mathew, T.; Sivaranjani, K.; Gnanakumar, E. S.; Yamada, Y.; Kobayashi, T.; Gopinath, C. S. γ-Al2−xMxO3±y (M = Ti4+ through Ga3+): Potential Pseudo-3D Mesoporous Materials with Tunable Acidity and Electronic Structure. J. Mater. Chem. 2012, 22, 13484–13493. (33) Lin, Q.; Su, Y.; Zhang, M.-J.; Yang, X.; Yuan, S.; Hu, J.; Lin, Y.; Liang, J.; Pan, F. A Novel P-Type and Metallic Dual-Functional Cu–Al2O3 Ultra-Thin Layer as the Back Electrode Enabling High Performance of Thin Film Solar Cells. Chem. Commun. 2016, 52, 10708–10711. (34) Wang, W.; Liu, W.; Chang, I.-Y.; Wills, L. A.; Zakharov, L. N.; Boettcher, S. W.; Cheong, P. H.-Y.; Fang, C.; Keszler, D. A. Electrolytic Synthesis of Aqueous Aluminum Nanoclusters and in Situ Characterization by Femtosecond Raman 117 Spectroscopy and Computations. Proc. Natl. Acad. Sci. 2013, 110, 18397–18401. (35) Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: Data Analysis for X-Ray Absorption Spectroscopy Using IFEFFIT. J. Synchrotron Radiat. 2005, 12, 537–541. (36) Bachmann, H. G.; Ahmed, F. R.; Barnes, W. H. The Crystal Structure of Vanadium Pentoxide. Zeitschrift für Krist. - Cryst. Mater. 1961, 115, 110–131. (37) Sawada, H. Residual Electron Density Study of Chromium Sesquioxide by Crystal Structure and Scattering Factor Refinement. Mater. Res. Bull. 1994, 29, 239–245. (38) Stephens, J. S.; Cruickshank, D. W. J. The Crystal Structure of (CrO3)∞. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1970, 26, 222–226. (39) Geller, S. Structure of α-Mn2O3, (Mn0.983Fe0.017)2O3 and (Mn0.37Fe0.63)2O3 and Relation to Magnetic Ordering. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1971, 27, 821–828. (40) Blake, R. L.; Hessevick, R. E.; Zoltai, T.; Finger, L. W. Refinement of the Hematite Structure. Am. Mineral. 1966, 51, 123–129. (41) Will, G.; Masciocchi, N.; Parrish, W.; Hart, M. Refinement of Simple Crystal Structures from Synchrotron Radiation Powder Diffraction Data. J. Appl. Crystallogr. 1987, 20, 394–401. (42) Sasaki, S.; Fujino, K.; Takeuchi, Y. X-Ray Determination of Electron-Density Distributions in Oxides, MgO, MnO, CoO, and NiO, and Atomic Scattering Factors of Their Constituent Atoms. Proc. Jpn. Acad. 1979, 55, 43–48. (43) Åsbrink, S.; Norrby, L. J. A Refinement of the Crystal Structure of Copper(II) Oxide with a Discussion of Some Exceptional E.s.d.’s. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1970, 26, 8–15. (44) Abrahams, S. C.; Bernstein, J. L. Remeasurement of the Structure of Hexagonal ZnO. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1969, 25, 1233– 1236. (45) Geller, S. Crystal Structure of β-Ga2O3. J. Chem. Phys. 1960, 33, 676–684. (46) Ravel, B. ATOMS: Crystallography for the X-Ray Absorption Spectroscopist. J. Synchrotron Radiat. 2001, 8, 314–316. (47) Newville, M. IFEFFIT : Interactive XAFS Analysis and FEFF Fitting. J. Synchrotron Radiat. 2001, 8, 322–324. (48) Buchholz, D. B.; Liu, J.; Marks, T. J.; Zhang, M.; Chang, R. P. H. Control and Characterization of the Structural, Electrical, and Optical Properties of Amorphous Zinc-Indium-Tin Oxide Thin Films. ACS Appl. Mater. Interfaces 2009, 1, 2147– 2153. (49) Tauc, J.; Menth, A. States in the Gap. J. Non. Cryst. Solids 1972, 8–10, 569–585. 118 (50) Wulfsberg, G. Inorganic Chemistry; University Science Books: Sausalito, CA, 2000. (51) Hammann, B. A.; Kast, M. G.; Plassmeyer, P. N.; Ma, Z. L.; Woods, K. N.; Keszler, D. A.; Page, C. J.; Boettcher, S. W.; Hayes, S. E. Elucidation of Structure- Property Relationships in Aqueous-Processed Aluminum Oxide Thin Films: Insights from Solid-State NMR and Dielectric Response. Chem. Mater. 2018, submitted for publication. (52) Pauling, L. The Principles Determining the Structure of Complex Ionic Crystals. J. Am. Chem. Soc. 1929, 51, 1010–1026. (53) Zachariasen, W. H. The Atomic Arrangement in Glass. J. Am. Chem. Soc. 1932, 196, 3841–3851. (54) Thimsen, E.; Biswas, S.; Lo, C. S.; Biswas, P. Predicting the Band Structure of Mixed Transition Metal Oxides: Theory and Experiment. J. Phys. Chem. C 2009, 113, 2014–2021. (55) Dunitz, J. D.; Orgel, L. E. Electronic Properties of Transition-Metal Oxides-II. 1957, 3, 318–323. (56) Gao, X.; Wachs, I. E. Investigation of Surface Structures of Supported Vanadium Oxide Catalysts by UV−vis−NIR Diffuse Reflectance Spectroscopy. J. Phys. Chem. B 2000, 104, 1261–1268. (57) Akl, A. A. Effect of Solution Molarity on the Characteristics of Vanadium Pentoxide Thin Film. Appl. Surf. Sci. 2006, 252, 8745–8750. (58) Lee, C.-H.; Lin, T.-S.; Mou, C.-Y. Preparation and Characterization of Vanadium Oxide Species Supported on Mesoporous Silica for the Hydroxylation of Benzene. J. Phys. Chem. C 2007, 111, 3873–3882. (59) Schweitzer, G. K.; Pesterfield, L. L. The Aqueous Chemistry of the Elements; Oxford University Press: New York, 2009. (60) Peterson, M. L.; Brown, G. E.; Parks, G. A.; Stein, C. L. Differential Redox and Sorption of Cr(III/VI) on Natural Silicate and Oxide Minerals: EXAFS and XANES Results. Geochim. Cosmochim. Acta 1997, 61, 3399–3412. (61) Shriver, D.; Atkins, P. Shriver & Atkins’ Inorganic Chemistry, 5th ed.; W. H. Freeman and Company: New York, 2009. (62) Murata, T.; Torisaka, M.; Takebe, H.; Morinaga, K. Compositional Dependence of the Valency State of Cr Ions in Oxide Glasses. J. Non. Cryst. Solids 1997, 220, 139–146. (63) Bondioli, F.; Ferrari, A. M.; Leonelli, C.; Manfredini, T.; Linati, L.; Mustarelli, P. Reaction Mechanism in Alumina/chromia (Al2O3-Cr2O3) Solid Solutions Obtained by Coprecipitation. J. Am. Ceram. Soc. 2000, 83, 2036–2040. (64) Gaudry, É.; Sainctavit, P.; Juillot, F.; Bondioli, F.; Ohresser, P.; Letard, I. From 119 the Green Color of Eskolaite to the Red Color of Ruby: An X-Ray Absorption Spectroscopy Study. Phys. Chem. Miner. 2006, 32, 710–720. (65) Duret, A.; Grätzel, M. Visible Light-Induced Water Oxidation on Mesoscopic α- Fe2O3 Films Made by Ultrasonic Spray Pyrolysis. J. Phys. Chem. B 2005, 109, 17184–17191. (66) Shinde, S. S.; Bansode, R. A.; Bhosale, C. H.; Rajpure, K. Y. Physical Properties of Hematite α-Fe2O3 Thin Films: Application to Photoelectrochemical Solar Cells. J. Semicond. 2011, 32, No. 013001. (67) Biesinger, M. C.; Payne, B. P.; Grosvenor, A. P.; Lau, L. W. M.; Gerson, A. R.; Smart, R. S. C. Resolving Surface Chemical States in XPS Analysis of First Row Transition Metals, Oxides and Hydroxides: Cr, Mn, Fe, Co and Ni. Appl. Surf. Sci. 2011, 257, 2717–2730. (68) Chalmin, E.; Farges, F.; Brown, G. E. A Pre-Edge Analysis of Mn K-Edge XANES Spectra to Help Determine the Speciation of Manganese in Minerals and Glasses. Contrib. to Mineral. Petrol. 2009, 157, 111–126. (69) Eisenberger, P.; Brown, G. S. The Study of Disordered Systems by EXAFS: Limitations. Solid State Commun. 1979, 29, 481–484. (70) McKeown, D. A.; Kot, W. K.; Gan, H.; Pegg, I. L. X-Ray Absorption Studies of Manganese Valence and Local Environment in Borosilicate Waste Glasses. J. Non. Cryst. Solids 2003, 328, 71–89. (71) Hålenius, U.; Bosi, F.; Skogby, H. Galaxite, MnAl2O4, a Spectroscopic Standard for Tetrahedrally Coordinated Mn2+ in Oxygen-Based Mineral Structures. Am. Mineral. 2007, 92, 1225–1231. (72) Smith, W. L.; Hobson, A. D. The Structure of Cobalt Oxide, Co3O4. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1973, 29, 362–363. (73) Rangappa, D.; Naka, T.; Kondo, A.; Ishii, M.; Kobayashi, T.; Adschiri, T. Transparent CoAl2O4 Hybrid Nano Pigment by Organic Ligand-Assisted Supercritical Water. J. Am. Chem. Soc. 2007, 129, 11061–11066. (74) Cheng, C.-S.; Serizawa, M.; Sakata, H.; Hirayama, T. Electrical Conductivity of Co3O4 Films Prepared by Chemical Vapour Deposition. Mater. Chem. Phys. 1998, 53, 225–230. (75) Gulino, A.; Fiorito, G.; Fragala, I. Deposition of Thin Films of Cobalt Oxides by MOCVD. J. Mater. Chem. 2003, 13, 861–865. (76) Mahmoud, S. A.; Akl, A. A.; Kamal, H.; Abdel-Hady, K. Opto-Structural, Electrical and Electrochromic Properties of Crystalline Nickel Oxide Thin Films Prepared by Spray Pyrolysis. Phys. B Condens. Matter 2002, 311, 366–375. (77) Kamal, H.; Elmaghraby, E. K.; Ali, S. A.; Abdel-Hady, K. Characterization of Nickel Oxide Films Deposited at Different Substrate Temperatures Using Spray 120 Pyrolysis. J. Cryst. Growth 2004, 262, 424–434. (78) Praliaud, H.; Mikhailenko, S.; Chajar, Z.; Primet, M. Surface and Bulk Properties of Cu-ZSM-5 and Cu/Al2O3 Solids during Redox Treatments. Correlation with the Selective Reduction of Nitric Oxide by Hydrocarbons. Appl. Catal. B Environ. 1998, 16, 359–374. (79) Mindru, I.; Gingasu, D.; Patron, L.; Marinescu, G.; Calderon-Moreno, J. M.; Preda, S.; Oprea, O.; Nita, S. Copper Aluminate Spinel by Soft Chemical Routes. Ceram. Int. 2016, 42, 154–164. (80) Koffyberg, F. P.; Benko, F. A. A Photoelectrochemical Determination of the Position of the Conduction and Valence Band Edges of P-Type CuO. J. Appl. Phys. 1982, 53, 1173–1177. (81) Marabelli, F.; Parravicini, G. B.; Salghetti-Drioli, F. Optical Gap of CuO. Phys. Rev. B 1995, 52, 1433–1436. (82) Meyer, B. K.; Polity, A.; Reppin, D.; Becker, M.; Hering, P.; Klar, P. J.; Sander, T.; Reindl, C.; Benz, J.; Eickhoff, M.; et al. Binary Copper Oxide Semiconductors: From Materials towards Devices. Phys. Status Solidi 2012, 249, 1487–1509. (83) Anufrienko, V. F.; Yashnik, S. A.; Bulgakov, N. N.; Larina, T. V.; Vasenin, N. T.; Ismagilov, Z. R. A Study of Linear Copper Oxide Structures in the Channels of the ZSM-5 Zeolite by Electronic Diffuse Reflectance Spectroscopy. Dokl. Phys. Chem. 2003, 392, 207–211. (84) Miron, I.; Enache, C.; Vasile, M.; Grozescu, I. Optical Properties of ZnAl2O4 Nanomaterials Obtained by the Hydrothermal Method. Phys. Scr. 2012, T149, 14064. (85) Hoffman, R. Solids and Surfaces: A Chemist’s View of Bonding in Extended Structures, 1st ed.; Wiley-VCH Inc.: New York, 1988. (86) Liu, Z.; Zhu, J.; Li, J.; Ma, L.; Woo, S. I. Novel Mn-Ce-Ti Mixed-Oxide Catalyst for the Selective Catalytic Reduction of NOx with NH3. ACS Appl. Mater. Interfaces 2014, 6, 14500–14508. (87) Zhang, C.; Fagan, R. D.; Smith, R. D. L.; Moore, S. A.; Berlinguette, C. P.; Trudel, S. Mapping the Performance of Amorphous Ternary Metal Oxide Water Oxidation Catalysts Containing Aluminium. J. Mater. Chem. A 2015, 3, 756–761. (88) Kim, H. S.; Park, J. S.; Jeong, H. K.; Son, K. S.; Kim, T. S.; Seon, J. B.; Lee, E.; Chung, J. G.; Kim, D. H.; Ryu, M.; et al. Density of States-Based Design of Metal Oxide Thin-Film Transistors for High Mobility and Superior Photostability. ACS Appl. Mater. Interfaces 2012, 4, 5416–5421. (89) Grzybowska, B.; Słoczyński, J.; Grabowski, R.; Wcisło, K.; Kozłowska, A.; Stoch, J.; Zieliński, J. Chromium Oxide/Alumina Catalysts in Oxidative Dehydrogenation of Isobutane. J. Catal. 1998, 178, 687–700. 121 (90) Rida, K.; Benabbas, A.; Bouremmad, F.; Peña, M. A.; Martínez-Arias, A. Surface Properties and Catalytic Performance of La1-xSrxCrO3 Perovskite-Type Oxides for CO and C3H6 Combustion. Catal. Commun. 2006, 7, 963–968. (91) Smith, J.; Zeng, L.; Khanal, R.; Stallings, K.; Facchetti, A.; Medvedeva, J. E.; Bedzyk, M. J.; Marks, T. J. Cation Size Effects on the Electronic and Structural Properties of Solution-Processed In–X–O Thin Films. Adv. Electron. Mater. 2015, 1, 1–11. (92) Choi, C.-H.; Han, S.-Y.; Su, Y.-W.; Fang, Z.; Lin, L.-Y.; Cheng, C.-C.; Chang, C. Fabrication of High-Performance, Low-Temperature Solution Processed Amorphous Indium Oxide Thin-Film Transistors Using a Volatile Nitrate Precursor. J. Mater. Chem. C 2015, 3, 854–860. CHAPTER III (1) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci. 2006, 103, 15729–15735. (2) Ursua, A.; Gandia, L. M.; Sanchis, P. Proc. IEEE 2012, 100, 410–426. (3) Pinaud, B. A.; Benck, J. D.; Seitz, L. C.; Forman, A. J.; Chen, Z.; Deutsch, T. G.; James, B. D.; Baum, K. N.; Baum, G. N.; Ardo, S.; Wang, H.; Miller, E.; Jaramillo, T. F. Energy Environ. Sci. 2013, 6, 1983–2002. (4) Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S. W.; Mi, Q.; Santori, E. A.; Lewis, N. S. Chem. Rev. 2010, 110, 6446–6473. (5) Trotochaud, L.; Boettcher, S. W. Scr. Mater. 2014, 74, 25–32. (6) Rossmeisl, J.; Logadottir, A.; Nørskov, J. K. Chem. Phys. 2005, 319, 178–184. (7) Galán-Mascarós, J. R. ChemElectroChem 2015, 2, 37–50. (8) Dau, H.; Limberg, C.; Reier, T.; Risch, M.; Roggan, S.; Strasser, P. ChemCatChem 2010, 2, 724–761. (9) Rossmeisl, J.; Qu, Z.-W.; Zhu, H.; Kroes, G.-J.; Nørskov, J. K. J. Electroanal. Chem. 2007, 607, 83–89. (10) Man, I. C.; Su, H.-Y.; Calle-Vallejo, F.; Hansen, H. A.; Martínez, J. I.; Inoglu, N. G.; Kitchin, J.; Jaramillo, T. F.; Nørskov, J. K.; Rossmeisl, J. ChemCatChem 2011, 3, 1159–1165. (11) Ruetschi, P.; Delahay, P. J. Chem. Phys. 1955, 23, 556–560. (12) Sabatier, F. La catalyse en chimie organique; Berauge: Paris, 1920; pp. 48-57. (13) Trotochaud, L.; Ranney, J. K.; Williams, K. N.; Boettcher, S. W. J. Am. Chem. Soc. 2012, 134, 17253–17261. (14) Trotochaud, L.; Young, S. L.; Ranney, J. K.; Boettcher, S. W. J. Am. Chem. Soc. 2014, 136, 6744–6753. 122 (15) Burke, M. S.; Kast, M. G.; Trotochaud, L.; Smith, A. M.; Boettcher, S. W. J. Am. Chem. Soc. 2015, 137, 3638–3648. (16) Smith, A. M.; Trotochaud, L.; Burke, M. S.; Boettcher, S. W. Chem. Commun. 2015, 51, 5261–5263. (17) Hong, W. T.; Risch, M.; Stoerzinger, K. A.; Grimaud, A.; Suntivich, J.; Shao- Horn, Y. ENERGY Environ. Sci. 2015, 8, 1404–1427. (18) Suntivich, J.; May, K. J.; Gasteiger, H. A.; Goodenough, J. B.; Shao-Horn, Y. Science 2011, 334, 1383–1385. (19) Smith, R. D. L.; Prévot, M. S.; Fagan, R. D.; Zhang, Z.; Sedach, P. A.; Siu, M. K. J.; Trudel, S.; Berlinguette, C. P. Science 2013, 340, 60–63. (20) Matsumoto, Y.; Sato, E. Mater. Chem. Phys. 1986, 14 (5), 397–426. (21) Burke, M. S.; Enman, L. J.; Batchellor, A. S.; Zou, S.; Boettcher, S. W. Chem. Mater. 2015, 27 (22), 7549–7558. (22) Burke, M. S.; Zou, S.; Enman, L. J.; Kellon, J. E.; Gabor, C. A.; Pledger, E.; Boettcher, S. W. J. Phys. Chem. Lett. 2015, 6, 3737–3742. (23) Batchellor, A. S.; Boettcher, S. W. ACS Catal. 2015, 5, 6680–6689. (24) Lu, X.; Zhao, C. Nat. Commun. 2015, 6, No. 6616. (25) Ma, W.; Ma, R.; Wang, C.; Liang, J.; Liu, X.; Zhou, K.; Sasaki, T. ACS Nano 2015, 9, 1977–1984. (26) Hunter, B. M.; Blakemore, J. D.; Deimund, M.; Gray, H. B.; Winkler, J. R.; Müller, A. M. J. Am. Chem. Soc. 2014, 136, 13118–13121. (27) Gong, M.; Li, Y.; Wang, H.; Liang, Y.; Wu, J. Z.; Zhou, J.; Wang, J.; Regier, T.; Wei, F.; Dai, H. J. Am. Chem. Soc. 2013, 135, 8452–8455. (28) Corrigan, D. A.; Bendert, R. M. J. Electrochem. Soc. 1989, 136, 723–728. (29) Landon, J.; Demeter, E.; İnoğlu, N.; Keturakis, C.; Wachs, I. E.; Vasić, R.; Frenkel, A. I.; Kitchin, J. R. ACS Catal. 2012, 2, 1793–1801. (30) Louie, M. W.; Bell, A. T. J. Am. Chem. Soc. 2013, 135 (33), 12329–12337. (31) Cordoba, S. I.; Carbonio, R. E.; Teijelo, M. L.; Macagno, V. A. Electrochim. Acta 1986, 31, 1321–1332. (32) Lu, P. W. T.; Srinivasan, S. J. Electrochem. Soc. 1978, 125, 265–270. (33) McCrory, C. C. L.; Jung, S.; Ferrer, I. M.; Chatman, S. M.; Peters, J. C.; Jaramillo, T. F. J. Am. Chem. Soc. 2015, 137, 4347–4357. (34) McCrory, C. C. L.; Jung, S.; Peters, J. C.; Jaramillo, T. F. J. Am. Chem. Soc. 2013, 135, 16977–16987. (35) Qian, L.; Lu, Z.; Xu, T.; Wu, X.; Tian, Y.; Li, Y.; Huo, Z.; Sun, X.; Duan, X. Adv. Energy Mater. 2015, 5, No. 1500245. 123 (36) Niu, K.-Y.; Lin, F.; Jung, S.; Fang, L.; Nordlund, D.; McCrory, C. C. L.; Weng, T.-C.; Ercius, P.; Doeff, M. M.; Zheng, H. Nano Lett. 2015, 15, 2498–2503. (37) Haber, J. A.; Guevarra, D.; Jung, S.; Jin, J.; Gregoire, J. M. ChemElectroChem 2014, 1, 1613–1617. (38) Diaz-Morales, O.; Ledezma-Yanez, I.; Koper, M. T. M.; Calle-Vallejo, F. ACS Catal. 2015, 5, 5380–5387. (39) Yeo, B. S.; Bell, A. T. J. Phys. Chem. C 2012, 116, 8394–8400. (40) Lyons, M. E. G.; Brandon, M. P. J. Electroanal. Chem. 2010, 641, 119–130. (41) Miles, M. H. J. Electroanal. Chem. Interfacial Electrochem. 1975, 60, 89–96. (42) Dincă, M.; Surendranath, Y.; Nocera, D. G. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 10337–10341. (43) Corrigan, D. A. J. Electrochem. Soc. 1987, 134, 377–384. (44) Lu, P. W. T.; Srinivasan, S. J. Electrochem. Soc. 1978, 125, 1416–1422. (45) Wehrens-Dijksma, M.; Notten, P. H. L. Electrochim. Acta 2006, 51, 3609–3621. (46) Godwin, I. J.; Lyons, M. E. G. Electrochem. Commun. 2013, 32, 39–42. (47) Klaus, S.; Cai, Y.; Louie, M. W.; Trotochaud, L.; Bell, A. T. J. Phys. Chem. C 2015, 119, 7243–7254. (48) Haber, J. A.; Xiang, C.; Guevarra, D.; Jung, S.; Jin, J.; Gregoire, J. M. ChemElectroChem 2014, 1, 524–528. (49) Corrigan, D. A.; Conell, R. S.; Fierro, C. A.; Scherson, D. A. J. Phys. Chem. 1987, 91, 5009–5011. (50) Friebel, D.; Louie, M. W.; Bajdich, M.; Sanwald, K. E.; Cai, Y.; Wise, A. M.; Cheng, M.-J.; Sokaras, D.; Weng, T.-C.; Alonso-Mori, R.; Davis, R. C.; Bargar, J. R.; Nørskov, J. K.; Nilsson, A.; Bell, A. T. J. Am. Chem. Soc. 2015, 137, 1305– 1313. (51) Zou, S.; Burke, M. S.; Kast, M. G.; Fan, J.; Danilovic, N.; Boettcher, S. W. Chem. Mater. 2015, 27, 8011–8020. (52) Chen, J. Y. C.; Dang, L.; Liang, H.; Bi, W.; Gerken, J. B.; Jin, S.; Alp, E. E.; Stahl, S. S. J. Am. Chem. Soc. 2015, 137, 15090–15093. (53) Oliver-Tolentino, M. A.; Vázquez-Samperio, J.; Manzo-Robledo, A.; González- Huerta, R. de G.; Flores-Moreno, J. L.; Ramírez-Rosales, D.; Guzmán-Vargas, A. J. Phys. Chem. C 2014, 118, 22432–22438. (54) Seitz, L. C.; Pinaud, B. A.; Nordlund, D.; Jaramillo, T. F. ECS Trans. 2013, 58, 285–291. (55) Seitz, L. C.; Pinaud, B. A.; Nordlund, D.; Gorlin, Y.; Gallo, A.; Jaramillo, T. F. J. Electrochem. Soc. 2015, 162, H841–H846. 124 (56) Trasatti, S.; Petrii, O. A. J. Electroanal. Chem. 1992, 327, 353–376. (57) Sauerbrey, G. Zeitschrift für Phys. A Hadron. Nucl. 1959, 155, 206–222. (58) Klaus, S.; Louie, M. W.; Trotochaud, L.; Bell, A. T. J. Phys. Chem. C 2015, 119, 18303–18316. (59) Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions; Pergamon Press: Oxford, 1966; pp. 194. (60) Schweitzer, G. K.; Pesterfield, L. L. The Aqueous Chemistry of the Elements; Oxford University Press, 2009; pp. 269. (61) Balasubramanian, M.; Melendres, C. A.; Mansour, A. N. J. Electrochem. Soc. 1999, 146, 607–614. (62) Balasubramanian, M.; Melendres, C. A.; Mansour, A. N. Thin Solid Films 1999, 347, 178–183. (63) Haber, J. A.; Anzenburg, E.; Yano, J.; Kisielowski, C.; Gregoire, J. M. Adv. Energy Mater. 2015, 5, No.1402307. (64) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. Handbook of X-ray photoelectron spectroscopy: a reference book of standard spectra for identification and interpretation of XPS data; Chastain, J., Ed.; Perkin-Elmer: Eden Prairie, MN, 1992; pp. 45 (65) Klaus, S.; Trotochaud, L.; Cheng, M.-J.; Head-Gordon, M.; Bell, A. T. ChemElectroChem 2016, 3, 66–73. (66) Doyle, R. L.; Godwin, I. J.; Brandon, M. P.; Lyons, M. E. G. Phys. Chem. Chem. Phys. 2013, 15, 13737–13783. (67) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, second.; John Wiley & Sons: New York, 2001; pp. 108-115. (68) Trześniewski, B. J.; Diaz-Morales, O.; Vermaas, D. A.; Longo, A.; Bras, W.; Koper, M. T. M.; Smith, W. A. J. Am. Chem. Soc. 2015, 137, 15112–15121. (69) Tung, C.-W.; Hsu, Y.-Y.; Shen, Y.-P.; Zheng, Y.; Chan, T.-S.; Sheu, H.-S.; Cheng, Y.-C.; Chen, H. M. Nat. Commun. 2015, 6, No. 8106. (70) Bergmann, A.; Martinez-Moreno, E.; Teschner, D.; Chernev, P.; Gliech, M.; de Araújo, J. F.; Reier, T.; Dau, H.; Strasser, P. Nat. Commun. 2015, 6, No. 8625. CHAPTER IV (1) Lewis, N. S. Science 2016, 351, aad1920. (2) Pinaud, B. A.; Benck, J. D.; Seitz, L. C.; Forman, A. J.; Chen, Z.; Deutsch, T. G.; James, B. D.; Baum, K. N.; Baum, G. N.; Ardo, S.; Wang, H.; Miller, E.; Jaramillo, T. F. Energy Environ. Sci. 2013, 6, 1983. 125 (3) Burke, M. S.; Enman, L. J.; Batchellor, A. S.; Zou, S.; Boettcher, S. W. Chem. Mater. 2015, 27, 7549. (4) Armaroli, N.; Balzani, V. Chem. - A Eur. J. 2016, 22, 32. (5) Zou, S.; Burke, M. S.; Kast, M. G.; Fan, J.; Danilovic, N.; Boettcher, S. W. Chem. Mater. 2015, 27, 8011. (6) Burke, M. S.; Zou, S.; Enman, L. J.; Kellon, J. E.; Gabor, C. A.; Pledger, E.; Boettcher, S. W. J. Phys. Chem. Lett. 2015, 6, 3737. (7) Lee, J.; Lee, H.; Lim, B. J. Ind. Eng. Chem. 2018, 58, 100. (8) Luo, W.; Jiang, C.; Li, Y.; Shevlin, S. A.; Han, X.; Qiu, K.; Cheng, Y.; Guo, Z.; Huang, W.; Tang, J. J. Mater. Chem. A 2017, 5, 2021. (9) Vassalini, I.; Borgese, L.; Mariz, M.; Polizzi, S.; Aquilanti, G.; Ghigna, P.; Sartorel, A.; Amendola, V.; Alessandri, I. Angew. Chemie - Int. Ed. 2017, 56, 6589. (10) Zhao, Q.; Li, D.; Gao, G.; Yuan, W.; Hao, G.; Li, J. Int. J. Hydrogen Energy 2016, 41, 17193. (11) Doyle, R. L.; Lyons, M. E. G. J. Electrochem. Soc. 2013, 160, H142. (12) Burke, M. S.; Kast, M. G.; Trotochaud, L.; Smith, A. M.; Boettcher, S. W. J. Am. Chem. Soc. 2015, 137, 3638. (13) Trotochaud, L.; Young, S. L.; Ranney, J. K.; Boettcher, S. W. J. Am. Chem. Soc. 2014, 136, 6744. (14) Benck, J. D.; Pinaud, B. A.; Gorlin, Y.; Jaramillo, T. F. PLoS One 2014, 9, 1. (15) Klaus, S.; Trotochaud, L.; Cheng, M.-J.; Head-Gordon, M.; Bell, A. T. ChemElectroChem 2016, 3, 66. (16) Chakthranont, P.; Kibsgaard, J.; Gallo, A.; Park, J.; Mitani, M.; Sokaras, D.; Kroll, T.; Sinclair, R.; Mogensen, M. B.; Jaramillo, T. F. ACS Catal. 2017, 7, 5399. (17) Frydendal, R.; Busch, M.; Halck, N. B.; Paoli, E. A.; Krtil, P.; Chorkendorff, I.; Rossmeisl, J. ChemCatChem 2015, 7, 149. (18) Hou, Y.; Liu, Y.; Gao, R.; Li, Q.; Guo, H.; Goswami, A.; Zboril, R.; Gawande, M. B.; Zou, X. ACS Catal. 2017, 7, 7038. (19) Kim, H.; Kim, Y.; Noh, Y.; Kim, W. B. Dalt. Trans. 2016, 45, 13686. (20) Sidhureddy, B.; Thiruppathi, A. R.; Chen, A. J. Electroanal. Chem. 2017, 794, 28. (21) Strickler, A. L.; Escudero-Escribano, M.; Jaramillo, T. F. Nano Lett. 2017, 17, 6040. (22) Qu, Q.; Zhang, J.-H.; Wang, J.; Li, Q.-Y.; Xu, C.-W.; Lu, X. Sci. Rep. 2017, 7, 41542. 126 (23) Yu, L.; Zhou, H.; Sun, J.; Qin, F.; Yu, F.; Bao, J.; Yu, Y.; Chen, S.; Ren, Z. Energy Environ. Sci. 2017, 10, 1820. (24) Zhou, Y.; Zeng, H. C. J. Phys. Chem. C 2016, 120, 29348. (25) Sayeed, M. A.; Herd, T.; O’Mullane, A. P. J. Mater. Chem. A 2016, 4, 991. (26) Yeo, B. S.; Bell, A. T. J. Phys. Chem. C 2012, 116, 8394. (27) Yeo, B. S.; Bell, A. T. J. Am. Chem. Soc. 2011, 133, 5587. (28) Gorlin, Y.; Chung, C. J.; Benck, J. D.; Nordlund, D.; Seitz, L.; Weng, T. C.; Sokaras, D.; Clemens, B. M.; Jaramillo, T. F. J. Am. Chem. Soc. 2014, 136, 4920. (29) Seitz, L. C.; Hersbach, T. J. P.; Nordlund, D.; Jaramillo, T. F. J. Phys. Chem. Lett. 2015, 6, 4178. (30) Corrigan, D. A. J. Electrochem. Soc. 1987, 134, 377. (31) Stevens, M. B.; Enman, L. J.; Batchellor, A. S.; Cosby, M. R.; Vise, A. E.; Trang, C. D. M.; Boettcher, S. W. Chem. Mater. 2017, 29, 120. (32) Roger, I.; Symes, M. D. J. Am. Chem. Soc. 2015, 137, 13980. (33) Gorlin, M.; De Araujo, J. F.; Schmies, H.; Bernsmeier, D.; Dresp, S.; Gliech, M.; Jusys, Z.; Chernev, P.; Kraehnert, R.; Dau, H.; Strasser, P. J. Am. Chem. Soc. 2017, 139, 2070. (34) Ng, J. W. D.; García-Melchor, M.; Bajdich, M.; Chakthranont, P.; Kirk, C.; Vojvodic, A.; Jaramillo, T. F. Nat. Energy 2016, 1, 1. (35) Frydendal, R.; Seitz, L. C.; Sokaras, D.; Weng, T. C.; Nordlund, D.; Chorkendorff, I.; Stephens, I. E. L.; Jaramillo, T. F. Electrochim. Acta 2017, 230, 22. (36) Lu, A.; Peng, D. L.; Chang, F.; Skeete, Z.; Shan, S.; Sharma, A.; Luo, J.; Zhong, C. J. ACS Appl. Mater. Interfaces 2016, 8, 20082. CHAPTER V [1] M. S. Burke, L. J. Enman, A. S. Batchellor, S. Zou, S. W. Boettcher, Chem. Mater. 2015, 27, 7549–7558. [2] N. Armaroli, V. Balzani, Chem. - A Eur. J. 2016, 22, 32–57. [3] J. H. Montoya, L. C. Seitz, P. Chakthranont, A. Vojvodic, T. F. Jaramillo, J. K. Nørskov, Nat. Mater. 2016, 16, 70–81. [4] L. Trotochaud, J. K. Ranney, K. N. Williams, S. W. Boettcher, J. Am. Chem. Soc. 2012, 134, 17253–61. [5] M. S. Burke, S. Zou, L. J. Enman, J. E. Kellon, C. A. Gabor, E. Pledger, S. W. Boettcher, J. Phys. Chem. Lett. 2015, 6, 3737–3742. 127 [6] D. Friebel, M. W. Louie, M. Bajdich, K. E. Sanwald, Y. Cai, A. M. Wise, M.-J. Cheng, D. Sokaras, T.-C. Weng, R. Alonso-Mori, et al., J. Am. Chem. Soc. 2015, 137, 1305–1313. [7] M. Gorlin, J. F. De Araujo, H. Schmies, D. Bernsmeier, S. Dresp, M. Gliech, Z. Jusys, P. Chernev, R. Kraehnert, H. Dau, et al., J. Am. Chem. Soc. 2017, 139, 2070–2082. [8] M. Görlin, P. Chernev, J. Ferreira de Araújo, T. Reier, S. Dresp, B. Paul, R. Krähnert, H. Dau, P. Strasser, J. Am. Chem. Soc. 2016, 138, 5603–5614. [9] R. D. L. Smith, C. Pasquini, S. Loos, P. Chernev, K. Klingan, P. Kubella, M. R. Mohammadi, D. Gonzalez-Flores, H. Dau, Nat. Commun. 2017, 8, 2022. [10] M. K. Bates, Q. Jia, H. Doan, W. Liang, S. Mukerjee, ACS Catal. 2016, 6, 155– 161. [11] L. Gong, X. Y. E. Chng, Y. Du, S. Xi, B. S. Yeo, ACS Catal. 2018, 8, 807–814. [12] N. Li, D. K. Bediako, R. G. Hadt, D. Hayes, T. J. Kempa, F. von Cube, D. C. Bell, L. X. Chen, D. G. Nocera, Proc. Natl. Acad. Sci. 2017, 114, 1486–1491. [13] J. Y. C. Chen, L. Dang, H. Liang, W. Bi, J. B. Gerken, S. Jin, E. E. Alp, S. S. Stahl, J. Am. Chem. Soc. 2015, 137, 15090–15093. [14] B. M. Hunter, N. B. Thompson, A. M. Müller, G. R. Rossman, M. G. Hill, J. R. Winkler, H. B. Gray, Joule 2018, 747–763. [15] M. S. Burke, M. G. Kast, L. Trotochaud, A. M. Smith, S. W. Boettcher, J. Am. Chem. Soc. 2015, 137, 3638–3648. [16] L. Trotochaud, S. L. Young, J. K. Ranney, S. W. Boettcher, J. Am. Chem. Soc. 2014, 136, 6744–53. [17] B. M. Hunter, W. Hieringer, J. Winkler, H. B. Gray, A. M. Müller, Energy Environ. Sci. 2016, 23, 1–22. [18] M. B. Stevens, C. D. M. Trang, L. J. Enman, J. Deng, S. W. Boettcher, J. Am. Chem. Soc. 2017, 139, 11361–11364. [19] H. Yang, Y. Liu, S. Luo, Z. Zhao, X. Wang, Y. Luo, Z. Wang, J. Jin, J. Ma, ACS Catal. 2017, 7, 5557–5567. [20] F. Song, X. Hu, Nat. Commun. 2014, 5, 4477. [21] J. Fester, M. García-Melchor, A. S. Walton, M. Bajdich, Z. Li, L. Lammich, A. Vojvodic, J. V. Lauritsen, L. R. Merte, M. A. Henderson, et al., Nat. Commun. 2017, 8, 14169. [22] B. Hunter, J. Winkler, H. Gray, Molecules 2018, 23, 903. [23] M. Favaro, J. Yang, S. Nappini, E. Magnano, F. M. Toma, E. J. Crumlin, J. Yano, I. D. Sharp, J. Am. Chem. Soc. 2017, 139, 8960–8970. 128 [24] E. Fabbri, M. Nachtegaal, T. Binninger, X. Cheng, B. J. Kim, J. Durst, F. Bozza, T. Graule, R. Schäublin, L. Wiles, et al., Nat. Mater. 2017, 16, 925–931. [25] L. Calvillo, F. Carraro, O. Vozniuk, V. Celorrio, L. Nodari, A. E. Russell, D. Debellis, D. Fermin, F. Cavani, S. Agnoli, et al., J. Mater. Chem. A 2018, 6, 7034- 7041. [26] B. Ravel, M. Newville, J. Synchrotron Radiat. 2005, 12, 537–541. [27] B. Ravel, J. Synchrotron Radiat. 2001, 8, 314–316. [28] M. Newville, J. Synchrotron Radiat. 2001, 8, 322–324. [29] M. Deliens, H. Goethals, Miner. Mag. 1973, 39, 152–157. [30] D. Gonzáles-Flores, K. Klingan, P. Chernev, S. Loos, M. R. Mohammadi, C. Pasquini, P. Kubella, I. Zaharieva, R. Smith, H. Dau, Sustain. Energy Fuels 2018, DOI 10.1039/C8SE00114F. [31] J. G. McAlpin, Y. Surendranath, M. Dinca, T. A. Stich, S. A. Stoian, W. H. Casey, D. G. Nocera, R. D. Britt, J. Am. Chem. Soc. 2010, 132, 6882–6883. [32] R. G. Hadt, D. Hayes, C. N. Brodsky, A. M. Ullman, D. M. Casa, M. H. Upton, D. G. Nocera, L. X. Chen, J. Am. Chem. Soc. 2016, 138, 11017–11030. [33] M. Balasubramanian, C. A. Melendres, S. Mini, J. Phys 2000, 104, 4300–4306. [34] S. P. Cramer, T. K. Eccles, F. W. Kutzler, K. O. Hodgson, L. E. Mortenson, J. Am. Chem. Soc. 1976, 98, 1287–1288. [35] R. L. Blake, R. E. Hessevick, T. Zoltai, L. W. Finger, Am. Mineral. 1966, 51, 123– 129. [36] H. Christensen, A. N. Christensen, U. Turpeinen, A. F. Andresen, O. Smidsrød, C.-O. Pontchour, P. Phavanantha, S. Pramatus, B. N. Cyvin, S. J. Cyvin, Acta Chem. Scand. 1978, 32a, 87–88. [37] G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558–561. [38] G. Kresse, J. Furthmüller, Comput. Mater. Sci. 1996, 6, 15–50. [39] S. Dudarev, G. Botton, Phys. Rev. B - Condens. Matter Mater. Phys. 1998, 57, 1505–1509. [40] I. C. Man, H.-Y. Su, F. Calle-Vallejo, H. A. Hansen, J. I. Martínez, N. G. Inoglu, J. Kitchin, T. F. Jaramillo, J. K. Nørskov, J. Rossmeisl, ChemCatChem 2011, 3, 1159–1165. [41] K. Momma, F. Izumi, J. Appl. Crystallogr. 2011, 44, 1272–1276. 129 APPENDIX A (1) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. Handbook of X-Ray Photoelectron Spectroscopy: A Reference Book of Standard Spectra for Identification and Interpretation of XPS Data; Chastain, J., Ed.; Perkin-Elmer: Eden Prairie, MN, 1992. (2) Biesinger, M. C.; Lau, L. W. M.; 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, 887–898. (3) Aronniemi, M.; Sainio, J.; Lahtinen, J. Chemical State Quantification of Iron and Chromium Oxides Using XPS: The Effect of the Background Subtraction Method. Surf. Sci. 2005, 578, 108–123. (4) Biesinger, M. C.; Brown, C.; Mycroft, J. R.; Davidson, R. D.; McIntyre, N. S. X- Ray Photoelectron Spectroscopy Studies of Chromium Compounds. Surf. Interface Anal. 2004, 36, 1550–1563. (5) Biesinger, M. C.; Payne, B. P.; Grosvenor, A. P.; Lau, L. W. M.; Gerson, A. R.; Smart, R. S. C. Resolving Surface Chemical States in XPS Analysis of First Row Transition Metals, Oxides and Hydroxides: Cr, Mn, Fe, Co and Ni. Appl. Surf. Sci. 2011, 257, 2717–2730. (6) Yamashita, T.; Hayes, P. Analysis of XPS Spectra of Fe2+ and Fe3+ Ions in Oxide Materials. Appl. Surf. Sci. 2008, 254, 2441–2449. (7) Biesinger, M. C.; Payne, B. P.; Lau, L. W. M.; Gerson, A.; Smart, R. S. C. X-Ray Photoelectron Spectroscopic Chemical State Quantification of Mixed Nickel Metal, Oxide and Hydroxide Systems. Surf. Interface Anal. 2009, 41, 324–332. (8) Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: Data Analysis for X-Ray Absorption Spectroscopy Using IFEFFIT. J. Synchrotron Radiat. 2005, 12, 537–541. (9) Newville, M. IFEFFIT : Interactive XAFS Analysis and FEFF Fitting. J. Synchrotron Radiat. 2001, 8, 322–324. APPENDIX B (1) Batchellor, A. S.; Boettcher, S. W. ACS Catal. 2015, 5, 6680–6689. (2) Corrigan, D. A.; Knight, S. L. J. Electrochem. Soc. 1989, 136, 613–619. (3) Mullins, D. R.; Overbury, S. H.; Huntley, D. R. Surf. Sci. 1998, 409, 307–319. (4) Wehrens-Dijksma, M.; Notten, P. H. L. Electrochim. Acta 2006, 51, 3609–3621. 130 APPENDIX D [1] M. S. Burke, M. G. Kast, L. Trotochaud, A. M. Smith, S. W. Boettcher, J. Am. Chem. Soc. 2015, 137, 3638–3648. [2] B. Ravel, M. Newville, J. Synchrotron Radiat. 2005, 12, 537–541. [3] B. Ravel, J. Synchrotron Radiat. 2001, 8, 314–316. [4] M. Newville, J. Synchrotron Radiat. 2001, 8, 322–324. [5] M. Deliens, H. Goethals, Miner. Mag. 1973, 39, 152–157. [6] G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558–561. [7] G. Kresse, J. Furthmüller, Comput. Mater. Sci. 1996, 6, 15–50. [8] S. Dudarev, G. Botton, Phys. Rev. B - Condens. Matter Mater. Phys. 1998, 57, 1505–1509. [9] Y. F. Li, A. Selloni, J. Phys. Chem. Lett. 2014, 5, 3981–3985. [10] Y. Li, A. Selloni, ACS Catal. 2014, 4, 1148–1153. [11] A. J. Tkalych, K. Yu, E. A. Carter, J. Phys. Chem. C 2015, 119, 24315–24322. [12] V. Fidelsky, M. Caspary Toroker, J. Phys. Chem. C 2016, 120, 8104–8108. [13] J. Zaffran, M. Caspary Toroker, J. Chem. Theory Comput. 2016, 12, 3807–3812. [14] P. E. Blöchl, Phys. Rev. B 1994, 50, 17953–17979. [15] G. Kresse, D. Joubert, Phys. Rev. B 1999, 59, 1758–1775. [16] M. Casas-Cabanas, J. Canales-Vázquez, J. Rodríguez-Carvajal, M. R. Palacín, J. Am. Chem. Soc. 2007, 129, 5840–5842. [17] P. Oliva, J. Leonardi, J. F. Laurent, C. Delmas, J. J. Braconnier, M. Figlarz, F. Fievet, A. d. Guibert, J. Power Sources 1982, 8, 229–255. [18] R. G. Delaplane, J. A. Ibers, J. R. Ferraro, J. J. Rush, J. Chem. Phys. 1969, 50, 1920–1927. [19] J. Chen, A. Selloni, J. Phys. Chem. C 2013, 117, 20002–20006. [20] D. Friebel, M. W. Louie, M. Bajdich, K. E. Sanwald, Y. Cai, A. M. Wise, M.-J. Cheng, D. Sokaras, T.-C. Weng, R. Alonso-Mori, et al., J. Am. Chem. Soc. 2015, 137, 1305–1313. [21] I. C. Man, H.-Y. Su, F. Calle-Vallejo, H. A. Hansen, J. I. Martínez, N. G. Inoglu, J. Kitchin, T. F. Jaramillo, J. K. Nørskov, J. Rossmeisl, ChemCatChem 2011, 3, 1159–1165. [22] M. Bajdich, M. García-Mota, A. Vojvodic, J. K. Nørskov, A. T. Bell, J. Am. Chem. Soc. 2013, 135, 13521–13530. 131 [23] C. Hareli, M. C. Toroker, J. Chem. Theory Comput. 2018, 14, 2380–2385. [24] M. C. Toroker, J. Phys. Chem. C 2014, 118, 23162–23167.